Voet D., Voet Ju.G. Biochemistry

Подождите немного. Документ загружается.

3. The primosome is propelled in the 5¿S3¿ direction

along the (⫹) strand by the PriA and DnaB helicases at the

expense of ATP hydrolysis. This motion, which displaces

the SSB in its path, is opposite in direction to that of tem-

plate reading during DNA chain propagation.

4. At randomly selected sites, the primosome reverses

its migration while primase synthesizes an RNA primer.

The initiation of primer synthesis requires the participation

of DnaB protein which, through concomitant ATP hydrol-

ysis, is thought to alter template DNA conformation in a

manner required by primase.

5. Pol III holoenzyme extends the primers to form

Okazaki fragments.

6. Pol I excises the primers and replaces them by DNA.

The fragments are then joined by DNA ligase and super-

coiled by DNA gyrase to form the ␾X174 RF I.

The primosome remains complexed with the DNA (Fig.30-25)

where it participates in (⫹) strand synthesis (see below).

b. ␾X174 (ⴙ) Strand Replication Serves as a Model

for Leading Strand Synthesis

One strand of a circular duplex DNA may be synthe-

sized via the rolling circle or ␴-replication mode (so called

because of the resemblance of the replicating structure to

the Greek letter sigma; Fig. 30-26). The ␾X174 (⫹) strand is

synthesized on an RF I template by a variation on this

process, the looped rolling circle mode (Fig. 30-27):

1. (⫹) strand synthesis begins with the primosome-

aided binding of the phage-encoded 513-residue enzyme

gene A protein to its ⬃30-bp recognition site. There, gene

A protein cleaves a specific phosphodiester bond on the

(⫹) strand nucleotide (near the beginning of gene A) by

forming a covalent bond between a Tyr residue and the

DNA’s 5¿-phosphoryl group, thereby conserving the

cleaved bond’s energy.

2. Rep helicase (Section 30-2Cb) subsequently attaches

to the (⫺) strand at the gene A protein and, with the aid of

Section 30-3. Prokaryotic Replication 1191

Figure 30-24 The synthesis of the ␾X174 (⫺) strand on a (⫹)

strand template to form ␾X174 RF I DNA. [After Arai, K., Low,

R., Kobori, J., Schlomai, J., and Kornberg, A., J. Biol. Chem. 256,

5280 (1981).]

Figure 30-25 Electron micrograph of a primosome bound to a

␾X174 RF I DNA. Such complexes always contain a single

primosome with one or two associated small DNA loops.

[Courtesy of Jack Griffith, Lineberger Cancer Research Center,

University of North Carolina.]

Single-strand

binding protein

(SSB)

1. Recognition

+ DnaC

(DnaB)

• (DnaC)

DnaT + ATP

ADP + P

ADP

RFII

RFI

ATP

+ 4rNTPs

DNA polymerase III

holoenzyme, 4dNTPs

Pol I, ligase,

gyrase, 4dNTPs,

NAD

, ATP

ATP

Primase

Primasome

PriA, PriB, PriC

pas

2. Assembly

3. Migration

4. Priming

5. Elongation

6. Excision,

Gap filling,

ligation,

supercoiling

PriA

JWCL281_c30_1173-1259.qxd 8/10/10 9:11 PM Page 1191

the primosome still associated with the (⫹) strand, com-

mences unwinding the duplex DNA from the (⫹) strand’s

5¿ end. The displaced (⫹) strand is coated with SSB, which

prevents it from reannealing to the (⫺) strand. Rep heli-

case is essential for the replication of ␾X174 DNA, but not

for the E. coli chromosome, as is demonstrated by the in-

ability of ␾X174 to multiply in rep

⫺

E. coli. Pol III holoen-

zyme extends the (⫹) strand from its free 3¿-OH group.

1192 Chapter 30. DNA Replication, Repair, and Recombination

Figure 30-26 The rolling circle mode of DNA replication. The

(⫹) strand being synthesized is extended from a specific cut

made at the replication origin (1) so as to strip away the old (⫹)

strand (2 and 3). The continuous synthesis of the (⫹) strand on a

circular (⫺) strand template produces a series of tandemly

linked (⫹) strands (4), which may later be separated by a specific

endonuclease.

Origin

(+)(–)

Replicative

form

Origin

5′

3′

(+)

(–)

(+)

(–)

(+)

(–)

RF with

associated primosome

Origin

Gene A protein

(–) (+)

1 Gene A protein

Rep, SSB,

Pol ,

III

SSB

(+)

(–)

(+)

etc.

