Voet D., Voet Ju.G. Biochemistry

Подождите немного. Документ загружается.

2. The code is read in a sequential manner starting from

a fixed point in the gene. The insertion or deletion of a nu-

cleotide shifts the frame (grouping) in which succeeding

nucleotides are read as codons (insertions or deletions of

nucleotides are therefore also known as frameshift muta-

tions).Thus the code has no internal punctuation that indi-

cates the reading frame; that is, the code is comma free.

3. The code is a triplet code.

4. All or nearly all of the 64 triplet codons code for an

amino acid; that is, the code is degenerate.

These principles are illustrated by the following anal-

ogy. Consider a sentence (gene) in which the words

(codons) each consist of three letters (bases).

(Here the spaces separating the words have no physical sig-

nificance; they are only present to indicate the reading

frame.) The deletion of the fourth letter, which shifts the

reading frame, changes the sentence to

so that all words past the point of deletion are unintelligi-

ble (specify the wrong amino acids). An insertion of any

letter, however, say an X in the ninth position,

restores the original reading frame. Consequently, only the

words between the two changes (mutations) are altered.As

in this example, such a sentence might still be intelligible

(the gene could still specify a functional protein), particu-

larly if the changes are close together.Two deletions or two

insertions, no matter how close together, would not sup-

press each other but just shift the reading frame. However,

three insertions, say X, Y, and Z in the fifth, eighth, and

twelfth positions, respectively, would change the sentence to

which, after the third insertion, restores the original read-

ing frame. The same would be true of three deletions. As

before, if all three changes were close together, the sen-

tence might still retain much of its meaning.

Crick and Brenner did not unambiguously demonstrate

that the genetic code is a triplet code because they had no

proof that their insertions and deletions involved only single

nucleotides. Strictly speaking,they showed that a codon con-

sists of 3r nucleotides where r is the number of nucleotides in

an insertion or deletion. Although it was generally assumed

at the time that r  1, proof of this assertion had to await the

elucidation of the genetic code (Section 32-1C).

C. Deciphering the Genetic Code

The genetic code could, in principle, be determined by sim-

ply comparing the base sequence of an mRNA with the

amino acid sequence of the polypeptide it specifies. In the

1960s, however, techniques for isolating and sequencing

THE BXI GYR EDZ FOX ATE THE EGG

THE IGR EDX FOX ATE THE EGG

THE IGR EDF OXA TET HEE GG

THE BIG RED FOX ATE THE EGG

Section 32-1. The Genetic Code 1341

mRNAs had not yet been developed. The elucidation of

the genetic code dictionary therefore proved to be a diffi-

cult task.

a. UUU Specifies Phe

The major breakthrough in deciphering the genetic code

came in 1961 when Marshall Nirenberg and Heinrich

Matthaei established that UUU is the codon specifying Phe.

They did so by demonstrating that the addition of poly(U) to

a cell-free protein synthesizing system stimulates only the

synthesis of poly(Phe). The cell-free protein synthesizing

system was prepared by gently breaking open E. coli cells by

grinding them with powdered alumina and centrifuging the

resulting cell sap to remove the cell walls and membranes.

This extract contained DNA, mRNA, ribosomes, enzymes,

and other cell constituents necessary for protein synthesis.

When fortified with ATP, GTP, and amino acids, the system

synthesized small amounts of proteins. This was demon-

strated by the incubation of the system with

C-labeled

amino acids followed by the precipitation of its proteins by

the addition of trichloroacetic acid. The precipitate proved

to be radioactive.

A cell-free protein synthesizing system, of course, pro-

duces proteins specified by the cell’s DNA. On addition of

DNase, however, protein synthesis stops within a few min-

utes because the system can no longer synthesize mRNA,

whereas the mRNA originally present is rapidly degraded.

Nirenberg found that crude mRNA-containing fractions

from other organisms were highly active in stimulating pro-

tein synthesis in a DNase-treated protein synthesizing sys-

tem.This system is likewise responsive to synthetic mRNAs.

The synthetic mRNAs that Nirenberg used in subse-

quent experiments were synthesized by the Azotobacter

vinelandii enzyme polynucleotide phosphorylase. This en-

zyme, which was discovered by Severo Ochoa and Mari-

anne Grunberg-Manago, links together nucleotides in the

reaction

In contrast to RNA polymerase, however, polynucleotide

phosphorylase does not utilize a template. Rather, it ran-

domly links together the available NDPs so that the base

composition of the product RNA reflects that of the reac-

tant NDP mixture.

Nirenberg and Matthaei demonstrated that poly(U)

stimulates the synthesis of poly(Phe) by incubating

poly(U) and a mixture of 1 radioactive and 19 unlabeled

amino acids in a DNase-treated protein synthesizing sys-

tem. Significant radioactivity appeared in the protein pre-

cipitate only when phenylalanine was labeled. UUU must

therefore be the codon specifying Phe. In similar experi-

ments using poly(A) and poly(C), it was found that

poly(Lys) and poly(Pro), respectively, were synthesized.

