Voet D., Voet Ju.G. Biochemistry

Подождите немного. Документ загружается.

phosphoSer group that it would have to nucleophilically

attack in the religation step and the space between them

is filled with protein. Clearly, the synaptic complex must

undergo a dramatic structural change to accomplish the

religation step.

Section 30-6. Recombination and Mobile Genetic Elements 1241

Figure 30-92 A model for the resolution of a cointegrate

containing two ␥␦ transposons to form two catenated dsDNA

circles. (1) The ␥␦ resolvase binds as six homodimers to its

binding sites, I, II, and III, in each of the cointegrate’s two res

sites (yellow and green circles represent the ␥␦ resolvase

monomers initially bound to the red and blue res sites,

respectively), which then associate to form synaptic tetramers.

Although not shown as such, the synaptic tetramers bound to

sites I associate with the synaptic tetramers bound to sites II and

III to form, as seen in the electron microscope, a compact globule

of unknown structure known as a synaptosome. (2) The dsDNA

at sites I both undergo staggered (by 2 bp) double-strand

scissions via the transient formation of phosphoSer bonds

between Ser 10 and the 5¿-phosphates at the cleavage sites.The

cleaved strands then exchange places (cross over) in a process

that apparently requires the rotation of one of the pairs of

resolvase monomers with respect to the other and are then

ligated. (3) The dissociation of the synaptosome yields the

catenated dsDNA circles. [Courtesy of Gregory Mullen,

University of Connecticut Health Center.]

Res

Cointegrate

III

Figure 30-93 X-ray structure of a ␥␦ resolvase synaptic

tetramer in complex with two 34-bp palindromic, site

I–containing dsDNAs. The D

-symmetric complex is viewed with

one of its 2-fold axes horizontal (the other 2-fold axes lie in the

vertical plane labeled “Flat interface”).The DNAs (dark and

light green and yellow and orange) have been cleaved into half-

sites labeled L, R, L¿, and R¿ through the nucleophilic attack of

the Ser 10 side chain (drawn in space-filling form with C blue

and O red) on the 5¿-phosphate group of A20 (drawn in stick

form in green).The subunits of the resolvase tetramer, whose

helices are drawn in tube form, are colored blue, green, red, and

pink.The intact DNAs initially bind to the L–R and L¿–R¿

dimers, the so-called site I dimers. The L–L¿ dimer and the

symmetrically equivalent R–R¿ dimer form so-called antiparallel

dimers whose D and E helices associate as four-helix bundles. In

addition, a Val 114 S Cys mutation (V114C) disulfide-cross-links

the E helices of the L–R and L¿–R¿ dimers across the so-called

flat interface. [Courtesy of Thomas Steitz, Yale University. PDBid

1ZR4.]

JWCL281_c30_1173-1259.qxd 8/10/10 9:12 PM Page 1241

The observation that the interface between the L–L¿ and

R–R¿ dimers is largely hydrophobic and unusually flat (Fig.

30-93) strongly suggests that the religation occurs after a

180° rotation of these dimers with respect to one another

(about the horizontal 2-fold axis in Fig. 30-93), thus exchang-

ing the positions of the R and R¿ subunits with respect to the

L and L¿ subunits. The rotation is presumably driven by the

superhelical tension in the naturally negatively supercoiled

cointegrate. This model is supported by energy calculations

and the observation that mutating Val 114 of the E helix to

Cys, which, under oxidizing conditions, disulfide-links the E

helices on the L and R (and L¿ and R¿) subunits (a mutation

that was present in the foregoing structure), yields a com-

plex that can form the covalent intermediate in Fig. 30-93

but cannot carry out the religation step—presumably be-

cause the disulfide bonds prevent the above rotation. How-

ever, reducing these disulfide bonds restores the complex’s

recombinational activity. Furthermore, mutation of Lys 136,

an E helix residue located at the flat interface, to Cys pre-

vents religation of the cleaved DNA when subjected to oxi-

dizing conditions, even though it requires a 75° rotation of

the L–L¿/R–R¿ interface in the structure in Fig. 30-93 to bring

the Cys 136 side chains on opposing subunits close enough

to form a disulfide bond.Of course, a detailed understanding

of the mechanism of the ␥␦ resolvase reaction will require

the knowledge of how all six ␥␦ resolvase dimers that form

the synaptosome participate in the reaction (Fig. 30-92).

e. Replicative Transposons Are Responsible for

Much Genetic Remodeling in Prokaryotes

In addition to mediating their own insertion into DNA,

replicative transposons promote inversions, deletions, and

rearrangements of the host DNA. Inversions can occur

when the host DNA contains two copies of a transposon in

inverted orientation. The recombination of these trans-

posons inverts the region between them (Fig. 30-94a). If, in-

stead, the two transposons have the same orientation, the

resolution of this cointegrate-like structure deletes the seg-

ment between the two transposons (Fig. 30-94b; if the

deleted segment lacks a replication origin, it will not be

propagated). The deletion of a chromosomal segment in

this manner, followed by its integration into the chromo-

some at a different site by a separate recombinational

event, results in chromosomal rearrangement.

