Voet D., Voet Ju.G. Biochemistry

Подождите немного. Документ загружается.

participate in ATP binding and are implicated in catalysis.

The Class II synthetases lack the foregoing sequences but

have three other sequences in common. Their X-ray struc-

tures reveal that these sequences occur in a so-called signa-

ture motif, a fold found only in Class II enzymes that con-

sists of a 7-stranded antiparallel  sheet with three flanking

helices, which forms the core of their catalytic domains.

Many Class I aaRSs require anticodon recognition to

aminoacylate their cognate tRNAs. In contrast, several

Class II enzymes, including AlaRS and SerRS, do not inter-

act with their bound tRNA’s anticodon. Indeed, several

class II aaRSs accurately aminoacylate “microhelices” de-

rived from only the acceptor stems of their cognate tRNAs.

Another difference between Class I and Class II syn-

thetases is that all Class I enzymes aminoacylate their

bound tRNA’s 3¿-terminal 2¿-OH group, whereas Class II

enzymes, with the exception of PheRS, all charge the 3¿-

OH group. The amino acids for which the Class I syn-

thetases are specific tend to be larger and more hydropho-

bic than those used by Class II synthetases. Finally, as Table

32-4 indicates, Class I aaRSs are mainly monomers,

whereas most Class II aaRSs are homodimers.

LysRS has been classified as a Class II aaRS. However, a

search of the genome sequences of Methanococcus jan-

naschii and Methanobacterium thermoautotrophicum failed

to reveal the presence of such a LysRS.This led to the discov-

ery that the LysRSs expressed by these archaebacteria are

Class I rather than Class II enzymes. This raises the interest-

ing question of how Class I LysRS evolved.

Prokaryotic aaRSs occur as individual protein molecules.

However, in many higher eukaryotes (e.g., Drosophila and

mammals), 9 aaRSs, some of each class, associate to form a

multienzyme particle in which the glutamyl and prolyl syn-

thetase functions are fused into a single polypeptide named

GluProRS. The advantages of these systems are unknown.

b. The Structural Features Recognized by

Aminoacyl–tRNA Synthetases May Be Quite Simple

As we shall see in Section 32-2D, ribosomes select

aminoacyl–tRNAs only via codon–anticodon interactions,

not according to the identities of their aminoacyl groups.

Accurate translation therefore requires not only that each

tRNA be aminoacylated by its cognate aaRS but that it not

be aminoacylated by any of its 19 noncognate aaRSs. More-

over, since most cells express only one aaRS for each

amino acid, each aaRS must aminoacylate all of the sev-

eral, if not many, isoaccepting tRNAs (different tRNAs

that are specific for the same amino acid) in each cell.

Considerable effort has therefore been expended, no-

tably by LaDonne Schulman, Paul Schimmel, Olke Uhlen-

beck, and John Abelson, in elucidating how aaRSs manage

this feat, despite the close structural similarities of nearly

all tRNAs. The experimental methods employed involved

the use of specific tRNA fragments, mutationally altered

tRNAs, chemical cross-linking agents, computerized se-

quence comparisons, and X-ray crystallography. The most

common synthetase contact sites on tRNA occur on the in-

ner (concave) face of the L. Other than that, there appears

to be little regularity in how the various tRNAs are recog-

nized by their cognate synthetases. Indeed, as we shall see,

some aaRSs recognize only their cognate tRNA’s acceptor

stem, whereas others also interact with its anticodon re-

gion.Additional tRNA regions may also be recognized.

Genetic manipulations by Schimmel revealed that the

tRNA features recognized by at least one type of aaRS are

surprisingly simple. Numerous sequence alterations of E.

coli tRNA

Ala

do not appreciably affect its capacity to be

aminoacylated with alanine. Yet, most base substitutions in

the G3  U70 base pair located in the tRNA’s acceptor stem

(Fig. 32-14a) greatly diminish this reaction. Moreover,

the introduction of a G  U base pair into the analogous

Section 32-2. Transfer RNA and Its Aminoacylation 1351

Figure 32-14 Major identity elements in four tRNAs. Each

base in the tRNA is represented by a filled circle. Red circles

indicate positions that have been shown to be identity elements

for the recognition of the tRNA by its cognate aminoacyl–tRNA

synthetase. The anticodon bases that are identity elements are

(b)

•

• •

•••••

•

• •

•

••••

•

tRNA

Asp

(yeast)

G73

U35

C36

U25

tRNA

Ala

(a)

•

• •

•••••

•

• •

•

••••

•

(E. coli

)

U70

C71

C72

A73

G20

•

• •

•••••

•

• •

•

••

•

(d) tRNA

Ser

G24

C11

C71

UA70

C72

G73

AU3

•

••

G10

U35

G37

C38

G34

•

• •

•••••

•

• •

•

•••

•

(E. coli

)tRNA

Gln

(c)

G36

C34

ψ38

A37

Um32

U33

G73

C71

C70

•

C25

G10

underlined. In each case, additional identity elements may yet

be discovered.The base at position 73, which is an identity

element in all four tRNAs shown here, is known as the

discriminator base.