Rep

Figure 30-27 The synthesis of the ␾X174 (⫹)

strand by the looped rolling circle mode. The

numbered steps are described in the text.

JWCL281_c30_1173-1259.qxd 8/10/10 9:11 PM Page 1192

Section 30-3. Prokaryotic Replication 1193

Leading strand

(a)

(b)

DNA polymerase III

holoenzyme

5′

3′

5′

3′

Lagging strand

Parental strand

Primosome

RNA primer

Growing Okazaki

fragment

5′

3′

5′

3′

Primosome making

new RNA primer

5′

3′

RNA primer to be replaced

with DNA by Pol

nick sealed by DNA ligase

Completed Okazaki fragment

5′

3′

(c)

5′

3′

Newly

initiated

Okazaki fragment

Old Okazaki

fragment

5′

3′

5′

3′

DnaB helicase

SSB

Sliding clamp

Figure 30-28 The replication of E. coli DNA. (a) The

E. coli DNA replisome, which contains two DNA

polymerase III holoenzyme complexes, synthesizes both

the leading and the lagging strands.The lagging strand

template must loop around to permit the holoenzyme to

extend the primosome-primed lagging strand.Although not

shown here, the DnaB helicase binds to ␶

and hence moves with

the replisome. (b) The holoenzyme releases the lagging strand

template when it encounters the previously synthesized Okazaki

fragment.This possibly signals the primosome to initiate the

synthesis of lagging strand RNA primer. (c) The holoenzyme

rebinds the lagging strand template and extends the RNA primer

to form a new Okazaki fragment. Note that in this model, leading

strand synthesis is always ahead of lagging strand synthesis.

3. The extension process generates a looped rolling cir-

cle structure in which the 5¿ end of the old (⫹) strand re-

mains linked to the gene A protein at the replication fork.

It is thought that as the old (⫹) strand is peeled off the RF,

the primosome synthesizes the primers required for the

later generation of a new (⫺) strand.

4. When it has come full circle around the (⫺) strand,

the gene A protein again makes a specific cut at the repli-

cation origin so as to form a covalent linkage with the new

(⫹) strand’s 5¿ end. Simultaneously, the newly formed 3¿-

terminal OH group of the old, looped-out (⫹) strand nu-

cleophilically attacks its 5¿-phosphoryl attachment to the

gene A protein, thereby liberating a covalently closed (⫹)

strand.This is possible because the gene A protein has two

closely spaced Tyr residues that alternate in their attach-

ment to the 5¿ ends of successively synthesized (⫹) strands.

The replication fork continues its progress about the du-

plex circle, producing new (⫹) strands in a manner reminis-

cent of linked sausages being pulled off a reel.

In the intermediate stages of a ␾X174 infection, each newly

synthesized (⫹) strand directs the synthesis of the (⫺)

strand to form RF I as described above. In the later stages

of infection, however, the newly formed (⫹) strands are

packaged into phage particles.

C. Escherichia coli

See Guided Exploration 25. The replication of DNA in E. coli The

E. coli chromosome replicates by the bidirectional ␪ mode from

a single replication origin (Section 30-1Aa).The most plausi-

ble model for events at the E. coli replication fork (Fig. 30-

28) is largely derived from studies on the simpler and more

experimentally accessible DNA replication mechanisms of

coliphages such as M13 and ␾X174. Duplex DNA is un-

wound by DnaB helicase on the lagging strand template,

JWCL281_c30_1173-1259.qxd 8/27/10 7:27 PM Page 1193

1194 Chapter 30. DNA Replication, Repair, and Recombination

where it is joined by the primosome. The separated single

strands are immediately coated by SSB. Leading strand syn-

thesis is catalyzed by Pol III holoenzyme, as is that of the

lagging strand after priming by primosome-associated pri-

mase. Both leading and lagging strand syntheses occur on a

single ⬃900-kD multisubunit particle, the replisome, which

contains two Pol III cores (␣ε␪) that are joined together by

a dimer of ␶ subunits that bridges the ␣ subunits. Hence, the

lagging strand template must be looped around (Fig. 30-28).

The ␶

dimer also binds the DnaB helicase (an interaction

that is not indicated in Fig. 30-28), thereby stimulating its

helicase action while holding it to the replication fork.After

completing the synthesis of an Okazaki fragment, the lag-

ging strand holoenzyme relocates to a new primer near the

replication fork, the primer heading the previously synthe-

sized Okazaki fragment is excised by Pol I-catalyzed nick

translation, and the nick is sealed by DNA ligase. Since lag-

ging strand synthesis is more complex and hence more time-

consuming than leading strand synthesis, the replisome func-

tions to coordinate these two processes.

a. E. coli DNA Replication Is Initiated at oriC in a

Process Mediated by DnaA Protein

The replication origin of the E. coli chromosome consists

of a unique 250-bp segment known as the oriC locus.This se-

quence, segments of which are highly conserved among

gram-negative bacteria, supports the bidirectional replica-

tion of the various plasmids into which it has been inserted.