Transposition appears to be important in chromosomal

and plasmid evolution. Indeed, it has been suggested that

transposons are nature’s genetic engineering “tools.” For

example, the rapid evolution, since antibiotics came into

common use, of plasmids that confer resistance to several

antibiotics (Section 5-5Ba) has resulted from the accumu-

lation of the corresponding antibiotic-resistance trans-

posons in these plasmids. Transposon-mediated rearrange-

ments may well have been responsible for organizing

originally distant genes into coordinately regulated oper-

ons (Section 5-4Aa) as well as for forming new proteins by

linking two formerly independent gene segments. More-

over, the occurrence of identical transposons in unrelated

bacteria indicates that the transposon-mediated transfer of

genetic information between organisms is not limited to re-

lated species, in contrast to genetic transfers mediated by ho-

mologous recombination.

f. Phase Variation Is Mediated by Site-Specific

Recombination

Phenotypic expression in bacteria can be regulated by

site-specific recombination. For example, certain strains of

1242 Chapter 30. DNA Replication, Repair, and Recombination

Figure 30-94 Chromosomal rearrangement via recombination.

(a) The inversion of a DNA segment between two identical

transposons with inverted orientations. (b) The deletion of a

Transposons with inverted orientations

(a) (b)

Pairing of inverted repeats

Recombination

Inverted segment

Transposons with the same orientation

Pairing of direct repeats

Recom-

bination

Chromosome

containing one

transposon

Deleted segment

containing

one transposon

DNA segment between two identical transposons with the same

orientation.This process parcels one transposon each to the

resulting two DNA segments.

JWCL281_c30_1173-1259.qxd 8/10/10 9:12 PM Page 1242

Salmonella typhimurium make two antigenically distinct

versions of the protein flagellin (the major component of

the whiplike flagella with which bacteria propel them-

selves; Section 35-3I) that are designated H1 and H2. Only

one of these proteins is expressed by any particular cell but

about once every 1000 cell divisions, in a process known as

phase variation, a cell switches the type of flagellin it syn-

thesizes. It is thought that phase variation helps Salmonella

evade its host’s immunological defenses.

What is the mechanism of phase variation? The two fla-

gellin genes reside on different parts of the bacterial chro-

mosome. H2 is linked to the rh1 gene that encodes a repres-

sor of H1 expression (Fig. 30-95; rh1, H2, and H1 are also

known as fljA, fljB, and fljC, respectively). Hence, when the

H2–rh1 transcription unit is expressed, H1 synthesis is

repressed; otherwise H1 is synthesized. Melvin Simon has

shown that the expression of the H2–rh1 unit is controlled

by the orientation of a 995-bp segment that lies upstream of

H2 (Fig. 30-95) and that contains the following elements:

1. A promoter for H2–rh1 expression.

2. The hin gene, which encodes the 190-residue Hin

DNA invertase. Hin,a serine recombinase, mediates the in-

version of the DNA segment in a manner similar to that di-

agrammed in Fig. 30-94a. In fact, Hin is ⬃40% identical in

sequence with the ␥␦ resolvase, which strongly suggests

that these proteins have similar structures.

3. Two closely related 26-bp sites, hixL and hixR, that

form the boundaries of the segment and hence contain its

cleavage sites. They each consist of two imperfect 12-bp in-

verted repeats separated by 2 nt.

In the Phase 2 orientation (Fig. 30-95a), the properly ori-

ented promoter is just upstream of H2, so this gene and rh1

are coordinately expressed, thereby repressing H1 synthe-

sis. In Phase 1 bacteria (Fig. 30-95b), however, this segment

has the opposite orientation. Consequently, neither H2 nor

rh1, which then lacks a promoter, is expressed so that H1 is

synthesized.

g. Cre-Mediated Site-Specific Recombination

Occurs via 3¿-PhosphoTyr Intermediates

Bacteriophages, as we have seen (Fig. 1-31), replicate

themselves within their host bacterial cells which, in most

cases, they then lyse to release the progeny phage, a

lifestyle that is therefore known as the lytic mode. How-

ever, certain bacteriophages can assume an alternative,

nondestructive lifestyle, the lysogenic mode, in which they

install their DNA, usually in the host chromosome via site-

specific recombination, so that the phage DNA is passively

replicated with the host DNA. However, if the bacterial

host encounters conditions in which it is unlikely to sur-

vive, the phage DNA is excised from the bacterial chromo-

some via a reversal of the site-specific recombination reac-

tion and it reenters the lytic mode so as to escape the

doomed host. We discuss the genetic factors that maintain

the balance between the lytic and lysogenic lifestyles in

bacteriophage ␭ in Section 33-3.

The enzymes that mediate the foregoing site-specific

recombination reactions are members of the ␭ integrase

(␭ Int; alternatively, tyrosine recombinase) family, whose

⬃1000 known members also occur in prokaryotes and eu-

karyotes. These include the XerC and XerD proteins of E.

coli which, operating in concert, function to decatenate the

two linked circular dsDNA products of homologous re-

combination (Fig. 30-69g, left), as well as type IB topoiso-

merases (Section 29-3Cc).