JWCL281_c32_1338-1428.qxd 9/7/10 2:22 PM Page 1351

1352 Chapter 32. Translation

ever,this altered tRNA

Ile

is also a much better substrate for

MetRS than it is for IleRS. Thus, both the codon and the

amino acid specificity of this tRNA are changed by a single

post-transcriptional modification. The N

-methylation of

G37 in yeast tRNA

Asp

(Fig. 32-14b) provides another ex-

ample of a base modification forming an identity element.

In the absence of this N

-methyl group, tRNA

Asp

is recog-

nized by ArgRS, largely via its C36 and G37, whereas

ArgRS normally recognizes only tRNA

Arg

, mainly via its

C35 and U36.

The available experimental evidence has largely located

the various tRNA identifiers in the acceptor stem and the an-

ticodon loop (Fig. 32-16). The X-ray structures of several

aaRS  tRNA complexes, which we consider next, have

structurally rationalized some of these observations.

c. The X-Ray Structure of GlnRS  tRNA

Gln

, a

Class I Complex

The X-ray structures of all 20 different amino acid–

specific aaRSs from numerous organisms have been deter-

mined, many of which are in complex with ATP, their cog-

nate amino acids, or their analogs. These structures reveal

that the active sites of these enzymes bind the ATP and

target amino acid in optimal positions for in-line nucle-

ophilic displacement (Section 16-2B) during amino acid

position of tRNA

Cys

and tRNA

Phe

causes them to be

aminoacylated with alanine even though there are few

other sequence identities between these mutant tRNAs and

tRNA

Ala

(e.g., Fig. 32-15). In fact, E. coli AlaRS even effi-

ciently aminoacylates a 24-nt “microhelix” derived from

only the G3  U70–containing acceptor stem of E. coli

tRNA

Ala

. Since the only E. coli tRNAs that normally have

a G3  U70 base pair are the tRNA

Ala

, and this base pair is

also present in the tRNA

Ala

from many organisms including

yeast (Fig. 32-8), the foregoing observations strongly sug-

gest that the G3  U70 base pair is a major feature recognized

by AlaRSs. These enzymes presumably recognize the distorted

shape of the G  U base pair (Fig. 32-12), a hypothesis corrob-

orated by the observation that base changes at G3  U70 which

least affect the acceptor identity of tRNA

Ala

yield base pairs

that structurally resemble G  U.

The elements of three other tRNAs, which are recog-

nized by their cognate tRNA synthetases, are indicated in

Fig. 32-14.As with tRNA

Ala

, these identity elements appear

to comprise only a few bases. Note that the anticodon forms

an identity element in two of these tRNAs. In another

example of an anticodon identifier, the E. coli tRNA

Ile

spe-

cific for the codon AUA has the anticodon LAU, where L is

lysidine, a modified cytosine whose 2-keto group is re-

placed by the amino acid lysine (Fig. 32-10). The L in this

context pairs with A rather than G, a rare instance of base

modification altering base pairing specificity. The replace-

ment of this L with unmodified C, as expected, yields a

tRNA that recognizes the Met codon AUG (codons bind

anticodons in an antiparallel fashion). Surprisingly, how-

Figure 32-15 Three-dimensional model of E. coli tRNA

Ala

This model is based on the X-ray structure of yeast tRNA

Phe

(Fig.

32-11b) in which the nucleotides that are different in E. coli

tRNA

Cys

are highlighted in cyan and the G3  U70 base pair is

highlighted in ivory. [Courtesy of Ya-Ming Hou, MIT.]

Figure 32-16 Experimentally observed identity elements of

tRNAs. The tRNA backbone is cyan and each of its nucleotides

is represented by a yellow circle whose diameter is proportional

to the fraction of the 20 tRNA acceptor types for which the

nucleotide is an observed determinant. [Courtesy of William

McClain, University of Wisconsin.]

JWCL281_c32_1338-1428.qxd 8/19/10 10:05 PM Page 1352

(a)

activation and that the specificity of an aaRS for its target

amino acid is determined by idiosyncratic contacts with

the side chain of the amino acid.

The X-ray structures of 16 different aaRSs in their com-

plexes with their cognate tRNAs (all but those of Gly,Ala,

Lys, and His) have so far been reported.The first of them to

be elucidated, that of E. coli GlnRS, a Class I synthetase, in

its complex with tRNA

Gln

and ATP (Fig. 32-17), was deter-

mined by Thomas Steitz. The tRNA

Gln

assumes an L-

shaped conformation that resembles those of tRNAs of

known structures (e.g., Fig. 32-11b). GlnRS, a 553-residue

monomeric protein that consists of four domains arranged

to form an elongated molecule, interacts with the tRNA

along the entire inside face of the L such that the anticodon

is bound near one end of the protein and the acceptor stem

is bound near its other end.