The oriC locus contains five highly conserved 9-bp seg-

ments with consensus sequence 5¿-TTATCCACA-3¿ known as

DnaA boxes because they are specifically bound by DnaA

protein (Fig. 30-29a).These are interspersed with several so-

called I-sites that deviate from this consensus sequence and

are bound by DnaA with lesser affinity. In addition, the

“left” boundary region of oriC contains three tandemly re-

peated, 13-bp, AT-rich segments (consensus sequence 5¿-

GATCTNTTNTTTT-3¿ where N marks nonspecific posi-

tions) that are known as DNA unwinding elements (DUEs).

DnaA (467 residues in E. coli) consists of four domains

that are, from N- to C-terminus, a helicase interaction domain

that mediates interactions with DnaB helicase (see below), a

Figure 30-29 DNA replication initiation at

oriC. (a) Diagram of oriC showing the relative

positions of its DnaA boxes (green) and its DNA

unwinding elements (DUEs; yellow). (b) X-ray

structure of the right-handed helical filament

formed by the two C-terminal domains of

A. aeolicus DnaA. It has eight subunits per turn

and a pitch of 178 Å.Twelve subunits are shown,

from right to left, in the alternating colors red,

green, yellow, and blue.The right subunits are

drawn as surface diagrams and the left subunits

are represented by their polypeptide back bones

with helices in tube form. (c) Model for initiation

at oriC.The green ovals represent the N-terminal

three domains of DnaA and the yellow ovals

represent the associated C-terminal DNA-binding

domains. See the text for an explanation. [Parts b

and c courtesy of James Berger, University of

California at Berkeley. PDBid 2HCB.]

Supercoiled

template

DnaA

boxes

13–bp

segments

oriC

(a)

(b)

(c)

JWCL281_c30_1173-1259.qxd 8/10/10 9:11 PM Page 1194

35 0 min

Division

Initiation

Termination

Multiforked

chromosome

TerminusOrigin

20 15

of DnaA. The AAA⫹ domains of DnaA and DnaC interact

in an ATP-dependent manner to recruit and properly position

the DnaB helicases, following which the DnaC is released.

In the presence of SSB and gyrase, DnaB helicase fur-

ther unwinds the DNA in the prepriming complex in both

directions so as to permit the entry of primase and RNA

polymerase. The participation of both these enzymes in

leading strand primer synthesis (Section 30-1D), together

with the limitation of this process to the oriC site, suggests

that the RNA polymerase activates primase to synthesize

the primer. This perhaps explains the similarity of oriC’s

DUEs to RNA polymerase’s transcriptional promoters

(Section 31-2Ba). The stage is thereby set for bidirectional

DNA replication by Pol III holoenzyme as described above.

b. The Initiation of E. coli DNA Replication Is

Strictly Regulated

Chromosome replication in E. coli occurs only once per

cell division, so this process must be tightly controlled.The

doubling (cell generation) time of E. coli at 37°C varies with

growth conditions from ⬍20 min to ⬃10 h.Yet the constant

⬃1000 nt/s rate of movement of each replication fork fixes

the 4.6 ⫻ 10

-bp E. coli chromosome’s replication time, C,at

⬃40 min. Moreover,the segregation of cellular components

and the formation of a septum between them, which must

precede cell division, requires a constant time, D ⫽ 20 min,

after the completion of the corresponding round of chro-

mosome replication. Cells with doubling times less than C ⫹

D ⫽ 60 min must consequently initiate chromosome replica-

tion before the end of the preceding cell division cycle. This

results in the formation of multiforked chromosomes as is

diagrammed in Fig. 30-30 for a cell division time of 35 min.

Even in cells that contain multiple oriC sites, DNA

replication is initiated at each such site once and only once

per cell generation. However, after initiation has occurred,

chain elongation proceeds at a uniform, largely uncon-

trolled rate. This suggests that a post-initiation oriC site is

somehow sequestered from (prevented from interacting

with) the replication initiation machinery, a phenomenon

called sequestration. There is extensive morphological evi-

dence, such as shown in Fig. 30-31, that the E. coli chromo-

some is associated with the cell membrane.This attachment

flexible and poorly conserved linker, an ATPase domain that

is a member of the AAA⫹ family (Section 30-2Ca), and a

DNA binding domain.The X-ray structure of the C-terminal

two domains of DnaA from Aquifex aeolicus (a thermophilic

bacterium), determined by Berger, unexpectedly revealed

that it forms a multisubunit right-handed helix (Fig. 30-29b).