The structurally best characterized member of the ␭ in-

tegrase family is the Cre recombinase of E. coli bacterio-

phage P1. In its lysogenic state, bacteriophage P1 is a sin-

gle-copy circular plasmid (rather than being inserted in the

host chromosome as is bacteriophage ␭), but in the phage

head (the lytic mode), P1 DNA is a linear dsDNA that has

a 34-bp loxP site at each end. The main function of Cre,

Section 30-6. Recombination and Mobile Genetic Elements 1243

Figure 30-95 The mechanism of phase variation in

Salmonella. (a) In Phase 2 bacteria, the H2–rh1 promoter is

oriented so that H2 flagellin and repressor are synthesized.

Repressor binds to the H1 gene, thereby preventing its expression.

(b) In Phase 1 bacteria, the segment preceding the H2–rh1

transcription unit has been inverted relative to its orientation in

Phase 2 bacteria. Hence this transcription unit cannot be

expressed because it lacks a promoter.This releases H1 from

repression and results in the synthesis of H1 flagellin.The

inversion of the segment preceding the H2–rh1 transcription unit

is mediated by the Hin protein, which is expressed in either

orientation by the hin gene.

hin

Phase 2(a)

rh1H2

hixR

Hin protein

Repressor

H2–rh1

promoter

Inactive

Phase 1(b)

hin rh1H2

Inactive

H1 flagellin

mRNA

hixL

H2 flagellin

JWCL281_c30_1173-1259.qxd 8/10/10 9:12 PM Page 1243

which is encoded by bacteriophage P1, is to mediate the

site-specific recombination between these two loxP sites so

as to circularize the linear DNA (Fig. 30-96).

The loxP site is palindromic except for its central 8-bp

crossover region, which confers directionality on the site.

In carrying out the recombination reaction, the 343-residue

Cre subunits form a homotetramer that binds two loxP sites

in an antiparallel orientation, with each Cre subunit binding

half of a loxP site. Then, as is diagrammed in Fig. 30-97,

oppositely located Cre subunits catalyze single-strand scis-

sions on the 5¿ side of the crossover region on one strand of

each of the two dsDNAs. This occurs through the nucle-

ophilic attack of each of these active Cre subunit’s con-

served Tyr 324 residues on the DNA’s scissile phosphoester

bond to yield a 3¿-phosphoTyr intermediate on one side of

the cleaved bond and a free 5¿-OH group on the other side

(as similarly occurs in the reactions catalyzed by type IB

topoisomerases; Section 29-3Cc). Each of the liberated 5¿-

OH groups then nucleophilically attacks the 3¿-phospho-

Tyr group on the opposite duplex to form a Holliday junc-

tion, thereby releasing the Tyr residues. The Holliday

junction is resolved into two recombined dsDNAs when

the two Cre subunits that had not yet participated in the re-

action mediate the same cleavage and strand exchange re-

actions on the two heretofore unreacted single strands.This

latter process must be preceded by a structural rearrange-

ment (isomerization) of the Cre tetramer that positions the

catalytic Tyr residues in the latter pair of subunits to partic-

ipate in the reaction while those in the former pair of sub-

units are similarly removed from the scene of the action.

Note that this mechanism differs from that mediated by

serine recombinases in that the latter cleave all four DNA

strands prior to initiating their exchange and hence do not

have a Holliday junction intermediate (Section 30-6Bd).

The X-ray structures of Cre tetramers in their complexes

with several loxP model DNAs, determined by Gregory Van

Duyne, have helped elucidate its mechanism.When the DNA

had a single-strand nick past the second nucleotide from the

5¿ end of the crossover region, Cre-catalyzed strand scission

yielded a free nucleotide (a CMP) that diffused away. Since

this nucleotide contained the otherwise reactive 5¿-OH

group, the 3¿-phosphoTyr intermediate was irreversibly

trapped, that is, Cre could not carry out the strand exchange

reaction in Fig. 30-97 (this nicked DNA is a suicide substrate

for Cre; Section 28-3Bc).The X-ray structure of the Cre com-

plex of this nicked DNA confirmed the presence of the 3¿-

phosphoTyr intermediate and indicated, through model

building, that the 5¿-OH group on the missing CMP residue

would be well positioned to nucleophilically attack the 3¿-

phosphoTyr bond on the opposite strand (Fig. 30-98a). Note

that this complex is only 2-fold symmetric although its four

Cre subunits and much of the DNA are related by pseudo-4-

fold symmetry. When the DNA was, instead, an immobile

Holiday junction (Fig. 30-98b),the complex was also pseudo-

4-fold symmetric with the single strands that had crossed

over noticeably kinked at their centers. These structures re-

vealed that the conformational changes necessary to carry

out the strand exchange and isomerization reactions (Fig. 30-

97) required surprisingly small movements on the part of the

Cre subunits and that only the sugar–phosphate backbones

of the strand-exchanged nucleotides needed to move in or-

der to form the Holliday junction.

h. Most Transpositions in Eukaryotes Involve

RNA Intermediates

Transposons similar to those in prokaryotes also occur in

eukaryotes, including yeast, maize, Drosophila, and humans.

