Voet D., Voet Ju.G. Biochemistry

Подождите немного. Документ загружается.

nucleophilic attack on the P atom that links atom O3¿ of

C-17 to atom O5¿ of C-1.1. The N1 of the invariant G-12

when deprotonated and the 2¿-OH of the invariant G-8 ap-

pear to be properly positioned to respectively act as base

and acid catalysts in this reaction (Fig. 31-57c), which

strongly suggests that the reaction occurs via a concerted

acid–base catalyzed mechanism. This reaction mechanism

does not involve the participation of metal ions and none

are observed in the ribozyme’s catalytic core. However, in

solution, the presence of divalent metal ions provides an

⬃50-fold enhancement rate over the presence of only

monovalent metal ions. Perhaps divalent metal ions stabi-

lize the negative charge on the trigonal bipyramidal inter-

mediate (Fig. 16-6b) and/or they may help position and

orient reactive groups.

h. Splicing of Pre-mRNAs Is Mediated by snRNPs

in the Spliceosome

How are the splice junctions of pre-mRNAs recognized

and how are the two exons to be joined brought together in

the splicing process? Part of the answer to this question

was established by Joan Steitz going on the assumption

that one nucleic acid is best recognized by another.The eu-

karyotic nucleus, as has been known since the 1960s,

contains numerous copies of several highly conserved

60- to 300-nucleotide RNAs called small nuclear RNAs

Section 31-4. Post-Transcriptional Processing 1311

Figure 31-57 Structure of the Schistosoma mansoni

hammerhead ribozyme. (a) The sequence and schematic

structural organization of the ribozyme colored to match the

X-ray structure drawn in Part b. Base pairs and tertiary

interactions are represented by dashes and dotted lines,

respectively. Nucleotides are labeled according to the universal

numbering system for hammerhead ribozymes. (b) X-ray

structure of the ribozyme drawn in paddle form with C atoms the

same color as in Part a except that those of G-12 are light green

and those of C-1.1 are light blue, N blue, O red, and P orange.

Adjacent P atoms in the same strand are connected by thin

orange rods. The P atom of the scissile phosphate group

(bridging C-17 and C-1.1) and the nucleophile (O2¿ of C-17) are

represented by small spheres. (c) The ribozyme’s active site

residues, which are drawn in stick form with C green, O red,

N blue, and P orange. Hydrogen bonds are represented by

dashed black lines.The proposed reaction mechanism is

indicated by the curved arrows with the dotted black line

marking the in-line trajectory taken by atom O2¿ of C-17 in

nucleophilically attacking the P atom of the scissile phosphate

group. [Part a courtesy of and Parts b and c based on an X-ray

structure by William Scott, University of California at Santa

Cruz. PDBid 2GOZ.]

See Interactive Exercise 43

10.1

10.110.111.1

1212

1313

1414

15.1

16.1

16.116.1

2.1

2.12.1

1.1

1.11.1

L3L3

L2L2

L6L6

L5L5

L4L4

B5B5

B4B4

11.2

11.3

10.2

10.210.2

10.3

10.310.3

scissile

bond

Uridine

turn

Stem I

Stem III

Stem I

Stem II

5'3'

(a)

(

)

JWCL281_c31_1260-1337.qxd 10/19/10 10:41 AM Page 1311

(snRNAs), which form protein complexes termed small

nuclear ribonucleoproteins (snRNPs; pronounced “snurps”).

Steitz recognized that the 5¿ end of one of these snRNAs,

U1-snRNA (so called because it is a member of a U-rich

subfamily of snRNAs), is partially complementary to the

consensus sequence of the 5¿ splice site. The consequent

hypothesis, that U1-snRNA recognizes the 5¿ splice site, was

corroborated by the observations that splicing is inhibited

by the selective destruction of the U1-snRNA sequences

that are complementary to the 5¿ splice site or by the pres-

ence of anti-U1-snRNP antibodies (produced by patients

suffering from systemic lupus erythematosus, an often fatal

autoimmune disease). Three other snRNPs are also impli-

cated in splicing: U2-snRNP, U4–U6-snRNP (in which

the U4- and U6-snRNAs associate via base pairing), and

U5-snRNP.

Splicing takes place in an as yet poorly characterized

⬃2700 kD particle dubbed the spliceosome (Fig. 31-58).

The spliceosome brings together a pre-mRNA, the forego-

ing four snRNPs, and a variety of pre-mRNA binding pro-

teins. Note that the spliceosome, which consists of 5 RNAs

and ⬃150 polypeptides, is comparable in size and complex-

ity to the ribosome (which in E. coli consists of 3 RNAs and

52 polypeptides with an aggregate mass of ⬃2500 kD; Sec-

tion 32-3A).

