Voet D., Voet Ju.G. Biochemistry

Подождите немного. Документ загружается.

activated expression platform forms an intrinsic transcrip-

tion termination site (Section 31-2Da) so that transcription

beyond this site proceeds only when the effector is absent.

In others, the activated expression platform cleaves itself,

thereby inactivating the mRNA (the ability of RNA to act

as an enzyme is discussed in Section 31-4Ae).

Riboswitches collectively regulate ⬎2% of the genes in

certain bacteria. Plants and fungi also contain riboswitches

(although the above TTP-sensing riboswitch could not

function in eukaryotes because eukaryotic ribosomes do

not bind Shine–Dalgarno sequences; Section 32-3Cd). The

fact that the interaction of riboswitches with their effectors

does not require the participation of proteins suggests that

they are relics of the RNA world (Section 1-5Ca) and

hence among the oldest regulatory systems.

I. Regulation of Ribosomal RNA Synthesis:

The Stringent Response

E. coli cells growing under optimal conditions divide every

20 min. Such cells contain up to 20,000 ribosomes and

hence must synthesize ⬃10,000 ribosomes per cell division

cycle. Yet RNAP can initiate the transcription of an rRNA

gene no faster than about once per second. If E. coli con-

tained only one copy of each of the three types of rRNA

genes (those specifying the so-called 23S, 16S, and 5S

rRNAs; Section 32-3A), fast-growing cells could synthesize

no more than ⬃1200 ribosomes during their cell division

cycle. However, the E. coli genome contains seven sepa-

rately located rRNA operons, all of which contain one

nearly identical copy of each type of rRNA gene. Moreover,

rapidly growing cells contain multiple copies of their repli-

cating chromosomes (Section 30-3Cb), thereby accounting

for the observed rRNA synthesis rate.

Cells have the remarkable ability to coordinate the rates

at which their thousands of components are synthesized.

For example,E. coli adjust their ribosome content to match

the rate at which they can synthesize proteins under the

prevailing growth conditions.The rate of rRNA synthesis is

therefore proportional to the rate of protein synthesis. One

mechanism by which this occurs is known as the stringent

response: A shortage of any species of amino acid–charged

tRNA (usually a result of “stringent” or poor growth condi-

tions) that limits the rate of protein synthesis triggers a

sweeping metabolic readjustment. A major facet of this

change is an abrupt 10- to 20-fold reduction in the rate of

rRNA and tRNA synthesis. This stringent control, more-

over, depresses numerous metabolic processes (including

DNA replication and the biosynthesis of carbohydrates,

lipids, nucleotides, proteoglycans, and glycolytic intermedi-

ates) while stimulating others (such as amino acid biosyn-

thesis). The cell is thereby prepared to withstand nutri-

tional deprivation.

a. (p)ppGpp Mediates the Stringent Response

The stringent response is correlated with a rapid intracel-

lular accumulation of two unusual nucleotides, ppGpp and

pppGpp [known collectively as (p)ppGpp], and their

prompt decay when amino acids become available. The ob-

servation that mutants, designated relA

⫺

, which do not ex-

hibit the stringent response (they are said to have relaxed

control) lack (p)ppGpp suggests that these substances me-

diate the stringent response.This idea was corroborated by

in vitro studies demonstrating, for example, that (p)ppGpp

inhibits the transcription of rRNA genes but stimulates the

transcription of the trp and lac operons as does the strin-

gent response in vivo. Apparently, (p)ppGpp acts by some-

how altering RNAP’s promoter specificity at stringently

controlled operons, a hypothesis that is supported by the

isolation of RNAP mutants that exhibit reduced responses

to (p)ppGpp. In addition, (p)ppGpp causes an increased

frequency of pausing in RNAPs engaged in elongation,

thereby reducing the rate of transcription.

The protein encoded by the wild-type relA gene, named

stringent factor (RelA), catalyzes the reaction

and, to a lesser extent,

However, several ribosomal proteins convert pppGpp to

ppGpp so that ppGpp is the stringent response’s usual effec-

tor. Stringent factor is only active in association with a ribo-

some that is actively engaged in translation. (p)ppGpp syn-

thesis occurs when a ribosome binds its mRNA-specified

but uncharged tRNA (lacking an appended amino acid

residue). The binding of a specified and charged tRNA

greatly reduces the rate of (p)ppGpp synthesis. The ribo-

some apparently signals the shortage of an amino acid by

stimulating the synthesis of (p)ppGpp which,acting as an in-

tracellular messenger, influences the rates at which a great

variety of operons are transcribed.

(p)ppGpp degradation is catalyzed by the spoT gene

product. The spoT

⫺

mutants show a normal increase in

(p)ppGpp level on amino acid starvation but an abnor-

mally slow decay of (p)ppGpp to basal levels when amino

acids again become available.The spoT

⫺

mutants therefore

exhibit a sluggish recovery from the stringent response.

The (p)ppGpp level is apparently regulated by the counter-

vailing activities of stringent factor and the spoT gene

product.