Genetic and biochemical data indicate that the identity

elements of tRNA

Gln

are largely clustered in its anticodon

loop and acceptor stem (Fig. 32-14c).The anticodon loop of

tRNA

Gln

is extended by two novel non-Watson–Crick base

pairs (2¿-O-methyl-U32 ⴢ ␺38 and U33 ⴢ m

A37), thereby

causing the bases of the anticodon to unstack and splay

outward in different directions so as to bind in separate

recognition pockets of GlnRS. These structural features

suggest that GlnRS uses all seven bases of the anticodon

loop to discriminate among tRNAs. Indeed, changes to any

one of the bases of residues C34 through ␺38 yield tRNAs

with decreases in k

cat

for aminoacylation by GlnRS by

factors ranging from 70 to 28,000.

The GCCA at the 3¿ end of the tRNA

Gln

makes a hair-

pin turn toward the inside of the L rather than continuing

helically onward (as does the ACCA at the 3¿ end in the X-

ray structure of tRNA

Phe

; Fig. 32-11b). This conformation

change is facilitated by the insinuation of a Leu side chain

between the 5¿ and 3¿ ends of the tRNA so as to disrupt the

first base pair of the acceptor stem (U1 ⴢ A72).The GlnRS

reaction is therefore relatively insensitive to base changes

in these latter two positions except when base pairing is

strengthened by their conversion to G1 ⴢ C72. The GCCA

end of the tRNA

Gln

plunges deeply into a protein pocket

Section 32-2. Transfer RNA and Its Aminoacylation 1353

Figure 32-17 X-ray structure of E. coli GlnRS ⴢ tRNA

Gln

ⴢ

ATP. (a) The tRNA is drawn in stick form colored as the ATP

but with the C atoms of the anticodon (UCG) and the 3¿-CCA

end yellow. An orange rod links its successive P atoms. The ATP

bound in the protein’s active site is drawn in space-filling form

with C green, N blue, O red, and P orange. The protein is

represented by a semitransparent light blue surface diagram that

reveals the buried portions of the tRNA and ATP. Note that both

the 3¿ end of the tRNA (top right) and its anticodon bases

(b)

(bottom) are inserted into deep pockets in the protein. (b) The

complex viewed as in Part a.The tRNA’s sugar–phosphate

backbone is represented by an orange worm and the bases

forming its identity elements (Fig. 32-14c) are drawn in stick form

colored according to atom type with C magenta, N blue, and O

red.The ATP is drawn as in Part a. The protein is shown in ribbon

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

C-terminus (red). [Based on an X-ray structure by Thomas Steitz,

Yale University. PDBid 1GTR.]

See Kinemage Exercise 20

JWCL281_c32_1338-1428.qxd 10/27/10 1:40 PM Page 1353

1354 Chapter 32. Translation

that also binds the enzyme’s ATP and glutamine substrates.

Three protein “fingers” are inserted into the minor groove

of the acceptor stem to make sequence-specific interac-

tions with base pairs G2  C71 and G3  C70 [recall that

double helical RNA has an A-DNA-like structure (Section

29-1Bc) whose wide minor groove readily admits protein

but whose major groove is normally too narrow to do so].

The GlnRS domain that binds glutamine, ATP, and the

GCCA end of tRNA

Gln

, the so-called catalytic domain, con-

tains, as we previously discussed, a dinucleotide-binding

fold. Much of this domain is nearly superimposable with and

thus evolutionarily related to the corresponding domains of

other Class I aaRSs.

d. The X-Ray Structure of AspRS  tRNA

Asp

, a

Class II Complex

Yeast AspRS, a Class II synthetase, is an 

dimer of

557-residue subunits. Its X-ray structure in complex with

tRNA

Asp

, determined by Moras, reveals that the protein

symmetrically binds two tRNA molecules (Fig. 32-18).Like

GlnRS, AspRS principally contacts its bound tRNA both

at the end of its acceptor stem and in its anticodon region.

The contacts in these two enzymes are, nevertheless, quite

different in character (Fig. 32-19): Although both tRNAs

approach their cognate synthetases along the inside of

their L shapes, tRNA

Gln

does so toward the direction of the

minor groove of its acceptor stem, whereas tRNA

Asp

does

so toward the direction of its major groove. The GCCA at

the 3¿ end of tRNA

Asp

thereby continues its helical track as

it plunges into AspRS’s catalytic site, whereas, as we saw,

the GCCA end of tRNA

Gln

bends backward into a hairpin

turn that opens up the first base pair (U1  A72) of its

acceptor stem. Although the deep major groove of an

A-RNA helix is normally too narrow to admit groups

larger than water molecules (Section 29-1Bc), the major

groove at the end of the acceptor stem in AspRS  tRNA

Asp

is sufficiently widened for its base pairs to interact with a

protein loop.

The anticodon arm of tRNA

Asp

is bent by as much as 20

Å toward the inside of the L relative to that in the X-ray

structure of uncomplexed tRNA

Asp

and its anticodon bases

are unstacked.The hinge point for this bend is a G30  U40

base pair in the anticodon stem which, in nearly all other

species of tRNA, is a Watson–Crick base pair. The anti-

codon bases of tRNA

Gln

are also unstacked in contacting

GlnRS but with a backbone conformation that differs from

that in tRNA

Asp

. Evidently, the conformation of a tRNA in

complex with its cognate synthetase appears to be dictated

more by its interactions with the protein (induced fit) than

by its sequence.