Thus AAA specifies Lys and CCC specifies Pro. [Poly(G)

cannot function as a synthetic mRNA because, even under

denaturing conditions, it aggregates to form a four-

stranded helix (Section 30-4De).An mRNA must be single

stranded to direct its translation; Section 32-2D.]

(RNA)

 NDP Δ (RNA)

n1

 P

JWCL281_c32_1338-1428.qxd 8/4/10 4:44 PM Page 1341

Nirenberg and Ochoa independently employed ribonu-

cleotide copolymers to further elucidate the genetic code. For

example, in a poly(UG) composed of 76% U and 24% G, the

probability of a given triplet being UUU is 0.76  0.76  0.76

 0.44. Likewise, the probability of a particular triplet con-

sisting of 2U’s and 1G, that is, UUG, UGU, or GUU, is 0.76 

0.76  0.24  0.14. The use of this poly(UG) as an mRNA

therefore indicated the base compositions, but not the se-

quences, of the codons specifying several amino acids (Table

32-1). Through the use of copolymers containing two, three,

and four bases, the base compositions of codons specifying

each of the 20 amino acids were inferred. Moreover,these ex-

periments demonstrated that the genetic code is degenerate

since, for example, poly(UA), poly(UC), and poly(UG) all di-

rect the incorporation of Leu into a polypeptide.

b. The Genetic Code Was Elucidated through Triplet

Binding Assays and the Use of Polyribonucleotides

with Known Sequences

In the absence of GTP, which is necessary for protein

synthesis, trinucleotides but not dinucleotides are almost

as effective as mRNAs in promoting the ribosomal bind-

ing of specific tRNAs.This phenomenon, which Nirenberg

and Philip Leder discovered in 1964, permitted the various

codons to be identified by a simple binding assay. Ribo-

somes, together with their bound tRNAs, are retained by a

nitrocellulose filter but free tRNA is not. The bound

tRNA was identified by using charged tRNA mixtures in

which only one of the pendent amino acid residues was ra-

dioactively labeled. For instance,it was found, as expected,

that UUU stimulates the ribosomal binding of only Phe

tRNA. Likewise, UUG, UGU, and GUU stimulate the

binding of Leu, Cys, and Val tRNAs, respectively. Hence

UUG, UGU, and GUU must be codons that specify Leu,

Cys, and Val, respectively. In this way, the amino acids

specified by some 50 codons were identified. For the re-

maining codons, the binding assay was either negative (no

tRNA bound) or ambiguous.

The genetic code dictionary was completed and previ-

ous results confirmed through H. Gobind Khorana’s syn-

thesis of polyribonucleotides with specified repeating se-

quences (Section 7-6A). In a cell-free protein synthesizing

system, UCUCUCUC, for example, is read

so that it specifies a polypeptide chain of two alternating

amino acid residues. In fact, it was observed that this

mRNA stimulated the production of

which indicates that either UCU or CUC specifies Ser and

the other specifies Leu.This information, together with the

tRNA-binding data, permitted the conclusion that UCU

codes for Ser and CUC codes for Leu. These data also

prove that codons consist of an odd number of nucleotides,

thereby relieving any residual suspicions that codons con-

sist of six rather than three nucleotides.

Alternating sequences of three nucleotides, such as

poly(UAC), specify three different homopolypeptides be-

cause ribosomes may initiate polypeptide synthesis on

these synthetic mRNAs in any of the three possible read-

ing frames (Fig. 32-5). Analyses of the polypeptides speci-

fied by various alternating sequences of two and three nu-

cleotides confirmed the identity of many codons and filled

out missing portions of the genetic code.

c. mRNAs Are Read in the 5¿S3¿ Direction

The use of repeating tetranucleotides indicated the

reading direction of the code and identified the chain

Ser — Leu — Ser — Leu — Ser — Leu —

UCU CUC UCU CUC UCU C

1342 Chapter 32. Translation

Table 32-1 Amino Acid Incorporation Stimulated by a

Random Copolymer of U and G in Mole

Ratio 0.76 : 0.24

Probability Relative Amount

of Relative Amino of Amino Acid

Codon Occurrence Incidence

Acid Incorporated

UUU 0.44 100 Phe 100

UUG 0.14 32 Leu 36

UGU 0.14 32 Cys 35

GUU 0.14 32 Val 37

UGG 0.04 9 Trp 14

GUG 0.04 9

GGU 0.04 9 Gly 12

GGG 0.01 2

Relative incidence is defined here as 100  probability of

occurrence/0.44.

Source: Matthaei, J.H., Jones, O.W., Martin, R.G., and Nirenberg, M.,

Proc. Natl. Acad. Sci. 48, 666 (1962).