4 POST-TRANSCRIPTIONAL

PROCESSING

The immediate products of transcription, the primary tran-

scripts, are not necessarily functional entities. In order to

acquire biological activity, many of them must be specifi-

cally altered in several ways: (1) by the exo- and endonu-

cleolytic removal of polynucleotide segments; (2) by ap-

pending nucleotide sequences to their 3¿ and 5¿ ends; and

(3) by the modification of specific nucleosides. The three

major classes of RNAs, mRNA, rRNA, and tRNA, are al-

tered in different ways in prokaryotes and in eukaryotes. In

this section we shall outline these post-transcriptional

modification processes.

ATP ⫹ GDP Δ AMP ⫹ ppGpp

ATP ⫹ GTP Δ AMP ⫹ pppGpp

Section 31-4. Post-Transcriptional Processing 1301

JWCL281_c31_1260-1337.qxd 8/11/10 9:48 PM Page 1301

A. Messenger RNA Processing

In prokaryotes, most primary mRNA transcripts function

in translation without further modification. Indeed, as we

have seen, ribosomes in prokaryotes usually commence

translation on nascent mRNAs. In eukaryotes, however,

mRNAs are synthesized in the cell nucleus, whereas trans-

lation occurs in the cytosol. Eukaryotic mRNA transcripts

can therefore undergo extensive post-transcriptional pro-

cessing while still in the nucleus.

a. Eukaryotic mRNAs Are Capped

Eukaryotic mRNAs have a peculiar enzymatically

appended cap structure consisting of a 7-methylguanosine

G) residue joined to the transcript’s initial (5¿) nucleo-

side via a 5¿–5¿ triphosphate bridge (Fig. 31-47).This m

cap,

which is added to the growing transcript before it is ⬃30

nucleotides long, defines the eukaryotic translational start

site (Section 32-3Cd). A cap may be O

2¿

-methylated at the

transcript’s leading nucleoside (cap-1, the predominant

cap in multicellular organisms), at its first two nucleosides

(cap-2), or at neither of these positions (cap-0, the pre-

dominant cap in unicellular eukaryotes). If the leading

nucleoside is adenosine (it is usually a purine), it may also

be N

-methylated.

Capping involves several enzymatic reactions: (1) the

removal of the leading phosphate group from the mRNA’s

5¿ terminal triphosphate group by an RNA triphosphatase;

(2) the guanylation of the mRNA by capping enzyme,

which requires GTP and yields the 5¿–5¿ triphosphate

bridge and PP

; (3) the methylation of guanine by guanine-

7-methyltransferase in which the methyl group is supplied

by S-adenosylmethionine (SAM); and possibly (4) the O2¿

methylation of the mRNA’s first and perhaps its second

nucleotide by a SAM-requiring 2ⴕ-O-methyltransferase.

Both the capping enzyme and the guanine-7-methyltrans-

ferase bind to RNAP II’s phosphorylated CTD (Section

31-2E). Hence it is likely that capping marks the comple-

tion of RNAP II’s switch from transcription initiation to

elongation.

b. Eukaryotic mRNAs Have Poly(A) Tails

Eukaryotic mRNAs, unlike those of prokaryotes, are

invariably monocistronic. Moreover, in contrast to the case

in bacteria (Section 31-2D), no eukaryotic transcriptional

termination sequence has been identified (but see below).

In fact, the eukaryotic termination process is imprecise;

that is, the primary transcripts of a given structural gene

have heterogeneous 3¿ sequences. Nevertheless, mature eu-

karyotic mRNAs have well-defined 3¿ ends; almost all of

them in mammals have 3¿-poly(A) tails of ⬃250 nucleotides

(⬃80 in yeast). The poly(A) tails are enzymatically ap-

pended to the primary transcripts in two reactions that are

mediated by a 500- to 1000-kD complex that consists of at

least six proteins:

1. A transcript is cleaved to yield a free 3¿-OH group at

a specific site that is 15 to 25 nucleotides past an

AAUAAA sequence and within 50 nucleotides before a

U-rich or G ⫹ U–rich sequence. The AAUAAA sequence

is highly conserved in higher eukaryotes (but not yeast) in

which its mutation abolishes cleavage and polyadenyla-

tion. The precision of the cleavage reaction has apparently

eliminated the need for accurate transcriptional termina-

tion. Nevertheless, the identity of the endonuclease that

cleaves the RNA is uncertain although cleavage factors I

and II (CFI and CFII) are required for this process.

2. The poly(A) tail is subsequently generated from ATP

through the stepwise action of poly(A) polymerase (PAP).

This enzyme, which by itself only weakly binds RNA, is re-

cruited by cleavage and polyadenylation specificity factor

(CPSF) on this heterotetramer’s recognition of the

AAUAAA sequence, which it does with almost no toler-

ance for sequence variation. The downstream G ⫹ U–rich

element is recognized by the heterotrimeric cleavage stim-

ulation factor (CstF), which increases the affinity with

which CPSF binds the AAUAAA sequence. However,

1302 Chapter 31. Transcription

Figure 31-47 The structure of the 5¿ cap of eukaryotic

mRNAs. It is known as cap-0, cap-1, or cap-2, respectively, if it

has no further modifications, if the leading nucleoside of the

transcript is O

2¿

-methylated, or if its first two nucleosides are

2¿

-methylated.

OH OH

–

O(CH

)

Base

O(CH

)

Base

...

