Voet D., Voet Ju.G. Biochemistry

Подождите немного. Документ загружается.

(uORFs), uORF1 to uORF4, in its 5¿ leader that precedes

the sequence encoding GCN4. Under the normal nutrient

conditions in which GCN2 is inactive, the ribosome binds

to the mRNA near its 5¿ cap and scans for the nearest AUG

initiation codon (which is in uORF1), where it forms the

48S preinitiation complex (Fig. 32-44) and commences the

translation of uORF1 (Section 32-3Cd). On terminating

translation at uORF1’s Stop codon,the presence of the sur-

rounding A  U–rich sequences causes the ribosome to re-

sume scanning for the next AUG codon, where it initiates

the translation of uORF2.This process repeats until the ri-

bosome terminates at the end of uORF4, where its Stop

codon’s surrounding G  C–rich sequences induce the ri-

bosome to disengage from the mRNA. Hence GCN4 is

only expressed at a low basal level. However, under the

low nutrient conditions in which GCN2 phosphorylates

eIF2 at its Ser 51, the resulting reduced level of the

eIF2 GTP ternary complex causes the 40S

subunit to scan longer distances before it can form the 48S

preinitiation complex. Consequently, ⬃50% of the ribo-

somes scan past uORF2, uORF3, and uORF4 and only ini-

tiate translation at the GCN4 AUG codon, which is there-

fore translated at a high level (uORF2 and uORF3 can be

mutationally eliminated without significantly affecting

translational control).

Mammalian homologs of GCN2 are activated under

conditions of amino acid starvation. This suggests that the

foregoing process has been conserved throughout eukary-

otic evolution.

B. Regulation of eIF4E

eIF4E (cap-binding protein) binds to the m

G cap of eu-

karyotic mRNAs and thereby participates in translational

initiation by helping to identify the initiating AUG codon

(Section 32-3Cd).When mammalian cells are treated with

hormones, cytokines, mitogens (substances that induce

mitosis), and/or growth factors, Ser 209 of human eIF4E is

phosphorylated via a Ras-activated MAP kinase cascade

(Sections 19-3C and 19-3D), thereby increasing eIF4E’s

affinity for capped mRNA and hence stimulating transla-

tional initiation. Ser 209 occupies a surface position on

eIF4E adjacent to the binding site for the  phosphate

group of the m

GDP and flanking the putative binding

cleft for mRNA (Fig. 32-45b).The structure of eIF4E sug-

gests that the phosphoryl group of phosphorylated Ser

209 forms a salt bridge with Lys 159, which occupies the

other side of the putative mRNA-binding cleft, so as to

form a clamp that would help stabilize the bound mRNA.

The importance of regulating eIF4E activity is indicated

by the observations that the overexpression of eIF4E

causes the malignant transformation of rodent cell lines

and that eIF4E expression is elevated in several human

cancers.

The homologous ⬃120-residue proteins known as 4E-

BP1, 4E-BP2, and 4E-BP3 (BP for binding protein; the

first two are also known as PHAS-I and PHAS-II) inhibit

cap-dependent translation. They do so by binding on the

Met–tRNA

Met

opposite side of eIF4E from its mRNA-binding site, pre-

sumably to a patch of seven highly conserved surface

residues, and hence do not prevent eIF4E from binding the

G cap. Rather, they block eIF4E from binding to eIF4G

and thereby interfere with the formation of the eIF4F com-

plex that positions the 40S ribosomal subunit-bound

Met–tRNA

Met

on the mRNA’s initiating AUG codon (Sec-

tion 32-3Cd). In fact, the 4E-BPs and eIF4G all possess the

sequence motif YXXXXL (where  is an aliphatic

residue, most often L but also M or F) through which they

bind to eIF4E.

The treatment of responsive cells with insulin or any of

several protein growth factors causes the 4E-BPs to disso-

ciate from eIF4E.This is because the presence of these hor-

mones induces the phosphorylation of the 4E-BPs at six

Ser/Thr residues via the signal transduction pathway in-

volving PI3K, PKB, and mTOR (Fig. 19-67). Evidently, the

phosphorylation of eIF4E and the 4E-BPs have similar if

not synergistic effects in the hormonal regulation of trans-

lation in eukaryotes.

C. mRNA Masking and Cytoplasmic

Polyadenylation

It has been known since the nineteenth century that early

embryonic development in animals such as sea urchins, in-

sects, and frogs is governed almost entirely by information

present in the oocyte (egg) before fertilization.Indeed, sea

urchin embryos exposed to sufficient actinomycin D (Sec-

tion 31-2Cc) to inhibit RNA synthesis without blocking

DNA replication develop normally through their early

stages without a change in their protein synthesis pro-

gram. This is in part because an unfertilized egg contains

large quantities of mRNA that is “masked” by associated

proteins to form ribonucleoprotein particles, thereby pre-

venting the mRNAs’ association with the ribosomes that

are also present. On fertilization, this mRNA is “un-

masked” in a controlled fashion, quite possibly by the de-

phosphorylation of the associated proteins, and com-

mences directing protein synthesis. Development of the

embryo can therefore start immediately on fertilization

rather than wait for the synthesis of paternally specified

mRNAs.Thus, gene expression in the early stages of devel-

opment is entirely translationally controlled; transcrip-

tional control only becomes important when transcription

is initiated.

a. Cytoplasmic Polyadenylation

Another mechanism of translational control in oocytes

and early embryos involves the polyadenylation of mRNAs

in the cytoplasm (polyadenylation usually occurs in the

nucleus, following which the mRNA is exported to the cy-

toplasm; Section 31-4Ab). A substantial number of mater-

nally supplied mRNAs in oocytes have relatively short

poly(A) tails (20–40 nt versus a usual length of ⬃250 nt).