Third reading

frame start

Third reading

frame

Second reading

frame start

Second reading

frame

First reading

frame start

First reading

frame

•

Tyr Tyr Tyr

...

Thr Thr Thr

...

Leu Leu Leu

...

Figure 32-5 The three potential reading frames of an mRNA.

Each reading frame would yield a different polypeptide.

JWCL281_c32_1338-1428.qxd 8/4/10 4:44 PM Page 1342

termination codons. Poly(UAUC) specifies, as expected,

a polypeptide with a tetrapeptide repeat:

The amino acid sequence of this polypeptide indicates that

the mRNA’s 5¿ end corresponds to the polypeptide’s N-

terminus; that is, mRNA is read in the 5¿S3¿ direction.

d. UAG, UAA, and UGA Are Stop Codons

In contrast to the above results, poly(AUAG) yields

only dipeptides and tripeptides. This is because UAG is a

signal to the ribosome to terminate protein synthesis:

Likewise, poly(GUAA) yields dipeptides and tripeptides

because UAA is also a chain termination signal:

UGA is a third stop signal.These Stop codons, whose exis-

tence was first inferred from genetic experiments, are

known, somewhat inappropriately, as nonsense codons be-

cause they are the only codons that do not specify amino

acids. UAG, UAA, and UGA are sometimes referred to as

amber, ochre, and opal codons. [They were so named as

the result of a laboratory joke: The German word for am-

ber is Bernstein, the name of an individual who helped dis-

cover amber mutations (mutations that change some other

codon to UAG); ochre and opal are puns on amber.]

e. AUG and GUG Are Chain Initiation Codons

The codons AUG, and less frequently GUG, form part of

the chain initiation sequence (Section 32-3Ca). However,

they also specify the amino acid residues Met and Val, re-

spectively, at internal positions of polypeptide chains.

(Nirenberg and Matthaei’s discovery that UUU specifies

Phe was only possible because ribosomes indiscriminately

initiate polypeptide synthesis on an mRNA when the Mg

2

concentration is unphysiologically high as it was, serendip-

itously, in their experiments.)

D. The Nature of the Code

The genetic code dictionary, as elucidated by the above

methods, is presented in Table 32-2 as well as in Table 5-3.

Examination of the table indicates that the genetic code

has several remarkable features:

1. The code is highly degenerate. Three amino acids,

Arg, Leu, and Ser, are each specified by six codons, and

Val Val

Ser Ser

Lys

Stop

...

GUA AGU AAG UAA GUA AGU

...

Ile Ile

Asp Asp

Arg

Stop

...

AUA GAU AGA UAG AUA GAU

...

Tyr Tyr

Leu Leu

Ser

Ile

...

UAU CUA UCU AUC UAU CUA

...

5

3

most of the rest are specified by either four, three, or two

codons. Only Met and Trp, two of the least common amino

acids in proteins (Table 4-1), are represented by a single

codon. Codons that specify the same amino acid are

termed synonyms.

2. The arrangement of the code table is nonrandom.

Most synonyms occupy the same box in Table 32-2; that is,

they differ only in their third nucleotide. The only excep-

tions are Arg, Leu, and Ser, which, having six codons each,

must occupy more than one box. XYU and XYC always

specify the same amino acid; XYA and XYG do so in all

but two cases. Moreover, changes in the first codon posi-

tion tend to specify similar (if not the same) amino acids,

whereas codons with second position pyrimidines encode

mostly hydrophobic amino acids (tan in Table 32-2), and

those with second position purines encode mostly polar

amino acids (blue, red, and purple in Table 32-2). Appar-

ently the code evolved so as to minimize the deleterious ef-

fects of mutations.

Many of the mutations causing amino acid substitutions

in a protein can be rationalized, according to the genetic

code, as single point mutations. As a consequence of the ge-

netic code’s degeneracy, however, many point mutations at a

third codon position are phenotypically silent; that is, the

Section 32-1. The Genetic Code 1343

Table 32-2 The “Standard” Genetic Code

Nonpolar amino acid residues are tan, basic residues are blue, acidic

residues are red, and nonpolar uucharged residues are purple.

AUG forms part of the initiation signal as well as coding for internal

Met residues.