Structural analyses of complexes of AspRS  tRNA

Asp

with ATP and aspartic acid, and of GlnRS  tRNA

Gln

with

ATP, have permitted models of the aminoacyl–AMP com-

plexes of these enzymes to be independently formulated.

Comparison of these models reveals that the 3¿-terminal A

residues of tRNA

Gln

and tRNA

Asp

(to which the aminoacyl

groups are appended; Fig. 32-13) are positioned on oppo-

site sides of the enzyme-bound aminoacyl–AMP interme-

diate (Fig. 32-20).The 3¿-terminal ribose residues are puck-

ered C2¿-endo for tRNA

Asp

and C3¿-endo for tRNA

Gln

(see

Fig. 29-8) such that the 2¿-hydroxyl group of tRNA

Gln

(Class I) is stereochemically positioned to attack the

Figure 32-18 X-ray structure of yeast AspRS  tRNA

Asp



ATP. (a) The homodimeric enzyme with its two symmetrically

bound tRNAs is viewed with its 2-fold axis approximately

vertical.The tRNAs are drawn in skeletal form colored according

to atom type with the C atoms of the anticodon (GUC) and the

3¿-CCA end yellow, the remaining C atoms green, N blue, O red,

and P orange. An orange rod connects its successive P atoms. The

two protein subunits are represented by semitransparent pink

and light blue surface diagrams that reveal the buried portions of

(a)

(b)

the tRNAs. (b) A ribbon diagram of the AspRS  tRNA

Asp

protomer.The tRNA’s sugar–phosphate backbone is represented

by an orange worm and the bases forming its identity elements

(Fig. 32-14b) are drawn in stick form colored according to atom

type with C magenta, N blue, and O red. The protein is shown in

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

to its C-terminus (red). [Based on an X-ray structure by Dino

Moras, CNRS/INSERM/ULP, Illkirch Cédex, France. PDBid

1ASY.]

JWCL281_c32_1338-1428.qxd 8/19/10 10:05 PM Page 1354

aminoacyl–AMP’s carboxyl group, whereas for tRNA

Asp

(Class II), only the 3¿-hydroxyl group is situated to do so.

This clearly explains the different aminoacylation specifici-

ties of the Class I and Class II aaRSs.

e. Proofreading Enhances the Fidelity of Amino Acid

Attachment to tRNA

The charging of a tRNA with its cognate amino acid is a

remarkably accurate process: aaRSs display an overall er-

ror rate of about 1 in 10,000.We have seen that aaRSs bind

only their cognate tRNAs through an intricate series of

specific contacts. But how do they discriminate among the

various amino acids, some of which are quite similar?

Experimental measurements indicate, for example, that

IleRS transfers as many as 40,000 isoleucines to tRNA

Ile

for every valine it so transfers. Yet, as Linus Pauling first

pointed out, there are insufficient structural differences be-

tween Val and Ile to permit such a high degree of discrimina-

tion in the direct generation of aminoacyl–tRNAs. The X-

ray structure of Thermus thermophilus IleRS, a monomeric

Section 32-2. Transfer RNA and Its Aminoacylation 1355

Figure 32-19 Comparison of the modes by which GlnRS and

AspRS bind their cognate tRNAs. The proteins and tRNAs are

represented by blue and red spheres centered on their C



and P

atom positions. Note how GlnRS (a), a Class I synthetase, binds

tRNA

Gln

from the minor groove side of its acceptor stem so as to

bend its 3¿ end into a hairpin conformation. In contrast, AspRS

(b), a Class II synthetase, binds tRNA

Asp

from the major groove

side of its acceptor stem so that its 3¿ end continues its helical

path on entering the active site. [After drawings by Dino Moras,

CNRS/INSERM/ULP, Illkirch Cédex, France. PDBids 1GTR

abd 1ASY.]

A76

tRNA

Asp

(Class II)

A76

tRNA

Gln

(Class I)

′

O3′

Aminoacyl–AMP

Figure 32-20 Comparison of the stereochemistries of aminoacylation by

Class I and Class II aaRSs. The positions of the 3¿ terminal adenosine residues

(A76) of AspRS (Class II, left) and GlnRS (Class I, right) are drawn relative

to that of the enzyme-bound aminoacyl–AMP (below; only the carbonyl

group of its aminoacyl residue is shown). Note how only O3¿ of tRNA

Asp

and

O2¿ of tRNA

Gln

are suitably positioned to attack the aminoacyl residue’s

carbonyl group and thereby transfer the aminoacyl residue to the tRNA.

[After Cavarelli, J., Eriani, G., Rees, B., Ruff, M., Boeglin, M., Mitschler,A.,

Martin, F., Gangloff, J., Thierry, J.-C., and Moras, D., EMBO J. 13, 335 (1994).]