In fact, ⬃3% of the human genome consists of DNA-based

transposons although, in most cases, their sequences have

mutated so as to render them inactive, that is, these trans-

posons are evolutionary fossils. However, many eukaryotic

transposons exhibit little similarity to those of prokaryotes.

Rather, their base sequences resemble those of retroviruses

(see below), which suggests that these transposons are de-

generate retroviruses. The transposition of these so-called

retrotransposons occurs via a pathway that resembles the

replication of retroviral DNA (Section 15-4C): (1) their tran-

scription to RNA, (2) the reverse transcriptase–mediated

copying of this RNA to cDNA (Section 30-4C), and (3) the

largely random insertion of this DNA into the host

1244 Chapter 30. DNA Replication, Repair, and Recombination

Figure 30-96 The circularization of linear bacteriophage P1

DNA. This occurs through the Cre-mediated site-specific

recombination between its two terminally located loxP sites (red

and green) to yield its lysogenic plasmid.

Figure 30-97 The mechanism of Cre–loxP site-specific

recombination. The dashed lines represent the nonpalindromic

crossover regions of the loxP sites. The green and magenta Cre

subunits are active for cleavage in the top and bottom parts of

the diagram, respectively, with their roles being switched by the

isomerization step. Note that the mechanism does not require

branch migration of the Holliday junction intermediate.

[Courtesy of Gregory Van Duyne, University of Pennsylvania

School of Medicine.]

Bacteriophage P1

linear

Bacteriophage P1

lysogenic plasmid

Cre

LoxP LoxP

JWCL281_c30_1173-1259.qxd 8/10/10 9:13 PM Page 1244

organism’s genome as mediated by enzymes known as

integrases (which catalyze reactions similar to and struc-

turally resemble cut-and-paste DNA transposases).

The involvement of RNA in retrotransposon-mediated

transposition was ingeniously shown by Gerald Fink

through his remodeling of Ty1, the most common transpos-

able element in budding yeast (which has ⬃35 copies of this

6.3-kb element comprising ⬃13% of its 1700 kb genome;Ty

stands for Transposon yeast), so that it contained a yeast in-

tron (a sequence that is excised from an RNA transcript

and hence is absent in the mature RNA; Section 5-4Ac) and

was preceded by a galactose-sensitive yeast promoter. The

transposition rate of this remodeled Ty1 element varied

with the galactose concentration in the medium and the

transposed elements all lacked the intron, thereby demon-

strating the participation of an RNA intermediate.

A retroviral genome (Fig. 30-99a) is flanked by direct

long terminal repeats (LTRs) of 250 to 600 bp and typically

contains the genes encoding three polyproteins:gag, which is

Section 30-6. Recombination and Mobile Genetic Elements 1245

(a)

(b)

Figure 30-98 X-ray structures of the Cre homotetramer in its

complexes with model loxP DNAs. (a) Two identical dsDNAs

that were nicked past the second nucleotide from the 5¿ end of

their crossover regions; and (b) an immobile Holliday junction.

The left panels show the Cre–DNA complexes as viewed along

their exact 2-fold and pseudo-4-fold axes, with the active and

inactive subunits green and magenta, respectively (as in Fig. 30-97),

and with the DNA gold. The right panels show only the DNAs in

the X-ray structures as viewed from below the left panels. In the

right panel of Part a, the active site Tyr that is covalently linked

to the 3¿-OH group of the cleaved DNA strand is shown in stick

form (red) and the modeled-in position of the cleaved CMP’s

5¿-OH group is shown positioned to nucleophilically attack the

3¿-phosphoTyr group on the opposite dsDNA (curved arrows).

In the right panel of Part b, the three base pairs that form as a

consequence of strand exchange are indicated. Note that the

vertical strands in the crossovers but not the horizontal strands

are distinctly kinked at their centers. [Courtesy of Gregory Van

Duyne, University of Pennsylvania School of Medicine. PDBids

2CRX, 3CRX, 4CRX, and 5CRX.]

Figure 30-99 Gene sequences of (a) retroviruses and (b) the

Ty1 retrotransposon.

Ty1

LTRLTR TYA TYB

(b)

Retrovirus

LTRLTR gag pol env

(a)

JWCL281_c30_1173-1259.qxd 8/10/10 9:13 PM Page 1245

cleaved to the proteins comprising the viral core (Fig. 15-34);

pol, which is cleaved to the above-mentioned reverse tran-

scriptase and integrase, as well as the protease that catalyzes

these cleavages; and env, which is cleaved to viral outer en-

velope proteins. Ty1 (Fig. 30-99b) is likewise flanked by

LTRs (of 330 bp) but expresses only two polyproteins: TYA

and TYB, the counterparts of gag and pol. Moreover, TYA

and TYB, together with Ty1 RNA, form viruslike particles

in the yeast cytoplasm. However,Ty1 lacks a counterpart of

the retroviral env gene.Hence Ty1 is an “internal virus” that

can only replicate within a genome, albeit at an extremely

low rate compared to that of real retroviral infections.

