Content uploaded by Anat Bashan
Author content
All content in this area was uploaded by Anat Bashan
Content may be subject to copyright.
31 Aug 2004 11:34 AR AR224-MI58-10.tex AR224-MI58-10.sgm LaTeX2e(2002/01/18)
P1: IKH
10.1146/annurev.micro.58.030603.123822
Annu. Rev. Microbiol. 2004. 58:233–51
doi: 10.1146/annurev.micro.58.030603.123822
Copyright
c
2004 by Annual Reviews. All rights reserved
First published online as a Review in Advance on May 14, 2004
RIBOSOMAL CRYSTALLOGRAPHY: Initiation,
Peptide Bond Formation, and Amino Acid
Polymerization are Hampered by Antibiotics
AdaYonath
1,2
and Anat Bashan
1
1
Department of Structural Biology, The Weizmann Institute, 76100 Rehovot,
Israel; email: ada.yonath@weizmann.ac.il; anat.bashan@weizmann.ac.il
2
Max-Planck-Research Unit for Ribosomal Structure, 22603 Hamburg, Germany
KeyWords ribosomes, initiation factor 3, exit tunnel, gating, peptidyl-transferase,
synergism
■ Abstract High-resolution structures of ribosomal complexes revealed that minute
amounts of clinically relevant antibiotics hamper protein biosynthesis by limiting ri-
bosomal mobility or perturbing its elaborate architecture, designed for navigating and
controlling peptide bond formation and continuous amino acid polymerization. To
accomplish this, the ribosome contributes positional rather than chemical catalysis,
provides remote interactions governing accurate substrate alignment within the flex-
ible peptidyl-transferase center (PTC) pocket, and ensures nascent-protein chirality
through spatial limitations. Peptide bond formation is concurrent with aminoacylated-
tRNA 3
end translocation and is performed by a rotatory motion around the axis of
a sizable ribosomal symmetry-related region, which is located around the PTC in all
known crystal structures. Guided by ribosomal-RNA scaffold along an exact pattern,
the rotatory motion results in stereochemistry that is optimal for peptide bond forma-
tion and for nascent protein entrance into the exit tunnel, the main target of antibi-
otics targeting ribosomes. By connecting the PTC, the decoding center, and the tRNA
entrance and exit regions, the symmetry-related region can transfer intraribosomal
signals, guaranteeing smooth processivity of amino acid polymerization.
CONTENTS
INTRODUCTION ..................................................... 234
INITIATION OF THE TRANSLATION PROCESS .......................... 235
Initiation Factor 3: A Flexible Protein Factor .............................. 235
Edeine: An Antibiotic Interfering with Initiation by Limiting
the Platform Mobility ............................................... 237
PEPTIDE BOND FORMATION AND AMINO ACID
POLYMERIZATION .................................................. 238
The PTC and Its Vicinity: A Sizable Symmetry-Related Region ............... 238
Clinically Relevant PTC Targeting ...................................... 240
Sparsomycin, Puromycin, and the “Fragment Reaction” ..................... 241
0066-4227/04/1013-0233$14.00
233
31 Aug 2004 11:34 AR AR224-MI58-10.tex AR224-MI58-10.sgm LaTeX2e(2002/01/18)
P1: IKH
234 YONATH
BASHAN
POST PEPTIDE BOND FORMATION .................................... 243
Macrolides and Ketolides: Resistance and Selectivity ....................... 243
Tunnel Mobility and Ribosomal Involvement in Cellular Regulation ........... 246
INTRODUCTION
Ribosomes are the universal cellular riboprotein assemblies that catalyze amino
acid polymerization into proteins. In this process, called protein biosynthesis, the
ribosomes interact with the cellular environment, select the genetic frame to be
translated, form the peptide bonds, and act as polymerases to ensure smooth and
efficient elongation of the newly born protein chains. The ribosomes are composed
of two riboprotein subunits of unequal size that associate upon the initiation of
the process and dissociate at its termination. Protein biosynthesis is performed
by a cooperative effort of both subunits, and the ribosome possesses features that
facilitate transmission between its various functional sites, leading to efficient
processivity of protein formation essential for cell vitality. The small ribosomal
subunit facilitates the initiation of the translation process and is involved in se-
lecting the frame to be translated, decoding the genetic message, and controlling
the fidelity of codon-anticodon interactions. The large ribosomal subunit forms
the peptide bond, ensures smooth amino acid polymerization, and channels the
nascent proteins through their exit tunnel.