7-Methyl-G

May be

-methylated

if A

JWCL281_c31_1260-1337.qxd 8/11/10 9:48 PM Page 1302

once the poly(A) tail has grown to ⬃10 residues, the

AAUAAA sequence is no longer required for further

chain elongation. This suggests that CPSF becomes disen-

gaged from its recognition site in a manner reminiscent of

the way ␴ factor is released from the transcriptional initia-

tion site once the elongation of prokaryotic mRNA is un-

der way (Section 31-2B). The final length of the poly(A)

tail is controlled by poly(A)-binding protein II (PAB II),

multiple copies of which bind to successive segments of

poly(A). PAB II also increases the processivity of PAP.

Both CPSF and CstF bind to the phosphorylated RNAP II

CTD (Section 31-2E); deleting the CTD inhibits

polyadenylation. Evidently, the CTD couples polyadenyla-

tion to transcription. The mechanism that controls the

length of a poly(A) tail is unclear.

PAP is a template-independent RNA polymerase that

elongates an mRNA primer with a free 3¿-OH group. The

X-ray structure of the 530-residue D154A mutant form of

yeast PAP (D154 is a catalytically essential active site

residue) in complex with A

and ATP, determined by An-

drew Bohm, reveals that this monomeric protein consists

of three domains that form a prominent U-shaped cleft

(Fig. 31-48). Hence it has the handlike domain arrange-

ment of template-directed polymerases (Section 30-2Ad).

Indeed, PAP’s N-terminal domain, which contains the en-

zyme’s active site, is homologous to the palm domain of

DNA polymerase ␤, although it forms the thumb side of

the cleft rather than its base. PAP’s central domain, which

forms the base of the cleft, is functionally but not struc-

turally analogous to the fingers domain of template-

directed polymerases in that it interacts with the ␤ and ␥

phosphates of the incoming ATP. However, the C-terminal

domain shows no resemblance to a fingers domain. Rather,

it is topologically similar to the RNA-recognition motif

[RRM; also known as the RNA-binding domain (RBD)]

that occurs in ⬎200 different RNA-binding proteins (see

below).The A

binds in the cleft in an extended conforma-

tion such that, in contrast to the nucleic acids bound to

template-dependent polymerases, its bases are not in con-

tact. However, the 3¿-terminal base of the A

stacks on the

base of the ATP.

In comparison to the X-ray structure of yeast PAP in

complex with 3¿-dATP, also determined by Bohm, the N-

terminal domain in the PAP–A

–ATP structure has under-

gone an ⬃23⬚ hingelike motion toward the C-terminal do-

main to form a closed conformation resembling that

observed in the structure of Taq DNA polymerase I in

complex with its substrates (Section 30-2Ae). This motion

forms several new interactions, both direct and water-

mediated, between PAP and its ATP substrate that helps

differentiate adenine from other bases. In contrast, in tem-

plate-dependent polymerases, the incoming base only

makes sequence-specific contacts with the template base

(Section 30-2Ae).

In vitro studies indicate that a poly(A) tail is not re-

quired for mRNA translation. Rather, the observations

that an mRNA’s poly(A) tail shortens as it ages in the

cytosol and that unadenylated mRNAs have abbreviated

cytosolic lifetimes suggest that poly(A) tails have a protec-

tive role. In fact, the only mature mRNAs that lack poly(A)

tails, those of histones (which, with few exceptions, lack the

AAUAAA cleavage–polyadenylation signal), have life-

times of ⬍30 min in the cytosol, whereas most other

mRNAs last hours or days.The poly(A) tails are specifically

complexed in the cytosol by poly(A) binding protein

(PABP; not related to PAB II), which organizes poly(A)-

bearing mRNAs into ribonucleoprotein particles. PABP is

thought to protect mRNA from degradation as is sug-

gested, for example, by the observation that the addition of

PABP to a cell-free system containing mRNA and mRNA-

degrading nucleases greatly reduces the rate at which the

mRNAs are degraded and the rate at which their poly(A)

tails are shortened.

All known PABPs contain four tandem and highly con-

served RNA-recognition motifs (RRMs) followed by a less

conserved Pro-rich C-terminal segment of variable length.

A variety of evidence suggests that PABP’s first two RRMs

support most of the biochemical functions of full-length

PABP. The X-ray structure of the first two RRMs of human

PABP (RRM1/2; the N-terminal 190 residues of this 636-

residue protein) in complex with A

, determined by

Stephen Burley, reveals that RRM1/2 forms a continuous

trough-shaped surface in which the poly(A) binds in an

extended conformation via interactions with conserved

Section 31-4. Post-Transcriptional Processing 1303

Figure 31-48 X-ray structure of the D154A mutant of yeast

poly(A) polymerase (PAP) in complex with A

and ATP. PAP is

drawn in ribbon form with its N-terminal domain lavender, its

central domain yellow-green, and its C-terminal domain light

blue. The A

and ATP are drawn in stick form with A

C white,

ATP C cyan, N blue, O red, and P orange. [Based on an X-ray

structure by Andrew Bohm,Tufts University School of Medicine.

PDBid 2Q66.]