In addition to its size and complexity, the spliceosome is a

highly dynamic entity, with its various components associ-

ating and dissociating during specific stages of the splicing

reaction (Fig. 31-59), while undergoing a variety of ATP-

driven conformational changes. For example, to carry out

the first transesterification reaction yielding the lariat

structure (Fig. 31-53), the spliceosome undergoes a com-

plex series of rearrangements that are schematically dia-

grammed in Fig. 31-60. Similarly extensive rearrangements

are required to carry out the second transesterification

reaction and to recycle the spliceosome for subsequent

splicing reactions.

Although spliceosomal transesterification reactions

were initially assumed to be mediated by protein catalysts,

their chemical resemblance to the reactions carried out by

the self-splicing group II introns suggests, as is noted

above, that it is really the snRNAs that catalyze the splicing

of pre-mRNAs (pre-mRNA introns have such varied se-

quences outside of their splice and branch sites that they

are unlikely to play an active role in splicing). In fact, James

Manley has shown that, in the absence of protein, segments

of human U2- and U6-snRNAs catalyze an Mg

2⫹

-depend-

ent reaction in an intron branch site sequence–containing

RNA that resembles splicing’s first transesterification

reaction.

i. Splicing Also Requires the Participation of

Splicing Factors

Around 170 different proteins known as splicing-associ-

ated factors that are extrinsic to spliceosomes also partici-

pate in splicing, with individual assembly intermediates

(e.g., complexes A, B, and C in Fig. 31-59) each associated

with ⬃125 such proteins. Among them are branch

point–binding protein [BBP; also known as splicing factor

1 (SF1)] and U2-snRNP auxiliary factor (U2AF), which

cooperate to select the intron’s branch point. U2AF binds

to the polypyrimidine tract upstream of the 3¿ splice site

(Fig. 31-52), whereas BBP recognizes the nearby branch

point sequence (Figs. 31-53 and 31-60). The NMR structure

of the 131-residue RNA-binding segment of the 638-residue

BBP in complex with an 11-nt RNA containing a branch

1312 Chapter 31. Transcription

Figure 31-58 An electron micrograph of spliceosomes in

action. A Drosophila gene that is ⬃6 kb long enters from the

upper left of the micrograph and exits at the lower left.

Transcription initiates near the point marked by an asterisk. The

growing RNA chains appear as fibrils of increasing lengths that

emanate from the DNA. The transcripts are undergoing

cotranslational splicing as revealed by the progressive formation

and loss of intron loops near the 5¿ ends of the RNA transcripts

(arrows).The beads at the base of each intron loop as well as

elsewhere on the transcripts are the spliceosomes.The large

arrow points to a transcript near the 3¿ end of the gene that is no

longer attached to the DNA template and hence appears to have

recently been terminated and released.The bar is 200 nm long.

[Courtesy of Ann Beyer and Yvonne Osheim, University of

Virginia.]

JWCL281_c31_1260-1337.qxd 8/11/10 9:48 PM Page 1312

Section 31-4. Post-Transcriptional Processing 1313

Figure 31-59 The spliceosomal assembly/disassembly cycle.

The sequential actions of the spliceosomal snRNPs (colored

circles), but not the non-snRNP proteins, are diagrammed in the

process of excising an intron from a pre-RNA containing two

exons (blue). Here 5¿SS, BP, and 3¿SS stand for the pre-mRNA’s

5¿ splice site, its branch point, and its 3¿ splice site, respectively.

Figure 31-60 Schematic diagram of six rearrangements

that the spliceosome undergoes in mediating the first

transesterification reaction in pre-mRNA splicing. The RNA is

color coded to indicate segments that become base-paired.The

black and green lines represent snRNA and pre-mRNA, and

BBP stands for branch point–binding protein. U5, which

participates in the second transesterification reaction, has been

omitted for clarity. (1) Exchange of U1 for U6 in base pairing to

the intron’s 5¿ splice site. (2) Exchange of BBP for U2 in binding

to the intron’s branch site. (3) Intramolecular rearrangement in

3' exon

5' exon

3' exon

BBP

Eight conserved DExD/H box–containing RNA-dependent

RNA ATPases/helicases as well as the GTPase Snu114 act in

specific steps of the splicing cycle to motivate RNA–RNA

rearrangements and RNP remodeling reactions. [Courtesy of

Reinhard Lührmann, Max-Planck-Institut für biophysikalische

Chemie, Göttingen, Germany.]

U2. (4) Disruption of a base-paired stem between U4 and U6 to

form a stem–loop in U6. (5) Disruption of a second stem

between U4 and U6 to form a stem between U2 and U6.

(6) Disruption of a stem–loop in U2 to form a second stem

between U2 and U6.The order of these rearrangements is

unclear.The transesterification reaction is represented by the

arrow from the A in the yellow segment of the pre-mRNA (right

panel) to the 3¿ end of the 5¿ exon. [Adapted from Staley, J.P. and

Guthrie, C., Cell 92, 315 (1998).]