Messenger RNA (mRNA) carries the genetic instructions to the ribosome, and
aminoacylated transfer RNA (tRNA) molecules deliver the amino acids within
a ternary complex containing GTP and the elongation factor Tu. The ribosome
possesses three tRNA binding sites. The A-site hosts the aminoacylated-tRNA, the
P-site hosts the peptidyl tRNA, and the E-site designates the location of the exiting
free tRNA once a peptide bond has been formed. Elongation of the polypeptide
chain is driven by GTPase activity. This motion is associated with A → P →
E translocation of the mRNA together with the tRNA molecules bound to it. In
each step of the elongation event, the mRNA advances by one codon and a new
peptide bond is formed between the amino acid bound to the A-site tRNA and the
growing peptidyl bound to the P-site tRNA. The leaving group of this reaction, a
free tRNA molecule, is released from the ribosome through the E-site as a part of
the overall mRNA-tRNA translocation. The entire cycle of elongation is extremely
fast, requiring about 50 ms for its completion from the stage of tRNA selection by
the mRNA until the leaving group is released. Peptide bond formation consumes
avery small portion of this time.
Two decades (84) of crystallographic attempts yielded several three-dimensional
structures of free and complexed ribosomal particles, despite the common belief
that ribosomal particles cannot crystallize owing to their instability, complexity,
and conformational variability. The key to obtaining crystals was the use of ri-
bosomes from robust bacteria, namely halophiles and thermophiles (79). These
yielded suitable crystals from the large subunit from Haloarcula marismortui,
31 Aug 2004 11:34 AR AR224-MI58-10.tex AR224-MI58-10.sgm LaTeX2e(2002/01/18)
P1: IKH
PROTEIN MAKING AND ITS OBSTRUCTION
235
H50S (36, 61, 74), and the small subunit from Thermus thermophilus, T30S (82).
However, because H. marismortui is an archaeon sharing properties with prokary-
otes as well as eukaryotes, the structural information obtained from its crystals is
only partially relevant to the vast amount of functional knowledge accumulated
over the years from the Escherichia coli ribosome. The addition of a new crystal
type, from the large ribosomal subunit from the robust mesophilic eubacterium
Deinococcus radiodurans, D50S (28), the ribosomes of which have a sequence
almost identical to that of E. coli ribosomes, addressed this need successfully.
Importantly, D50S crystallizes under conditions similar to those optimized for
its maximal functional activity, and its crystals yielded functionally meaningful,
well-ordered structures in contrast to the disorder of the functionally relevant fea-
tures observed in the structure of H50S (4, 46). Furthermore, typical of eubacteria,
D. radiodurans is a suitable pathogen model for antibiotics targeting the ribosome
(3, 7, 8, 27, 56, 58). A central element of the cell cycle, the ribosome is one of
the main targets for a broad range of antibiotics structurally diverse in nature that
inhibit different ribosomal functions. Because of its huge size, the ribosome offers,
theoretically, numerous different binding sites. Nevertheless, the crystal structures
of more than a dozen complexes of antibiotics bound to D50S show that practically
all the known drugs utilize a single or very few binding sites composed mainly of
ribosomal RNA (rRNA) (3, 7, 8, 10, 11, 24, 25, 27, 48, 56, 58).
We describe how the striking ribosomal architecture, designed for smooth pep-
tide bond formation and efficient amino acid polymerization, becomes useless by
minute amounts of clinically relevant antibiotics. This article is based on the high-
resolution structures of T30S (57) and D50S (28) as well as their complexes with
substrate analogs (1, 2, 5, 6, 80, 81), nonribosomal factors (48), and various an-
tibiotics targeting the ribosome (3, 7, 8, 27, 48, 56, 58). The functional analysis
presented here was aided by docking onto these structures the tRNA molecules
using the 5.5
˚
A crystal structure of their complex with T. thermophilus ribosome
T70S (85).
For assessing the medicinal relevance of the antibiotic modes of action as re-
vealed in their complexes with D50S, their binding sites were correlated with
biochemical, genetic, and functional results (3, 62, 72, 75). Antibiotics selectivity
was assessed by comparing the structural information obtained from complexes
of D50S (3, 7, 8, 27, 56, 58), which resembles pathogens, with that of H50S, since
the latter share properties also with eukaryotes (24, 25).
INITIATION OF THE TRANSLATION PROCESS
Initiation Factor 3: A Flexible Protein Factor
The initiation of protein biosynthesis has an important role in governing the ac-
curate setting of the reading frame. In prokaryotes the initiation complex contains
the small subunit, mRNA, three initiation factors (IF1, IF2, and IF3), and initiator
31 Aug 2004 11:34 AR AR224-MI58-10.tex AR224-MI58-10.sgm LaTeX2e(2002/01/18)
P1: IKH
236 YONATH
BASHAN
tRNA. IF3 binds strongly to the 30S subunit. It plays multiple roles in the formation
of the initiation complex. It influences the binding of the other factors, acts as a
fidelity factor by destabilizing noncanonical codon-anticodon interactions, selects
the first mRNA codon, discriminates against leaderless mRNA chains, and acts
as an antiassociation factor because it binds with high affinity to the 30S subunit
and shifts the dissociation equilibrium of the 70S ribosome toward free subunits,
thus maintaining a pool of 30S (23, 47). IF3 is a small, basic protein built of C-
and N-terminal domains (IF3C and IF3N, respectively) connected by a rather long
lysine-rich flexible linker region (Figure 1a). IF3C participates in most of the IF3
functional tasks, except for discrimination against noncognate tRNA, which is
performed by IF3N.
