Voet D., Voet Ju.G. Biochemistry

Подождите немного. Документ загружается.

Dale Wigley determined the X-ray structure of E. coli

RecBCD in complex with a 51-nt DNA that could form a

hairpin loop containing an up to 21-bp dsDNA stem (Fig. 30-

75). The structure shows that RecB (1180 residues) and

RecC (1122 residues) are intimately intertwined with

RecB’s C-terminal nuclease domain connected to the rest

of the subunit by an extended 21-residue polypeptide

tether. A 15-bp segment of dsDNA enters the protein

through a tunnel between RecB and RecC. There it en-

counters a loop from RecC that appears to wedge the two

strands apart, with the 6-nt 3¿-ending single strand of the

DNA binding to RecB and the 10-nt 5¿-ending single

strand binding to RecD (608 residues; the 5-nt loop con-

necting the two strands of the dsDNA at the top of Fig. 30-

75 is disordered).The structure explains the different rates

of cleavage of the two DNA strands. The 3¿-ending strand

emerges from a tunnel through RecC in the vicinity of the

RecB nuclease domain, which is positioned to proces-

sively cleave it.The 5¿-ending strand competes with the 3¿-

ending strand for the nuclease site, but since the 5¿-ending

strand is less favorably located, it is cleaved less fre-

quently. However, after RecD has bound a Chi sequence,

the 3¿-ending strand is no longer available for cleavage,

which permits the nuclease to cleave the 5¿-ending strand

more frequently.

RecBCD can only commence unwinding DNA at a free

duplex end. Such ends are not normally present in E. coli,

which has a circular genome, but become available during

such recombinational processes as bacterial transforma-

tion, conjugation, and viral transduction, as well as at col-

lapsed replication forks.

e. RuvABC Mediates the Branch Migration

and the Resolution of the Holliday Junction

The branch migration of the RecA-generated Holliday

junction (Fig. 30-67e,f) requires the breaking and reform-

ing of base pairs as the bases exchange partners in passing

from one double helical stem to the other.Since ⌬G ⫽ 0 for

this process, it was initially assumed that it occurs sponta-

neously. However, such a process would move forward and

backward at random and, moreover, would be blocked by

as little as a single mismatched base pair. In E. coli, and

most other bacteria, branch migration is an ATP-driven

unidirectional process that is mediated by two proteins

whose synthesis is induced by the SOS response (Section

30-5D): RuvB (336 residues; Ruv for repair of UV dam-

age), an ATP-powered pump that drives branch migration

but binds only weakly to DNA; and RuvA (203 residues),

which binds to both a Holliday junction and to RuvB,

thereby targeting RuvB to the DNA.

The X-ray structure of Mycobacterium leprae (the cause

of leprosy) RuvA in complex with a synthetic and immo-

bile Holliday junction (Fig. 30-76a), determined by

Morikawa, reveals that RuvA forms a homotetramer to

which the Holliday junction binds in its open-X conforma-

tion (Fig. 30-76b). The RuvA tetramer, which has the ap-

pearance of a four-petaled flower (it has C

symmetry

rather than the D

symmetry of the vast majority of ho-

motetramers), is relatively flat (80 ⫻ 80 ⫻ 45 Å) with one

square face concave and the other convex. The concave

face (that facing the viewer in Fig. 30-76b), which is highly

positively charged and is studded with numerous con-

served residues, has four symmetry-related grooves that

bind the Holliday junction’s four arms.This face’s centrally

located projection or “pin” is formed by the side chains of

Glu 55 and Asp 56 from each subunit, and hence the repul-

sive forces between them and the Holliday junction’s an-

ionic phosphate groups probably facilitate the separation

of the single-stranded DNA segments and guide them from

one double helix to another.

RuvB is a member of the AAA⫹ family of ATPases

(Section 30-2Ca). The X-ray structure of Thermus ther-

mophilus RuvB crystallized in the presence of both ADP

and AMPPNP, determined by Morikawa, reveals two mole-

cules of RuvB with somewhat different conformations: one

binding ADP and the other binding AMPPNP. Each RuvB

molecule consists of three consecutive domains arranged in

a crescentlike configuration with the adenine nucleotides

binding at the interface between its N-terminal and middle

domains. EM studies indicate that, in the presence of ds-

DNA, RuvB oligomerizes to form a hexamer (Fig. 30-77a),

Section 30-6. Recombination and Mobile Genetic Elements 1231

Figure 30-75 X-ray structure of E. coli RecBCD in complex

with a 51-nt DNA capable of forming a 21-bp hairpin loop. The

protein is drawn in semitransparent ribbon form with RecB

yellow, RecC cyan, and RecD magenta. Note how the RecB

nuclease domain is linked to the rest of the subunit by an

extended polypeptide tether.The DNA is shown in space-filling

form with C gray, N blue, O red, and P orange.A loop from

RecC, which is drawn in space-filling form in green, is situated so

as to wedge apart the two strands of the incoming dsDNA with

the 3¿-ending strand binding to the 3¿S5¿ helicase of RecB and

the 5¿-ending strand passing through RecC to bind to the 5’ S 3¿

helicase of RecD. [Based on an X-ray structure by Dale Wigley,

The London Research Institute, Herts, U.K. PDBid 3K70.]