JWCL281_c31_1260-1337.qxd 8/11/10 9:48 PM Page 1303

3′

5′

DNA

poly A

III

VII

mRNA

residues (Fig. 31-49). Each RRM, as also seen in the struc-

tures of a variety of other RNA-binding proteins, consists

of a compact globule made of a 4-stranded antiparallel

sheet that forms the RNA-binding surface backed by two

helices.

The cleavage of a transcript past its AAUAAA se-

quence does not, in itself, terminate transcription. How-

ever, in yeast, the protein Rtt103, which binds to the phos-

phorylated CTD of RNAP II, recognizes the AAUAAA

sequence and recruits the 5¿S3¿ exonuclease known as

Rat1 (Xrn2 in humans). Then, in what is termed the tor-

pedo model, the highly processive Rat1/Xrn2 loads onto

the newly liberated 5¿ end of the still nascent RNA and rap-

idly degrades it until it intercepts the RNAP and induces it

to terminate RNA synthesis. It has been hypothesized that

this occurs in much the same way as Rho factor terminates

bacterial transcription (Section 31-2Db). This frees the

RNAP to initiate a new round of transcription.

c. Eukaryotic Genes Consist of Alternating

Expressed and Unexpressed Sequences

The most striking difference between eukaryotic and

prokaryotic structural genes is that the coding sequences of

most eukaryotic genes are interspersed with unexpressed re-

gions. Early investigations of eukaryotic structural gene

transcription found, quite surprisingly, that primary tran-

scripts are highly heterogeneous in length (from ⬃2000 to

well over 20,000 nucleotides) and are much larger than was

expected from the known sizes of eukaryotic proteins.

Rapid labeling experiments demonstrated that little of this

so-called heterogeneous nuclear RNA (hnRNA) is ever

transported to the cytosol; most of it is quickly turned over

(degraded) in the nucleus. Yet, the hnRNA’s 5¿ caps and 3¿

tails eventually appear in cytosolic mRNAs. The straight-

forward explanation of these observations, that pre-mRNAs

are processed by the excision of internal sequences, seemed

so bizarre that it came as a great surprise in 1977 when

Phillip Sharp and Richard Roberts independently demon-

strated that this is actually the case. In fact, mammalian pre-

mRNAs typically contain eight noncoding intervening se-

quences (introns) whose aggregate length averages 4 to 10

times that of their flanking expressed sequences (exons).

This situation is graphically illustrated in Fig. 31-50, which

is an electron micrograph of chicken ovalbumin mRNA hy-

bridized to the antisense strand of the ovalbumin gene

(ovalbumin is the major protein component of egg white).

1304 Chapter 31. Transcription

Figure 31-49 X-ray structure of the N-terminal two

RNA-recognition motifs (RRMs) of human PABP in complex

with A

. RRM1 is cyan, RRM2 is gold, and their linking segment

is lavender.The poly(A), only nine of whose nucleotides are

observed, is drawn in stick form with C green, N blue, O red, and

P magenta. [Based on an X-ray structure by Stephen Burley, The

Rockefeller University. PDBid 1CVJ.]

Figure 31-50 An electron micrograph and its interpretive

drawing of a hybrid between the antisense strand of the

chicken ovalbumin gene and its corresponding mRNA. The

complementary segments of the DNA (purple line in the

drawing) and mRNA (red dashed line) have annealed to reveal

the exon positions (L, 1–7). The looped-out segments (I–VII),

which have no complementary sequences in the mRNA, are the

introns. [From Chambon, P., Sci. Am. 244(5), 61 (1981).]

JWCL281_c31_1260-1337.qxd 8/11/10 9:48 PM Page 1304

Exons have lengths that range up to 17,106 nt (in the

gene encoding the 34,350-residue muscle protein titin, the

largest known single-chain protein; Section 35-3Ae) but

with most ⬍300 nt (and averaging 150 nt in humans). In-

trons, in contrast, are usually much longer, with lengths av-

eraging ⬃3500 nt and as high as ⬃800,000 nt [in the gene

encoding the muscle protein dystrophin (Section 35-3Ae),

whose length is ⬃2400 kb and hence is the largest human

gene] with no obvious periodicity. Moreover, the corre-

sponding introns from genes in two vertebrate species can

vary extensively in both length and sequence so as to bear

little resemblance to one another.The number of introns in

a gene averages 7.8 in the human genome and varies from

none to 364 (with the latter number occurring in the gene

encoding titin).

The formation of eukaryotic mRNA begins with the

transcription of an entire structural gene, including its in-

trons, to form pre-mRNA (Fig. 31-51).Then, following cap-

ping, the introns are excised and their flanking exons are

connected, a process called gene splicing or just splicing,

that often occurs cotranscriptionally. The most striking as-

pect of gene splicing is its precision; if one nucleotide too few

or too many were excised, the resulting mRNA could not be

translated properly (Section 32-1B). Moreover, exons are

never shuffled; their order in the mature mRNA is exactly

the same as that in the gene from which it is transcribed.

d. Exons Are Spliced in a Two-Stage Reaction

Sequence comparisons of exon–intron junctions from a

diverse group of eukaryotes indicate that they have a high

degree of homology (Fig. 31-52), including, as Richard

Breathnach and Pierre Chambon first pointed out, an in-

variant GU at the intron’s 5¿ boundary and an invariant AG

at its 3¿ boundary. These sequences are necessary and suffi-

cient to define a splice junction: Mutations that alter the se-

quences interfere with splicing, whereas mutations that

change a nonjunction to a consensus-like sequence can

generate a new splice junction.