The initiation complex contains the small ribosomal subunit, initiation factors,
mRNA, and the first tRNA. The tRNAs are large L-shaped molecules built pri-
marily of double helices (Figure 1a), but their two functional tasks reside on short
single strands. The anticodon stem loop decodes the genetic information by base
pairing with the mRNA on the small subunit. The tRNA acceptor stem and the
universal tRNA 3
end—composed in all tRNAs of three nucleotides, CCA, which
binds the amino acid to be incorporated into the growing protein (at the A-site) and
the newly formed polypeptide chain (at the P-site)—interact with the large subunit
(Figure 1a). The initiator mRNA in prokaryotes includes an upstream purine-rich
sequence, called Shine-Dalgarno (SD), that pairs with a complementary region at
the 3
end of the 16S rRNA, thus anchoring the mRNA chain. Different positions
for the anti-SD region were observed in the two high-resolution structures of T30S.
In the 30S functionally activated conformation (57) the SD region is located on the
solvent side of the platform, whereas in the nontreated structure (78) the anti-SD
region was observed in the decoding region on the 30S front face acting as an
mRNA mimic, thus indicating clearly that this rather large structural element can
undergo a major swing around the platform (Figure 1b,c).
IF3 was also localized in two different positions (Figure 1d). Cryo-electron
microscopy (EM) of its complex with T30S suggested that both IF3C and IF3N
are located in the interface side of the 30S subunit (39) within a region involved
in extensive contacts with the 50S subunit in the entire 70S ribosome (85). In this
position the location of IF3C is consistent with results of hydroxyl radical cleavage
data (12) and provides a direct explanation for the role of IF3 in preventing subunit
association. However, it does not address the detachment of the SD anchor, which
is required for initiating the translocation process.
An alternative location for IF3C was observed in the crystal structure of a
complex of T30S with IF3C (48). In this complex IF3C is located in a position
that seems especially designed for it, within a cavity near the anti-SD region that
matches its size and shape on the solvent side of the upper edge of the platform
(Figure 1c). This location is consistent with results of NMR, mutagenesis, and
almost all crosslinks, footprints, and protection patterns reported for E. coli IF3C
(47), as well as with the location of the eukaryotic initiation factor 3 (eIF3), as
revealed by electron microscopy in a complex of rat liver small subunit (63).
31 Aug 2004 11:34 AR AR224-MI58-10.tex AR224-MI58-10.sgm LaTeX2e(2002/01/18)
P1: IKH
PROTEIN MAKING AND ITS OBSTRUCTION
237
As crystals containing the entire IF3 molecule or IF3N could not be produced
owing to tight crystal contacts, IF3N and IF3 linker regions were docked onto
the T30S complexed with IF3C with its known crystal structures (9), taking into
account the flexibility of the linker as indicated by NMR studies (19, 20). In
its docked location, IF3N is positioned close to the P-site (Figure 1a), leaving
sufficient space for cognate tRNA binding, but is too tight for allowing near-
cognate or noncognate interactions (48), indicating that IF3N discriminates against
noncognate tRNA by space exclusion principles. If the inherent flexibility of IF3
linker can be paralleled to that of the anti-SD region, it seems that IF3 exploits
its conformational mobility for fulfilling its various tasks. It is conceivable that
during the initial stages of the creation of the initiation complex, IF3 binds to
the interface side of the 30S subunit, as observed by cryo-EM (39) and hydroxyl
radical cleavage (12), while the linker that extends from the rim of the decoding
center to the subunit solvent side acts as a strap between the two IF3 domains,
presumably for transmitting signals between the two IF3 domains. Once cognate
interactions are created by the initiator tRNA, namely the formation of the right
initiation complex, IF3C swings to the platform backside (Figure 1d), facilitates
the detachment of the SD region, and enables subunit association. Such IF3C
swinging motion can also explain a prominent feature in the cryo-EM result (39)
called the “flying object” in Figure 1b, which so far has not been accounted for, as
this flying feature may represent a swinging IF3C domain trapped in the middle
of its motion.
Edeine: An Antibiotic Interfering with Initiation
by Limiting the Platform Mobility
The small ribosomal subunit seems to possess mobile elements for controlling
the fidelity of the decoding process. It is built of several domains radiating from
a junction located near the decoding center that seem to support correlated mo-
tions required for the protein biosynthesis process (17, 65), including the initial
accommodation of the ribosomal A-site (71). Comparative studies performed on
the available high-resolution structures of the free and the complexed ribosomal
particles also indicate significant mobility and show that specific conformations of
the small subunit can be associated with functional states and that antibiotics, such
as spectinomycin, can trap the small subunit at a particular conformation (51).