(a)

(b)

JWCL281_c32_1338-1428.qxd 8/4/10 4:44 PM Page 1355

(a product of Pseudomonas fluorescens that acts by specif-

ically binding to bacterial IleRS so as to inhibit bacterial

protein synthesis), determined by Steitz, suggests how

IleRS carries out its editing process. The X-ray struc-

ture (Fig. 32-21) reveals that this complex resembles the

GlnRS  tRNA

Gln

 ATP complex (Fig. 32-17) but with IleRS

having an additional editing domain (also called CP1 for

connective peptide 1) inserted in its Rossmann fold domain.

The two 3¿ terminal residues of the tRNA

Ile

, C75 and A76,

are disordered but, when modeled so as to continue the ac-

ceptor stem’s stacked A-form helix, extend into a cleft in

the editing domain that has been implicated as its hy-

drolytic site (Fig. 32-22a, left).Thus, this IleRS complex ap-

pears to resemble an “editing complex” instead of a “trans-

fer complex” as seen in the GlnRS structure. However, a

transfer complex would form if the 3¿ ending segment of

the tRNA

Ile

assumes a hairpin conformation (Fig. 32-22a,

right) similar to that in the GlnRS structure (Figs. 32-17b

and 32-19a; recall that IleRS and GlnRS are both Class I

aaRSs). Steitz has therefore postulated that the aminoacyl

group is shuttled between the IleRS’s aminoacylation site

and its editing site by such a conformational change (Fig.

32-22b). This process functionally resembles the way in

which DNA polymerase I edits its newly synthesized

strand (Section 30-2Ag), which Steitz also elucidated.

ValRS is a monomeric Class I aaRS that resembles IleRS.

The X-ray structure of the complex of T. thermophilus

ValRS, tRNA

Va l

, and the nonhydrolyzable valyl–adenylate

analog 5-O-[N-(

L-valyl)sulfamoyl]adenosine (Val-AMS),

5-O-[N-(L-Valyl)sulfamoyl]adenosine (Val-AMS)

OH OH

NH S

NCH

1356 Chapter 32. Translation

Class I aaRS, in complex with isoleucine, determined by

Shigeyuki Yokoyama and Schimmel, indicates that

isoleucine fits snugly into its binding site in the enzyme’s

Rossmann fold domain and hence that this binding site

would sterically exclude leucine as well as larger amino

acids. However, valine, which differs from isoleucine by

only the lack of a single methylene group, fits into this

isoleucine-binding site. The binding free energy of a meth-

ylene group is estimated to be ⬃12 kJ  mol

1

. Equation

[3.17] indicates that the ratio f of the equilibrium constants,

and K

, with which two substances bind to a given bind-

ing site is given by

[32.1]

where G°G

  G

 is the difference between the

free energies of binding of the two substances. It is there-

fore estimated that isoleucyl–tRNA synthetase could dis-

criminate between isoleucine and valine by no more than a

factor of ⬃100.

Berg resolved this apparent paradox by demonstrating

that, in the presence of tRNA

Ile

, IleRS catalyzes the nearly

quantitative hydrolysis of valyl–adenylate to valine  AMP

rather than forming Val–tRNA

Ile

. Moreover, the few

Val–tRNA

Ile

molecules that do form are hydrolyzed to va-

line  tRNA

Ile

.Thus, IleRS subjects both aminoacyl–adeny-

late and aminoacyl–tRNA

Ile

to a proofreading or editing

step that occurs at a separate catalytic site.This site binds Val

residues but excludes the larger Ile residues. The enzyme’s

overall selectivity is therefore the product of the selectivities

of its synthesis and proofreading steps, thereby accounting

for the high fidelity of aminoacylation. Note that in this so-

called double-sieve mechanism, editing occurs at the ex-

pense of ATP hydrolysis, the thermodynamic price of high

fidelity (increased order).

The X-ray structure of Staphylococcus aureus IleRS in

complex with tRNA

Ile

and the clinically useful antibiotic

mupirocin

f 



¢G

°¿



¢G

°¿



 e

¢¢G°¿



(CH

)

Mupirocin

Figure 32-21 X-ray structure of S. Aureus isoleucyl–tRNA

synthetase in complex with tRNA

Ile

and mupirocin. The tRNA is

white, the protein is colored by domain, and the mupirocin is

shown in stick form in pink. [Courtesy of Thomas Steitz,Yale

University. PDBid 1QU2.]