Section 31-4. Post-Transcriptional Processing 1305

Figure 31-51 The sequence of steps in the production of

mature eukaryotic mRNA as shown for the chicken ovalbumin

gene. Following transcription, the primary transcript is capped

Figure 31-52 The consensus sequence at the exon–intron

junctions of vertebrate pre-mRNAs. The subscripts indicate the

percentage of pre-mRNAs in which the specified base(s) occurs.

Note that the 3¿ splice site is preceded by a tract of 11

Splicing

Capping and polyadenylation

Transcription

Primary transcript (hnRNA)

7156

Ovalbumin mRNA

23 4

1872 nucleotides

5′ 3′

1L 24

Ovalbumin gene, 7700 bp

7531642L

VIIVIIV VII IIII

I IIIII IV

576

V VI VII

Exon

Intron

Cap

1L 24

I IIIII IV

576

V VI VII Poly(A)

tail

pre-mRNA

mRNA

DNA

5′ A

100

77–91

100

3′

Intron ExonExon

3′ splice

site

5′ splice

site

... ... ...

and polyadenylated.The introns are then excised and the exons

spliced together to form the mature mRNA. However, splicing

may also occur cotranscriptionally.

predominantly pyrimidine nucleotides. [Based on data from

Padgett, R.A., Grabowski, P.J., Konarska, M.M., Seiler, S.S., and

Sharp, P.A., Annu. Rev. Biochem. 55, 1123 (1986).]

JWCL281_c31_1260-1337.qxd 8/11/10 9:48 PM Page 1305

Investigations of both cell-free and in vivo splicing sys-

tems by Argiris Efstratiadis, Tom Maniatis, Michael Ros-

bash, and Sharp established that intron excision occurs via

two transesterification reactions that are remarkably simi-

lar from yeast to humans (Fig. 31-53):

1. The formation of a 2¿,5¿-phosphodiester bond be-

tween an intron adenosine residue and its 5¿-terminal phos-

phate group with the concomitant liberation of the 5¿

exon’s 3¿-OH group. The intron thereby assumes a novel

lariat structure. The adenosine residue at the lariat branch

has been identified in yeast as the last A in the highly con-

served sequence UACUAAC and in vertebrates as the A in

the equivalent but more permissive sequence YNCURAY

[where R represents purines (A or G),Y represents pyrim-

idines (C or U), and N represents any nucleotide]. In yeast

and vertebrates, the branch point A occurs ⬃50 and 18 to

40 residues upstream of the associated 3¿ splice site, respec-

tively. In yeast, which have relatively few introns, mutations

that change this branch point A residue abolish splicing at

that site. However, in higher eukaryotes, the mutation or

deletion of a branch site often activates a so-called cryptic

branch site that is also near the 3¿ splice site. Evidently, the

branch site functions to identify the nearest 3¿ splice site as

a target for linkage to the 5¿ splice site.

2. The now free 3¿-OH group of the 5¿ exon forms a

phosphodiester bond with the 5¿-terminal phosphate of the

3¿ exon yielding the spliced product and releasing the in-

tron lariat with a free 3¿-OH group.The intron lariat is then

debranched (linearized) and, in vivo, is rapidly degraded.

Mutations that alter the conserved AG at the 3¿ splice site

block this second step, although they do not interfere with

lariat formation.

Note that the splicing process proceeds without free en-

ergy input; its transesterification reactions preserve the

free energy of each cleaved phosphodiester bond through

the concomitant formation of a new one.

The sequences required for splicing are the short con-

sensus sequences at the 3¿ and 5¿ splice sites and at the

branch site. Nevertheless, these sequences are poorly con-

served. However, other short sequence elements within ex-

ons and introns that are known as exonic and intronic splic-

ing enhancers (ESEs and ISEs) and silencers (ESSs and

ISSs) also play important roles in splice site selection al-

though their characteristics are poorly understood (even

highly sophisticated computer programs are only ⬃50%

successful in predicting actual splice sites over apparently

equally good candidates that are not). In contrast, large

portions of most introns can be deleted without impeding

splicing.

e. Some Eukaryotic Genes Are Self-Splicing

It is now recognized that there are eight distinct types of

introns, seven of which occur in eukaryotes (Table 31-4).

Group I introns occur in the nuclei, mitochondria, and

chloroplasts of diverse eukaryotes (but not vertebrates),

and even in some bacteria. Thomas Cech’s study of how

1306 Chapter 31. Transcription

Figure 31-53 The sequence of

transesterification reactions that splice

together the exons of eukaryotic

pre-mRNAs. The exons and introns are

drawn in blue and orange, and R and Y

represent purine and pyrimidine residues.

(1) The 2¿-OH group of a specific intron A

residue nucleophilically attacks the

5¿-phosphate at the 5¿ intron boundary to

yield an unusual 2¿,5¿-phosphodiester bond

and thus form a lariat structure. (2) The

liberated 3¿-OH group forms a 3¿,5¿-

phosphodiester bond with the 5¿ terminal

residue of the 3¿ exon, thereby splicing the

two exons together and releasing the

intron in lariat form with a free 3¿-OH.