JWCL281_c30_1173-1259.qxd 8/10/10 9:12 PM Page 1231

as do most other AAA⫹ family members, including the D2

domain of NSF (Fig. 12-78). A hexameric model of RuvB

(Fig. 30-77b), constructed by superimposing the N-terminal

domain of the RuvB monomer on the ATPase domains of

the NSF D2 hexamer,agrees well with the EM-based image

and contains no serious steric clashes. This 130-Å-diameter

1232 Chapter 30. DNA Replication, Repair, and Recombination

Figure 30-76 X-ray structure of a RuvA tetramer in complex

with a Holliday junction. (a) A schematic drawing of the

synthetic and immobile Holliday junction in this structure

showing its base sequence. The two A ⴢ T base pairs that are

disrupted at the crossover (and which, if the Holliday junction

consisted of two homologous dsDNAs, as it normally does, would

exchange base pairing partners) are magenta. (b) The

RuvA–Holliday junction complex as viewed along the protein

tetramer’s 4-fold axis. The protein is represented as its molecular

surface with its subunits alternately colored pink and light blue.

The DNA is drawn in space-filling form colored according to

atom type with C atoms in different chains in different colors,

N blue, O red, and P orange. [Part a courtesy of and Part b based

on an X-ray structure by Kosuke Morikawa, Biomolecular

Engineering Research Institute, Osaka, Japan. PDBid 1C7Y.]

Figure 30-77 Proposed structure of the T. thermophilus RuvB

hexamer. (a) An EM-based image reconstruction of RuvB

complexed with a 30-bp DNA (not visible) as viewed along its

6-fold axis. The image resolution is 30 Å. (b) A model of the

RuvB hexamer that was constructed from the X-ray structure of

RuvB monomers by superimposing their N-terminal domains on

the homologous ATPase domains of the NSF D2 homohexamer

(Fig. 12-78). The N-terminal, middle, and C-terminal domains are

blue, yellow, and green, respectively, and its bound AMPPNP is

drawn in stick form in red. [Courtesy of Kosuke Morikawa,

Biomolecular Engineering Research Institute, Osaka, Japan.

PDBid 1HQC.]

(a)

(b)

(a)

JWCL281_c30_1173-1259.qxd 8/10/10 9:12 PM Page 1232

hexameric model contains a 30-Å-diameter hole through

which a single dsDNA can readily be threaded (see below).

Moreover the six ␤ hairpins, one per monomer, that have

been implicated in binding to RuvA are located on the top

face of the hexamer (as pictured in Fig. 30-77b).

The EM images of the RuvAB–Holliday junction com-

plex indicate that RuvA binds two oppositely located

RuvB hexamers. This has led to the model of their interac-

tion depicted in Fig. 30-78 in which RuvA binds the Holli-

day junction and helps load the RuvB hexameric rings onto

two opposing arms of the Holliday junction.The two hexa-

meric rings are postulated to counter-rotate,each in the an-

ticlockwise direction looking toward the center of the junc-

tion, so as to screw the horizontal DNA strands through

the center of the junction and into the top and bottom dou-

ble helices, thereby effecting branch migration (although

rather than actually rotating relative to RuvA, a RuvB

hexamer might pull the dsDNA through its central hole by

“walking” up its grooves in a manner resembling that pos-

tulated for hexagonal helicases; Section 30-2Ca).The direc-

tion of branch migration depends on which pair of arms the

RuvB hexamers are loaded.

The final stage of homologous recombination is the res-

olution of the Holliday junction into its two homologous

dsDNAs. This process is carried out by RuvC, a homo-

dimeric endonuclease of 173-residue subunits whose X-ray

structure indicates that its active sites are located ⬃30 Å

apart on the same face of the protein. This suggests that

RuvC sits down on the open face of the RuvAB–Holliday

junction complex, that facing the viewer in Fig. 30-78, to

cleave oppositely located strands at the Holliday junction.

The resulting single-strand nicks in the now resolved

dsDNAs are sealed by DNA ligase.