Second position

First

position

(5 end)

Third

position

(3 end)

Stop

Trp

UUU

UUC

UUA

UUG

Phe

Leu

UCU

UCC

UCA

UCG

Ser

UAU

UAC

UAA

UAG

UGU

UGC

UGA

UGG

CysTyr

CUU

CUC

CUA

CUG

Leu Pro

CCU

CCC

CCA

CCG

CAU

CAC

CAA

CAG

His

Gln

CGU

CGC

CGA

CGG

Arg

AUU

AUC

AUA

AUG

ACU

ACC

ACA

ACG

Ile

Met

Thr

AAU

AAC

AAA

AAG

Asn

Lys

AGU

AGC

AGA

AGG

Ser

Arg

GUU

GUC

GUA

GUG

Val

GCU

GCC

GCA

GCG

Ala

GAU

GAC

GAA

GAG

Asp

Glu

GGU

GGC

GGA

GGG

Gly

JWCL281_c32_1338-1428.qxd 8/4/10 4:44 PM Page 1343

mutated codon specifies the same amino acid as the wild

type. Degeneracy may account for as much as 33% of the 25

to 75% range in the G  C content among the DNAs of dif-

ferent organisms (Section 5-1Ba). The frequent occurrence

of Arg, Ala, Gly, and Pro also tends to give a high G  C

content, whereas Asn, Ile, Lys, Met, Phe, and Tyr contribute

to a low G  C content.

a. Some Phage DNA Segments Contain Overlapping

Genes in Different Reading Frames

Since any nucleotide sequence may have three reading

frames, it is possible, at least in principle, for a polynu-

cleotide to encode two or even three different polypep-

tides. This idea was never seriously entertained, however,

because it seemed that the constraints on even two over-

lapping genes in different reading frames would be too

great for them to evolve so that both could specify sensi-

ble proteins. It therefore came as a great surprise, in 1976,

when Frederick Sanger reported that the 5386-nucleotide

DNA of bacteriophage X174 (which,at the time,was the

largest DNA to have been sequenced) contains two genes

that are completely contained within larger genes of dif-

ferent reading frames (Fig. 32-6). Moreover, the end of

the overlapping D and E genes contains the control se-

quence for the ribosomal initiation of the J gene so that

this short DNA segment performs triple duty. Bacteria

also exhibit such coding economy; the ribosomal initia-

tion sequence of one gene in a polycistronic mRNA often

overlaps the end of the preceding gene. Nevertheless,

completely overlapping genes have only been found in

small single-stranded DNA phages, which presumably

must make maximal use of the little DNA that they can

pack inside their capsids.

b. The “Standard” Genetic Code Is Widespread but

Not Universal

For many years it was thought that the “standard” ge-

netic code (that given in Table 32-2) is universal. This as-

sumption was, in part, based on the observations that one

kind of organism (e.g., E. coli) can accurately translate the

genes from quite different organisms (e.g., humans). This

phenomenon is, in fact, the basis of genetic engineering.

Once the “standard” genetic code had been established,

presumably during the time of prebiotic evolution (Section

1-5B), any mutation that would alter the way the code is

translated would result in numerous, often deleterious, pro-

tein sequence changes. Undoubtedly there is strong selec-

tion against such mutations.

Despite the foregoing, DNA sequencing studies in

1981 revealed that the genetic codes of certain mitochon-

dria (mitochondria contain their own genes and protein

synthesizing systems but produce only a few mitochondrial

proteins; Section 12-4E) are variants of the “standard” ge-

netic code (Table 32-3). For example, in mammalian mito-

chondria,AUA, as well as the standard AUG, is a Met/ini-

tiation codon; UGA specifies Trp rather than “Stop”; and

AGA and AGG are “Stop” rather than Arg. Note that all

mitochondrial genetic codes except those of plants sim-

plify the “standard” code by increasing its degeneracy. For

example, in the mammalian mitochondrial code, each

amino acid is specified by at least two codons that differ

only in their third nucleotide. Apparently the constraints

preventing alterations of the genetic code are eased by

the small sizes of mitochondrial genomes. More recent

1344 Chapter 32. Translation

Table 32-3 Mitochondrial Deviations from the “Standard”

Genetic Code

Mitochondrion UGA AUA CUN

CGG

Mammalian Trp Met

Stop

Baker’s yeast Trp Met

Thr

Neurospora crassa Tr p

Drosophila Trp Met

Ser

Protozoan Trp

Plant Trp

“Standard” code Stop Ile Leu Arg Arg

N represents any of the four nucleotides.

Also acts as part of an initiation signal.

AGA only; no AGG codons occur in Drosophila mitochondrial DNA.

Source: Mainly Breitenberger, C.A. and RajBhandary, U.L., Trends

Biochem. Sci. 10, 481 (1985).

Figure 32-6 Genetic map of bacteriophage X174 as

determined by DNA sequence analysis. Genes are labeled A, B,

C, etc. Note that gene B is wholly contained within gene A and

gene E is wholly contained within gene D. These pairs of genes

are read in different reading frames and therefore specify

unrelated proteins.The unlabeled regions correspond to

untranslated control sequences.

JWCL281_c32_1338-1428.qxd 8/4/10 4:44 PM Page 1344

GAGCC

5′ p

Anticodon

Amino acid

attachment site

CGG

CGCG

3′

studies, however, have revealed that in ciliated protozoa,

the codons UAA and UAG specify Gln rather than

“Stop.” Perhaps UAA and UAG were sufficiently rare

codons in a primordial ciliate (which molecular phyloge-

netic studies indicate branched off very early in eukary-

otic evolution) to permit the code change without unac-

ceptable deleterious effects. At any rate, the “standard”

genetic code,although very widely utilized,is not universal.