Copia (Latin for abundance), the most abundant retro-

transposon in the Drosophila genome (which contains

20–60 copies of copia), resembles Ty1.

The LTRs in retroviruses and retrotransposons such as

Ty1 and copia are essential elements for their transcription

and hence for their transposition. Yet, vertebrate genomes

also contain retrotransposons that lack LTRs and hence

cannot be transcribed analogously to retroviruses. A com-

mon family of these nonviral retrotransposons, the 1- to 7-

kb long interspersed nuclear elements (LINEs), each con-

tain two open reading frames: ORF1, which contains

sequences similar to those in gag; and ORF2, which con-

tains sequences similar to those encoding reverse tran-

scriptase. A proposed mechanism for the transposition of

LINEs is diagrammed in Fig. 30-100.

Different types of transposons, DNA-only, retroviral,

and nonviral, predominate in different organisms.Thus bac-

teria, as we have seen, contain nearly exclusively DNA-only

transposons, yeast have mainly retroviral retrotransposons,

Drosophila have all three types, and in humans LINEs pre-

dominate. In fact, the human genome contains an estimated

1.4 million LINEs or LINE fragments that comprise ⬃20%

of the 3.0-billion-bp human genome (genomic organization

is discussed in Section 34-2). The great majority of these

molecular parasites have mutated to the point of inactivity

but a few still appear capable of further transposition. In-

deed, several hereditary diseases are caused by the inser-

tion of a LINE into a gene. Several other types of retro-

transposons also comprise significant fractions of the

human genome as we shall see in Section 34-2.

7 DNA METHYLATION AND

TRINUCLEOTIDE REPEAT EXPANSIONS

The A and C residues of DNA may be methylated, in a

species-specific pattern, to form N

-methyladenine (m

A),

-methylcytosine (m

C), and 5-methylcytosine (m

residues, respectively.

-Methyladenine (m

residue

5-Methylcytosine (m

residue

-Methylcytosine (m

residue

1246 Chapter 30. DNA Replication, Repair, and Recombination

Figure 30-100 Proposed mechanism for the transposition of

nonviral retrotransposons. (1) The retrotransposon-encoded

reverse transcriptase/endonuclease nicks one strand of the target

DNA and then recruits the RNA transcript of the retrotransposon

to this site. (2) The DNA-primed reverse transcription of the

retrotransposon RNA. (3) The RNA is degraded, and the second

DNA strand is synthesized using the first strand as its template

(normal reverse transcriptase reactions; Section 30-4C), followed

by the insertion of the resulting nonviral retrotransposon into

the target DNA via a poorly understood process.

Target DNA

RNA

Transposed nonviral retrotransposon

Strand cleavage by the

endonuclease activity

and binding of transcribed

nonviral retrotransposon RNA

DNA-primed

reverse transcription

Degradation of RNA,

synthesis of second

DNA strand, and insertion

of the dsDNA into the

target DNA

Reverse transcriptase/endonuclease

JWCL281_c30_1173-1259.qxd 8/10/10 9:13 PM Page 1246

Cytosine residue

H S

–

DNA

–

S-Adenosylmethionine (SAM)

–

COO

–

DNA

B +

C residue

DNA

S-Adenosylhomocysteine

COO

–

DNA

These are the only types of modifications to which DNA is

subjected in cellular organisms (although all the C residues

of T-even phage DNAs are converted to 5-hydroxymethyl-

cytosine residues,

which may, in turn, be glycosylated). These methyl groups

project into B-DNA’s major groove, where they can inter-

act with DNA-binding proteins. In most cells, only a few

percent of the susceptible bases are methylated, although

this figure rises to ⬎30% of the C residues in some plants.

Bacterial DNAs are methylated at their own particular

restriction sites, thereby preventing the corresponding re-

striction endonucleases from degrading the DNA (Section

5-5A). These restriction–modification systems, however,

account for only part of the methylation of bacterial

DNAs. In E. coli, most DNA methylation is catalyzed by

the products of the dam and dcm genes. The Dam methyl-

transferase (Dam MTase) methylates the A residue in all

GATC sequences, whereas the Dcm MTase methylates

both C residues in GG at their C5 positions. NoteCC

5-Hydroxymethylcytosine residue

that both of these sequences are palindromic. We have

seen that E. coli uses Dam Mtase–mediated methylation to

differentiate parental from newly synthesized DNA in mis-

match repair (Section 30-5C) and in limiting oriC-based

DNA replication initiation to once per cell generation via

sequestration (Section 30-3Cb).