A head-to-shoulder motion in the small subunit may facilitate the threading
of the incoming mRNA through a dynamic pore, and the platform, located at the
opposite side of the particle, is likely to perform the motions that facilitate the
mRNA exit (57, 79). These motions are hindered by edeine, a universal antibi-
otic that inhibits the initiation of protein synthesis in all phylogenetic kingdoms.
Cocrystals of edeine and T30S showed that edeine binds in the solvent side of the
platform in a position that might affect the binding of IF3N, IF3 linker, and the
mRNA (Figure 1e). It also limits the platform mobility by inducing the forma-
tion of a new base pair and physically linking the decoding site, the P-site tRNA,
31 Aug 2004 11:34 AR AR224-MI58-10.tex AR224-MI58-10.sgm LaTeX2e(2002/01/18)
P1: IKH
238 YONATH
BASHAN
and the features of the platform critical for tRNA, IF3, and mRNA binding (25).
Pactamycin, an antibiotic agent that shares a protection pattern with edeine, acts
in a similar fashion (10).
PEPTIDE BOND FORMATION AND AMINO ACID
POLYMERIZATION
The PTC and Its Vicinity: A Sizable Symmetry-Related Region
The interface surfaces of both subunits are rich in rRNA, which comprises about
two thirds of the ribosomal mass. These RNA-rich surfaces contain both ribosomal
active sites, the decoding center in the small subunit and the PTC in the large subunit
(Figure1f), thus indicating that the ribosome is a ribozyme. A flexible intersubunit
bridge called B2a combines these two active sites and is composed solely of rRNA,
the tip of helix H69. Its strategic location (Figure 1f) hints at a possible role as
a platform carrying the tRNA acceptor stem from the A- to the P-site within the
frame of the overall translocation process. The extension of this bridge, namely
helix H69, forms part of the PTC rim. The PTC is located close to the subunit
interface, at the bottom of a cavity in the large ribosomal subunit, which hosts the
ribosomal protein L16 (E. coli numbering system is used throughout) in its upper
rim (Figure 2a,b).
The ribosome is a giant asymmetric riboprotein particle. However, a large
symmetry-related region containing about 180 nucleotides was revealed in all
known structures of the large ribosomal subunit (Figure 2c) (1, 2, 5, 6), with an
axis positioned close to the center of the PTC, between the A-site and the P-site
loops, pointing into the protein exit tunnel (Figure 2d–f ). This symmetry region
relates the backbone fold and the nucleotides’ orientations, rather than their types,
and includes several nucleotides that do not obey the symmetry, mostly bulged
nucleotides with functional relevance. The existence of this twofold symmetry can
be justified by the need to offer comparable supportive environments to two similar
chemical moieties that have to face each other to allow participation of the A-site
amine and the P-site carbonyl-carbon in peptide bond formation.
In the structure of D50S in complex with ASM, a 35-ribonucleotide chain
mimicking the aminoacylated-tRNA acceptor stem and its universal 3
end (5),
the ribosomal symmetry axis nearly coincides with the bond connecting the ASM
double helical features to its single-stranded 3
end, the moiety carrying the amino
acid. Hence, the A-site to P-site passage is likely to involve two independent,
albeit correlated, motions: a shift of the A-site tRNA helical regions, performed as
a part of the overall mRNA-tRNA translocation, and a spiral rotation of the tRNA
3
end (Figure 2f ), consistent with results of footprinting experiments, indicating
spontaneous movement of the A-tRNA acceptor stem into the P-site (77).
The outer contour of the symmetry-related region at the PTC surrounds an
arched void of a size sufficient to accommodate the tRNA-3
end, with a shape
and position that seem to be designed as the path along which the rotatory motion
31 Aug 2004 11:34 AR AR224-MI58-10.tex AR224-MI58-10.sgm LaTeX2e(2002/01/18)
P1: IKH
PROTEIN MAKING AND ITS OBSTRUCTION
239
proceeds (Figure 2d,g,h). The PTC wall located away from the subunit interface,
called the rear wall, forms the scaffold that guides the rotatory motion. Two uni-
versally conserved nucleotides, A2602 and U2585, that bulge into the center of
the PTC from its front wall (Figure 2d,f,g) seem to have a pivotal role in directing
the rotatory motion. A2602 is located at the upper side of the PTC, very close to
the rotation axis, and can interact with the bond connecting the 3
end and the rest
of the tRNA and U2585, which resides at the bottom of the PTC and points into
the ribosomal exit tunnel. The rotatory motion could be simulated with no space
constraints or steric hindrances. While rotating, the A-site 3
end slides along the
backbone of two rear-wall nucleotides and may interact with the rear-wall bases
that point inward. In this way the PTC walls confine the precise path for the ro-
tating moiety, guarantee smooth A-site to P-site translocation, and ensure that
the rotating moiety arrives at the P-site at a distance, configuration, orientation,
and stereochemistry optimal for peptide bond formation (Figure 2h). Importantly,
the interactions of ASM, mimicking the A-site tRNA and the derived P-site with
the PTC, which represent the beginning and the end of the rotatory motion, are
consistent with most of the available biochemical data (22, 32, 42, 77).