Indeed, as we shall see in Section 32-2De, under the

proper context in mRNA, certain codons can specify

“nonstandard” amino acids.

2 TRANSFER RNA AND ITS

AMINOACYLATION

The establishment of the genetic function of DNA led to

the realization that cells somehow “translate” the language

of base sequences into the language of polypeptides. Yet,

nucleic acids originally appeared unable to bind specific

amino acids [more recently RNA aptamers for specific

amino acids have been generated; aptamers are nucleic

acids that have been selected for their ability to bind spe-

cific ligands (Section 7-6C)]. In 1955, Crick, in what became

known as the adaptor hypothesis, postulated that transla-

tion occurs through the mediation of “adaptor” molecules.

Each adaptor was postulated to carry a specific enzymati-

cally appended amino acid and to recognize the correspon-

ding codon (Fig. 32-7). Crick suggested that these adaptors

contain RNA because codon recognition could then occur

by complementary base pairing. At about this time, Paul

Zamecnik and Mahlon Hoagland discovered that in the

course of protein synthesis,

C-labeled amino acids be-

came transiently bound to a low molecular mass fraction of

RNA. Further investigations indicated that these RNAs,

which at first were called “soluble RNA” or “sRNA” but

are now known as transfer RNA (tRNA), are, in fact,

Crick’s putative adaptor molecules.

A. Primary and Secondary Structures of tRNA

See Guided Exploration 26: The structure of tRNA In 1965, after

a 7-year effort, Robert Holley reported the first known

base sequence of a biologically significant nucleic acid, that

of yeast alanine tRNA (tRNA

Ala

; Fig. 32-8).To do so Holley

had to overcome several major obstacles:

1. All organisms contain many species of tRNAs (usu-

ally at least one for each of the 20 amino acids) which, be-

cause of their nearly identical properties (see below), are

not easily separated. Preparative techniques had to be de-

veloped to provide the gram or so of pure yeast tRNA

Ala

Holley required for its sequence determination.

2. Holley had to invent the methods that were initially

used to sequence RNA (Section 7-2).

3. Ten of the 76 bases of yeast tRNA

Ala

are modified

(see below).Their structural formulas had to be elucidated

although they were never available in more than milligram

quantities.

Since 1965, the techniques for tRNA purification and se-

quencing have vastly improved. A tRNA may now be se-

quenced in a few hours’ time with only ⬃1 g of material.

Presently, the base sequences of many thousands of

tRNAs from nearly 800 organisms are known, most from

genomic sequences (they are compiled at the Genomic

Section 32-2. Transfer RNA and Its Aminoacylation 1345

Figure 32-7 The adaptor hypothesis. It postulates that the

genetic code is read by molecules that recognize a particular

codon and carry the corresponding amino acid.

Figure 32-8 Base sequence of yeast tRNA

Ala

drawn in the

cloverleaf form. The symbols for the modified nucleosides

(color) are explained in Fig. 32-10.

Amino

acid 1

Amino

acid 2

Amino

acid 3

Polypeptide

Adaptors

Nucleic acid

Codon 1 Codon 2 Codon 3

JWCL281_c32_1338-1428.qxd 8/19/10 10:04 PM Page 1345

tRNA Database, http://gtrnadb.ucsc.edu/). They vary in

length from 54 to 100 nucleotides (18–28 kD) although

most have ⬃76 nucleotides.

Almost all known tRNAs, as Holley first recognized,

may be schematically arranged in the so-called cloverleaf

secondary structure (Fig. 32-9). Starting from the 5¿ end,

they have the following common features:

1. A 5¿-terminal phosphate group.

2. A 7-bp stem that includes the 5¿-terminal nucleotide

and that may contain non-Watson–Crick base pairs such as

G  U. This assembly is known as the acceptor or amino acid

stem because the amino acid residue carried by the tRNA is

appended to its 3¿-terminal OH group (Section 32-2C).

3. A 3- or 4-bp stem ending in a loop that frequently con-

tains the modified base dihydrouridine (D; see below). This

stem and loop are therefore collectively termed the D arm.

4. A 5-bp stem ending in a loop that contains the anti-

codon, the triplet of bases that is complementary to the

codon specifying the tRNA. These features are known as

the anticodon arm.

5. A 5-bp stem ending in a loop that usually contains

the sequence TC (where  is the symbol for pseudouri-

dine; see below).This assembly is called the TC or T arm.

6. All tRNAs terminate in the sequence CCA with a free

3¿-OH group.The CCA may be genetically specified or enzy-

matically appended to immature tRNA (Section 31-4Cc).

7. There are 15 invariant positions (always have the

same base) and 8 semi-invariant positions (only a purine or

only a pyrimidine) that occur mostly in the loop regions.