The X-ray structure of RuvC in complex with DNA has

not been determined, although model building studies sug-

gest that it binds Holliday junction DNA in its stacked-X

conformation. However, Dietrich Suck determined the X-

ray structure of bacteriophage T4 endonuclease VII in

complex with a Holliday junction in the stacked-X confor-

mation (Fig. 30-79). RuvC and the 157-residue T4 endonu-

clease VII exhibit no structural similarity but both are

homodimers of relatively small subunits that have similar

Section 30-6. Recombination and Mobile Genetic Elements 1233

Figure 30-78 Model of the RuvAB–Holliday junction

complex. The model is based on electron micrographs such as

that in the inset.The proteins are represented by their surface

diagrams with the RuvA tetramer, as seen in its X-ray structure,

green and the two oppositely oriented RuvB hexamers white. The

DNA of the Holliday junction is drawn in space-filling form with

its homologous blue and pink strands complementary to its light

blue and red strands.The complex is postulated to drive branch

migration via the ATP-driven counter-rotation of the RuvB

hexamers relative to the RuvA tetramer.This pumps (screws)

the horizontal dsDNAs through the RuvB hexamers to the

center of the Holliday junction, where their strands separate and

then base-pair with their homologs to form new dsDNAs, which

are pumped out vertically. [Courtesy of Peter Artymiuk,

University of Sheffield, U.K.]

Figure 30-79 X-ray structure of bacteriophage T4

endonuclease VII resolving a Holliday junction as viewed along

its pseudo-2-fold axis. The Holliday junction DNA is drawn in

ladder form with each of its four different 24-nt strands

differently colored.The protein, a homodimer of 157-residue

subunits, is shown in ribbon form embedded in its semitransparent

molecular surface with one subunit red and the other blue. The

arrows indicate the symmetrically located DNA cleavage sites.

Compare the DNA in this structure to that in Fig. 30-68b. [Based

on an X-ray structure by Dietrich Suck, European Molecular

Biology Laboratory, Heidelberg, Germany. PDBid 2QNC.]

JWCL281_c30_1173-1259.qxd 8/26/10 8:22 PM Page 1233

functions: the resolution of Holliday junctions into two du-

plex DNAs by introducing symmetrically placed nicks in

equivalent strands (Fig. 30-67j).

The forgoing model of the RuvABC resolvosome pro-

vides a satisfying mechanism for branch migration and Holl-

iday junction resolution. However,there is a fly in this partic-

ular ointment. The X-ray structure of an M. leprae

RuvA–Holliday junction complex crystallized under condi-

tions different from that in Fig. 30-76, determined by Lau-

rence Pearl, resembles the complex in Fig. 30-76b but with a

second RuvA tetramer in face-to-face contact with the con-

cave (DNA-binding) side of the first. Hence, the Holliday

junction is contained in two intersecting tunnels running

through the resulting RuvA octamer.Are both RuvA–Holli-

day junction structures biologically relevant, or is one an arti-

fact of crystallization? Pearl argues that the extensive comple-

mentary contacts between the two RuvA tetramers, which are

strongly conserved, are unlikely to be artifactual and that a

single RuvA tetramer is unlikely to withstand the torque ex-

erted by the two (in effect) counter-rotating RuvB hexamers.

However,if the RuvA octamer is biologically relevant, one of

its tetramers would at some point have to dissociate in order

to allow RuvC access to the Holliday junction.Yet, modeling

studies indicate that the RuvC dimer cannot properly contact

the RuvB tetramer-bound Holliday junction without it chang-

ing from its open-X to its stacked-X conformation.Further in-

vestigations are necessary to resolve these inconsistencies.

f. Recombination Repair Reconstitutes Damaged

Replication Forks

Transformation, transduction, and conjugation are such

rare events that the vast majority of bacterial cells never par-

ticipate in these processes. Similarly, the only place in the

metazoan life cycle at which gene shuffling through homolo-

gous recombination occurs is in meiosis (Section 1-4A).Why

then do nearly all cells have elaborate systems for mediating

homologous recombination? It is because damaged replica-

tion forks occur at a frequency of at least once per bacterial

cell generation and perhaps 10 times per eukaryotic cell cy-

cle. The DNA lesions that damage the replication forks can

be circumvented via homologous recombination in a process

named recombination repair [translesion synthesis, which is

highly mutagenic, is a process of last resort (Section 30-

5Db)]. Indeed, the rates of synthesis of RuvA and RuvB are

greatly enhanced by the SOS response.Thus, as Michael Cox

pointed out, the primary function of homologous recombina-

tion is to repair damaged replication forks. In what follows, we

describe recombination repair as it occurs in E. coli.

Recombination repair is called into play when a replica-

tion fork encounters an unrepaired single-strand lesion

(Fig. 30-80):

1. DNA replication is arrested at the lesion but contin-

ues on the opposing undamaged strand for some distance

before the replisome fully collapses (Section 30-5Db).

2. The replication fork regresses to form a type of Hol-

liday junction dubbed a “chicken foot.” This process may

occur spontaneously as driven by the positive supercoiling

that has built up ahead of the replication fork, it may be

mediated by RecA, or it may be promoted by RecG, an

ATP-driven helicase that catalyzes branch migration at

DNA junctions with three or four branches.