Experiments with oriC-containing plasmids, pioneered

by Kornberg, together with the X-ray structure of DnaA,

indicate that replication initiation in E. coli occurs via the

following process (Fig. 30-29c):

1. In the presence of ATP, DnaA, which is normally

bound to three of oriC’s five DnaA boxes throughout

E. coli’s lifetime, recruits additional DnaA subunits to the re-

maining DnaA boxes and to the I-sites so as to form a right-

handed helix of DnaA subunits that is bound to the DNA.

This generates local positive supercoils in the DNA.The su-

perhelical strain resulting from the compensating negative

supercoils (recall that the linking number of a covalently

closed circular DNA such as an E. coli chromosome is invari-

ant; Section 29-3A) melts the DUE-containing segment [Fig.

30-29c,middle left;recall that bacterial chromosomes are nor-

mally already negatively supercoiled (Section 29-2Bb)]. Al-

ternatively, or in addition, the DnaA’s ATPase domains may

actively unwind the DNA (Fig. 30-29c, middle right). This

process is facilitated by two homologous DNA-binding pro-

teins, HU and integration host factor (IHF), that induce

DNA bending (IHF is discussed in Section 33-3Ca).

2. The oriC–DnaA complex recruits two DnaB

ⴢ DnaC

complexes to opposite ends of the melted region to form the

prepriming complex. DnaC, an ATPase that is a homolog of

DnaA, functions to facilitate the loading of the DnaB hexam-

ers onto the DNA. Its, X-ray structure, also determined by

Berger, shows that it forms a helical assembly similar to that

Section 30-3. Prokaryotic Replication 1195

Figure 30-30 Multiforked chromosomes in E. coli. In cells

that are dividing every 35 min, the fixed 60-min interval between

the initiation of replication and cell division results in the

production of multiforked chromosomes. [After Lewin, B., Genes

VII, p. 370, Oxford University Press (2000).]

Figure 30-31 Electron micrograph of an intact and

supercoiled E. coli chromosome attached to two fragments of

the cell membrane. [From Delius, H. and Worcel, A., J. Mol. Biol.

82, 108 (1974).]

JWCL281_c30_1173-1259.qxd 8/10/10 9:11 PM Page 1195

c. The Clamp Loader Loads the Sliding Clamp onto

the DNA

Extensive investigations in many laboratories have led

to the model of the E. coli replisome drawn in Fig. 30-32.

The sliding clamp, which is responsible for Pol III’s high

processivity, is a ring-shaped dimer of ␤ subunits through

which the DNA strand being replicated is threaded (Sec-

tion 30-2Bb). The two tightly associated ␤ subunits (K

⬍

50 nM) that form the sliding clamp dissociate with a half-

life of ⬃100 min at 37°C.Yet, since each replisome synthe-

sizes around one Okazaki fragment per second, a sliding

clamp must be loaded onto the lagging strand template at

this frequency. This loading function is carried out in an

ATP-dependent process by the ␥ complex (␥␶

␦␦¿␹⌿). The

␶ and ␥ subunits are both encoded by the dnaX gene with ␶

(643 residues) the full-length product and ␥ (431 residues)

its C-terminally truncated form; the C-terminal 122

residues of ␶ are known as ␶

.The ␥ complex, of which only

␶

is diagrammed in Fig. 30-28, bridges the replisome’s two

Pol III cores via its ␶

segments, which also bind the DnaB

helicase (Fig. 30-32). The ␹ and ⌿ subunits form a het-

erodimer in which ␹ competes with primase for its binding

site on SSB and hence functions to accelerate the dissocia-

tion of primase from the RNA primer it synthesized as well

as link the ␥ complex to SSB. However, ␹ and ⌿ are not es-

sential participants in the clamp loading process and, there-

fore, we shall refer to the ␥␶

␦␦¿ complex as the clamp

loader. How does the clamp loader do its job?

Of the clamp loader’s five subunits, only ␦ is capable of

binding to and opening up the sliding clamp on its own.

Kuriyan and O’Donnell determined the X-ray structure of the

␦ subunit in 1:1 complex with a ␤ subunit that had two residues

in its dimerization interface mutated so as to prevent its dimer-

ization.The structure reveals (Fig. 30-33) that ␦, which consists

of three domains, inserts its ␤ interaction element,a hydropho-

1196 Chapter 30. DNA Replication, Repair, and Recombination

would help explain how replicated chromosomes are seg-

regated into different cells during cell division. But what is

the mechanism of sequestration?