The 3¿ untranslated region of these mRNAs contains both

the AAUAAA polyadenylation signal (which is required

for polyadenylation in the nucleus; Section 31-4Ab)

Section 32-4. Control of Eukaryotic Translation 1401

JWCL281_c32_1338-1428.qxd 8/4/10 4:45 PM Page 1401

(a) (b)

Translationally dormant Translationally active

CPEB

CPEB CPSF

PAP

PABP PABP

40S

CRF

4E-Cap

Maskin

UUUUUAU AAUAAA

-AAAAAAAAAAAAAAAAAAA

UUUUUAU AAUAAA

-Cap

-A

together with a so-called cytoplasmic polyadenylation

element (CPE), which has the consensus sequence

UUUUUAU. The CPE is recognized by CPE-binding pro-

tein (CPEB), which contains two RNA recognition motifs

(RRMs) as well as a zinc finger motif (Section 34-3Bl) that

contribute to its binding to the mRNA. Joel Richter discov-

ered that CPEB recruits a 931-residue protein named

maskin which, in turn, binds the eIF4E (cap-binding pro-

tein) that is bound to the mRNA’s 5¿ cap (Fig. 32-69a).

Maskin contains the same YXXXXL motif through

which the 4E-BPs and eIF4G bind to eIF4E (Section 32-4B),

thereby blocking the binding of eIF4G to eIF4E and hence

preventing the formation of the 48S preinitiation complex

(Fig. 32-44).

In the maturation of Xenopus laevis oocytes, a process

that precedes fertilization and is stimulated by the steroid

hormone progesterone (Section 19-1Gb), a variety of

mRNAs, including those encoding several cyclins (which

participate in cell cycle control;Section 34-4Da) are transla-

tionally activated. Soon after exposure to progesterone, a

protein kinase named aurora phosphorylates the mRNA-

bound CPEB at its Ser 174. This increases CPEB’s affinity

for cleavage and polyadenylation specificity factor (CPSF;

Section 31-4Ab), which then binds to the mRNA’s

AAUAAA sequence, where it recruits poly(A) polymerase

(PAP) to lengthen the mRNA’s poly(A) tail (Fig. 32-69b).

Translational initiation and cytoplasmic polyadenyla-

tion occur simultaneously, which suggests that these

processes are linked. Indeed, Richter has shown that this

occurs through the binding to poly(A) of poly(A)-binding

protein (PABP; Section 31-4Ab), which as we saw (Section

32-3Cd), also binds to eIF4G to circularize the mRNA.The

eIF4G in this complex displaces maskin from eIF4E,

thereby permitting the formation of the 48S preinitiation

complex and hence the mRNA’s translation (Fig. 32-69b).

Mammalian cells also exhibit cell cycle–dependent cyto-

plasmic polyadenylation of mRNAs. This suggests that

translational control by polyadenylation is a general fea-

ture in animal cells.

D. Antisense Oligonucleotides

Since ribosomes cannot translate double-stranded RNA

or DNA–RNA hybrid helices, the translation of a given

mRNA can be inhibited by a segment of its complemen-

tary RNA or DNA, that is, an antisense RNA or an anti-

sense oligodeoxynucleotide, which are collectively known

as antisense oligonucleotides. Moreover, endogenous

RNase H’s (enzymes that cleave the RNA strand of

an RNA–DNA duplex; Section 31-4C) cleave an

mRNA–oligodeoxynucleotide duplex on its mRNA

strand, leaving the antisense oligodeoxynucleotide intact

for binding to another mRNA.

Since the human genome consists of ⬃3 billion bp, an

⬃15-nt oligonucleotide (which is easily synthesized; Sec-

tion 7-6Aa) should ideally be able to target any segment of

the human genome. This exquisite specificity provides the

delivery of an antisense oligonucleotide to, or its expres-

sion in, a selected tissue or organism with enormous bio-

medical and biotechnological potential. However, care

must be taken that an antisense oligonucleotide does not

also eliminate nontarget mRNAs.

Methods for the delivery of a therapeutically useful an-

tisense oligonucleotide to a target tissue are as yet in their

infancy. This is in large part because oligonucleotides are

readily degraded by the many nucleases present in an or-

ganism and because they do not readily pass through cell

membranes. Moreover, a target mRNA is likely to be asso-

ciated with cellular proteins and hence not available for

binding to other molecules. The nuclease resistance of

oligonucleotides can be increased by derivatizing them, for

example, by replacing a nonbridging oxygen at each phos-

phate group with a methyl group or an S atom so as to

yield methylphosphonate or phosphorothioate oligonu-

cleotides, although this reduces their antisense activity.

The expression of antisense oligonucleotides in the speci-

fied tissues would, of course,circumvent the delivery prob-

lem but has all the difficulties associated with gene ther-

apy (Section 5-5H).