3. The single-strand gap at the collapsed replication

fork, now an overhang, is filled in by Pol I.

4. Reverse branch migration mediated by RuvAB or

RecG yields a reconstituted replication fork, which sup-

ports replication restart (see below).

Note that this process does not actually repair the single-

strand lesion that has caused the problem but instead re-

constructs the replication fork in a way that permits the

previously discussed DNA repair systems (Section 30-5) to

eventually eliminate the lesion.

1234 Chapter 30. DNA Replication, Repair, and Recombination

Figure 30-80 The recombination repair of a replication fork

that has encountered a single-strand lesion. Thick lines indicate

parental DNA, thin lines indicate newly synthesized DNA, the

cyan lines indicate DNA that was synthesized by Pol I, and

the arrows point in the 5¿S 3¿ direction. [After Cox, M.M.,

Annu. Rev. Genet. 35, 53 (2001).]

1 Replication fork

collapse

DNA lesion

Chicken foot

2 Fork regressionRecA

3 ReplicationPol I

4 Reverse branch

migration

RuvAB

JWCL281_c30_1173-1259.qxd 8/10/10 9:12 PM Page 1234

1 Rliti fk

DNA nick

2 DNA synthesis and

ligation

1 Formation of two

Holliday junctions

RAD51

3 Holliday junction

resolution

A second situation that requires recombination repair is

the encounter of a replication fork with an unrepaired

single-strand nick (Fig. 30-81):

1. When a single-strand nick is encountered, the repli-

cation fork collapses.

2. The repair process begins via the RecBCD plus

RecA–mediated invasion of the newly synthesized and un-

damaged 3¿-ending strand into the homologous dsDNA

starting at its broken end.

3. Branch migration, as mediated by RuvAB, then

yields a Holliday junction, which exchanges the replication

fork’s 3¿-ending strands.

4. RuvC then resolves the Holliday junction yielding a

reconstituted replication fork ready for replication restart.

Thus, the 5¿-ending strand of the nick has, in effect, become

the 5¿ end of an Okazaki fragment.

The final step in the recombination repair process is the

restart of DNA replication. This process is, of necessity, dis-

tinct from the replication initiation that occurs at oriC (Sec-

tion 30-3Ca).Origin-independent replication restart is medi-

ated by the same seven-protein primosome that initiates the

minus strand replication of bacteriophage ␾X174 (Table 30-

4), which has therefore been named the restart primosome.

g. Recombination Repair Reconstitutes

Double-Strand Breaks

We have seen that double-strand breaks (DSBs) in

DNA can be rejoined, often mutagenically, by nonhomolo-

gous end-joining (NHEJ; Section 30-5E). DSBs may also

be nonmutagenically repaired through a recombination re-

pair process known as homologous end-joining, which oc-

curs via two Holliday junctions (Fig. 30-82):

1. The DSB’s double-stranded ends are resected to pro-

duce single-stranded ends. One of the 3¿-ending strands

invades the corresponding sequence of a homologous

Section 30-6. Recombination and Mobile Genetic Elements 1235

Figure 30-81 The recombination repair of a replication fork

that has encountered a single-strand nick. Thick lines indicate

parental DNA, thin lines indicate newly synthesized DNA, and

the arrows point in the 5¿S3¿ direction. [After Cox, M.M.,

Annu. Rev. Genet. 35, 53 (2001).]

Figure 30-82 The repair of a double-strand break in DNA by

homologous end-joining. Thick lines indicate parental DNA, thin

lines indicate newly synthesized DNA, and the arrows point in

the 5¿S3¿ direction. [After Haber, J.E., Trends Genet. 16, 259

(2000).]

1 Replication fork

collapse

2 Strand invasionRecBCD + RecA

3 Branch migrationRuvAB

4 Holliday junction

resolution

RuvC

JWCL281_c30_1173-1259.qxd 8/10/10 9:12 PM Page 1235

chromosome to form a Holliday junction, a process that, in

eukaryotes, is mediated by the RecA homolog RAD51.

The other 3¿-ending strand pairs with the displaced strand

segment on the homologous chromosome to form a second

Holliday junction.

2. DNA synthesis and ligation fills in the gaps and seals

the joints.

3. Both Holliday junctions are resolved to yield two in-

tact double strands.

Thus, the sequences that may have been expunged in the

formation of the DSB are copied from the homologous

chromosome. Of course, a limitation of homologous end-

joining, particularly in haploid cells, is that a homologous

chromosomal segment may not be available.

The importance of recombination repair in humans is

demonstrated by the observation that defects in the pro-

teins BRCA1 (1863 residues) and BRCA2 (3418

residues), both of which interact with RAD51, are associ-

ated with a greatly increased incidence of breast, ovarian,

prostate, and pancreatic cancers. Indeed, individuals with

mutant BRCA1 or BRCA2 genes have up to an 80% life-

time risk of developing cancer. Recombination can also

function to elongate shortened telomeres without the

need for telomerase.