A nucleophilic attack by the amino nitrogen of the aminoacylated-tRNA can
readily occur at the pH optimal for protein biosynthesis in almost all species,
which is also the pH within the D50S crystals, as these are grown and maintained
under close to physiological conditions (28, 79). This attack is performed while
both the A-site and P-site are still occupied by the aminoacylated-tRNA and the
peptidyl-tRNA, respectively, generating a tetrahedral oxyanion intermediate. The
next steps of this reaction, namely sp
3
→ sp
2
reorganization and a transfer of a
hydrogen bond to the leaving group, which leads to the deacylation of the tRNA
in the P-site, seem to occur in concert with the rotatory motion. The passage of the
A-site tRNA 3
end into the P-site should assist the release of the leaving group,
thus vacating the space for a new aminoacyl tRNA and securing the processivity
of protein biosynthesis.
Four nucleotides of the P-site region that are located near the tunnel entrance
are positioned about 2
˚
Alower than their A-site mates, thus acquiring a spiral
nature to the rotary motion and ensuring the entrance of the nascent proteins to
their exit tunnel. Hence, the rotatory motion is the key component of a unified ribo-
somal mechanism for peptide bond formation, translocation, and nascent protein
progression. According to this integrated ribosomal machinery, the peptide bonds
are formed spontaneously in an exothermic reaction that drives the elongation
(1, 2, 5, 6). By providing the architectural means for accurate substrate positioning
and alignment, by the design of an exact path for the A-site to P-site passage,
and by the guidance of the rotatory motion along this path, the ribosome con-
tributes positional catalysis to peptide bond formation as well as to polypeptide
elongation. No ribosomal component is required for the mere chemical events of
peptide bond formation, but specific ribosomal moieties may have a major influ-
ence on the rate of protein biosynthesis, which plays a key role in cell vitality
(31, 52).
31 Aug 2004 11:34 AR AR224-MI58-10.tex AR224-MI58-10.sgm LaTeX2e(2002/01/18)
P1: IKH
240 YONATH
BASHAN
Elements of the PTC architecture seem to be designed for specific tasks, such
as directing the initiator tRNA to the P-site. Theoretically, the first tRNA could
be accommodated in the A-site or the P-site, since both are free and available.
However, analyzing the potential interactions in the two PTC binding sites shows
that compared with the A-site, the P-site contains an extra potential interaction
that can be exploited by a tRNA 3
end, indicating the preference for this mode of
binding. Hence, the sole stereochemical requirement for extending the initiation
process toward the creation of a peptide bond is that the initial P-site tRNA be at
the orientation acquired at the end of the rotatory motion.
The PTC architecture also ensures the exclusive incorporation of L-amino acids
and the rejection of the D-isomers, despite the occurrence of minute amounts of D-
amino acids in cells. Owing to the PTC significant tolerance for substrate binding
(1, 2, 5, 6, 81), one would expect that D-amino acid could be accommodated by it.
Hence, the exclusivity of the L-isomer in proteins points to the existence of a ribo-
somal mechanism for avoiding this mistake, which should function satisfactorily
despite the natural conformational flexibility of the PTC. U2506 and U2585 are
the two key nucleotides responsible for D-amino acid rejection (86). Among them,
U2585 is the nucleotide involved in directing the rotatory motion by interacting
with the amino acid bound to the CCA, and in the placement of the first tRNA
(27), consistent with the dominant lethal phenotypes produced by its mutations.
Overall, the PTC exclusion of D-amino acid incorporation into the growing chain
is due to space considerations as well as the creation of unproductive interactions
that may be irreversible (86). Substantial conformational rearrangements within
the PTC may bypass both mechanisms, consistent with the suggestion based on
mutation experiments (14).
Finally, the location of the symmetry-related region suggests its participation in
intraribosomal functional signaling, since it is positioned between the two lateral
protuberances of the large ribosomal subunit. Both stalks possess significant con-
formational dynamics and are presumed to undergo cooperative motions enabling
the entrance and the exit of aminoacylated and deacylated free tRNA, respectively
(Figure 2e). Thus, the symmetry-related region connects the four most important
functional sites of the ribosome: the PTC, the decoding center (via H69, located at
the PTC upper rim and serving as the intersubunit bridge B2a), the GTPase center,
and the L1 arm, which was suggested to serve as the “door” of the leaving tRNA
molecules (2, 6, 28).
Clinically Relevant PTC Targeting
The peptidyl transferase center is a major target site for universal inhibitors pro-
duced by microorganisms as well as for clinically relevant antibiotics. Among
the latter are chloramphenicol and clindamycin, which bind at the PTC and com-
pete with A-site and P-site tRNAs (41, 53). Chloramphenicol is widely used be-
cause it is highly specific for eubacteria and has a low level of toxicity in humans.