1402 Chapter 32. Translation

Figure 32-69 CPEB-mediated translational control. (a) In

immature Xenopus oocytes, an mRNA containing the CPE

(UUUUUAU) is bound by CPEB, which binds maskin, which

then binds eIF4E so as to prevent it from binding eIF4G, thereby

maintaining the mRNA in a translationally dormant (masked)

state. (b) In the maturation process, CPEB is phosphorylated by

an aurora protein kinase. The phosphorylated CPEB binds CPSF,

which recruits PAP to extend the mRNA’s heretofore short

poly(A) tail. PABP binds to the newly lengthened poly(A) tail

and simultaneously binds to eIF4G so as to displace maskin.This

permits the 48S preinitiation complex to assemble and hence the

translation of the mRNA to proceed. [Based on a drawing by

Mendez, R. and Richer, J.D., Nature Rev. Mol. Cell Biol. 2, 521

(2001).]

JWCL281_c32_1338-1428.qxd 9/7/10 2:23 PM Page 1402

Despite the foregoing, antisense technology is begin-

ning to show success. Fomivirsen (trade name Vitravene),

a 21-nt phosphorothioate oligonucleotide that is comple-

mentary to an mRNA expressed by cytomegalovirus

(CMV), is effective in the treatment of retinitis (inflam-

mation of the retina) caused by CMV infection in individ-

uals with AIDS (CMV is an opportunistic pathogen that

rarely infects individuals with normally functioning im-

mune systems). It was approved for human use in 1998 by

the FDA, the first antisense drug so approved. A number

of antisense oligonucleotides that are mainly targeted

against genes that are overexpressed in specific cancers

and autoimmune diseases as well as other viral infections

are in clinical trials (Section 15-4Bb), although additional

antisense drugs have not yet been approved for human

use.

Antisense technology has also had some success in the

arena of biotechnology. For example, in tomatoes and

other fruits, the enzyme polygalacturonase (PG), which is

expressed during ripening, depolymerizes the pectin

(mainly polygalacturonic acid) in the cell wall. This results

in a softening of tomatoes to the point that vine-ripened

(and hence better tasting) tomatoes are unable to with-

stand the rigors of shipping and hence must be picked be-

fore they are ripe. The introduction into a tomato, via ge-

netic engineering techniques, of a gene expressing

antisense PG RNA yielded the so-called Flavr Savr tomato

that had substantially reduced PG expression and hence

remained firm after vine ripening.

5 POST-TRANSLATIONAL

MODIFICATION

To become mature proteins, polypeptides must fold to

their native conformations, their disulfide bonds, if any,

must form, and, in the case of multisubunit proteins, the

subunits must properly combine. Moreover, as we have

seen throughout this text, many proteins are modified in

enzymatic reactions that proteolytically cleave certain

peptide bonds and/or derivatize specific residues. In this

section we shall review some of these post-translational

modifications.

A. Proteolytic Cleavage

Proteolytic cleavage is the most common type of post-

translational modification. Probably all mature proteins

have been so modified, if by nothing else than the prote-

olytic removal of their leading Met (or fMet) residue

shortly after it emerges from the ribosome. Many proteins,

which are involved in a wide variety of biological

processes, are synthesized as inactive precursors that are

activated under proper conditions by limited proteolysis.

Some examples of this phenomenon that we have encoun-

tered are the conversion of trypsinogen and chy-

motrypsinogen to their active forms by tryptic cleavages

of specific peptide bonds (Section 15-3E), and the forma-

tion of active insulin from the 84-residue proinsulin by the

excision of its internal 33-residue C chain (Section 9-1Aa).

Inactive proteins that are activated by removal of

polypeptides are called proproteins, whereas the excised

polypeptides are termed propeptides.

a. Propeptides Direct Collagen Assembly

Collagen biosynthesis is illustrative of many facets of

post-translational modification. Recall that collagen, a ma-

jor extracellular component of connective tissue, is a fi-

brous triple-helical protein whose polypeptides each con-

tain the amino acid sequence (Gly-X-Y)

where X is often

Pro, Y is often 4-hydroxyproline (Hyp), and 340 (Sec-

tion 8-2B). The polypeptides of procollagen (Fig. 32-70)

differ from those of the mature protein by the presence of

both N-terminal and C-terminal propeptides of ⬃100

residues whose sequences, for the most part, are unlike

those of mature collagen. The procollagen polypeptides

rapidly assemble, in vitro as well as in vivo, to form a colla-

gen triple helix. In contrast, polypeptides extracted from

mature collagen will reassemble only over a period of days,

if at all. The collagen propeptides are apparently necessary

for proper procollagen folding.

The N- and C-terminal propeptides of procollagen are

respectively removed by amino- and carboxyprocollagen

n ⬇

Section 32-5. Post-Translational Modification 1403

Figure 32-70 Electron micrograph of procollagen aggregates

that have been secreted by fibroblasts into the extracellular

medium. [Courtesy of Jerome Gross, Massachusetts General

Hospital, Harvard Medical School.]

JWCL281_c32_1338-1428.qxd 8/4/10 4:45 PM Page 1403

peptidases (Fig. 32-71), which may also be specific for the

different collagen types. An inherited defect of aminopro-

collagen peptidase in cattle and sheep results in a bizarre

condition, dermatosparaxis, that is characterized by ex-

tremely fragile skin. An analogous disease in humans,

Ehlers–Danlos syndrome VII, is caused by a mutation in

one of the procollagen polypeptides that inhibits the enzy-

matic removal of its aminopropeptide. Collagen molecules

normally spontaneously aggregate to form collagen fibrils

(Figs. 8-31 and 8-32). However, electron micrographs of

dermatosparaxic skin show sparse and disorganized colla-

gen fibrils. The retention of collagen’s aminopropeptides ap-

parently interferes with proper fibril formation. (The der-

matosparaxis gene was bred into some cattle herds because

heterozygotes produce tender meat.)

b. Signal Peptides Are Removed from Nascent

Proteins by a Signal Peptidase

Many transmembrane proteins or proteins that are des-

tined to be secreted are synthesized with an N-terminal sig-

nal peptide of 13 to 36 predominantly hydrophobic residues.