B. Transposition and Site-Specific Recombination

In the early 1950s, on the basis of genetic analysis, Bar-

bara McClintock reported that the variegated pigmenta-

tion pattern of maize (Indian corn) kernels results from

the action of genetic elements that can move about the

maize genome. This proposal was resoundingly ignored

because it was contrary to the then held genetic ortho-

doxy that chromosomes consist of genes linked in fixed

order. Another 20 years were to pass before evidence of

mobile genetic elements was found in another organism,

E. coli.

It is now known that transposable elements or trans-

posons are common in both prokaryotes and eukaryotes,

where they influence the variation of phenotypic expres-

sion over the short term and evolutionary development

over the long term. Each transposon codes for the enzymes

that specifically insert it into the recipient DNA. This

process has been described as illegitimate recombination

because it requires no homology between donor and

recipient DNAs. Since the insertion site is chosen largely

at random, transposition is a potentially dangerous

process; the insertion of a transposon into an essential

gene will kill a cell together with its resident transposons.

Hence transposition is tightly regulated; it occurs at a rate

of only 10

⫺5

to 10

⫺7

events per element per generation.

The conditions that trigger transposition are, for the most

part, unknown.

a. Prokaryotic Transposons

Prokaryotic transposons with three levels of complexity

have been characterized:

1. The simplest transposons, and the first to be char-

acterized, are named insertion sequences or IS elements.

They are designated by “IS” followed by an identifying

number. IS elements are normal constituents of bacterial

chromosomes and plasmids. For example, a common E.

coli strain has eight copies of IS1 and five copies of IS2.

IS elements generally consist of ⬍2000 bp. These com-

prise a so-called transposase gene, and in some cases a

regulatory gene, flanked by short inverted (having oppo-

site orientation) terminal repeats (Fig. 30-83 and Table

30-6). The inverted repeats are essential for transposi-

tion; their genetic alteration invariably prevents this

process. An inserted IS element is flanked by a directly

(having the same orientation) repeated segment of host

DNA (Fig. 30-83). This suggests that an IS element is in-

serted in the host DNA at a staggered cut that is later

filled in (Fig. 30-84). The length of this target sequence

(most commonly 5 to 9 bp), but not its sequence, is char-

acteristic of the IS element.

2. More complex transposons carry genes not involved

in the transposition process, for example, antibiotic resist-

ance genes. Such transposons are designated “Tn” followed

1236 Chapter 30. DNA Replication, Repair, and Recombination

Figure 30-83 Structure of IS elements. These and other

transposons have inverted terminal repeats (numerals) and are

flanked by direct repeats of host DNA target sequences (letters).

•

A′

B′

C′

D′

A′

B′

C′

D′

1′

2′

3′

4′

5′

4′

3′

2′

1′

IS element

Target

sequence

Target

sequence

Inverted Direct Number of

Insertion Length Terminal Repeat at Copies in E. coli

Element (bp) Repeat (bp) Target (bp) Chromosome

IS1 768 23 9 5–8

IS2 1327 41 5 5

IS4 1428 18 11–13 5

IS5 1195 16 4 1–2

Table 30-6 Properties of Some Insertion Elements

Source: Mainly Lewin, B., Genes IX, p. 524, Oxford University Press (2008).

JWCL281_c30_1173-1259.qxd 8/10/10 9:12 PM Page 1236

by an identifying number. For example, Tn3 (Fig. 30-85)

consists of 4957 bp and has inverted terminal repeats of

38 bp each. The central region of Tn3 codes for three pro-

teins: (1) a 1015-residue transposase named TnpA; (2) a

185-residue protein known as TnpR, which mediates the

site-specific recombination reaction necessary to complete

the transposition process (see below) and also functions as

a repressor for the expression of both tnpA and tnpR; and

(3) a ␤-lactamase that inactivates ampicillin (Section 11-3Bb).

The site-specific recombination occurs in an AT-rich region

known as the internal resolution site that is located be-

tween tnpA and tnpR.

3. The so-called composite transposons (Fig. 30-86)

consist of a gene-containing central region flanked by two

identical or nearly identical IS-like modules that have either

the same or an inverted relative orientation. It therefore

seems that composite transposons arose by the association

of two originally independent IS elements. Since the IS-like

modules are themselves flanked by inverted repeats, the

ends of either type of composite transposon must also

be inverted repeats. Experiments demonstrate that com-

posite transposons can transpose any sequence of DNA in

their central region.