The crystal structure of D50S in complex with each of these two drugs at clini-
cally relevant drug concentrations is consistent with these findings. Furthermore,
31 Aug 2004 11:34 AR AR224-MI58-10.tex AR224-MI58-10.sgm LaTeX2e(2002/01/18)
P1: IKH
PROTEIN MAKING AND ITS OBSTRUCTION
241
although their binding sites differ, the sites of these two antibiotic drugs appear to
partially overlap (58).
The case of chloramphenicol is of special interest because it illuminates the
differences between medically meaningful and nonrelevant binding. In the crystal
structure of D50S in complex with chloramphenicol obtained at clinically relevant
drug concentrations, chloramphenicol is located well within the PTC, at the A-site
region. In this location (58), one of the chloramphenicol OH groups can make
ahydrogen bond with G2061, a nucleotide implicated in chloramphenicol resis-
tance in rat mitochondria (73). G2061 is also proximal to A2062, the mutation of
which confers resistance in Halobacterium halobium (38), and to U2506, which
is affected by in vivo modifications of chloramphenicol (29, 60).
Owing to the prominent selectivity of chloramphenicol, it binds to the H. maris-
mortui large subunit only at extremely high concentrations, about 1000 times higher
than the concentration commonly used in therapeutic treatment. Surprisingly, even
these large drug amounts did not lead to chloramphenicol binding in the site that
is consistent with the known biochemical and functional findings that are consis-
tent with chloramphenicol position at the PTC. Instead, in H50S it was localized
in the entrance of the ribosome tunnel, in the secondary site (25), held by con-
tacts between its aromatic ring and the tunnel’s hydrophobic crevice that hosts the
common macrolides.
Sparsomycin, Puromycin, and the “Fragment Reaction”
When discovered over half a century ago, sparsomycin and puromycin were con-
sidered antibiotics because microorganisms produced them. However, owing to
their universality, they are not useful clinically. Sparsomycin is a potent ribosome-
targeted inhibitor that acts on all cell types. Although sparsomycin does not compet-
itively inhibit A-site substrate binding, A-site antibiotics such as chloramphenicol
compete with it for binding to bacterial ribosomes, and mutants of A-site nu-
cleotides increase the tolerance to sparsomycin (34, 67).
Consistent with crosslink data (50), in complexes of D50S (5) and H50S (25)
sparsomycin forms stacking contacts with the nucleotide A2602, positioned in
the middle of the PTC. When sparsomycin is bound to an empty PTC, as in the
complex D50S, it occupies the A-site side of A2602 and alters the PTC confor-
mation, whereas when attached to the large subunit together with the P-site tRNA
analog, as in its complex with H50S (25), sparsomycin points at the P-site side
of this nucleotide. A2602 is a conserved nucleotide that seems to play a key role
in substrate positioning, translocation, and E-site tRNA release (49). This base
adopts a different orientation in each of the known complexes of the large subunit,
and its striking conformational variability appears to be synchronized with the
tRNA translocation within the PTC (2, 5, 6). Consequently, nucleotide A2602 was
proposed to act as a conformational switch within the PTC, propelling the rotating
moiety in concert with the action of helix H69, the feature that seems to assist the
shift of the tRNA acceptor stem at the subunit interface (2, 6, 80, 81).
31 Aug 2004 11:34 AR AR224-MI58-10.tex AR224-MI58-10.sgm LaTeX2e(2002/01/18)
P1: IKH
242 YONATH
BASHAN
Puromycin resembles the 3
terminus of aminoacyl-tRNA, but its aminoacyl
residue is linked via an amide bridge rather than an ester bond. For over four
decades it played a central role in biochemical experiments aimed at understanding
the mechanism of peptide bond formation (69). Puromycin structure resembles that
of the tip of aminoacylated-tRNA (Figure 2b); hence, it competes with A-site tRNA
binding. Consequently, it was used as a part of substrate analogs designed to mimic
different portions of aminoacylated-tRNA. These include compounds representing
the CCA moiety, such as a tetra-oligonucleotide called ACCP (ACC-puromycin),
and polynucleotides mimicking the entire portion of the tRNA that interacts with
the large subunit, such as ASM (Figure 2b). The ends of these analogs contain
puromycin, a universal ribosomal inhibitor whose structure mimics tyrosylated
adenine (Figure 2b) (5).
A striking net of extensive interactions between the tRNA acceptor stem and
the upper rim of the PTC cavity was observed in the 70S ribosome (85) as well
as in complexes of D50S with tRNA acceptor stem mimics (5). These remote
interactions, together with a universal base pair between the tRNA, govern the
accurate positioning of the tRNA substrates (1, 2, 5, 6, 80, 81), whereas the global
localization of the tRNA molecules is based on the match between the overall
ribosomal architecture and the size and shape of the tRNA molecules.