As we saw in our discussion of the secretory pathway (Sec-

tion 12-4B), a signal peptide is recognized by a signal recog-

nition particle (SRP). The SRP binds a ribosome synthesiz-

ing a signal peptide to a protein pore known as the

translocon that is embedded in the membrane [the rough

endoplasmic reticulum (RER) in eukaryotes and the plasma

membrane in bacteria] and conducts the signal peptide and

its following nascent polypeptide through the translocon.

Proteins bearing a signal peptide are known as prepro-

teins or, if they also contain propeptides, as preproproteins.

Once the signal peptide has passed through the membrane,

it is specifically cleaved from the nascent polypeptide by a

membrane-bound signal peptidase. Both insulin and colla-

gen are secreted proteins and are therefore synthesized

with leading signal peptides in the form of preproinsulin

and preprocollagen. These and many other proteins are

therefore subject to three sets of sequential proteolytic

cleavages: (1) the deletion of their initiating Met residue,

(2) the removal of their signal peptides, and (3) the excision

of their propeptides.

c. Polyproteins

Some proteins are synthesized as segments of polypro-

teins, polypeptides that contain the sequences of two or

more proteins. Examples include many polypeptide hor-

mones (Section 34-3C); the proteins synthesized by many

viruses, including those causing polio (Section 33-2C) and

AIDS (Section 15-4Cb); and ubiquitin, a highly conserved

eukaryotic protein involved in protein degradation (Section

32-6Bb). Specific proteases post-translationally cleave

polyproteins to their component proteins, presumably

through the recognition of the cleavage site sequences. Some

of these proteases are conserved over remarkable evolu-

tionary distances. For instance, ubiquitin is synthesized as

several tandem repeats (polyubiquitin) that E. coli properly

cleave even though prokaryotes lack ubiquitin. Other pro-

teases have more idiosyncratic cleavage sequences. This has

allowed medicinal chemists to design inhibitors of HIV pro-

tease (which catalyzes an essential step in the viral life cycle)

that have been highly effective in attenuating if not prevent-

ing the progression of AIDS (Section 15-4Cd).

1404 Chapter 32. Translation

Figure 32-71 Schematic representation of the procollagen

molecule. Gal, Glc, GlcNAc, and Man, respectively, denote

galactose, glucose, N-acetylglucosamine, and mannose residues.

Note that the N-terminal propeptide has intrachain disulfide

Glc

Gal

GlcNAc

(Man)

N-terminal

propeptide

(150 Å)

Collagen molecule

(3000 Å)

C-terminal

propeptide

(100 Å)

Gal

15 Å

Non–triple

helical domain

Non–triple

helical domain

Triple helical

domain

Globular

domain

Triple helical domain

Non–triple

helical domain

Aminoprocollagen

peptidase

cleavage site

Carboxyprocollagen

peptidase

cleavage site

20 Å

100 Å

S-S

= OH

S-S

bonds while the C-terminal propeptide has both intrachain and

interchain disulfide bonds. [After Prockop, D.J., Kivirikko, K.I.,

Tuderman, L., and Guzman, N.A., New Engl. J. Med. 301, 16

(1979).]

JWCL281_c32_1338-1428.qxd 8/4/10 4:45 PM Page 1404

B. Covalent Modification

Proteins are subject to specific chemical derivatizations,

both at the functional groups of their side chains and at

their terminal amino and carboxyl groups. Over 150 differ-

ent types of side chain modifications, involving all side

chains but those of Ala, Gly, Ile, Leu, Met, and Val, are

known (Section 4-3A).These include acetylations, glycosy-

lations, hydroxylations, methylations, nucleotidylations,

phosphorylations, and ADP-ribosylations as well as nu-

merous “miscellaneous” modifications.

Some protein modifications, such as the phosphoryla-

tion of glycogen phosphorylase (Section 18-1A) and the

ADP-ribosylation of eEF2 (Section 32-3Ge), modulate

protein activity. Several side chain modifications covalently

bond cofactors to enzymes, presumably to increase their

catalytic efficiency. Examples of linked cofactors that we

have encountered are N

-lipoyllysine in dihydrolipoyl

transacetylase (Section 21-2Ac) and 8-histidylflavin in

succinate dehydrogenase (Section 21-3F). The attachment

of complex carbohydrates, which occur in almost infinite

variety, alter the structural properties of proteins and form

recognition markers in various types of targeting and

cell–cell interactions (Sections 11-3C, 12-3E, and 23-3B).

Modifications that cross-link proteins, such as occur in col-

lagen (Section 8-2Bc), stabilize supramolecular aggregates.