There are two modes of transposition: (1) direct or simple

transposition, in which the transposon, as the name im-

plies, physically moves from one DNA site to another;

and (2) replicative transposition, in which the transposon

remains at its original site and a copy of it is inserted at a

target site. The two modes, as we shall see, have similar

mechanistic features and, indeed, some transposons can

move by either mode.

b. Direct Transposition of Tn5 Occurs by a

Cut-and-Paste Mechanism

Tn5 is a 5.8-kb composite transposon that contains the

gene encoding the 476-residue Tn5 transposase together

with three antibiotic resistance genes. It is flanked by in-

verted IS-like modules ending in 19-bp sequences called

outside end (OE) sequences. Tn5 undergoes direct trans-

position via a “cut-and-paste” mechanism that was eluci-

dated in large part by William Reznikoff (Fig. 30-87):

1. Each of Tn5’s two OE sequences on the donor DNA

is bound by a monomer of Tn5 transposase.

2. The transposase dimerizes to form a catalytically ac-

tive synaptic complex in which the transposon is held be-

tween the two transposase subunits.

3. Each transposase subunit activates a water molecule

to nucleophilically attack the outermost nucleotide of its

bound OE sequence, yielding a free 3¿-OH group. This 3¿-

OH group is then activated to attack the opposite strand

on the DNA to form a hairpin structure, thereby excising

the transposon from the DNA. The hairpin is then hy-

drolyzed to yield a blunt-ended dsDNA at each end of the

transposon, thus completing the “cut” portion of the trans-

position mechanism.

4. The synaptic complex binds to the target DNA.

Section 30-6. Recombination and Mobile Genetic Elements 1237

Figure 30-84 A model for the generation of direct repeats of

the target sequence by transposon insertion.

Figure 30-86 A composite transposon. This element consists

of two identical or nearly identical IS-like modules (green)

flanking a central region carrying various genes.The IS-like

modules may have either (a) direct or (b) inverted relative

orientations.

TATTA

ATAAT

TATTA

ATAA

Transposon

Inverted

repeat

Filling in and ligation

Staggered cut

within host DNA

IR IR

Inverted

repeat

Internal

resolution site

Transposase β-Lactamase

tnpA amptnpR

Inverted

repeat

Repressor and

resolvase

Central region

IS-like

module

(a)

IS-like

module

Central region

Inverted

repeats

Inverted

repeats

(b)

Figure 30-85 A map of transposon Tn3.

JWCL281_c30_1173-1259.qxd 8/10/10 9:12 PM Page 1237

5. The transposon’s 3¿-OH groups nucleophilically at-

tack the target DNA on opposite strands spaced 9 bp apart,

thereby installing the transposon at the target site. Re-

markably, this reaction and the three preceding lytic reac-

tions are all mediated by the same catalytic site. The repair

of the oppositely located single-strand gaps (Fig. 30-84)

completes the “paste” portion of the mechanism.

Although, strictly speaking, not part of the transposition

process, the double-strand break in the donor DNA left by

the excision of the transposon must be repaired if the donor

DNA is to be propagated (in bacteria, the donor DNA is of-

ten a plasmid so that its loss has little effect on the cell since

plasmids are generally present in multiple copies).

The X-ray structure of a Tn5 synaptic complex (Fig. 30-88),

determined by Reznikoff and Ivan Rayment, provides a

model of the synaptic complex at the stage following its

cleavage from the donor DNA (the product of Step 3 in

Fig. 30-87). This 2-fold symmetric complex consists of a

dimer of Tn5 transposase subunits binding two 20-bp DNA

segments containing the Tn5 transposon’s 19-bp OE se-

quence with the outer end of each OE sequence bound to

the protein (and whose opposite ends would, in vivo, be

connected by the looped around transposon; Fig. 30-87).

Both transposase subunits extensively participate in bind-

ing each DNA segment, thereby explaining why the indi-

vidual subunits cannot cleave their bound DNA segments

before forming the synaptic complex.The protein holds the

DNA in a distorted B-DNA conformation with its two end

pairs of nucleotides no longer base paired. Indeed, the

penultimate base on the nontransferred strand is flipped

out of the double helix and binds in a hydrophobic pocket.

The transferred strand’s free 3¿-OH group, which occupies

the active site, is bound in the vicinity of a cluster of three

catalytically essential acidic residues, the so-called DDE

motif, which is shared with other transposases. In the X-ray

structure the DDE motif binds two Mn

2⫹

ions, although

physiologically it probably binds two Mg

2⫹

ions. This sug-

gests that transposases employ a metal-activated catalytic

1238 Chapter 30. DNA Replication, Repair, and Recombination

Figure 30-87 The cut-and-paste transposition mechanism

catalyzed by Tn5 transposase. The reactions comprising Steps 3

and 5 are indicated beside the braces to the right of these steps.

[After Davies, D.R., Goryshin, I.Y., Reznikoff, W.S., and

Rayment, I., Science 289, 77 (2000).]