JWCL281_c32_1338-1428.qxd 8/4/10 4:44 PM Page 1356

determined by Yokoyama, reveals that the Val-AMS is

bound in the aminoacylation pocket in the Rossmann fold

domain, which accommodates the isosteric Val and Thr

moieties but sterically excludes Ile. Modeling studies based

on the IleRS  tRNA

Ile

 mupirocin structure indicate that

Valyl-adenylate

OH OH

O P

the Thr side chain would fit into the ValRS editing pocket

with its side chain hydroxyl group hydrogen bonded to the

side chain of Asp 279 of ValRS, which protrudes into the

pocket in contrast to the corresponding Asp 328 of IleRS,

which does not. Consequently, a Val side chain would be

excluded from the ValRS editing pocket because it cannot

form such a hydrogen bond, thereby explaining why this

editing pocket hydrolyzes threonyl–adenylate and Thr–

tRNA

Va l

but not the corresponding Val derivatives. The

ValRS  tRNA

Val

structure also indicates that ValRS and

tRNA

Va l

together form a tunnel connecting the ValRS’s

aminoacylation pocket with its editing pocket. Improperly

formed threonyl–adenylate is proposed to be channeled

through this tunnel for hydrolysis in the editing pocket,

thereby explaining why tRNA

Va l

must be bound to ValRS for

this pretransfer editing reaction to occur.Valyl–adenylate is

presumably channeled through the similar IleRS  tRNA

Ile

complex for its hydrolysis.

Section 32-2. Transfer RNA and Its Aminoacylation 1357

Figure 32-22 Comparison of the putative aminoacylation and

editing modes of IleRS  tRNA

Ile

. (a) The superposition of

tRNA

Ile

in these two binding modes on the solvent-accessible

surface of IleRS (green).The acceptor strand of tRNA

Ile

in the

editing mode observed in the X-ray structure of IleRS 

tRNA

Ile

 mupirocin (Fig. 32–21) is drawn in ribbon form in white

with the modeled positions of C75 and A76 in red.This places

the tRNA’s 3¿ end in the editing site. In contrast, the three 3¿

terminal residues of tRNA

Ile

, as positioned through homology

modeling based on the X-ray structure of GlnRS  tRNA

Gln



ATP (Fig. 32-17) and drawn in ball-and-stick form with C yellow,

N blue, O red, and P magenta, places the tRNA’s 3¿ end in the

synthetic (aminoacylation) site, 34 Å distant from its position in

the editing site. Note that there is a cleft running between the

editing and synthetic sites and that the 3¿ end of the tRNA

continues its A-form helical path in the editing mode but

assumes a hairpin conformation in the synthetic mode. (b) A

cartoon comparing the positions of the 3¿ end of tRNA

Ile

in its

complex with IleRS in its synthetic mode (left) and in its editing

mode (right). [Part a courtesy of and Part b based on a drawing

by Thomas Steitz, Yale University.]

Synthetic

domain

Synthetic mode Editing mode

Editing

domain

5′

3′

5′

(a)

(b)

JWCL281_c32_1338-1428.qxd 8/4/10 4:44 PM Page 1357

Class I

Class II

Catalytic

domain

Catalytic

domain

Editing

domain

Editing

domain

tRNA acceptor stem

ThrRS, a Class II homodimer, has the opposite problem

of ValRS: It must synthesize Thr–tRNA

Thr

but not Val–

tRNA

Thr

.The X-ray structure of E. coli ThrRS that lacks its

N-terminal domain but remains catalytically active in a

complex with either threonine or the threonyl–adenylate

analog Thr-AMS, determined by Moras, reveals that

ThrRS’s aminoacylation pocket contains a Zn

2

ion that is

coordinated by the side chain hydroxyl and amino groups

of the threonyl group as well as by three protein side

chains. The isosteric valine could not coordinate the Zn

2

ion in this way and hence does not undergo adenylylation

by ThrRS. However, what prevents ThrRS from synthesiz-

ing Ser–tRNA

Thr

? In fact, the truncated ThrRS synthesizes

Ser–tRNA

Thr

at more than half the rate it synthesizes

Thr–tRNA

Thr

, thereby indicating that the N-terminal do-

main of wild-type ThrRS contains the enzyme’s editing site.

Mutational analysis of ThrRS has localized this editing site

to a cleft in the N-terminal domain of wild-type ThrRS,

whose X-ray structure in complex with tRNA

Thr

was also

determined by Moras. In this latter structure, the tRNA’s 3¿

end follows a regular helical path similar to that seen in the

X-ray structure of AspRS  tRNA

Asp

 ATP (Fig. 32-18) so

as to enter the aminoacylation site. However, if the 3¿ end

of the bound tRNA

Thr

assumed a hairpin conformation

similar to that seen in X-ray structure of tRNA

Gln

in com-

plex with the Class I enzyme GlnRS and ATP (Fig. 32-17),

its covalently linked aminoacyl group would enter the edit-

ing site. This indicates an intriguing “mirror symmetry”

(Fig. 32-23): In Class I aaRSs that mediate a double-sieve

editing mechanism, the 3¿ end of the bound cognate tRNA

assumes a hairpin conformation when it enters the aminoa-

cylation site and a helical conformation when it enters the

editing site, whereas the converse holds for Class II aaRSs.

Finally, ThrRS does not appear to mediate pretransfer ed-

iting (does not hydrolyze seryl–adenylate), and, in fact, the

ThrRS  tRNA

Thr

complex lacks a channel connecting its

aminoacylation and editing sites such as is seen in the

ValRS  tRNA

Va l

complex.