5′

...

ApG OH

3′

5′

...

ApG pGpUp YpNpCpUpRpApYp ApG pGp

...

3′

Exon 1 Intron

Exon 2

3′ splice

site

5′ splice

site

(2′,5′)

Exon 2

ApG pGp

Intron

Exon 1

pGpUp

(2′,5′)

ApG

3′

+5′ ApG pGp 3′

Exon 2Exon 1

Splice junction

Excised intron in lariat form

......

+ Spliced exons

pre-mRNA

3′

...

2′

pGpUp

CpUpRpApYp

JWCL281_c31_1260-1337.qxd 8/11/10 9:48 PM Page 1306

group I introns are spliced in the ciliated protozoan

Tetrahymena thermophila led to an astonishing discovery:

RNA can act as an enzyme. When the isolated pre-rRNA of

this organism is incubated with guanosine or a free guanine

nucleotide (GMP, GDP, or GTP), but in the absence of pro-

tein, its single 421-nucleotide intron excises itself and splices

together its flanking exons; that is, this pre-rRNA is self-

splicing. The three-step reaction sequence of this process

(Fig. 31-54) resembles that of mRNA splicing:

1. The 3¿-OH group of the guanosine forms a phospho-

diester bond with the intron’s 5¿ end, liberating the 5¿ exon.

2. The 3¿-terminal OH group of the newly liberated 5¿

exon forms a phosphodiester bond with the 5¿-terminal

phosphate of the 3¿ exon, thereby splicing together the two

exons and releasing the intron.

3. The 3¿-terminal OH group of the intron forms a phos-

phodiester bond with the phosphate of the nucleotide 15

residues from the intron’s 5¿ end, yielding the 5¿-terminal

fragment with the remainder of the intron in cyclic form.

This self-splicing process consists of a series of transesteri-

fications and therefore does not require free energy input.

Cech further established the enzymatic properties of the

Section 31-4. Post-Transcriptional Processing 1307

Figure 31-54 The sequence of reactions in the self-splicing of

Tetrahymena group I intron. (1) The 3¿-OH group of a guanine

nucleotide attacks the intron’s 5¿-terminal phosphate so as to

form a phosphodiester bond and release the 5¿ exon. (2) The

newly generated 3¿-OH group of the 5¿ exon attacks the

5¿-terminal phosphate of the 3¿ exon, thereby splicing the two

exons and releasing the intron. (3) The 3¿-OH group of the intron

attacks the phosphate of the nucleotide that is 15 residues from

the 5¿ end so as to cyclize the intron and release its 5¿-terminal

fragment.Throughout this process, the RNA maintains a folded,

internally hydrogen bonded conformation that permits the

precise excision of the intron.

Table 31-4 Types of Introns

Intron Type Where Found

GU–AG introns Eukaryotic nuclear pre-mRNA

AU–AC introns Eukaryotic nuclear pre-mRNA

Group I Eukaryotic nuclear pre-mRNA,

organelle RNAs, a few

bacterial RNAs

Group II Organelle RNAs, a few

prokaryotic RNAs

Group III Organelle RNAs

Twintrons (composites Organelle RNAs

of two and/or more

group II or III introns)

Pre-tRNA introns Eukaryotic nuclear pre-tRNAs

Archaeal introns Various RNAs

Source: Brown,T.A., Genomes (3rd ed.), Garland Science, p. 355 (2007).

3′

Left exon

Right

exon

Intron

5′

3′

5′

Left exon Right exon

5′ 3′

CUCUCU pU

Spliced exons

3′

GGGAGG

Cyclized intron

pA UUU

3′5′

5′

JWCL281_c31_1260-1337.qxd 8/11/10 9:48 PM Page 1307

Tetrahymena intron, which stem from its three-dimensional

structure, by demonstrating that it catalyzes the in vitro

cleavage of poly(C) with an enhancement factor of 10

over the rate of spontaneous hydrolysis. Indeed, this RNA

catalyst even exhibits Michaelis–Menten kinetics (K

⫽ 42

␮M and k

cat

⫽ 0.033 s

⫺1

for C

). Such RNA enzymes have

been named ribozymes.

Although the idea that an RNA can have enzymatic

properties may seem unorthodox, there is no fundamental

reason why an RNA, or any other macromolecule, cannot

have catalytic activity (recall that it was likewise once gen-

erally accepted that nucleic acids lack the complexity to

carry hereditary information; Section 5-2). Of course, in or-

der to be an efficient catalyst, a macromolecule must be

able to assume a stable structure but, as we shall see below

and in Sections 32-2B and 32-3Ae, RNAs, including tRNAs

and rRNAs, can do so. In fact, the Tetrahymena intron un-

dergoes a series of well-defined conformational changes

during its reaction sequence. [Synthetic ssDNAs are also

known to have catalytic properties although such deoxyri-

bozymes are unknown in biology.]