The functions of most side chain modifications, however,

remain enigmatic.

a. Collagen Assembly Requires

Chemical Modification

Collagen biosynthesis (Fig. 32-72) is illustrative of pro-

tein maturation through chemical modification.As the nas-

cent procollagen polypeptides pass into the RER of the fi-

broblasts that synthesized them, the Pro and Lys residues

are hydroxylated to Hyp, 3-hydroxy-Pro, and 5-hydroxy-Lys

(Hyl). The enzymes that do so are sequence specific: Prolyl

4-hydroxylase and lysyl hydroxylase act only on the Y

residues of the Gly-X-Y sequences, whereas prolyl 3-

hydroxylase acts on the X residues but only if Y is Hyp.Gly-

cosylation, which also occurs in the RER, subsequently at-

taches sugar residues to Hyl residues (Section 8-2Bb). The

folding of three polypeptides into the collagen triple helix

must follow hydroxylation and glycosylation because the

hydroxylases and glycosyl transferases do not act on helical

substrates. Moreover,the collagen triple helix denatures be-

low physiological temperatures unless stabilized by hydro-

gen bonding interactions involving Hyp residues (Section 8-

2B). Folding is also preceded by the formation of specific

interchain disulfide bonds between the carboxylpropep-

tides. This observation bolsters the previously discussed

conclusion that collagen propeptides help select and align

the three collagen polypeptides for proper folding.

The procollagen molecules pass into the Golgi appara-

tus where they are packaged into secretory vesicles (Sec-

tions 12-4C and 12-4D) and secreted into the extracellular

spaces of connective tissue. The aminopropeptides are ex-

cised just after procollagen leaves the cell and the car-

boxypropeptides are removed sometime later. The colla-

gen molecules then spontaneously assemble into fibrils,

which suggests that an important propeptide function is to

prevent intracellular fibril formation. Finally, after the ac-

tion of the extracellular enzyme lysyl oxidase, the collagen

molecules in the fibrils spontaneously cross-link (Fig. 8-33).

C. Protein Splicing: Inteins and Exteins

Protein splicing is a post-translational modification process

in which an internal protein segment (an intein) excises it-

self from a surrounding external protein, which it ligates to

form the mature extein. The portions of the unspliced

Section 32-5. Post-Translational Modification 1405

Figure 32-72 Schematic representation of procollagen

biosynthesis. Saccharides are represented as in Fig. 32-71.The

diagram does not indicate the removal of signal peptides. [After

Procollagen

Rough endoplasmic

reticulum

Prockop, D.J., Kivirikko, K.I., Tuderman, L., and Guzman, N.A.,

New Engl. J. Med. 301, 18 (1979).]

JWCL281_c32_1338-1428.qxd 8/4/10 4:45 PM Page 1405

1406 Chapter 32. Translation

C-extein

Intein

His

N-extein

Ser 1

Ser

Asn

COO

–

C-extein

Ser

COO

–

C-extein

Intein

Ser 1

Ser

Asn

COO

–

Precursor

Linear ester intermediate

Branched intermediate

Ligated exteins

His

C-extein

Intein

Ser 1

Ser

Asn

COO

–

His

Intein

His

N-extein

C-extein

Ser

COO

–

Excised intein

Intein

His

Excised intein

Figure 32-73 Series of reactions catalyzed by inteins to splice themselves out of a polypeptide chain.

See the text for details.

JWCL281_c32_1338-1428.qxd 8/4/10 4:45 PM Page 1406

extein on the N- and C-terminal sides of the intein are

called the N-extein and the C-extein. Over 500 putative in-

teins, ranging in length from 100 to 1650 residues, have so far

been identified in archaebacteria, eubacteria, single-celled

eukaryotes, and viruses (and are registered in the Intein

Database at http://www.neb.com/neb/inteins.html). The var-

ious exteins in which these inteins are embedded exhibit no

significant sequence similarity and, in fact, can be replaced

by other polypeptides, thereby indicating that exteins do

not contain the catalytic elements that mediate protein

splicing. In contrast, the ⬃130-residue splicing elements of

inteins exhibit significant sequence similarity. All of them

have four conserved splice-junction residues: (1) a

Ser/Thr/Cys at the intein’s N-terminus; and (2 and 3) a

His–Asn/Gln dipeptide at the intein’s C-terminus; which is

immediately followed by (4) a Ser/Thr/Cys at the N-termi-

nus of the C-extein.

Protein splicing occurs via a reaction sequence that in-

volves four successive nucleophilic displacements, the first

three of which are mediated by the intein (Fig. 32-73):

1. Attack by the N-terminal intein residue (Ser, Thr, or

Cys; shown in Fig. 32-73 as Ser) on its preceding carbonyl

group, yielding a linear (thio)ester intermediate.

2. A transesterification reaction in which the ¬OH or

¬SH group on the C-extein’s N-terminal residue (shown

in Fig. 32-73 as Ser) attacks the above (thio)ester linkage,

thereby yielding a branched intermediate in which the N-

extein has been transferred to the C-extein.

3. Cleavage of the amide linkage connecting the intein

to the C-extein by cyclization of the intein’s C-terminal

Asn or Gln (shown in Fig. 32-73 as Asn). The succinimide

ring of the excised intein then spontaneously hydrolyzes to

regenerate Asn (or iso-Asn).

4. Spontaneous rearrangement of the (thio)ester link-

age between the ligated exteins to yield the more stable

peptide bond.

The X-ray structure of the 198-residue GyrA intein

from Mycobacterium xenopi, determined by James Sac-

chetini, indicates how this intein catalyzes the foregoing

splicing reactions. This intein’s N-terminal residue, Cys 1,

was replaced by an Ala–Ser dipeptide with the expectation

that the mutant protein would resemble the intein’s pre-

splicing state (the new N-terminal residue, Ala 0, presum-

ably represents the C-terminal residue of the N-extein).

The X-ray structure reveals that this monomeric protein

consists primarily of  strands, two of which curve about

the periphery of the entire protein to give it the shape of a

flattened horseshoe (Fig. 32-74). The intein’s catalytic site

is located at the bottom of a broad and shallow cleft near

the center of this so-called -horseshoe, where the intein’s

N-terminal and C-terminal residues are in close proximity.