JWCL281_c31_1260-1337.qxd 8/11/10 9:48 PM Page 1313

point sequence, determined by Michael Sattler, reveals that

the RNA assumes an extended conformation and is largely

buried in a groove that is lined with both aliphatic and basic

residues (Fig. 31-61). The branch point adenosine, whose

mutation abolishes BBP binding,is deeply buried and binds

to BBP via hydrogen bonds that mimic Watson–Crick base

pairing with uracil.

Other splicing factors include SR proteins and several

members of the heterogeneous nuclear ribonucleoprotein

(hnRNP) family. SR proteins each have one or more

RRMs (RNA recognition motifs) near their N-terminus

and a distinctive C-terminal RS domain that contains nu-

merous Ser-Arg (SR) repeats and which participates in

protein–protein interactions. SR proteins, when appropri-

ately phosphorylated on their RS domains, specifically

bind to their corresponding exonic splicing enhancers

(ESEs) via their RRMs and thereby recruit the splicing

machinery to the flanking 5¿ and 3¿ splice sites. The hnRNP

proteins, which are highly abundant RNA-binding pro-

teins, lack RS domains and hence cannot recruit the splic-

ing machinery. Instead, they bind to their corresponding

ESSs and ISSs (exonic and intronic splicing silencers) so as

to block the binding of the splicing machinery at the flank-

ing splice sites.

A simplistic interpretation of Fig. 31-53 suggests that

any 5¿ splice site could be joined with any following 3¿

splice site, thereby eliminating all the intervening exons

together with the introns joining them. However, such

exon skipping does not normally take place (but see be-

low). Rather, all of a pre-mRNA’s introns are individually

excised in what appears to be a largely fixed order that

more or less proceeds in the 5¿S3¿ direction.This occurs,

at least in part, because splicing takes place cotranscription-

ally (Fig. 31-58).Thus, as a newly synthesized exon emerges

from an RNAP II, it is bound by splicing factors that are

also bound to the RNAP II’s highly phosphorylated C-

terminal domain (CTD; Section 31-2E). This tethers the

exon and its associated spliceosome to the CTD so as to

ensure that splicing occurs when the next exon emerges

from the RNAP II.

j. Spliceosomal Structures

All four snRNPs involved in pre-mRNA splicing con-

tain the same so-called snRNP core protein, which consists

of seven Sm proteins (so called because they react with au-

toantibodies of the Sm serotype from patients with sys-

temic lupus erythematosis), which are named B, D1, D2,

D3, E, F, and G proteins. Each of these Sm proteins con-

tains two conserved segments, Sm1 and Sm2, that are sepa-

rated by a linker of variable length.The seven Sm proteins

collectively bind to a conserved RNA sequence, the Sm

RNA motif, which occurs in U1-,U2-,U4-, and U5-snRNAs

and which has the single-stranded sequence AAUU-

UGUG. However, in the absence of a U-snRNA, the Sm

proteins form three stable complexes D1–D2, D3–B, and

E–F–G. None of these complexes alone bind U-snRNA.

However, the D1–D2 and E–F–G complexes form a stable

subcore snRNP with U-snRNA, to which D3–B binds to

form the complete Sm core domain.

The X-ray structures of the D3–B and D1–D2 het-

erodimers, determined by Reinhard Lührmann and

Kiyoshi Nagai, reveals that these four proteins share a

common fold which consists of an N-terminal helix fol-

lowed by a 5-stranded antiparallel ␤ sheet that is strongly

bent so as to form a hydrophobic core (Fig. 31-62a). The

subunits of both dimers associate in a similar manner

with the ␤5 strands of D3 and D1 binding to the ␤4

strands of B and D2, respectively, so as to join their ␤

1314 Chapter 31. Transcription

Figure 31-61 The NMR structure of the RNA-binding portion

of human branch point–binding protein (BBP) in complex with

its target RNA. The 11-nt RNA contains the sequence

5¿-UAU

ACUAACAA-3¿ in which the branch site sequence for

both yeast and vertebrates is underlined and the branch point A

is in bold.The protein is drawn in cartoon form colored in

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

The RNA is drawn in paddle form with C green except for the C

atoms of the branch point A, which are yellow, N blue, and O red

and with successive P atoms connected by an orange rod.The

branch point O2¿ is represented by a small red sphere. [Based on

an NMR structure by Michael Sattler, European Molecular

Biology Laboratory, Heidelberg, Germany. PDBid 1K1G.]

JWCL281_c31_1260-1337.qxd 8/11/10 9:48 PM Page 1314

sheets (Fig. 31-62b). This, together with biochemical and

mutagenic experiments, indicates that the Sm proteins

form a closed heteroheptameric ring whose subunits are

arranged in the order –B–D3–G–E–F–D2–D1–. This

model is corroborated by the X-ray structure of an Sm-

like protein from the hyperthermophilic archeon Py-

robaculum aerophilum, determined by David Eisenberg,

that forms a homoheptameric ring that is structurally

similar to the heteroheptameric model. This structure

also supports the hypothesis that the seven eukaryotic

Sm proteins arose through a series of duplications of an

archaeal Sm-like protein gene.