These regions also contain correlated invariants, that is,

pairs of nonstem nucleotides that are base paired in all

tRNAs.The purine on the 3¿ side of the anticodon is invari-

ably modified. The structural significance of these features

is examined below.

The site of greatest variability among the known tRNAs

occurs in the so-called variable arm. It has from 3 to 21 nu-

cleotides and may have a stem consisting of up to 7 bp.The

D loop also varies in length from 5 to 7 nucleotides.

a. Transfer RNAs Have Numerous Modified Bases

One of the most striking characteristics of tRNAs is

their large proportion, up to 25%, of post-translationally

modified or hypermodified bases. Nearly 80 such bases,

found at 60 different tRNA positions, have been char-

acterized.A few of them, together with their standard ab-

breviations, are indicated in Fig. 32-10. Hypermodified

nucleosides, such as i

A, are usually adjacent to the anti-

codon’s 3¿ nucleotide when it is A or U. Their low polari-

ties probably strengthen the otherwise relatively weak

pairing associations of these bases with the codon,

thereby increasing translational fidelity. Conversely, cer-

tain methylations block base pairing and hence prevent

inappropriate structures from forming. Some of these

modifications form important recognition elements for

the enzyme that attaches the correct amino acid to a

tRNA (Section 32-2Cb). However, none of them are es-

sential for maintaining a tRNA’s structural integrity (see

below) or for its proper binding to the ribosome. Never-

theless, mutant bacteria unable to form certain modified

bases compete poorly against the corresponding normal

bacteria.

B. Tertiary Structure of tRNA

See Guided Exploration 26: The structure of tRNA The earliest

physicochemical investigations of tRNA indicated that it

has a well-defined conformation. Yet, despite numerous

hydrodynamic, spectroscopic, and chemical cross-linking

studies, its three-dimensional structure remained an

enigma until 1974. In that year, the 2.5-Å resolution X-ray

crystal structure of yeast tRNA

Phe

was separately eluci-

dated by Alexander Rich in collaboration with Sung-Hou

Kim and, in a different crystal form, by Aaron Klug. The

molecule assumes an L-shaped conformation in which one

leg of the L is formed by the acceptor and T stems folded

into a continuous A-RNA-like double helix (Section 29-

1Ba) and the other leg is similarly composed of the D and

1346 Chapter 32. Translation

OHA

3′

5′ p

Anticodon

•

Anticodon

arm

Variable arm

D arm T C arm ψ

Acceptor

stem

Figure 32-9 Cloverleaf secondary structure of tRNA. Filled

circles connected by dots represent Watson–Crick base pairs, and

open circles in the double-helical regions indicate bases involved

in non-Watson–Crick base pairing. Invariant positions are

indicated: R and Y represent invariant purines and pyrimidines

and

 signifies pseudouridine. The starred nucleosides are often

modified.The dashed regions in the D and variable arms contain

different numbers of nucleotides in the various tRNAs.

JWCL281_c32_1338-1428.qxd 8/4/10 4:44 PM Page 1346

Section 32-2. Transfer RNA and Its Aminoacylation 1347

Uracil derivatives

Ribose

Pseudouridine ( )

Dihydrouridine (D)

Cytosine derivatives

Ribose

3-Methylcytidine (m C)

Adenine derivatives

Ribose Ribose

Ribose Ribose Ribose

1-Methyladenosine (m A)

Inosine (I)

Guanine derivatives

-Dimethylguanosine (m

(CH

)

-Methylguanosine (m

(CH

)

Ribose

CH COO

–

Lysidine (L)

Ribose

-Acetylcytidine (ac

CCHCH

Ribose

Ribothymidine (T)

Ribose

4-Thiouridine (s

Ribose

-Isopentenyladenosine (i

NHCOCH

R  H

R  CH

CHCOCH

Wyosine (Wyo)

Figure 32-10 A selection of the modified nucleosides that

occur in tRNAs together with their standard abbreviations. Note

that although inosine chemically resembles guanosine, it is

biochemically derived from adenosine. Nucleosides may also be

methylated at their ribose 2¿ positions to form residues

symbolized, for instance, by Cm, Gm, and Um.