The Ala 0¬Ser 1 peptide bond, the bond cleaved in Reac-

tion 1 of the protein splicing process (Fig. 32-73) assumes

the cis conformation (Fig. 8-2), a rare high-energy confor-

mation (except when the peptide bond is followed by Pro)

that destabilizes this bond. Its amide nitrogen atom is hy-

drogen bonded to the side chain of the highly conserved

His 75. Hence His 75 is well positioned to donate a proton

that would promote the breakdown of the tetrahedral in-

termediate in Reaction 1.The side chains of Thr 72 and Asn

74 appear well positioned to stabilize this tetrahedral inter-

mediate in a manner resembling that of the oxyanion hole

in serine proteases (Section 15-3Db). The position of Ser 1

and a modeled Thr at the intein’s C-terminus is consistent

with Reaction 2 of the splicing process. The side chain of

the invariant His 197 is hydrogen bonded to the carboxy-

late of the C-terminal Asn 198 and hence is positioned to

protonate the peptide bond cleaved in Reaction 3.

a. Most Inteins Encode a Homing Endonuclease

What is the biological function of inteins? Nearly all in-

teins contain polypeptide inserts forming so-called homing

endonucleases. These are site-specific endonucleases that

make a double-strand break in genes that are homologous to

their corresponding extein but which lack inteins. The break

initiates the double-strand break repair of the DNA via re-

combination (Section 30-6Ag). Since the intein-containing

gene is likely to be the only other gene in the cell containing

extein-like sequences, the intein gene is copied into the

break.Thus, most inteins mediate a highly specific transposi-

tion or “homing” of the genes that insert them in similar

sites. The intein’s protease and endonuclease activities ap-

pear to have a symbiotic relationship: The protease activity

excises the intein from the host protein, thereby preventing

deleterious effects on the host, whereas the endonuclease

Section 32-5. Post-Translational Modification 1407

Figure 32-74 X-ray structure of the M. xenopi Gyr A intein in

which Cys 1 was replaced by an Ala 0–Ser 1 dipeptide. The

protein is drawn in ribbon form with its N-terminal Ala 0–Ser 1

dipeptide and its C-terminal His 197–Asn 198 dipeptide as well

as the side chains of residues 72 through 75 drawn in stick form

colored according to atom type (C of residues 0–1 magenta, C of

residues 72–75 green, C of residues 197–198 cyan, N blue, and O

red). Hydrogen bonds are represented by thin gray bonds.

[Based on an X-ray structure by James Sacchetini, Texas A&M

University. PDBid 1AM2.]

JWCL281_c32_1338-1428.qxd 8/19/10 10:06 PM Page 1407

activity assures the mobility of the intein gene. Thus intein

genes appear to be molecular parasites (junk DNA) that

only function to propagate themselves. Indeed, homing en-

donucleases are also encoded by certain types of introns.

6 PROTEIN DEGRADATION

The pioneering work of Henry Borsook and Rudolf

Schoenheimer around 1940 demonstrated that the compo-

nents of living cells are constantly turning over. For exam-

ple, adult humans normally turn over ⬃2% of their pro-

teins per day. Proteins have lifetimes that range from as

short as a few minutes to weeks or more. In any case, cells

continuously synthesize proteins from and degrade them to

their component amino acids. The function of this seem-

ingly wasteful process is 2-fold: (1) to eliminate abnormal

proteins whose accumulation would be harmful to the cell,

and (2) to permit the regulation of cellular metabolism by

eliminating superfluous enzymes and regulatory proteins.

Indeed, since the level of an enzyme depends on its rate of

degradation as well as its rate of synthesis, controlling a

protein’s rate of degradation is as important to the cellular

economy as is controlling its rate of synthesis. In this section

we consider the processes of intracellular protein degrada-

tion and their consequences.

A. Degradation Specificity

Cells selectively degrade abnormal proteins. For example,

hemoglobin that has been synthesized with the valine ana-

log -amino--chlorobutyrate

has a half-life in reticulocytes of ⬃10 min, whereas normal

hemoglobin lasts the 120-day lifetime of the red cell (which

makes it perhaps the longest lived cytoplasmic protein).

Likewise, unstable mutant hemoglobins are degraded soon

after their synthesis, which, for reasons explained in Sec-

tion 10-3A, results in the hemolytic anemia characteristic

of these molecular disease agents. Bacteria also selectively

degrade abnormal proteins. For instance, amber and ochre

mutants of -galactosidase have half-lives in E. coli of only

a few minutes, whereas the wild-type enzyme is almost in-

definitely stable. Most abnormal proteins, however, proba-

bly arise from the chemical modification and/or sponta-

neous denaturation of these fragile molecules in the cell’s

reactive environment rather than by mutations or the rare

errors in transcription or translation. The ability to elimi-

nate damaged proteins selectively is therefore an essential re-

cycling mechanism that prevents the buildup of substances

that would otherwise interfere with cellular processes.

Normal intracellular proteins are eliminated at rates

that depend on their identities. A given protein is elimi-

COO

-Amino--chlorobutyrate Valine

nated with first-order kinetics, indicating that the mole-

cules being degraded are chosen at random rather than ac-

cording to their age. The half-lives of different enzymes in

a given tissue vary substantially as is indicated for rat liver

in Table 32-11. Remarkably, the most rapidly degraded en-

zymes all occupy important metabolic control points,

whereas the relatively stable enzymes have nearly constant

catalytic activities under all physiological conditions. The

susceptibilities of enzymes to degradation have evidently

evolved together with their catalytic and allosteric properties

so that cells can efficiently respond to environmental

changes and metabolic requirements. The criteria through

which native proteins are selected for degradation are con-

sidered in Section 32-6B.