The sequence most commonly methylated in E. coli is

the palindrome GATC, which is methylated at N6 of both

its A bases by Dam methyltransferase (Section 30-7).

GATC occurs 11 times in oriC, including at the beginning of

all four of its 13-bp DUEs (see above). Newly replicated

GATC segments are hemimethylated, that is, the GATC

sequences on the newly synthesized strand are unmethyl-

ated. Although Dam methyltransferase begins methylat-

ing most hemimethylated GATC segments immediately

after their synthesis (within ⬃1.5 min), those on oriC re-

main hemimethylated for around one-third of a cell gener-

ation. Consequently, the observation that membranes bind

hemimethylated oriC, but not unmethylated or fully methyl-

ated oriC, suggests that hemimethylated oriC is bound to

the membrane in a way that makes it inaccessible to both

the initiation machinery and Dam methyltransferase.

The association of hemimethylated oriC with mem-

brane requires the presence of the 181-residue SeqA pro-

tein, the product of seqA gene. Thus in seqA

⫺

cells: (1) the

time to fully methylate hemimethylated GATC sites in

oriC is reduced to 5 min, whereas the time to do so for

other GATC sites is unaffected; (2) the synchrony of initia-

tion of multiple oriC sites is lost; and (3) in the absence of

functional Dam methyltransferase, fully methylated oriC-

containing plasmids are replicated numerous times per cell

generation, whereas in the presence of SeqA they are repli-

cated only once. Evidently, sequestration occurs via the

SeqA-mediated binding of hemimethylated oriC to the

membrane. The hemimethylated promoter of the dnaA

gene is similarly sequestered so as to repress its transcrip-

tion, thereby providing an additional mechanism for pre-

venting promiscuous initiation of DNA replication.

Figure 30-32 Architecture of the E. coli replisome. See the text for details. Compare

this to Fig. 30-28. [Courtesy of Charles Richardson, Harvard Medical School.]

JWCL281_c30_1173-1259.qxd 8/10/10 9:11 PM Page 1196

Section 30-3. Prokaryotic Replication 1197

bic plug that forms the tip of its N-terminal domain, into a hy-

drophobic pocket on one face of ␤. Comparison of ␦ in this

structure with that in the ␥

␦␦¿ complex (see below) reveals

that the ␤ interaction element undergoes a dramatic confor-

mational change on binding to ␤ in which its ␣4 helix rotates

by 45° and translates by 5.5 Å. Moreover,in forming the ␤–␦

complex, the ␤ subunit increases its radius of curvature rela-

tive to that in the ␤ dimer (Fig. 30-13) such that the ␤–␦ in-

teraction would induce the opening of one of the sliding

clamp’s ␤–␤ interfaces by ⬃15 Å. Such a gap is large enough

to permit the passage of ssDNA but not dsDNA. Appar-

ently, the clamp loader functions by trapping one ␤ subunit

of the sliding clamp in a conformation that prevents ring clo-

sure rather than actively pulling apart the two halves of the

ring. This is corroborated by molecular dynamics simula-

tions (Section 9-4a) suggesting that a ␤

dimer has a stable

conformation but that an isolated ␤ subunit with the confor-

mation it has in the ␤

dimer rapidly (in ⬃1.5 ns) converts to

a conformation resembling that in the ␤–␦ complex. Thus,

the conformational change of the ␦ subunit’s ␤ interaction

element on binding to a ␤ subunit is reminiscent of the ac-

tion of a plumber’s wrench in unlatching the nearby ␤–␤ in-

terface so as to allow the sliding clamp to spring open.

The X-ray structure of the ␥

␦␦¿ complex (the clamp

loader with both its ␶ subunits lacking ␶

; ␥ and ␶ are inter-

changeable in terms of their clamp loading functions) in

complex with a primer–template DNA and ADP ⴢ BeF

(an

ATP analog), also determined by Kuriyan and O’Donnell,

suggests how the clamp loader functions. The ␥, ␦, and ␦¿

subunits all have similar folds; their N-terminal domains

are all members of the widely distributed AAA⫹ family

(DnaA and DnaC proteins are also members of this fam-

ily) even though only the ␥ (and ␶) subunits bind and hy-

drolyze ATP. The conserved regions of AAA⫹ proteins

consist of two domains, an N-terminal ATP-binding do-

main and a smaller domain composed of a 3-helix bundle,

whose relative orientations vary with ATP binding. The

␥

␦␦¿ pentamer’s C-terminal domains form a ring-shaped

collar (Fig. 30-34a) in which the subunits are arranged in

clockwise order ␦¿–␥1–␥2–␥3–␦ (Fig. 30-34b). The AAA⫹

domains are arranged in a right-handed spiral that tracks

the minor groove of the dsDNA. Nevertheless, the clamp

Figure 30-33 X-ray structure of the ␤–␦ complex. A second ␤

subunit taken from the X-ray structure of the sliding clamp (Fig.