Synthetases that have adequate selectivity for their cor-

responding amino acid lack editing functions. Thus, for ex-

ample, the TyrRS aminoadenylylation site discriminates

between tyrosine and phenylalanine through hydrogen

bonding with the tyrosine ¬OH group. The cell’s other

amino acids, standard as well as nonstandard, have even

less resemblance to tyrosine, which rationalizes why TyrRS

lacks an editing site.

f. Gln–tRNA

Gln

May Be Formed via an

Alternative Pathway

Although it was long believed that each of the 20 stan-

dard amino acids is covalently linked to a tRNA by its corre-

sponding aaRS, it is now clear that gram-positive bacteria,

archaebacteria, cyanobacteria, mitochondria, and chloro-

plasts all lack GlnRS. Rather glutamate is linked to tRNA

Gln

by the same GluRS that synthesizes Glu–tRNA

Glu

. The re-

sulting Glu–tRNA

Gln

is then transamidated to Gln–tRNA

Gln

by the enzyme Glu–tRNA

Gln

amidotransferase (Glu-AdT)

in an ATP-requiring reaction in which glutamine is the

amide donor. Some microorganisms use a similar transami-

dation pathway for the synthesis of Asn–tRNA

Asn

from

Asp–tRNA

Asn

The overall reaction catalyzed by Glu-AdT occurs in

three stages (Fig. 32-24): (1) Glutamine is hydrolyzed to

glutamate and the resulting NH

sequestered; (2) ATP re-

acts with the Glu side chain of Glu–tRNA

Gln

to yield an ac-

tivated acylphosphate intermediate and ADP; and (3) the

acylphosphate intermediate reacts with the NH

to yield

Gln–tRNA

Gln

 P

. Glu-AdT from Bacillus subtilis, which

was characterized by Dieter Söll, is a heterotrimeric pro-

tein, none of whose subunits exhibit significant sequence

similarity to GlnRS. The genes encoding these subunits,

gatA, gatB, and gatC, form a single operon whose disrup-

tion is lethal, thereby demonstrating that B. subtilis has no

alternative pathway for Gln–tRNA

Gln

production. The

GatA subunit of Glu-AdT appears to catalyze the activa-

tion of the side chain carboxyl of glutamic acid via a reac-

tion resembling that catalyzed by carbamoyl phosphate

synthetase (Section 26-2A). Nevertheless, GatA exhibits

no sequence similarity with other known glutamine amido-

transferases (members of the triad or Ntn families; Section

26-5Aa). The GatB subunit may be used to select the cor-

rect tRNA substrate. The role of the GatC subunit is un-

clear, although the observation that its presence is neces-

sary for the expression of GatA in E. coli suggests that it

participates in the modification, folding, and/or stabiliza-

tion of GatA.

Since Glu is not misincorporated into B. subtilis proteins

in place of Gln, the Glu–tRNA

Gln

product of the above

aminoacylation reaction must not be transported to the ri-

bosome in the same way as Gln–tRNA

Gln

. It is likely that

this occurs because, as has been shown in chloroplasts, EF-

Tu, the elongation factor that binds and transports most

bacterial aminoacyl–tRNAs to the ribosome in a GTP-

dependent process (Section 32-3D), does not bind

Glu–tRNA

Gln

. It is unclear why two independent routes

have evolved for the synthesis of Gln–tRNA

Gln

1358 Chapter 32. Translation

Figure 32-23 Schematic diagram of the aminoacylation and

editing mechanisms of Class I and Class II aaRSs emphasizing

the “mirror symmetry” of their overall mechanisms. With Class I

aaRSs (left; e.g., IleRS), the 3¿ end of the bound tRNA’s acceptor

stem assumes a hairpin conformation in the synthetic mode and

a helical conformation in the editing mode, whereas the converse

occurs with Class II aaRSs (right; e.g.,ThrRS). [Courtesy of Dino

Moras, CNRS/INSERM/ULP, Illkirch Cedex, France.]

JWCL281_c32_1338-1428.qxd 8/4/10 4:44 PM Page 1358

g. Some Archaebacteria Lack a Separate CysRS

The genomes of certain archaebacteria such as M. jan-

naschii lack an identifiable gene for CysRS. This is because

the enzyme responsible for synthesizing Pro–tRNA

Pro

these organisms also synthesizes Cys–tRNA

Cys

. Interest-

ingly, this enzyme, which is named ProCysRS, does not syn-

thesize Pro–tRNA

Cys

or Cys–tRNA

Pro

.Although ProCysRS

synthesizes cysteinyl–adenylate only in the presence of

tRNA

Cys

, it synthesizes prolyl–adenylate in the absence of

tRNA

Pro

.The binding of tRNA

Cys

to ProCysRS blocks the

activation of proline so that only cysteine can be activated.

Conversely, the activation of proline facilitates the binding

of tRNA

Pro

while preventing the binding of tRNA

Cys

. How-

ever, the mechanism through which ProCysRS carries out

these mutually exclusive syntheses is unknown. In any case,

it appears that some organisms can get by with as few as 17

different aaRSs; they may lack GlnRS, AspRS, and a sepa-

rate CysRS.