a. The MTase Reaction Occurs via a Covalent

Intermediate in Which the Target Base Is Flipped Out

The Dam and Dcm MTases, as do all known DNA

MTases, use S-adenosylmethionine (SAM) as their methyl

donor. Indeed, all m

C-MTases share a set of conserved se-

quence motifs. Daniel Santi has proposed that the catalytic

mechanism of these m

C-MTases (Fig. 30-101) is similar to

that of thymidylate synthase (Fig. 28-19) in that both types

of enzymes transfer methyl groups to pyrimidine C5 atoms

via a reaction that is initiated by the nucleophilic attack of

a Cys thiolate group on the pyrimidine’s C6 position. The

pyrimidine’s C5 atom is thereby activated as a resonance-

stabilized carbanion that nucleophilically attacks the

methyl donor’s methyl group (which in thymidylate syn-

thase is donated by N

-methylene-THF rather than

SAM) to yield a covalent intermediate. This intermediate

subsequently decomposes to products through the enzy-

matic abstraction of the proton substituent to C5 and elim-

ination of the enzyme. The Cys thiolate nucleophile is a

component of a Pro-Cys dipeptide that is invariant in all

known m

C-MTases and thymidylate synthases.

Section 30-7. DNA Methylation and Trinucleotide Repeat Expansions 1247

Figure 30-101 The catalytic mechanism of 5-methylcytosine

methyltransferases (m

C-MTases). The methyl group is supplied

by SAM, which thereby becomes S-adenosylhomocysteine. In

M.HhaI, the DNA MTase from Haemophilus haemolyticus, the

active site thiolate group,

⫺

S¬E, is on Cys 81, the enzyme

general acid, E¬A, is Glu 119, and the enzyme general base,

E¬B, has not been identified. [After Verdine, G.L., Cell 76, 198

(1994).]

JWCL281_c30_1173-1259.qxd 8/10/10 9:13 PM Page 1247

This mechanism is supported by the observation that

the action of m

C-MTases on a 5-fluorocytosine (f

residue

irreversibly traps the covalent intermediate (and hence in-

activates the enzyme) because the enzyme cannot abstract

fluorine, the most electronegative element, as an F

⫹

ion (5-

fluorodeoxyuridylate is likewise a suicide substrate for

thymidylate synthase; Section 28-3Bc). Stereochemical

principles dictate that the enzyme’s Cys thiolate group can

nucleophilically attack cytosine’s C5 position only from

above or below the ring. This is possible because, as we

shall see below, the enzyme induces its cytosine target to

flip out of the DNA double helix.

The DNA MTase from Haemophilus haemolyticus

(M.HhaI), a 327-residue monomer, is a component of this

bacterium’s restriction–modification system. M.HhaI

methylates its recognition sequence, 5¿-GCGC-3¿ in double-

stranded DNA, to yield 5¿-G-m

C-GC-3¿. Richard Roberts

and Xiaodong Cheng determined the X-ray structure of the

inactivated M.HhaI–DNA complex formed by incubating

the enzyme with the self-complementary sequence

d(TGATAG-f

C-GCTATC) (in which the enzyme’s recog-

nition sequence is in bold) in the presence of SAM. The

DNA binds to the enzyme in a large cleft between its two

unequally sized domains (Fig. 30-102). The structure’s most

striking feature is that the f

C nucleotide has flipped out of

the minor groove in the otherwise largely undistorted B-

DNA helix and has inserted into the enzyme’s active site.

5-Fluorocytosine (f

C) residue

There, the f

C has reacted with SAM so as to yield adeno-

sylhomocysteine (SAM without its methyl group) and the

methylated intermediate covalently linked to Cys 81. The

side chain of Gln 237 fills the cavity in the DNA double he-

lix left by the departure of the f

C by hydrogen-bonding to

the opposing G base. Comparison of this structure with that

of M.HhaI in complex only with SAM indicates that on

binding the DNA the protein’s so-called active site loop

(residues 80–99) swings around to contact the DNA, a

movement of up to 25 Å. Nearly all base-specific interac-

tions are made in the major groove by two Gly-rich loops

(residues 233–240 and 250–257), the so-called recognition

loops. The protein also makes extensive sequence-nonspe-

cific contacts with DNA phosphate groups.

Base flipping was first observed in the above X-ray

structure. However, as is now clear from the structures of

other MTases as well as those of variety of DNA repair en-

zymes (e.g., Sections 30-5Aa and 30-5Be), base flipping is a

common mechanism through which enzymes gain access to

the bases in dsDNA on which they perform chemistry.

b. DNA Methylation in Eukaryotes Functions

in Gene Regulation

5-Methylcytosine is the only methylated base in most

eukaryotic DNAs, including those of vertebrates. This

modification occurs largely in the CG dinucleotide of vari-

ous palindromic sequences. CG is present in the vertebrate

genome at only about one-fifth its randomly expected

1248 Chapter 30. DNA Replication, Repair, and Recombination

Figure 30-102 X-ray structure of the M.HhaI DNA methyltransferase

in complex with S-adenosylhomocysteine and a dsDNA containing a

methylated 5-fluorocytosine base at the enzyme’s target site. The

protein is represented by a semitransparent ribbon.The DNA is

drawn in stick form colored according to atom type (C green,

N blue, O red, and P orange). Its methylated 5-fluorocytosine

residue (C atoms cyan) has swung out of the DNA helix into the

enzyme’s active site pocket, where its C6 forms a covalent bond with

the S atom of an enzyme Cys residue (C atoms magenta and S yellow).