Mammalian U1-snRNP consists of U1-snRNA and ten

proteins, the seven Sm proteins that are common to all

U-snRNPs as well as three that are specific to U1-snRNP:

U1-70K, U1-A, and U1-C (437, 282, and 159 residues, re-

spectively, in humans). The predicted secondary structure

of the 165-nt U1-snRNA contains five double helical stems,

four of which come together at a 4-way junction (Fig.31-63a).

U1-70K and U1-A bind directly to RNA stem–loops 1 and

2 (SL1 and SL2), respectively, whereas U1-C is bound by

other proteins.

Nagai determined the X-ray structure of human U1

snRNP at 5.5 Å resolution.At this low resolution the major

and minor grooves of the double-stranded RNA stems are

visible. Moreover, the helices and sheets of the proteins are

apparent, which allows the placement of protein folds of

known structure. SL2 of the U1-snRNA had been altered

and shortened to promote crystallization. This eliminated

the binding site for U1-A. However, U1 snRNP in which

U1-A is depleted is active in a splicing assay.

SL1 and SL2 are stacked coaxially as are SL3 and helix

H, and these two stacked helices cross at an angle of ⬃90⬚.

The Sm proteins form the predicted heteroheptameric

ring, which is ⬃70 Å in diameter (Fig. 31-63b). The Sm

RNA motif together with SL4 are threaded through the

ring’s funnel-shaped central hole such the Sm RNA motif

interacts with the Sm proteins (Fig. 31-63c). U1-C associ-

ates with the Sm ring via an interaction between its zinc fin-

ger domain and the D3 subunit (Fig. 31-63c). An RRM in

the C-terminal segment of U1-70K interacts with the loop

of SL1 and this protein’s N-terminal segment is draped

across the external face of the Sm ring.

Fortuitously, the 5¿ ends of two neighboring U1-snRNAs

in the crystal form a double helical segment. Since, in the

spliceosome, the 5¿ end of U1-snRNA base-pairs with the 5¿

splice site of the pre-RNA (Fig. 31-60), this interaction pro-

vides a model of how the spliceosome recognizes the 5¿

splice site. The zinc finger domain of U1-C interacts with

this double helix (left side of Fig. 31-63c) and presumably

stabilizes it. This is consistent with the observation that

mutants of U1-C in this region cannot initiate spliceosome

formation (the first reaction in Fig. 31-59).

Difficulties in obtaining homogeneous preparations of

spliceosomal subassemblies and their poor stabilities

have hindered their crystallization and limited the resolu-

tion of their cryo-EM–based images. Nevertheless, the

cryo-EM–based structures of a variety of spliceosomal

components have been reported. For example, the ⬃40-Å

resolution structure of human complex B after it has re-

leased U1-snRNP but before it has released U4-snRNP

(Fig. 31-59) was determined by Lührmann and Holger

Section 31-4. Post-Transcriptional Processing 1315

Figure 31-62 X-ray structures of Sm proteins. (a) The

structure of D3 protein.The N-terminal helix and the ␤ strands

of its Sm1 domain are red and blue and the ␤ strands of its Sm2

domain are yellow. The B, D1, and D2 Sm proteins have similar

structures with their L4 loops and N-terminal segments,

including helix A, comprising their most variable portions.

Several highly conserved residues are shown in stick form (with

C gray, N blue, and O red), and a conserved hydrogen bonding

network is represented by green dotted lines. (b) The D3–B

dimer with D3 gold and B blue. The ␤5 strand of D3 associates

with the ␤4 strand of B to form a continuous antiparallel ␤ sheet.

Note that their corresponding loops extend in similar directions.

[Courtesy of Kiyoshi Nagai, MRC Laboratory of Molecular

Biology, Cambridge, U.K. PDBid 1D3B.]

(a)

(b)

JWCL281_c31_1260-1337.qxd 8/11/10 9:48 PM Page 1315

Stark (Fig. 31-64). This 370 ⫻ 270 ⫻ 170 Å particle, which

is named B¢U1, has a roughly triangular body connected

to a head domain. Other splicesome components are

equally irregular.

k. The Significance of Gene Splicing

The analysis of the large body of known DNA se-

quences reveals that introns are rare in prokaryotic struc-

tural genes, uncommon in lower eukaryotes such as yeast

(which has a total of 239 introns in its ⬃6600 genes and,

with two exceptions, only one intron per polypeptide), and

abundant in higher eukaryotes (the only known vertebrate

structural genes lacking introns are those encoding his-

tones and the antiviral proteins known as interferons). Pre-

mRNA introns, as we have seen, can be quite long and

many genes contain large numbers of them. Consequently,

unexpressed sequences constitute ⬃80% of a typical verte-

brate structural gene and ⬃99% of a few of them.