The rate of protein degradation in a cell also varies with

its nutritional and hormonal state. Under conditions of nu-

tritional deprivation, cells increase their rate of protein

degradation so as to provide the necessary nutrients for in-

dispensable metabolic processes. The mechanism that in-

creases degradative rates in E. coli is the stringent response

(Section 31-3I). A similar mechanism may be operative in

eukaryotes since, as happens in E. coli, increased rates of

degradation are prevented by antibiotics that block protein

synthesis.

B. Degradation Mechanisms

Eukaryotic cells have dual systems for protein degradation:

lysosomal mechanisms and ATP-dependent cytosolically

based mechanisms. We consider both mechanisms below.

a. Lysosomes Mostly Degrade Proteins

Nonselectively

Lysosomes are membrane-encapsulated organelles

(Section 1-2Ad) that contain ⬃50 hydrolytic enzymes, in-

cluding a variety of proteases known as cathepsins. The

lysosome maintains an internal pH of ⬃5 and its enzymes

have acidic pH optima. This situation presumably protects

1408 Chapter 32. Translation

Table 32-11 Half-Lives of Some Rat Liver Enzymes

Enzyme Half-Life (h)

Short-Lived Enzymes

Ornithine decarboxylase 0.2

RNA polymerase I 1.3

Tyrosine aminotransferase 2.0

Serine dehydratase 4.0

PEP carboxylase 5.0

Long-Lived Enzymes

Aldolase 118

GAPDH 130

Cytochrome b 130

LDH 130

Cytochrome c 150

Source: Dice, J.F. and Goldberg,A.L., Arch. Biochem. Biophys. 170,

214 (1975).

JWCL281_c32_1338-1428.qxd 8/4/10 4:45 PM Page 1408

the cell against accidental lysosomal leakage since lysoso-

mal enzymes are largely inactive at cytosolic pH’s.

Lysosomes recycle intracellular constituents by fusing

with membrane-enclosed bits of cytoplasm known as au-

tophagic vacuoles and subsequently breaking down their

contents. They similarly degrade extracellular substances

that the cell takes up via endocytosis (Section 12-5Bc).The

existence of these processes has been demonstrated

through the use of lysosomal inhibitors. For example, the

antimalarial drug chloroquine

is a weak base that, in uncharged form, freely penetrates

the lysosome where it accumulates in charged form,

thereby increasing the intralysosomal pH and inhibiting

lysosomal function. The treatment of cells with chloro-

quine reduces their rate of protein degradation. Similar ef-

fects arise from treatment of cells with cathepsin inhibitors

such as the polypeptide antibiotic antipain.

Lysosomal protein degradation in well-nourished cells

appears to be nonselective. Lysosomal inhibitors do not af-

fect the rapid degradation of abnormal proteins or short-

lived enzymes. Rather, they prevent the acceleration of

nonselective protein breakdown on starvation. However,

the continued nonselective degradation of proteins in

starving cells would rapidly lead to an intolerable deple-

tion of essential enzymes and regulatory proteins. Lyso-

somes therefore also have a selective pathway, which is

activated only after a prolonged fast, that takes up and de-

grades proteins containing the pentapeptide Lys-Phe-Glu-

Arg-Gln (KFERQ) or a closely related sequence. Such

KFERQ proteins are selectively lost in fasting animals

from tissues that atrophy in response to fasting (e.g., liver

and kidney) but not from tissues that do not do so (e.g.,

brain and testes). KFERQ proteins are specifically bound

in the cytosol and delivered to the lysosome by a 73-kD

peptide recognition protein (prp73), a member of the 70-

kD heat shock protein (Hsp70) family (Section 9-2C).

Both normal and pathological processes are associated

with increased lysosomal activity. Diabetes mellitus (Sec-

tion 27-4B) stimulates the lysosomal breakdown of pro-

teins. Similarly, muscle wastage caused by disuse, denerva-

tion, or traumatic injury arises from increased lysosomal

activity. The regression of the uterus after childbirth, in

which this muscular organ reduces its mass from 2 kg to

OOCCHNH

Phe

NHCHC

NHCHC NHCHC H

Arg

ArgVal

Antipain

CH CH

N(C

)

Chloroquine

50 g in 9 days, is a striking example of this process. Many

chronic inflammatory diseases, such as rheumatoid arthri-

tis, involve the extracellular release of lysosomal enzymes

that break down the surrounding tissues.

b. Ubiquitin Marks Proteins Selected

for Degradation

It was initially assumed that protein degradation in eu-

karyotic cells is primarily a lysosomal process.Yet, reticulo-

cytes, which lack lysosomes, selectively degrade abnormal

proteins. The observation that protein breakdown is inhib-

ited under anaerobic conditions led to the discovery of a

cytosolically based ATP-dependent proteolytic system that

is independent of the lysosomal system. This phenomenon

was thermodynamically unexpected since peptide hydroly-

sis is an exergonic process.