The methyl group and a fluorine atom at C5 (which prevents the

methylation reaction from going to completion) are represented

by magenta and yellow-green spheres, respectively. The position of

the flipped-out cytosine base in the DNA double helix is occupied

by the side chain of a Gln residue (shown in space-filling form

with C yellow), which hydrogen bonds to the “orphaned” guanine

base. The S-adenosylmethionine, which has given up its methyl group,

is drawn in stick form with its C atoms pink. [Based on an X-ray structure

by Richard Roberts, New England Biolabs, Beverly, Massachusetts, and Xiaodong

Cheng, Cold Spring Harbor Laboratory, New York. PDBid 1MHT.]

JWCL281_c30_1173-1259.qxd 8/10/10 9:13 PM Page 1248

GC CH

replication

GC CH

GC +

CGCH

GC CH

maintenance

methylation

frequency. The upstream regions of many genes, however,

have normal CG frequencies and are therefore known as

CpG islands.

The degree of eukaryotic DNA methylation and its pat-

tern are conveniently assessed by comparing the Southern

blots (Section 5-5D) of DNA cleaved by the restriction en-

donucleases HpaII (which cleaves CCGG but not C-m

C-

GG) and MspI (which cleaves both). Such studies indicate

that eukaryotic DNA methylation varies with the species,

the tissue, and the position along a chromosome.

The m

C residues in a given DNA segment can be iden-

tified through bisulfite sequencing, in which the DNA is re-

acted with bisulfite ion which selectively deami-

nates C (but not m

C) residues to U, followed by PCR

amplification (Section 5-5F), which copies these U’s to T’s

and the m

C’s to C’s. Comparison of the sequences of the

amplified DNA with that of untreated DNA reveals which

C’s in the untreated DNA are methylated.

There is clear evidence that DNA methylation switches

off eukaryotic gene expression, particularly when it occurs

in the promoter regions upstream of a gene’s transcribed

sequences. For example, globin genes are less methylated

in erythroid cells than they are in nonerythroid cells and,

in fact, the specific methylation of the control region in a

recombinant globin gene inhibits its transcription in

transfected cells. In further support of the inhibitory ef-

fect of DNA methylation is the observation that 5-aza-

cytosine (5-azaC),

a base analog that cannot be methylated at its N5 position

and that inhibits DNA MTases, stimulates the synthesis of

several proteins and changes the cellular differentiation

patterns of cultured eukaryotic cells. The observation that

repetitive intragenic parasites such as LINEs are highly

methylated in somatic tissues has led to the hypothesis that

CpG methylation in mammals arose to prevent the spuri-

ous transcriptional initiation of these retrotransposons.

The way in which DNA methylation prevents gene ex-

pression is poorly understood. One possibility is that DNA

methylation can directly block the binding of transcrip-

tional activators to their target sequences. However, in

many cases, DNA methylation is recognized by a family of

proteins that contain a conserved methyl-CpG binding

domain (MBD). Since the methyl groups of m

C residues

extend into dsDNA’s major groove, MBDs can bind to

them without perturbing DNA’s double helical structure.

MBD-containing proteins inhibit the transcription of their

bound promoter-methylated genes by recruiting protein

complexes that induce the alteration of the local chromo-

5-Azacytosine

(5-azaC)

N N

(HSO

⫺

some structure in a way that prevents the transcription of

the associated gene (eukaryotic chromosome structure is

discussed in Section 34-1). Another possibility has been

raised by the observation that the methylation of synthetic

poly(GC) stabilizes its Z-DNA conformation. Quite possi-

bly, the formation of Z-DNA, which has been detected

in vivo (Section 29-1Bb), acts as a conformational switch to

turn off local gene expression.

c. DNA Methylation in Eukaryotes

Is Self-Perpetuating

The palindromic nature of DNA methylation sites in eu-

karyotes permits the methylation pattern on a parental

DNA strand to direct the generation of the same pattern in

its daughter strand (Fig. 30-103).This maintenance methyl-

ation would result in the stable “inheritance” of a methyla-

tion pattern in a cell line and hence cause these cells to all

have the same differentiated phenotype. Such changes to

the genome are described as being epigenetic (Greek: epi,

Section 30-7. DNA Methylation and Trinucleotide Repeat Expansions 1249

Figure 30-103 Maintenance methylation. The pattern of

methylation on a parental DNA strand induces the

corresponding methylation pattern in the complementary strand.

In this way, a stable methylation pattern may be maintained in a

cell line.