Two major participants in this interaction net are protein L16 and helix H69
(Figure 2b). The prominent contributions of protein L16 to accurate substrate
placement are reflected in the conservation of its three-dimensional structure in
eubacteria and archaea, represented by D50S (28) and H50S (4), respectively,
although the amino acid sequences of L16 and L10e, from the archaeon H. maris-
mortui, show poor homology. Similarly, structural observations manifested the
crucial contribution of H69 interactions to the key functional task of productive
alignment of the tRNA substrates. Lack of remote interactions may result from the
disorder of the intersubunit bridge B2a (the tip of H69), as observed for most of
the functionally relevant features in the H50S high-resolution structure (Figure 2a)
(4, 46). Similarly, tRNA analogs that are too short to reach the upper rim of the
PTC cavity bind in a mode similar, albeit distinctly different, to that within the PTC
(5, 26, 46, 59), requiring conformational rearrangements to participate in peptide
bond formation (43).
The binding of puromycin and its derivatives to the PTC in the presence of
an active donor substrate can result in peptide bond formation by a yet unknown
mechanism. This so-called fragment reaction is much (×10
3
) slower than the
normal formation of peptide bonds (43) and can be performed under various con-
ditions, some of which are neither similar to the optimal conditions for in vitro
protein biosynthesis, nor resemble the in vivo environment. A dipeptide formed
within H50S crystals by the fragment reaction, detected at the A-site, stimulated a
hypothesis that this finding represents normal amino acid polymerization, which
suggests that the A-site to P-site passage follows peptide bond formation (59). Such
a sequence of events seems to be impractical because it implies that each time a
new peptide bond is being formed, the entire newly formed polypeptide has to be
rotated by 180
◦
within the tunnel. However, because the entrance to the tunnel is
31 Aug 2004 11:34 AR AR224-MI58-10.tex AR224-MI58-10.sgm LaTeX2e(2002/01/18)
P1: IKH
PROTEIN MAKING AND ITS OBSTRUCTION
243
rather narrow, it poses severe space limitations, and rotations of bulky side chains
may lead to significant clashes. Similarly, the alternative suggestion, that for each
new peptide bond the nascent chain exploits free rotations around its main chain
bonds to compensate for the 180
◦
rotation of the tRNA 3
end, seems to create
problems because such rearrangements rule out incorporation of rigid residues,
such as proline. Furthermore, postpeptide bond rotation or rearrangement should
be exceedingly space and energy consuming, two assets that are hardly provided
during chain elongation and may not be compensated by the only energy source
of protein biosynthesis, namely the EF-G-induced GTP hydrolysis.
Thus, it seems that the attachment of this dipeptide to the A-site resulted either
from improper positioning of the reactant, as observed for similar compounds
in the absence of remote interactions, together with the lack of the bond around
which the rotatory motion should occur; from the creation of a peptide bond under
far from optimal reaction conditions [e.g., pH ∼5.6 instead of ∼8, and relatively
low salt concentrations (4) instead of the high salt required for efficient protein
production by H. marismortui (61, 79)]; or from the low affinity of puromycin
derivatives to the P-site (43, 59). Peptide bond formation in the A-site within H50S
crystals may reflect a special event, consistent with the variability of binding modes
observed in the PTC of these crystals. Thus, five complexes of the puromycin-like
substrate analogs with H50S (43) were required for reaching the conclusion that the
ribosomal catalytic contribution is positional rather than chemical. Remarkably,
this conclusion was evident by inspecting the structure of D50S (28) and verified
by a single complex of it (5). Compared with all H50S bound analogs, the D50S
complex is unique because D50S crystals are capable of providing the crucial
remote interactions, as their helix H69 is well ordered (28).
The ability of the ribosome to accommodate semireactive substrate analogs, in-
hibitors, and compounds supposed to represent reaction intermediates (46) does not
imply that each of the bound compounds could participate in amino acid polymer-
ization, although under specific circumstances a single peptide bonds formation
can be formed by the fragment reaction (59). Thus, it appears that in general the
orientations of reactants that are not placed by remote interactions may be incom-
patible with polypeptide elongation. Furthermore, almost all substrate analogs
mimicking only the tRNA 3
end necessitated conformational rearrangement to
participate in peptide bond formation and/or chain elongation (2, 6, 26, 43). Such
rearrangements should require time, justifying the lower rate of the fragment reac-
tions (43) compared with normal protein biosynthesis, and may be assisted by the
PTC mobility, in which the conserved A2602 seems to play a key role by adopting
various orientations (5, 6, 80, 81).
POST PEPTIDE BOND FORMATION
Macrolides and Ketolides: Resistance and Selectivity
A tunnel providing the path along which nascent proteins progress is located with
its entrance adjacent to the PTC. It was first observed in the mid-1980s (40, 83) and
31 Aug 2004 11:34 AR AR224-MI58-10.tex AR224-MI58-10.sgm LaTeX2e(2002/01/18)
P1: IKH
244 YONATH
BASHAN
although inferred by biochemical experiments (37, 54), cryo-EM revisualization
in the mid-1990s (16, 64) confirmed by X-ray crystallography (4, 28, 85) was
required for its final recognition. This tunnel spans the large ribosomal subunit
and has a nonuniform diameter. At its entrance the tunnel’s diameter matches
almost precisely the size of bulky amino acid residues; hence it is likely to restrict
motions of large side chains as well as conformational rearrangements of rigid
residues, but may allow the passage of flexible amino acid residues to which small
compounds are bound (33).