The analysis of a cell-free rabbit reticulocyte system

demonstrated that ubiquitin (Fig. 32-75) is required for

ATP-dependent protein degradation. This 76-residue

monomeric protein, so named because it is ubiquitous as well

as abundant in eukaryotes, is the most highly conserved pro-

tein known: It is identical in such diverse organisms as hu-

mans, toad, trout, and Drosophila and differs in only three

residues between humans and yeast. Evidently, ubiquitin is

all but uniquely suited to an essential cellular function.

Proteins that are selected for degradation are so marked

by covalently linking them to ubiquitin. This process, which

is reminiscent of amino acid activation (Section 32-2C),

Section 32-6. Protein Degradation 1409

Figure 32-75 X-ray structure of human ubiquitin. The

polypeptide is drawn in ribbon form colored in rainbow order

from its N-terminus (blue) to its C-terminus (red) with the side

chain of Lys 48 shown in stick form with C green and N blue. This

␣⫹␤architecture is known as a ␤-Grasp fold because its ␤

sheets appear to grasp its ␣ helix. [Based on an X-ray structure

by Charles Bugg, University of Alabama at Birmingham. PDBid

1UBQ.]

See Interactive Exercise 46.

JWCL281_c32_1338-1428.qxd 10/19/10 8:24 AM Page 1409

occurs in a three-step pathway that was elucidated notably

by Avram Hershko, Aaron Ciechanover, and Irwin Rose

(Fig. 32-76):

1. In an ATP-requiring reaction, ubiquitin’s terminal

carboxyl group is conjugated, via a thioester bond, to

ubiquitin-activating enzyme (E1), a homodimer of

⬃1050-residue subunits. In this process, the substrate

protein’s terminal carboxyl group is initially adenyly-

lated and then transferred to the E1 Cys¬SH group with

the elimination of AMP. Most organisms, including yeast

and humans, have only one type of E1.

2. The ubiquitin is then transferred to a specific Cys

sulfhydryl group on one of numerous proteins named

ubiquitin-conjugating enzymes (E2s; 11 in yeast and over

20 in mammals). The various E2’s are characterized by

⬃150-residue catalytic cores containing the active site Cys

that exhibit at least 25% sequence identities and which

mainly vary by the presence or absence of N- and/or C-

terminal extensions that exhibit little sequence identity to

each other. The X-ray and NMR structures of several

species of E2 reveal that their catalytic cores all assume

closely similar / structures (e.g., Fig. 32-77) in which

most of the identical residues are clustered on one surface

near the ubiquitin-accepting Cys residue, where they pre-

sumably interact with ubiquitin and E1.

3. Ubiquitin–protein ligase (E3) transfers the activated

ubiquitin from E2 to a Lys ε-amino group of its target pro-

tein, thereby forming an isopeptide bond. Each of the many

E3s present in eukaryotic cells mediates the ubiquitination

(alternatively, ubiquitylation) of a specific set of proteins

and thereby marks them for degradation. Each E3 is

served by one or a few specific E2s. Most E3s are members

of two unrelated families, those containing a HECT do-

main (HECT for homologous to E6AP C-terminus) and

those containing a so-called RING domain (also called a

RING-finger domain; RING for really interesting new

gene), although some E2s react well with members of both

families. The human genome contains 28 HECT genes and

616 RING genes, more than its number of protein kinase

genes (518), which is indicative of the specialized and

varied functions of E3s (although not all RING domain–

containing proteins are E3s). HECT domain E3s are

modularly constructed with a unique N-terminal domain

that interacts with its target proteins via their so-called

ubiquitination signals (usually short polypeptide segments;

see below) and an ⬃350-residue HECT domain that medi-

ates E2 binding and catalyzes the ubiquitination reaction.

RING domains, which are implicated in recognizing a sub-

strate protein’s ubiquitination signal, are 40- to 60-residue

motifs that each bind two structurally but not catalytically

implicated Zn

2

ions via a total of 8 Cys and His residues in

a characteristic consensus sequence (much like the zinc fin-

ger motifs in certain DNA-binding proteins, Section 34-

3Bl). RING domain–containing E3s may consist of a sin-

gle subunit or may be multisubunit proteins in which the

RING domain is contained in one subunit. HECT

E3–mediated ubiquitination occurs via the transfer of

ubiquitin from E2 to a conserved Cys residue on the

HECT domain followed by its transfer to the substrate

protein Lys side chain. In contrast, RING domain E3s act

1410 Chapter 32. Translation

Figure 32-76 Reactions involved in the attachment of

ubiquitin to a protein. In the first part of the process, ubiquitin’s

terminal carboxyl group is joined, via a thioester linkage, to E1 in

a reaction driven by ATP hydrolysis. The activated ubiquitin is

subsequently transferred to a sulfhydryl group of an E2 and

then, in a reaction catalyzed by an E3, to a Lys ε-amino group on

a condemned protein, thereby flagging the protein for

proteolytic degradation by the 26S proteasome.

Figure 32-77 X-ray structure of an E2 protein from

Arabidopsis thaliana. The protein is drawn in ribbon form

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

C-terminus (red).The side chain of Cys 88, to which ubiquitin is

covalently linked, is shown in ball-and-stick form with C green

and S yellow. [Based on an X-ray structure by William Cook,

University of Alabama at Birmingham. PDBid 2AAK.]

Ubiquitin C O

–

+ E1 SH

ATP

AMP

+ PP

Ubiquitin C NH Lys

Ubiquitin C SE1

SHE2

SHE1

Ubiquitin C SE2

Isopeptide

bond

Condemned

protein

Condemned

protein

JWCL281_c32_1338-1428.qxd 9/7/10 2:23 PM Page 1410