JWCL281_c30_1173-1259.qxd 8/10/10 9:13 PM Page 1249

upon or beside) because they provide an additional layer

of information that specifies when and where specific por-

tions of the otherwise fixed genome are expressed (an epi-

genetic change that we have already encountered is the

lengthening of telomeres in germ cells; Section 30-4D).

Epigenetic characteristics, as we shall see, are not bound by

the laws of Mendelian inheritance.

There is considerable experimental evidence favoring

the existence of maintenance methylation, including the

observation that artificially methylated viral DNA, on

transfection into eukaryotic cells, maintains its methylation

pattern for at least 30 cell generations. Maintenance

methylation in mammals appears to be mediated mainly by

the protein DNMT1 (for DNA methyltransferase 1), which

has a strong preference for methylating hemimethylated

substrate DNAs. In contrast, prokaryotic DNA MTases

such as M.HhaI do not differentiate between hemimethy-

lated and unmethylated substrate DNAs. The importance

of maintenance methylation is demonstrated by the obser-

vation that mice homozygous for deletion of the DNMT1

gene die early in embryonic development.

The interaction energy provided a single methyl group

in the major groove of DNA seems insufficient to permit

a protein to reliably differentiate between an m

C and a

C residue. Nevertheless, the SRA (for SET and RING-as-

sociated) domain of the protein UHRF1 (for ubiquitin-

like, containing PHD and RING finger domains 1) does

so and thereon recruits DNMT1 the site. How does the

SRA domain distinguish a single m

C residue from un-

methylated C?

The X-ray structure of the 210-residue SRA domain in

complex with a 12-bp DNA containing a centrally located

hemimethylated CpG was independently determined by

Masahiro Shirakawa and Sirano Dhe-Paganon. It reveals

that the DNA’s m

C residue has flipped out from the minor

groove side of the largely straight B-form DNA to tightly

bind deep within in a protein pocket (Fig. 30-104). Two

loops reach around the resulting gap in the DNA helix

from the major and minor grooves to interact with the

other three bases of the hemimethylated CpG site. The

C base is replaced in the double helix by the side chain

of an Arg residue extending from the minor groove loop,

which hydrogen-bonds to the “orphaned” G residue with-

out greatly disturbing its conformation.

DNMT1 has both a catalytic domain and a UHRF1-

binding domain. The catalytic domain is thought to have

a similar structure and mode of DNA binding as bacter-

ial MTases such as M.HhaI (Fig. 30-102). However, a

model of the DNMT1 catalytic domain–DNA–SRA do-

main complex based on the structures of the

M.HhaI–DNA and SRA–DNA complexes indicates that

it is unlikely that both the m

C and C residues of a dou-

ble helical CpG island could be simultaneously flipped

out of the DNA. This suggests that the binding of the

UHRF1-binding domain of DNMT1 to the SRA–DNA

complex causes the m

C to flip back into the DNA dou-

ble helix while the C to be methylated flips out to bind to

the catalytic domain of DNMT1.

The pattern of DNA methylation in mammals varies in

early embryological development. DNA methylation lev-

els are high in mature gametes (sperm and ova) but are

nearly eliminated by the time a fertilized ovum has become

a blastocyst (a hollow ball of cells, the stage at which the

embryo implants into the uterine wall; embryonic develop-

ment is discussed in Section 34-4A). After this stage, how-

ever, the embryo’s DNA methylation levels globally rise

until, by the time the embryo has reached the developmen-

tal stage known as a gastrula, its DNA methylation levels

have risen to adult levels, where they remain for the life-

time of the animal. This de novo (new) methylation ap-

pears to be mediated by two DNA MTases distinct from

DNMT1 named DNMT3a and DNMT3b. An important

exception to this remethylation process is that the CpG is-

lands of germline cells (cells that give rise to sperm or ova)

remain unmethylated. This ensures the faithful transmis-

sion of the CpG islands to the succeeding generation in the

face of the strong mutagenic pressure of m

C deamination

(which yields T, a mutation that mismatch repair occasion-

ally fails to correct).

The change in DNA methylation levels (epigenetic re-

programming) during embryonic development suggests

that the pattern of genetic expression differs in embryonic

and somatic cells. This explains the observed high failure

1250 Chapter 30. DNA Replication, Repair, and Recombination

Figure 30-104 X-ray structure of the SRA domain of mouse

UHRF1 in complex with a hemimethylated 12-bp DNA. The

protein is drawn in ribbon form (green) embedded in its

semitransparent molecular surface. The DNA is shown in stick

form colored according to atom type (DNA C gray except for

the m

C residue C which is magenta, N blue, O red, and P orange)

with successive P atoms in the same strand connected by orange

rods.The methyl group of the m

C residue is represented by a

magenta sphere. An Arg side chain (C cyan) fills the space in the

DNA double helix vacated by the flipped-out m

C. [Based on an

X-ray structure by Masahiro Shirakawa, Kyoto University, Japan.

PDBid 2ZDK.]

JWCL281_c30_1173-1259.qxd 8/10/10 9:13 PM Page 1250