1316 Chapter 31. Transcription

Figure 31-64 Cryo-EM–based structure of the human

spliceosome at its B⌬U1 stage of assembly/disassembly at 40 Å

resolution. The bar represents 100 Å. [Courtesy of Holger Stark,

Max-Planck Institute for Biophysical Chemistry, Göttingen,

Germany.]

Figure 31-63 X-ray structure of human U1-snRNP. (a) The

predicted secondary structure of U1-snRNA with the RNA

segments at which the proteins U1-70K and U1-A bind

indicated. ⌿ is the symbol for pseudouridine residues (Section

30-5Be). (b) View of the X-ray structure along the axis of the Sm

ring.The proteins are drawn in ribbon form in different colors

and the RNA is shown in paddle form in gray. SL4 has been

deleted for clarity. The yellow spheres represent the selenium

atoms in selenoMet residues (Met with its S atom replaced by

Se), which were mutagenically inserted into the proteins to aid in

solving the structure. Their known positions in the polypeptide

chains helped trace the polypeptides’ paths. (c) The structure as

viewed from above in Part a.The Zn

2⫹

ion bound to the zinc

finger motif of U1-C is represented by a green sphere. The 5¿

terminal portion of the U1-snRNA forms a double helical

segment with the 5¿ end of a neighboring U1-snRNA, which is

drawn in paddle form in purple. [Part a modified from a drawing

by, and Parts b and c courtesy of, Kiyoshi Nagai, MRC

Laboratory of Molecular Biology, Cambridge, U.K. PDBid

3CW1.]

(a)

(b)

(c)

JWCL281_c31_1260-1337.qxd 8/11/10 9:48 PM Page 1316

The argument that introns are only molecular parasites

(junk DNA) seems untenable since it would then be difficult

to rationalize why the evolution of complex splicing machin-

ery offered any selective advantage over the elimination of

the split genes. What then is the function of gene splicing?

Although, since its discovery, the significance of gene splic-

ing has been often vehemently debated, two important roles

for it have emerged: (1) It is an agent for rapid protein evo-

lution; and (2) through alternative splicing, it permits a sin-

gle gene to encode several (sometimes many) proteins that

may have significantly different functions. In the following

paragraphs, we discuss these aspects of gene splicing.

l. Many Eukaryotic Proteins Consist of Modules

That Also Occur in Other Proteins

The 839-residue LDL receptor is a plasma membrane

protein that functions to bind low-density lipoprotein

(LDL) to coated pits for transport into the cell via endocyto-

sis (Section 12-5Bc). LDL receptor’s 45-kb gene contains 18

exons, most of which encode specific functional domains of

the protein. Moreover, 13 of these exons specify polypeptide

segments that are homologous to segments in other proteins:

1. Five exons encode a 7-fold repeat of a 40-residue se-

quence that occurs once in complement C9 (an immune

system protein; Section 35-2F).

2. Three exons each encode a 40-residue repeat similar

to that occurring four times in epidermal growth factor

(EGF; Section 19-3A) and once each in three blood clot-

ting system proteins: factor IX, factor X, and protein C

(Section 35-1).

3. Five exons encode a 400-residue sequence that is

33% identical with a polypeptide segment that is shared

only with EGF.

Evidently, the LDL receptor gene is modularly constructed

from exons that also encode portions of other proteins. Nu-

merous other eukaryotic proteins are similarly constituted

including, as we have seen, many of the proteins involved

in signal transduction (e.g., those containing SH2 and SH3

domains; Section 19-3C). Moreover, many exons encode

complete domains that frequently have independent func-

tions. It therefore appears that the genes encoding these

modular proteins arose by the stepwise collection of exons

that were assembled by (aberrant) recombination between

their neighboring introns.

m. Alternative Splicing Greatly Increases the Number

of Proteins Encoded by Eukaryotic Genomes

The expression of numerous cellular genes is modulated

by the selection of alternative splice sites. Thus, certain exons

in one type of cell may be introns in another. For example,

a single rat gene encodes seven tissue-specific isoforms

(splice variants) of the muscle protein ␣-tropomyosin (Sec-

tion 35-3Ca) through the selection of alternative splice

sites (Fig. 31-65).