JWCL281_c32_1338-1428.qxd 8/19/10 10:04 PM Page 1347

anticodon stems (Fig. 32-11). Each leg of the L is ⬃60 Å

long and the anticodon and amino acid acceptor sites are at

opposite ends of the molecule, some 76 Å apart. The nar-

row 20- to 25-Å width of native tRNA is essential to its bi-

ological function: During protein synthesis, three RNA

molecules must simultaneously bind in close proximity at

adjacent codons on mRNA (Section 32-3Ae).

a. tRNA’s Complex Tertiary Structure Is Maintained

by Hydrogen Bonding and Stacking Interactions

The structural complexity of yeast tRNA

Phe

is reminis-

cent of that of a protein. Although only 42 of its 76 bases

occur in double helical stems, 71 of them participate in

stacking associations (Fig. 32-12). The structure also con-

tains 9 base pairing interactions that cross-link its tertiary

structure (Figs. 32-11a and 32-12). Remarkably, all but one

of these tertiary interactions, which appear to be the main-

stays of the molecular structure, are non-Watson–Crick as-

sociations. Moreover, most of the bases involved in these

interactions are either invariant or semi-invariant, which

strongly suggests that all tRNAs have similar conforma-

tions (see below).The structure is also stabilized by several

1348 Chapter 32. Translation

unusual hydrogen bonds between bases and either phos-

phate groups or the 2¿-OH groups of ribose residues.

The compact structure of yeast tRNA

Phe

results from its

large number of intramolecular associations, which renders

most of its bases inaccessible to solvent. The most notable

exceptions to this are the anticodon bases and those of the

amino acid–bearing¬CCA terminus. Both of these group-

ings must be accessible in order to carry out their biologi-

cal functions.

The observation that the molecular structures of yeast

tRNA

Phe

in two different crystal forms are essentially iden-

tical lends much credence to the supposition that its crystal

structure closely resembles its solution structure. Transfer

RNAs other than yeast tRNA

Phe

have, unfortunately, been

notoriously difficult to crystallize. As yet, the X-ray struc-

tures of only three other uncomplexed tRNAs have been re-

ported (although the X-ray structures of numerous tRNAs

in complex with the enzymes that append their correspon-

ding amino acids and with ribosomes have been elucidated;

Sections 32-2C and 32-3D). The major structural differ-

ences among them result from an apparent flexibility in the

anticodon loop and the¬CCA terminus as well as from a

(a)

Constant nucleotide

Constant purine

or pyrimidine

3′

5′

GUC

ACA

D loop

TψC loop

Variable

loop

Acceptor stem

Anticodon loop

(b)

Figure 32-11 Structure of yeast tRNA

Phe

. (a) The base

sequence drawn in cloverleaf form.Tertiary base pairing

interactions are represented by thin red lines connecting the

participating bases. Bases that are invariant or semi-invariant in

all tRNAs are circled by solid and dashed lines, respectively.The

5¿ terminus is colored bright green, the acceptor stem is yellow,

the D arm is white, the anticodon arm is light green, the variable

arm is orange, the T

␺C arm is cyan, and the 3¿ terminus is red.

(b) The X-ray structure drawn to show how its base paired stems

form an L-shaped molecule. The tRNA is drawn in stick form

with C atoms colored as in Part a, N blue, and O red. Adjacent P

atoms are connected by rods colored as in Part a. [Based on an

X-ray structure by Sung-Hou Kim, PDBid 6TRNA.]

See

Kinemage Exercises 19-1 and 19-2

JWCL281_c32_1338-1428.qxd 10/29/10 9:04 PM Page 1348

O OH

OPO

–

3⬘ 2⬘

Adenine

C RH

Aminoacyl–tRNA

tRNA

Section 32-2. Transfer RNA and Its Aminoacylation 1349

hingelike mobility between the two legs of the L that gives,

for instance, yeast tRNA

Asp

a boomerang-like shape. Such

observations are in accord with the expectation that all

tRNAs fit into the same ribosomal cavities.

C. Aminoacyl–tRNA Synthetases

See Guided Exploration 27: The structures of aminoacyl–tRNA syn-

thetases and their interactions with tRNAs

Accurate translation re-

quires two equally important recognition steps: (1) the

choice of the correct amino acid for covalent attachment to a

tRNA; and (2) the selection of the amino acid–charged

tRNA specified by mRNA. The first of these steps, which is

catalyzed by amino acid–specific enzymes known as

aminoacyl–tRNA synthetases (aaRSs), appends an amino

acid to the 3¿-terminal ribose residue of its cognate tRNA to

form an aminoacyl–tRNA (Fig. 32-13).This otherwise unfa-

vorable process is driven by the hydrolysis of ATP in two se-

quential reactions that are catalyzed by a single enzyme.

Figure 32-12 Tertiary base pairing interactions in yeast

tRNA

Phe

. Note that all but one of these nine interactions involve

non-Watson–Crick pairs and that they are all located near the

corner of the L. [After Kim, S.H., in Schimmel, P.R., Söll, D., and

Figure 32-13 An aminoacyl–tRNA. The amino acid residue is

esterified to the tRNA’s 3¿-terminal nucleoside at either its 3¿-

OH group, as shown here, or its 2¿-OH group.

Guanine

Guanine 15

Guanine 22

Thymine 54

Cytosine 56

Cytosine 48

Cytosine 25

Cytosine 13

Ribose

ψ 55

1-Methyladenine 58

7-Methyl-

guanine 46

Guanine 18

Guanine 4

Guanine 10

Guanine 45

Ribose

...