30-13), the “Reference ␤ monomer,” is drawn in gray. The view is

along the edge of the ␤ ring.The ␦ subunit’s ␤ interaction

element (yellow) consists largely of the ␣4 helix and two

hydrophobic residues, Leu 73 and Phe 74, whose side chains are

drawn in stick form. [Courtesy of John Kuriyan, University of

California at Berkeley. PDB 1JQJ.]

Figure 30-34 X-ray structure of the ␥

␦␦ⴕ clamp loader in

complex with a primer–template DNA and ADP ⴢ BeF

. (a)

View between the ␦¿ and ␦ subunits approximately perpendicular

to the dsDNA’s helical axis. The protein is drawn in tube-and-

arrow form with its subunits colored as indicated and with the

C-terminal domain of each subunit a lighter shade. The DNA

consists of 10 bp with a 5¿ overhang of 5 nt and, together with the

ADP ⴢ BeF

, is drawn in space-filling form with primer C cyan,

template C green,ADP C magenta, N blue, O red, and P orange,

Be light green, and F light blue. (b) View rotated 90° about the

horizontal axis relative to Part a.The C-terminal domain of each

subunit has been deleted for clarity. Note how the single-

stranded portion of the template strand turns by ⬃90° to avoid

colliding with the collar formed by the clamp loader’s C-terminal

domains.The stiffness of dsDNA makes it unlikely that it could

make such a turn. [Based on an X-ray structure by Mike

O’Donnell,The Rockefeller University, and John Kuriyan,

University of California at Berkeley. PDBid 3GLF.]

(a)

(b)

JWCL281_c30_1173-1259.qxd 9/2/10 9:02 AM Page 1197

1198 Chapter 30. DNA Replication, Repair, and Recombination

Figure 30-35 Schematic diagram of the clamp loader cycle.

This speculative model is based on a combination of structural

and biochemical information.The “arginine finger,” an Arg side

chain that interacts with the ␥-phosphate group of an ATP bound

tacts between the core and the DNA). Thus once the slid-

ing clamp has been loaded onto the primer–template, the

clamp loader is replaced by the Pol III core, which thereby

blocks the clamp loader from unloading the clamp. Instead,

the clamp loader loads a new clamp onto the lagging strand

template in association with the primer that the primo-

some had synthesized in preparation for the next round of

Okazaki fragment synthesis (Fig. 30-28b).

The X-ray structure of the sliding clamp in complex with

primer–template DNA (Fig. 30-13) indicates that its

ssDNA segment binds to the same site as do the ␣ and ␦

subunits. This may serve to attract the primer–template

DNA to the inside of the open clamp, which in turn may

facilitate the release of the clamp loader and hence the clo-

sure of the sliding clamp. The binding of ssDNA to the

sliding clamp may also prevent it from sliding away before

it can be bound by the ␣ subunit.

When the Pol III core has completed its synthesis of the

Okazaki fragment, that is, when the gap between the two

successively synthesized Okazaki fragments has been re-

duced to a nick, it releases the DNA and the sliding clamp.

The Pol III core then binds to the newly primed template

and its associated clamp (displacing the clamp loader),

where it commences the synthesis of the next Okazaki

fragment. Thus, a series of switches that are activated by

ATP and DNA structure ensure the vectorial progression

of lagging strand replication. Throughout this process, the

Pol III holoenzyme is held at the replication fork by the

leading strand Pol III core, which remains tethered to

the DNA by its associated sliding clamp.

The sliding clamp that remains around the completed

Okazaki fragment probably functions to recruit Pol I and

DNA ligase so as to replace the RNA primer on the previ-

ously synthesized Okazaki fragment with DNA and seal

the remaining nick. However, the sliding clamp must even-

tually be recycled. It was initially assumed that this was the

job of the clamp loader. However, it is now clear that the

release of the sliding clamp from its associated DNA is

largely carried out by free ␦ subunit (the “wrench” in the

clamp loader that cracks apart the ␤ subunits forming the

sliding clamp), which is synthesized in 5-fold excess over

that required to populate the cell’s few clamp loaders.

loader associates with the DNA almost entirely through

contacts with the phosphate groups of the template strand

alone. Thus, this structure is reminiscent of that of the E1

helicase in complex with ssDNA (Section 30-2Ca) with one

of its six subunits missing.