D. Codon–Anticodon Interactions

In protein synthesis, the proper tRNA is selected only

through codon–anticodon interactions; the aminoacyl group

does not participate in this process. This phenomenon was

demonstrated as follows. Cys–tRNA

Cys

, in which the Cys

residue was

C labeled, was reductively desulfurized with

Raney nickel so as to convert the Cys residue to Ala:

The resulting

C-labeled hybrid, Ala–tRNA

Cys

, was added

to a cell-free protein synthesizing system extracted from

rabbit reticulocytes. The product hemoglobin  chain’s

only radioactive tryptic peptide was the one that normally

contains the subunit’s only Cys. No radioactivity was found

in the peptides that normally contain Ala but no Cys.

C C

tRNA

Cys

Ni(H)

Cys–tRNA

Cys

C C tRNA

Cys

Ala–tRNA

Cys

Raney nickel

S Ni

Section 32-2. Transfer RNA and Its Aminoacylation 1359

COO

–

COO

–

Glutamine

Glutamate

–

Glu–tRNA

Gln

tRNA

Gln

Gln–tRNA

Gln

tRNA

Gln

Carboxyphosphate intermediate

tRNA

Gln

ATP

ADP

2–

Figure 32-24 The Glu-AdT–mediated synthesis of

Gln–tRNA

Gln

from Glu–tRNA

Gln

. The reaction involves the

ATP-activated transfer of a glutamine-derived NH

to the

glutamate moiety of Glu–tRNA

Gln

JWCL281_c32_1338-1428.qxd 8/4/10 4:44 PM Page 1359

Evidently, only the anticodons of aminoacyl–tRNAs partic-

ipate in codon recognition.

a. Genetic Code Degeneracy Is Largely Due to

Variable Third Position Codon–Anticodon Interactions

One might naively guess that each of the 61 codons

specifying an amino acid would be read by a different

tRNA. Yet, even though most cells contain several groups

of isoaccepting tRNAs, many tRNAs bind to two or three of

the codons specifying their cognate amino acids. For exam-

ple, yeast tRNA

Phe

,which has the anticodon GmAA, recog-

nizes the codons UUC and UUU (remember that the anti-

codon pairs with the codon in an antiparallel fashion),

and yeast tRNA

Ala

, which has the anticodon IGC, recog-

nizes the codons GCU, GCC, and GCA.

It therefore seems that non-Watson–Crick base pairing can

occur at the third codon–anticodon position (the anti-

codon’s first position is defined as its 3¿ nucleotide), the site

of most codon degeneracy (Table 32-2). Note also that the

third (5¿) anticodon position commonly contains a modi-

fied base such as Gm or I.

b. The Wobble Hypothesis Structurally Accounts for

Codon Degeneracy

By combining structural insight with logical deduction,

Crick proposed, in what he named the wobble hypothesis,

how a tRNA can recognize several degenerate codons. He

assumed that the first two codon–anticodon pairings have

normal Watson–Crick geometry. The structural constraints

that this places on the third codon–anticodon pairing en-

sure that its conformation does not drastically differ from

that of a Watson–Crick pair. Crick then proposed that

there could be a small amount of play or “wobble” in the

third codon position which allows limited conformational

adjustments in its pairing geometry. This permits the for-

mation of several non-Watson–Crick pairs such as U  G

and I  A (Fig. 32-25a). The allowed “wobble” pairings are

indicated in Fig. 32-25b. Then, by analyzing the known pat-

tern of codon–anticodon pairing, Crick deduced the most

plausible sets of pairing combinations in the third

codon–anticodon position (Table 32-5).Thus, an anticodon

...

3 3 55

nticodon: C G I C GI

5 5 33

Codon: G C U

...

3 5

Anticodon: C G I

5 3

Codon: G C A

G C C

...

3 3 55

nticodon: A A Gm A AGm

5 5 33

Codon: U U C U U U

with C or A in its third position can only pair with its Wat-

son–Crick complementary codon. If U, G, or I occupies the

third anticodon position, two, two, or three codons are rec-

ognized, respectively.

1360 Chapter 32. Translation

Table 32-5 Allowed Wobble Pairing Combinations in the

Third Codon–Anticodon Position

5¿-Anticodon Base 3¿-Codon Base

U A or G

G U or C

I U,C,or A

Figure 32-25 Wobble pairing. (a) U  G and I  A wobble

pairs. Both have been observed in X-ray structures. (b) The

geometry of wobble pairing.The spheres and their attached

bonds represent the positions of ribose C1¿ atoms with their

accompanying glycosidic bonds. X (left) designates the

nucleoside at the 5¿ end of the anticodon (tRNA). The positions

on the right are those of the 3¿ nucleoside of the codon (mRNA)

in the indicated wobble pairings. [After Crick, F.H.C., J. Mol.

Biol. 19, 552 (1966).]

C1'

(a)

...

Anticodon

...

Standard:

Codon

(b)

JWCL281_c32_1338-1428.qxd 8/4/10 4:44 PM Page 1360