Tn5 transposon DNA

Cleavage

donor DNA

Target capture

Strand transfer

Transposase binding

Target DNA

Synaptic complex

Donor DNADonor DNA

Dimerization

3′

5′

3′

5′

3′

5′

3′

5′

Tn5 Donor

Tn5

H–O–H

O–H

Tn5

3′

5′

3′

5′

Target

3′

5′

3′

Tn5

O–H

H–O

Figure 30-88 X-ray structure of Tn5 transposase in complex

with a 20-bp DNA containing the OE sequence. The complex,

which represents the product of Step 3 in Fig. 30-87, is viewed

along its 2-fold axis with one of its two identical subunits colored

in rainbow order from N-terminus (blue) to C-terminus (red)

and the other subunit pink.The DNA is drawn in stick form with

C green, N blue, O red, and P orange and with successive P

atoms on the same polynucleotide connected by orange rods.The

bound Mn

2⫹

ions, which mark the enzyme’s active site, are

represented by purple spheres.The DNAs’ reactive 3¿-OH

groups are located at these active sites. [Based on an X-ray

structure by William Reznikoff and Ivan Rayment, University of

Wisconsin. PDBid 1MUS.]

JWCL281_c30_1173-1259.qxd 8/10/10 9:12 PM Page 1238

mechanism similar to that of the DNA polymerases (Sec-

tion 30-2Af).The facing surface of the protein in Fig. 30-88

is positively charged with a prominent groove running

from upper left to lower right that forms the apparent

binding site for the target DNA.

Wild-type Tn5 transposase has such low catalytic activ-

ity that it is undetectable in vitro. However, that in the

X-ray structure is a hyperactive mutant form that contains

the mutations E54K and L372P (an unusual circumstance

in that it is far more common to mutationally inhibit an en-

zyme under crystallographic study so as to trap it at some

specific stage along its reaction pathway). Lys 54 is hydro-

gen bonded to O4 of a thymine base on the transferred

strand. In the wild-type transposase, Glu 54 would proba-

bly have an unfavorable charge–charge repulsion with a

nearby phosphate group, thus providing a structural basis

for the increased activity of the E54K mutant. The L372P

mutation disorders the peptide segment between residues

373 and 391 (it is ordered in the X-ray structure of wild-

type Tn5 transposase lacking its N-terminal 55 residues),

thereby suggesting that this mutation facilitates a confor-

mational change required for substrate binding.

c. Replicative Transposition Occurs via Cointegrates

If a plasmid carrying a transposon resembling Tn3 is in-

troduced into a bacterial cell carrying a plasmid that lacks

the transposon, in some of the progeny cells both types of

plasmid will contain the transposon (Fig. 30-89). Evidently,

such transposition involves the replication of the transposon

into the recipient plasmid rather than its transfer from donor

to recipient.

Two plasmids, one containing a replicative transposon,

will occasionally fuse to form a so-called cointegrate con-

taining like-oriented copies of the transposon at both junc-

tions of the original plasmids (Fig. 30-90). Yet, some of the

progeny of a cointegrate-containing cell lack the cointegrate

and instead contain both original plasmids, each with one

copy of the transposon (Fig. 30-89). The cointegrate must

therefore be an intermediate in the transposition process.

Although the mechanism of replicative transposition has

not been fully elucidated, a plausible model for this process

(and there are several) that accounts for the foregoing ob-

servations consists of the following steps (Fig. 30-91):

1. A pair of staggered single-strand cuts, such as is dia-

grammed in Fig. 30-84, is made by the transposon-encoded

transposase at the target sequence of the recipient plasmid

so as to liberate 3¿-OH ends. Similarly, single-strand cuts

are made on opposite strands to either side of the transpo-

son. Note that these reactions resemble those catalyzed by

Tn5 transposase (Fig. 30-87).

2. Each of the transposon’s free ends is ligated to a pro-

truding single strand at the insertion site.This forms a repli-

cation fork at each end of the transposon.

3. The transposon is replicated, thereby yielding a

cointegrate.

4. Through a site-specific recombination between the

internal resolution sites of the two transposons, the cointe-

grate is resolved into the two original plasmids, each of

which contains a transposon. This crossover process is cat-

alyzed by transposon-encoded recombinases (TnpR in

Tn3; also known as resolvases) rather than RecA; transpo-

sition proceeds normally in recA

⫺

cells (although RecA

will resolve a cointegrate containing a transposon with a

mutant resolvase and/or an altered internal resolution site,

albeit at a much reduced rate).

Site-specific recombinases fall into only two protein

families, serine recombinases and tyrosine recombinases,

which are named after the amino acid residue that forms a

transient covalent linkage to the DNA during the recombi-

nase reaction.As we shall, these two types of recombinases

function via different mechanisms.

d. ␥␦ Resolvase Catalyzes Site-Specific

Recombination

The ␥␦ resolvase, a serine recombinase that forms a ho-

modimer in solution, is a TnpR homolog that is encoded

by the ␥␦ transposon (a member of the Tn3 family of

replicative transposons; Fig. 30-85). It catalyzes a site-

specific recombination event in which a cointegrate con-

taining two copies of the ␥␦ transposon is resolved, via

double-strand DNA cleavage, strand exchange, and

Section 30-6. Recombination and Mobile Genetic Elements 1239

Figure 30-89 Replicative transposition. This type of

transposition inserts a copy of the transposon at the target site

while another copy remains at the donor site.