Section 31-4. Post-Transcriptional Processing 1317

Figure 31-65 The organization of the rat ␣-tropomyosin gene

and the seven alternative splicing pathways that give rise to

cell-specific ␣-tropomyosin isoforms. The thin kinked lines

indicate the positions occupied by the introns before they are

spliced out to form the mature mRNAs. Tissue-specific exons are

indicated together with the amino acid (aa) residues they

encode:“constitutive” exons (those expressed in all tissues) are

1–385′ UT

Striated muscle

mRNA transcripts

Smooth muscle

Striated muscle′

Myoblast

Nonmuscle/

fibroblast

Hepatoma

Brain

′ UT

STR

SM STR

3′ UT39–80 39–80 81–125 126–164 188

165–

213

189–

234

214–

257

235–

284

258–

284

258–

green, those expressed only in smooth muscle (SM) are brown,

those expressed only in striated muscle (STR) are purple, and

those variably expressed are yellow. Note that the smooth and

striated muscle exons encoding amino acid residues 39 to 80 are

mutually exclusive; likewise, there are alternative 3¿-untranslated

(UT) exons. [After Breitbart, R.E., Andreadis, A., and

Nadal-Ginard, B., Annu. Rev. Biochem. 56, 481 (1987).]

JWCL281_c31_1260-1337.qxd 8/11/10 9:48 PM Page 1317

Alternative splicing occurs in all metazoa and is espe-

cially prevalent in vertebrates. In fact, microarray-based

comparisons of the cDNAs obtained from various tissues

indicate that ⬃95% of human structural genes are subject

to at least one alternative splicing event. This rationalizes

the discrepancy between the ⬃23,000 genes identified in

the human genome (Section 7-2Bc) and earlier estimates

that it contains 50,000 to 140,000 structural genes.

The variation in mRNA sequence can take several dif-

ferent forms: Exons can be retained in an mRNA or they

can be skipped; introns may be excised or retained; and the

positions of 5¿ and 3¿ splice sites can be shifted to make ex-

ons shorter or longer. Alterations in the transcriptional

start site and/or the polyadenylation site can further con-

tribute to the diversity of the mRNAs that are transcribed

from a single gene. In a particularly striking example, the

Drosophila protein Dscam (for Down syndrome cell-adhe-

sion molecule), which functions in neuronal development,

is encoded by 24 exons of which there are 12 mutually ex-

clusive variants of exon 4, 48 of exon 6, 33 of exon 9, and 2

of exon 17 (which are therefore known as cassette exons)

for total of 38,016 possible isoforms of this protein (com-

pared to ⬃14,000 identified genes in the Drosophila

genome).Although it is unknown if all possible Dscam iso-

forms are produced, experimental evidence suggests that

the Dscam gene expresses many thousands of them.

[Dscam is a membrane-anchored cell-surface protein of

the immunoglobulin superfamily. The specific isoform ex-

pressed in a given neuron binds to itself but rarely to other

isoforms. This permits the neuron to distinguish its own

processes (axons and dendrites) from those of other neu-

rons and thereby plays an essential role in neural pattern-

ing. However, the precise identity of a given isoform ap-

pears to be unimportant.] Clearly, the number of genes in

an organism’s genome does not by itself provide an ade-

quate assessment of its protein diversity. Indeed, it has

been estimated that, on average, each human structural

gene encodes three different proteins.

The types of changes that alternative splicing confers on

expressed proteins spans the entire spectrum of protein

properties and functions. Entire functional domains or

even single amino acid residues may be inserted into or

deleted from a protein, and the insertion of a stop codon

may truncate the expressed polypeptide. Splice variations

may, for example, control whether a protein is soluble or

membrane bound, whether it is phosphorylated by a spe-

cific kinase, the subcellular location to which it is targeted,

whether an enzyme binds a particular allosteric effector,

and the affinity with which a receptor binds a ligand.

Changes in an mRNA, particularly in its noncoding re-

gions, may also influence the rate at which it is transcribed

and its susceptibility to degradation. Since the selection of

alternative splice sites is both tissue- and developmental

stage-specific,splice site choice must be tightly regulated in

both space and time. In fact, it is estimated that from ⬃15%

to 50% of human genetic diseases are caused by point muta-

tions that result in pre-mRNA splicing defects. Some of these

mutations delete functional splice sites, thereby activating

nearby pre-existing cryptic splice sites. Others generate

new splice sites that are used instead of the normal ones

and yet others are in the genes encoding components of

the splicing machinery. In addition, tumor progression is

correlated with changes in levels of proteins implicated in

alternative splice site selection.

How are alternative splice sites selected? Well-understood

examples of such processes occur in the pathway responsible

for sex determination in Drosophila, two of which we dis-

cuss here:

1. Exon 2 of transformer (tra) pre-mRNA contains two

alternative 3¿ splice sites (which succeed the excised in-

tron),with the proximal (close; to exon 1) site used in males

and the distal (far) site used in females (Fig. 31-66a). The

region between these two sites contains a Stop codon

(UAG). In males, the splicing factor U2AF binds to the

proximal 3¿ splice site to yield an mRNA containing this

premature stop codon, which thereby directs the synthesis

of truncated and hence nonfunctional TRA protein. In fe-

males, however, the proximal 3¿ splice site is bound by the

female-specific SXL protein, the product of the sex-lethal

(sxl) gene (which is only expressed in females), so as to

block the binding of U2AF, which then binds to the distal 3¿

splice site, thereby excising the UAG and inducing the ex-

pression of functional TRA protein (U2AF and TRA both

contain RS domains but not RRMs so that neither is an SR

protein).