The group II introns, which occur in the mitochondria of

fungi and plants and comprise the majority of the introns in

chloroplasts, are also self-splicing. They generally employ

an internal A residue as their initial attacking nucleophile

(instead of an external G) to form a lariat intermediate, a

process that resembles the splicing of nuclear pre-mRNAs

(Fig. 31-53). We shall see below that nuclear pre-mRNA

splicing is mediated by complex ribonucleoprotein parti-

cles known as spliceosomes. The chemical similarities of

the pre-mRNA and group II intron splicing reactions

therefore suggest that spliceosomes are ribozymal systems

whose RNA components have evolved from primordial self-

splicing RNAs and that their protein components serve

mainly to fine-tune ribozymal structure and function. Simi-

larly, the RNA components of ribosomes, which are two-

thirds RNA and one-third protein, clearly have a catalytic

function in addition to the structural and recognition roles

traditionally attributed to them (Section 32-3). Thus, the

observations that nucleic acids but not proteins can direct

their own synthesis, that cells contain batteries of protein-

based enzymes for manipulating DNA but relatively few

for processing RNA, and that many coenzymes are ribonu-

cleotides (e.g., ATP, NAD

⫹

, and CoA), led to the hypothe-

sis that RNAs were the original biological catalysts in

precellular times (the RNA world) and that the chemically

more versatile proteins were relative latecomers in macro-

molecular evolution (Section 1-5Ca).

f. The X-Ray Structures of a Group I Ribozyme

Group I introns are the most abundant self-splicing in-

trons, with ⬎2000 such sequences known. The sequence of

the 413-nt Tetrahymena group I intron, together with phy-

logenetic comparisons, indicates that it contains nine dou-

ble helical segments that are designated P1 through P9

(Fig. 31-55a; P for base-paired segment). Such analysis fur-

ther indicates that the conserved catalytic core of group I

introns consists of sets of coaxially stacked helices inter-

spersed with internal loops that are organized into two

domains, the P4-P5-P6 domain (also called P4-P6) and the

P3-P7-P8-P9 domain (also called P3-P9).

Cech designed a 247-nt RNA (Fig. 31-55a) that encom-

passes both the P4-P6 and P3-P9 domains of the Tetrahy-

mena group I intron (it lacks the P1-P2 domain and the at-

tached exons), with the addition of a 3¿ G (␻G), which

functions as an internal guanosine nucleophile. This RNA

is catalytically active; it binds the P1-P2 domain via tertiary

interactions and, with the assistance of ␻G, cleaves P1 in a

manner similar to the intact intron.

The X-ray structure of this RNA (Fig. 31-55b,c) reveals

that it is largely composed of three coaxially stacked sets of

A-RNA-like helices with P4-P6 consisting of two pseudo-

continous and straight parallel helices connected by a

sharp bend and P3-P9 consisting of a curved helix that

wraps around one side of P4-P6 through extensive interdo-

main interactions that form the ribozyme’s active site. Of

particular note are its so-called A-rich bulge, a 7-nt se-

quence about halfway along the short arm of the U-shaped

P4-P6, and the 6-nt sequence at the tip of the short arm of

the U, whose central GAAA assumes a characteristic con-

formation known as a tetraloop. In both of these substruc-

tures, the bases are splayed outward so as to stack on each

other and to associate in the minor groove of specific seg-

ments of the long arm of the U via hydrogen bonding inter-

actions involving ribose residues as well as bases. In many

such interactions, the close packing of phosphate groups is

mediated by hydrated Mg

2⫹

ions. Throughout this struc-

ture, the defining characteristic of RNA, its 2¿-OH group, is

both a donor and an acceptor of hydrogen bonds to phos-

phates, bases, and other 2¿-OH groups. Interestingly, al-

though this overall fold is highly conserved among group I

introns, their sequences are poorly conserved with the

exception of a few crucial active site residues.

In the initial reaction catalyzed by group I ribozymes,

the 3¿-OH group of ␻G nucleophilically attacks the phos-

phate group linking the 5¿ exon to the ribozyme (Fig. 31-54).

But how is only this 3¿-OH group activated as a nucle-

ophile? The binding site for the ␻G substrate is composed

of four coplanar base triples with the ␻G–G264–C311

triple sandwiched by three other base triples (Fig. 31-56).

Consequently, the base of ␻G is stacked between those of

A261 and C262 (Fig. 31-56a), which stabilizes the binding

of ␻G to this site through base stacking.

Divalent metal ions, usually Mg

2⫹

, are often required for

both the structural stability and the catalytic activity of ri-

bozymes. Unfortunately, the relatively low (3.8 Å) resolu-

tion of the ribozyme X-ray structure precluded the direct

observation of Mg

2⫹

ions (which have the same number of

electrons as water molecules). However, there is good evi-

dence that the heavy metal ions (Eu and Ir) used to solve

the X-ray structure occupied many of the same positions in

the ribozyme as do Mg

2⫹

ions and moreover, several of

these sites were observed to contain Mg

2⫹

ions in other

group I introns whose X-ray structures are known. In par-

ticular, an Mg

2⫹

ion is in contact with the 2¿-OH group of

␻G as well as being liganded by the phosphate groups of

1308 Chapter 31. Transcription

JWCL281_c31_1260-1337.qxd 8/11/10 9:48 PM Page 1308

three surrounding nucleotides. This both orients the ribose

group of ␻G and nucleophilically activates its 3¿-OH

group. Biochemical studies indicate that a second Mg

2⫹

ion, which accompanies the phosphate group of the RNA

substrate, also participates in the catalytic reaction. Note

that two Mg

2⫹

ions similarly participate in the phosphoryl-

transfer reactions catalyzed by protein enzymes such as

DNA polymerase (Section 30-2Af).