Uracil 69

Uracil 12

...

Adenine 9

Adenine 23

Adenine 44

.....

.......

Dimethylguanine 26

Ribose

ψ stem

ψ loop

Acceptor stem

D stem

Anticodon stem

Anticodon

loop

Anticodon

D loop

Variable

loop

3′ end

5′ end

Abelson, J.N. (Eds.), Transfer RNA: Structure, Properties and

Recognition, p. 87, Cold Spring Harbor Laboratory Press (1979).

Illustration, Irving Geis. Image from the Irving Geis Collection,

Howard Hughes Medical Institute. Reprinted with permission.]

See Kinemage Exercise 19-3

JWCL281_c32_1338-1428.qxd 10/27/10 1:40 PM Page 1349

1350 Chapter 32. Translation

Table 32-4 Characteristics of Bacterial Aminoacyl–tRNA

Synthetases

Amino Acid Quaternary Structure Number of Residues

Class I

Arg  577

Cys  461

Gln  553

Glu  471

Ile  939

Leu  860

Met , 

676

Tr p 

325

Ty r 

424

Va l  951

Class II

Ala , 

875

Asn 

467

Asp 

590

Gly 



303/689

His 

424

Lys 

505

Pro 

572

Phe 



,  327/795

Ser 

430

Thr 

642

Source: Mainly Carter, C.W., Jr., Annu. Rev. Biochem. 62, 715 (1993).

all aaRSs evolved from a common ancestor and should

therefore be structurally related. This is not the case. In

fact, the aaRSs form a diverse group of enzymes. The over

1000 such enzymes that have been characterized each have

one of four different types of subunit structures, , 

(the

predominant forms), 

, and 



, with known subunit sizes

ranging from ⬃300 to ⬃1200 residues. Moreover, there is

little sequence similarity among synthetases specific for

different amino acids. Quite possibly, aminoacyl–tRNA

synthetases arose very early in evolution, before the devel-

opment of the modern protein synthesis apparatus other

than tRNAs.

Detailed sequence and structural comparisons of

aminoacyl–tRNA synthetases by Dino Moras indicate that

these enzymes form two unrelated families, termed Class I

and Class II aaRSs, that each have the same 10 members in

nearly all organisms (Table 32-4). The Class I enzymes, al-

though of largely dissimilar sequences, share two homolo-

gous polypeptide segments, not present in other proteins,

that have the consensus sequences His-Ile-Gly-His

(HIGH) and Lys-Met-Ser-Lys-Ser (KMSKS). The X-ray

structures of Class I enzymes indicate that both of these

segments are components of a dinucleotide-binding fold

(Rossmann fold, which is also possessed by many NAD



and ATP-binding proteins; Section 8-3Bi) in which they

1. The amino acid is first “activated” by its reaction with

ATP to form an aminoacyl–adenylate

which, with all but three aaRSs, can occur in the absence of

tRNA. Indeed, this intermediate may be isolated although

it normally remains tightly bound to the enzyme.

2. This mixed anhydride then reacts with tRNA to form

the aminoacyl–tRNA

Some aaRSs exclusively append an amino acid to the ter-

minal 2¿-OH group of their cognate tRNAs, and others do

so at the 3¿-OH group. This selectivity was established with

the use of chemically modified tRNAs that lack either the

2¿- or 3¿-OH group of their 3¿-terminal ribose residue. The

use of these derivatives was necessary because, in solution,

the aminoacyl group rapidly equilibrates between the 2¿

and 3¿ positions.

The overall aminoacylation reaction is

These reaction steps are readily reversible because the free

energies of hydrolysis of the bonds formed in both the

aminoacyl–adenylate and the aminoacyl–tRNA are com-

parable to that of ATP hydrolysis. The overall reaction is

driven to completion by the inorganic pyrophosphatase-

catalyzed hydrolysis of the PP

generated in the first reac-

tion step. Amino acid activation therefore chemically re-

sembles fatty acid activation (Section 25-2A); the major

difference between these two processes, which were both

elucidated by Paul Berg, is that tRNA is the acyl acceptor

in amino acid activation, whereas CoA performs this func-

tion in fatty acid activation.

a. There Are Two Classes of Aminoacyl–tRNA

Synthetases

Most cells have one aaRS for each of the 20 amino acids.

The similarity of the reactions catalyzed by these enzymes

and the structural resemblance of all tRNAs suggests that

aminoacyl–tRNA  AMP  PP

Amino acid  tRNA  ATP Δ

Aminoacyl–AMPtRNA Δ aminoacyl–tRNAAMP

ATP

–

CCR

O P

–

O Ribose + PP

Amino acid

Aminoacyl–adenylate

(aminoacyl–AMP)

Adenine

JWCL281_c32_1338-1428.qxd 8/4/10 4:44 PM Page 1350