2. In doublesex (dsx) pre-mRNA, the first three exons

are constitutively spliced in both males and females. How-

ever, the branch site immediately upstream of exon 4 has a

suboptimal pyrimidine tract to which U2AF does not bind

(Fig. 31-66b). Hence in males, exon 4 is not included in dsx

mRNA, leading to the synthesis of male-specific DSX-M

protein that functions as a repressor of female-specific

genes. However, in females, TRA protein promotes the co-

operative binding of the SR protein RBP1 and the SR-like

protein

TRA2 [the product of the transformer 2 (tra-2)

gene] to six copies of an exonic splice enhancer (ESE)

within exon 4. This heterotrimeric complex recruits the

splicing machinery to the upstream 3¿ splice site of exon 4,

leading to its inclusion in dsx mRNA.The resulting female-

specific DSX-F protein is a repressor of male-specific

genes.

Thus, the synthesis of functional TRA protein involves the

repression of a splice site, whereas the synthesis of female-

specific DSX-F protein involves the activation of a splice

site. Similar mechanisms of alternative splice site selection

have been identified in vertebrates.

In general, the decision as to whether an alternative

exon is kept or eliminated is determined by the activities

and concentrations of its various regulators, many of which

are SR proteins and hnRNPs. Hence the tissue-specific ex-

pression of these regulators and the phosphorylation state

of the SR proteins are important contributors to the com-

plex regulation of mRNA splicing. Moreover, extensive

analysis of the sequences of numerous alternative splice

sites has revealed the existence of a “splicing code” that

uses combinations of over 200 RNA features that are

1318 Chapter 31. Transcription

JWCL281_c31_1260-1337.qxd 8/11/10 9:48 PM Page 1318

present in both introns and exons and which are recog-

nized by the foregoing regulators.

Riboswitches (Section 31-3H) have been implicated in

controlling alternative splicing in eukaryotes. For example,

as Breaker has shown, in the NMT1 gene of the bread mold

Neurospora crassa [which expresses an enzyme that partic-

ipates in the metabolism of TPP (thiamine pyrophos-

phate)], a TPP-sensing riboswitch is contained in an intron

that is located upstream of the mRNA’s normal AUG

translational start codon. This intron contains one 3¿ splice

site and two 5¿ splice sites with two AUG start codons be-

tween them. When the concentration of TPP is low so that

it does not bind to the riboswitch, one strand of the

riboswitch’s P4–P5 segment base-pairs with the second

(downstream) splice site so as to inactivate it (Fig. 31-67a).

The spliceosome then efficiently excises the entire intron

yielding an mRNA that is readily translated (I-3; Fig.31-67b).

However, when the concentration of TPP is high so that it

binds to the riboswitch, the riboswitch assumes a confor-

mation that activates the second splice site but occludes the

branch point A (Fig. 31-67c). Consequently, the spliceo-

some inefficiently excises only the downstream portion of

the intron.The two upstream AUG codons, which are pres-

ent in both the unspliced mRNA (I-1) and in the spliced

mRNA containing only the upstream portion of the intron

(I-2), compete for ribosomes with mRNA’s normal AUG

start codon and thereby repress the mRNA’s translation

(Fig. 31-67c).

n. AU–AC Introns Are Excised by

a Novel Spliceosome

A small fraction of introns (⬃0.3%) have AU rather

than GU at their 5¿ ends and AC rather than AG at their 3¿

ends, but are nevertheless excised via a lariat structure to

an internal intron A. These so-called AU–AC introns (al-

ternatively, AT–AC introns after their DNA sequences),

which occur in organisms as diverse as Drosophila, plants,

and humans, are excised by a novel so-called AU–AC

spliceosome (alternatively, an AT–AC spliceosome) that

has one snRNP, U5, in common with the major (GU–AG)

spliceosome, and three others, U11, U12, and

U4atac–U6atac, which are distinct from but structurally

and functionally analogous to U1, U2, and U4–U6. Curi-

ously, all genes known to contain AU–AC introns also con-

tain multiple major class introns. Moreover, AU–AC in-

trons are not conserved in either length or position in their

host genes. Thus, the functional and evolutionary signifi-

cance of the AU–AC spliceosome and introns is obscure.