Section 31-4. Post-Transcriptional Processing 1309

Figure 31-55 The group I intron from Tetrahymena

thermophila. (a) The secondary structure of the 414-nt ribozyme.

Its P4-P6 and P3-P9 domains are shaded in blue and green,

respectively, with the catalytically active 3¿ ␻G residue shaded in

yellow, the base triples of the P7 domain shaded in red, and the

A-rich bulge, the tetraloop, and the tetraloop receptor on the

P4-P6 domain shaded in brown, pink, and cyan, respectively.

Watson–Crick and non-Watson–Crick base pairing interactions

are represented by short horizontal lines and small filled circles,

whereas interdomain interactions are indicated by magenta lines.

Every tenth residue is marked by an outwardly pointing dash.

The positions of five residues that have been mutated to stabilize

the ribozyme structure are circled and those of the seven

mutations that facilitated crystallization are bracketed (this

mutant form retains its catalytic activity). (b) The X-ray structure

of the ribozyme, drawn in cartoon form with its bases shown as

paddles, and colored as in Part a. The inferred positions of Mg

2⫹

ions are represented by orange spheres. (c) As in Part b but

rotated 140° about the vertical axis to better show the A-rich

bulge and the interaction between the tetraloop and the

tetraloop receptor. [Part a modified from a drawing by and Parts

b and c based on an X-ray structure by Thomas Cech, University

of Colorado. PDBid 1X8W.]

See Interactive Exercise 42

(a)

(b)

(c)

JWCL281_c31_1260-1337.qxd 10/19/10 11:26 AM Page 1309

g. Hammerhead Ribozymes Catalyze

an In-Line Nucleophilic Attack

One of the best-characterized types of ribozymes is the

hammerhead ribozyme, so called because of the superficial

resemblance of its secondary structure, as it was originally

laid out, to a hammer.This minimally ⬃40-nt RNA partici-

pates in the replication of certain viruslike RNAs that in-

fect plants and also occurs in schistosomes (species of par-

asitic flatworms).The hammerhead ribozyme catalyzes the

site-specific cleavage of one of its own phosphodiester

bonds with an ⬃10

-fold rate enhancement. However, it is

not a true catalyst because it cannot return to its original

state.

The secondary structure of the 63-nt hammerhead ri-

bozyme from Schistosoma mansoni has three duplex

stems and an active site core of two nonhelical segments

(Fig. 31-57a). This ribozyme cleaves itself between its C-17

and C-1.1 nucleotides to yield a cyclic 2¿,3¿-phosphodiester

on C-17 with inversion of configuration about the P atom,

together with a free 5¿-OH on C-1.1, much like the interme-

diate product in the RNA hydrolysis reaction catalyzed by

RNase A (Section 15-1Ab). This suggests that the reaction

proceeds via an “in-line” mechanism such as that dia-

grammed in Fig. 16-6b with the transition state forming a

trigonal bipyramidal intermediate in which the attacking

nucleophile, the 2¿-OH group (Y in Fig. 16-6b), and the

leaving group, which forms the free 5¿-OH group (X in Fig.

16-6b), occupy the axial positions.

The X-ray structure of the S. mansoni hammerhead ri-

bozyme, determined by William Scott, reveals that its stem

II, stem III and the terminal end of stem I are coaxially

stacked.The remainder of stem I curves around in a manner

that more closely resembles the handle of a suitcase than

that of a hammer to form a junction with stems II and III

that contains the ribozyme’s active site core (Fig. 31-57b).

The nucleotides in the helical stems mainly form normal

Watson–Crick base pairs, whereas the nucleotides of the

active site core participate in non-Watson–Crick base pairs.

This explains the observations that most helical positions

can be occupied by any Watson–Crick base pair but that

few core bases can be changed without significantly reduc-

ing ribozymal activity.

The bases of the active site core participate in a hydro-

gen bonded network (Fig. 31-57c). This helps position C-17

such that its O2¿ atom is properly oriented for an in-line

1310 Chapter 31. Transcription

Figure 31-56 Structure of the guanosine binding site in the Tetrahymena

group I intron. (a) Side view of the four coplanar base triples at the

ribozymal active site in P7 (colored in red in Fig. 31-55). The sugar–phosphate

backbone is drawn in stick form and the bases are drawn in paddle form.The

nucleotides at each level have the same color with the exception of the

C413–␻G dinucleotide, which is yellow. (b) Top view of the base triple

interactions between ␻G and the G264 ⴢ C311 base pair. (c) Top view of the

base triple of C262,A263, and G312, which is located above that containing

␻G. (d) Top view of the base triple of A261, A265, and U310, which is located

below that containing ␻G. [Courtesy of Thomas Cech, University of

Colorado. PDBid 1X8W.]

(a)

(b)

(c)

(d)

JWCL281_c31_1260-1337.qxd 8/11/10 9:48 PM Page 1310