The clamp loader must tightly bind the sliding clamp

prior to its loading on the template DNA but must subse-

quently release the clamp to avoid interfering with its

binding to the Pol III core (␣ε␪). The structures of the

clamp loader and the ␤–␦ complex, together with a variety

of biochemical evidence, suggest a model of how this

might occur (Fig. 30-35): The binding of ATP to ␥1 (the ␥

subunit that contacts ␦¿) results in a conformational

change that exposes the otherwise occluded ATP-binding

site of ␥2;ATP binding to ␥2 likewise exposes ␥3; and ATP

binding to ␥3 exposes the ␦ subunit’s ␤ interaction ele-

ment, thereby permitting it to bind to a ␤ subunit so as to

spring open the sliding clamp. Primer–template DNA then

inserts itself through the resulting gap in the sliding clamp.

This process is facilitated by the gap between the AAA⫹

domains of ␦ and ␦¿ subunits, which permits the clamp

loader to track the template strand while avoiding contact

with the primer strand. Eventually, ␤- and DNA-stimu-

lated hydrolysis of the bound ATPs releases the ␤ subunit

from the clamp loader, whereon the sliding clamp closes

around the DNA.

The departure of the clamp loader permits the Pol III

core to bind to the sliding clamp. However, when the syn-

thesis of an Okazaki fragment has been completed, the Pol

III core must dissociate from the sliding clamp so that it

can initiate the synthesis of the next Okazaki fragment.

How does this occur?

Pol III’s ␣ subunit binds to the same hydrophobic

pocket on the ␤

sliding clamp as does the ␦ subunit. This

was shown by the observations that the phosphorylation of

a kinase recognition sequence that had been engineered

into the C-terminal segment of ␤ is inhibited by both ␣ and

␦. The ␤ subunit has an ⬃30-fold greater affinity for the ␥

complex in the presence of ATP than it has for the Pol III

core. However, when primer–template DNA is also pres-

ent, this order of affinity is reversed with ␤ preferring to

bind to the Pol III core (possibly due to the additional con-

to a neighboring subunit, is a common feature of AAA⫹

ATPases that form ringlike structures. [Modified from a drawing

by Mike O’Donnell,The Rockefeller University, and John

Kuriyan, University of California at Berkeley.]

JWCL281_c30_1173-1259.qxd 8/10/10 9:11 PM Page 1198

form a deep positively charged cleft that largely envelops

the bound DNA (Fig. 30-37). A 5-bp segment of the DNA

near the side of Tus that permits the passage of the replica-

tion fork (the lower side of Fig. 30-37) is deformed and un-

derwound relative to canonical (ideal) B-DNA such that its

major groove becomes deeper and its minor groove is sig-

nificantly expanded.The protein makes polar contacts with

more than two-thirds of the phosphate groups in a 13-bp re-

gion and its interdomain ␤ sheet penetrates the deepened

major groove to make sequence-specific contacts with the

exposed bases. The importance of this interdomain region

for Tus function is demonstrated by the observation that

most single residue mutations that reduce the ability of Tus

to arrest replication occur in this interdomain region.

When Tus is fused to another DNA-binding protein,

replication is inhibited at the other protein’s binding site.

This suggests that Tus does not act as a simple DNA-bind-

ing clamp, but interacts with DnaB helicase, the leading

component of a replication fork (Fig. 30-32), to inhibit its

helicase action. Apparently, Tus prevents the progress of

DnaB in unwinding DNA from one side of Tus but not the

other. Indeed, the encounter of DnaB with a Tus–Ter com-

plex in the permissive direction causes Tus to rapidly disso-

ciate from the DNA, whereas such an encounter from the

nonpermissive direction generates a so-called locked

Tus–Ter complex. Nevertheless, the way Tus and DnaB in-

teract is unknown. Curiously, however, this termination sys-

tem is not essential for termination. When the replication

Section 30-3. Prokaryotic Replication 1199

Figure 30-36 Map of the E. coli chromosome showing the

positions of the Ter sites and the oriC site. The TerJ, TerG, TerF,

TerB, and TerC sites, in combination with Tus protein, allow a

counterclockwise-moving replisome to pass but not a clockwise-

moving replisome. The opposite is true of the TerH,TerI, TerE,

TerD, and TerA sites. Consequently, two replication forks that

initiate bidirectional DNA replication at oriC will meet between

the oppositely facing Ter sites.