o. Trans-Splicing

The types of splicing we have so far considered occur

within single RNA molecules and hence are known as

Section 31-4. Post-Transcriptional Processing 1319

Figure 31-66 Mechanisms of alternative splice site selection in

the Drosophila sex-determination pathway as described in the

text. In all panels, exons are represented by colored rectangles

and introns are shown as pale gray lines. (a) Alternative splicing

in tra pre-mRNA. UAG is a Stop codon. (b) Alternative splicing

in dsx pre-mRNA. The six ESEs (exonic splice enhancers) in

UAG

U2AF

Male

Female

Male

Female

SXL

U2AF

TRA

RBP1

TRA2

123

112233

456

pA pA

(a)

(b)

exon 4 are indicated by green rectangles and S represents the

splicing machinery. In females, polyadenylation (pA) of dsx

mRNA occurs downstream of exon 4, whereas in males, it occurs

downstream of exon 6. [After a drawing by Maniatis, T. and Tasic,

B., Nature 418, 236 (2002).]

JWCL281_c31_1260-1337.qxd 8/11/10 9:48 PM Page 1319

cis-splicing. The chemistry of the spliceosomal cis-splicing

reaction, however, is the same as would occur if the two ex-

ons to be joined initially resided on two different RNA

molecules, a process called trans-splicing. This, in fact, oc-

curs in trypanosomes (kinetoplastid protozoa; the cause of

African sleeping sickness).Trypanosomal mRNAs all have

the same 35-nt noncoding leader sequence, although this

leader sequence is not present in the corresponding genes.

Rather, this sequence is part of a so-called spliced leader

(SL) RNA that is transcribed from an independent gene.

The 5¿ splice site that succeeds the SL RNA leader se-

quence, and the branch site and 3¿ splice site that precede

the exon sequence have the same consensus sequences as

occur in the RNAs spliced by the major spliceosome. Con-

sequently, the SL RNA leader and the pre-mRNA are

joined in a trans-splicing reaction that resembles the

spliceosomal cis-splicing reaction (Fig. 31-53) with the ex-

ception that the product of the first transesterification re-

action is necessarily Y-shaped rather than lariat-shaped

(Fig. 31-68). Trypanosomes, whose pre-mRNAs lack in-

trons, nevertheless have U2- and U4–U6-snRNPs but lack

U1- and U5-snRNPs. However, the SL RNA, which is pre-

dicted to fold into three stem–loops and a single-stranded

Sm RNA–like motif as does U1-snRNA (Fig. 31-63a), ap-

parently carries out the functions of U1-snRNA in the

trans-splicing reaction.

Trans-splicing has been shown to occur in nematodes

(roundworms; e.g., C. elegans) and flatworms. These organ-

isms also carry out cis-splicing and, indeed, perform both

types of splicing on the same pre-mRNA. There are also

several reports that trans-splicing occurs in higher eukary-

otes such as Drosophila and vertebrates, but if it does occur,

it does so in only a few pre-mRNAs and at a very low level.

p. mRNA Is Methylated at Certain

Adenylate Residues

During or shortly after the synthesis of vertebrate pre-

mRNAs, ⬃0.1% of their A residues are methylated at their

N6 atoms. These m

A’s tend to occur in the sequence

RRm

ACX, where X is rarely G. Although the functional

significance of these methylated A’s is unknown, it should

be noted that a large fraction of them are components of

the corresponding mature mRNAs.

q. RNA Can Be Edited by the Insertion or Deletion

of Specific Nucleotides

Certain mRNAs from a variety of eukaryotic organisms

have been found to differ from their corresponding genes

in several unexpected ways, including C S U and U S C

changes, the insertion or deletion of U residues, and the in-

sertion of multiple G or C residues. The most extreme ex-

amples of this phenomenon, which occur in the mitochon-

1320 Chapter 31. Transcription

Figure 31-67 Control of translation by the N. crassa

TPP-sensing riboswitch through alternative splicing. (a) The

predicted secondary structure of the TPP aptamer, which resides

in the 5¿ untranslated region of the NMT1 mRNA. The sequence

of one strand of the P4-P5 stem is complementary (orange

shading) to a segment that overlaps the second 5¿ splice site.

(b) At low TPP concentrations, the aptamer inhibits (red tee)

splicing from the second 5¿ splice site while activating (green

arrow) the branch point A so that the spliceosome excises the

RNA between the first 5¿ splice site and the 3¿ splice site, yielding

an ORF (open reading frame; I-3) that is normally translated

from its AUG start codon. (c) At high TPP concentrations, the

binding of TPP to the aptamer activates the second 5¿ splice site

but occludes the branch point A.The spliceosome therefore

inefficiently excises the intron’s downstream portion yielding an

mRNA (I-2) that contains two uORFs (upstream ORFs) that

compete with the translation of the primary ORF. The unspliced

mRNA (I-1) also contains the two uORFs and hence is likewise

inefficiently translated. [Courtesy of Ronald Breaker,Yale

University.]

JWCL281_c31_1260-1337.qxd 8/11/10 9:48 PM Page 1320