Figure 30-90 A cointegrate. This structure forms by the fusion

of two plasmids, one carrying a transposon, such that both

junctions of the original plasmid are spanned by transposons

with the same orientation (arrows).

many

generations

Transposon

Bacterium

Transposon

Cointegrate

JWCL281_c30_1173-1259.qxd 8/10/10 9:12 PM Page 1239

religation (the last step in Fig. 30-91), into two catenated

(linked) dsDNA circles that each contain one copy of the

␥␦ transposon (it also serves as its own transcriptional re-

pressor as does TnpR).The ␥␦ transposon contains a 114-bp

res site that includes three binding sites for ␥␦ resolvase

dimers, each of which contains an inverted repeat of the ␥␦

resolvase’s 12-bp recognition sequence. The resolution of

the cointegrate involves the binding of a ␥␦ resolvase ho-

modimer to all six of these binding sites in the cointegrate

(three from each of its two transposons) as is diagrammed

in Fig. 30-92. The reaction proceeds via the formation of a

transient phosphoSer bond between Ser 10 and the 5¿-

phosphate at each cleavage site.

The X-ray structure of the ␥␦ resolvase homotetramer

in complex with two 34-bp palindromic dsDNA segments

containing an inverted repeat of the 12-bp recognition se-

quence separated by an 4-bp spacer (Fig. 30-93), deter-

mined by Nigel Grindley and Steitz, reveals that this synap-

tic tetramer has D

symmetry. Each 183-residue resolvase

monomer consists of an N-terminal catalytic domain

(residues 1–120) and a C-terminal DNA-binding domain

(residues 148–183) connected by an extended arm

(residues 121–147). Both dsDNAs, which are located at the

periphery of the protein core, have been cleaved with each

of the resulting four 5¿ ends in phosphoSer linkage to the

resolvase. This preserves the free energy of the cleaved

phosphodiester bond so that it can later be reformed with

a different partner, much as occurs with topoisomerases

(Section 29-3C).

Each centrally located catalytic domain approaches its

bound DNA from its minor groove side with its C-terminal

helix (helix E) binding over the minor groove (the segment

of the E helix that contacts the DNA is disordered in the

absence of the DNA). Each C-terminal domain binds over

the major groove of its recognition sequence on the oppo-

site side of the DNA from its attached catalytic domain

with the extended arm that connects them running more or

less along the DNA’s minor groove.The two C-terminal do-

mains of the resolvase subunits labeled L and R (and the

symmetry-related L¿ and R¿ subunits) in Fig. 30-93 are

thereby separated by two helical turns along the cleaved

DNA, the segments of which closely assume the B-DNA

conformation. Each C-terminal helix binds in the DNA’s

major groove and, together with its preceding helix, forms

a helix–turn–helix (HTH) motif, a common sequence-spe-

cific DNA-binding motif that occurs mainly in prokaryotic

transcriptional repressors and activators (Section 31-3Da).

The structure of the L–R dimer closely resembles that in

the X-ray structure of the dimer bound to uncleaved site I

DNA (the presynaptic dimer). This, and the short (17 Å)

distance between the free 3¿-OH group in the L subunit

and the phosphoSer bond in the R subunit compared to

other such distances in the complex (L–L¿ and L–R¿), indi-

cates that the L–R and L¿–R¿ dimers correspond to the ini-

tial site I–bound dimers soon after cleavage or just before

religation (Fig. 30-92). Consequently, the interface between

the L–R and L¿–R¿ dimers must be the newly formed

synaptic interface.

How are the DNA strands in the synaptic complex

exchanged, that is, how is the DNA bound to the L sub-

unit ligated to the DNA bound to either the L¿ subunit or

the R¿ subunit (and R to either R¿ or L¿)? In either case,

the free 3¿-OH group on each subunit is ⬃50 Å from the

1240 Chapter 30. DNA Replication, Repair, and Recombination

Transposon

Cleavage

sites

Donor plasmid

Target sequence

Recipient plasmid

Site-specific cleavage

and ligation

5′

3′

Replication forks

Replication

Recombination sites

Resolution

Donor plasmid

restored

Recipient plasmid

containing

transposon

and duplicated

target sequence

Cointegrate

Figure 30-91 A model for transposition involving the

intermediacy of a cointegrate. Here more lightly shaded bars

represent newly synthesized DNA. [After Shapiro, J.A., Proc.

Natl.Acad. Sci. 76, 1934 (1979).]

JWCL281_c30_1173-1259.qxd 8/10/10 9:12 PM Page 1240