Figure 30-37 X-ray structure of E. coli Tus protein in complex

with a 15-bp Ter-containing DNA. The protein is drawn in ribbon

form colored in rainbow order from its N-terminus (blue) to its

C-terminus (red).The DNA is shown in stick form with C gray,

N blue, O red, and P orange and with successive P atoms in the

same strand joined by orange rods. [Based on an X-ray structure

by Kosuke Morikawa, Protein Engineering Research Institute,

Osaka, Japan. PDBid 1ECR.]

See Interactive Exercise 34

d. Replication Termination Is Facilitated

by Tus Protein

The E. coli replication terminus is a large (350 kb) re-

gion flanked by ten nearly identical nonpalindromic ⬃23-

bp terminator sites, TerH, TerI, TerE, TerD, and TerA on

one side and TerJ, TerG, TerF, TerB, and TerC on the other

(Fig. 30-36; note that oriC is directly opposite the terminus

region on the E. coli chromosome).A replication fork trav-

eling counterclockwise as drawn in Fig. 30-36 passes

through TerJ, TerG, TerF, TerB, and TerC but stops on en-

countering either TerA, TerD, TerE, TerI, or TerH (TerD,

TerE, TerI, and TerH are presumably backup sites for

TerA). Similarly, a clockwise-traveling replication fork

transits TerH, TerI, TerE, TerD, and TerA but halts at TerC

or, failing that, TerB or TerF or TerG or TerI. Thus, these

termination sites act as one-way valves that allow replica-

tion forks to enter the termination region but not to leave

it. This arrangement guarantees that the two replication

forks generated by bidirectional initiation at oriC will meet

in the replication terminus even if one of them arrives

there well ahead of its counterpart.

The arrest of replication fork motion at Ter sites requires

the action of Tus protein,a 309-residue monomer that is the

product of the tus gene (for terminator utilization sub-

stance).Tus specifically binds to a Ter site, where it prevents

strand displacement by DnaB helicase, thereby arresting

replication fork motion. The X-ray structure of Tus in com-

plex with a 15-bp Ter sequence-containing DNA with a sin-

gle T overhang at each 5¿ end, determined by Kosuke

Morikawa, reveals that Tus consists of two domains that

100/0

TerFTerG

oriC

TerB

TerE

TerH

TerI

TerD

TerA

TerC

TerJ

JWCL281_c30_1173-1259.qxd 10/19/10 10:29 AM Page 1199

1200 Chapter 30. DNA Replication, Repair, and Recombination

D. Fidelity of Replication

Since a single polypeptide as small as the Pol I Klenow

fragment can replicate DNA by itself, why does E. coli

maintain a battery of ⬎20 intricately coordinated proteins

to replicate its chromosome? The answer apparently is to

ensure the nearly perfect fidelity of DNA replication re-

quired to preserve the genetic message’s integrity from gen-

eration to generation.

The rates of reversion of mutant E. coli or T4 phage to

the wild type indicates that only one mispairing occurs per

to 10

base pairs replicated. This corresponds to ⬃1

error per 1000 bacteria per generation. Such high replica-

tion accuracy arises from four sources:

1. Cells maintain balanced levels of dNTPs through the

mechanism discussed in Section 28-3Ad. This is an impor-

tant aspect of replication fidelity because a dNTP present

terminus is deleted, replication simply stops, apparently

through the collision of opposing replication forks. Never-

theless, this termination system is highly conserved in

gram-negative bacteria.

As two oppositely moving replication forks collide at

the termination site, the newly synthesized strands become

covalently linked to yield two covalently closed double-

stranded chromosomes. However, since the parental DNA

strands remain wound about each other by several turns

(presumably, DNA gyrase cannot gain access to the DNA

when the colliding replication forks closely approach each

other), the product dsDNA strands must be wound about

each other by the same number of turns (Fig. 30-38). The

resulting catenated circular dsDNAs must be separated so

that each can be passed to a different daughter cell. This is

the job of the type II topoisomerase named topoisomerase

IV (Section 29-3Cd).

Termination region

Replication without

unwinding

Catenated chromosomes

DNA topoisomerase IV

Figure 30-38 The formation and separation of catenated

dsDNAs at the replication termination site. The parental strands

are red and black and the daughter strands are green and blue.

For clarity, the double helical character of the newly formed

dsDNA molecules is not shown.

JWCL281_c30_1173-1259.qxd 8/10/10 9:11 PM Page 1200