This tunnel is the preferred target of a large number of antimicrobial drugs from
the macrolide family, which ranks the highest in clinical usage. All macrolides
studied so far bind close to the entrance of the ribosomal exit tunnel, at a distance
from the PTC that permits the formation of a five- to six-amino-acid polypeptide
chain before reaching the drug. All macrolides block the exit tunnel, but their
binding modes and their exact positions may vary even when only one eubacterium
is considered (7, 8, 27, 56, 58).
Macrolides and ketolides possess a central lactone ring, composed of 12 to 22
atoms, to which at least one sugar moiety, called desosamine, is bound (Figure 3a).
Erythromycin, a macrolide of a 14-membered lactone ring, and its two semisyn-
thetic derivatives, clarithromycin and roxithromycin (designed to gain stability in
acidic environments and to have a broad spectrum, respectively), bind to the tunnel
in a similar manner (3, 58). All three interact exclusively with rRNA, and the crys-
tallographically observed contacts of the desosamine sugar with nucleotide A2058
(Figure 3a–c) clarify the mechanism whereby drug resistance in clinical pathogens
is acquired by A → G substitution or by the methylation of A2508 (15, 75). These
interactions also clarify why the replacement of the common prokaryotic adenine
by guanine, the typical eukaryotic moiety in this position, attains selectivity to
14-membered macrolides.
The rapidly acquired antibiotic resistance is a major handicap in modern
medicine and stimulated the search for improved drugs. Under special conditions,
15- and 16-member macrolides developed to combat resistance bind to ribosomes
with guanine in position 2058. An example is H50S, which possesses a guanine in
position 2058. Therefore, in this critical aspect the archaeon H. marismortui resem-
bles eukaryotes more than prokaryotes. A few macrolides belonging to the 15- and
16-member macrolide families could bind to H50S crystals when those crystals
were soaked in solutions containing extremely high concentrations of antibiotics
(24). However, the binding mode of azithromycin, a 15-membered ring macrolide,
to the archaeal H50S differs significantly from that to the eubacterial D50S (56).
In D50S azithromycin lies almost perpendicular to the tunnel long axis, similar to
all other D50S complexes of clinically relevant macrolides, ketolides, and azalides
studied so far. In H50S, however, the azithromycin binding mode leaves a relatively
large unoccupied space within the tunnel (Figure 3c), thus justifying the usage of
this drug for treating mammals possessing guanine in position 2058. Interestingly,
in D50S complex azithromycin binds cooperatively to two sites within the exit tun-
nel, and similar to the ketolides ABT-773 and telithromycin, it reaches the other
31 Aug 2004 11:34 AR AR224-MI58-10.tex AR224-MI58-10.sgm LaTeX2e(2002/01/18)
P1: IKH
PROTEIN MAKING AND ITS OBSTRUCTION
245
side of the tunnel wall (23S RNA domain II) in addition to the typical macrolide
interactions with domain V (7, 56). Thus, it seems that the recently developed
ketolides and the azalides have improved pharmacokinetics and enhanced activity
against certain macrolide-resistant pathogens presumably because they can form
additional interactions that may compensate for the loss of the contacts with the
A2058 (Figure 3b), which is the major site acquiring macrolide resistance.
Dual binding was observed also for streptogramins, an antibiotic family consist-
ing of two components capable of cooperative conversion of weak bacteriostatic
effects into lethal bactericidal activity. Crystals of the complex D50S-Synercid
®
,
a recently approved injectable streptogramin with excellent synergistic activity,
obtained at clinically relevant concentrations yielded the structural basis for its
action (27). In this complex the Synercid S
B
component binds to the same crevice
of the exit tunnel that is favored by common macrolides, and interacts with the
key nucleotide, A2058, in macrolide binding. Its slight inclination, compared with
that of most macrolides (3, 7, 56, 58), rationalizes its reduced antibacterial effects
in the absence of its S
A
mate. The S
B
component binds to the PTC and induces re-
markable conformational alterations, including a 180
◦
flip of U2585 base (Figure
3g), hence paralyzing the ability of U2585 to anchor the rotatory motion and to
direct the nascent protein into the exit tunnel. When bound to D50S, both Synercid
components interact with each other via the nucleotide A2062, which plays a key
role in chloramphenicol (38, 58) and tylosin (24) binding. These interactions seem
to stabilize the U2585 flipped conformation and consequently enhance the drug
antimicrobial activity. Importantly, U2585 conformation is hardly changed in an
H50S-S
A
complex (25), indicating a different binding mode (Figure 3g) consistent
with the inherent differences in PTC conformations between H50S and eubacterial
ribosomes (28, 85). Similarly, the S
B
interactions with A2058 in D50S explain why
S
B
components do not bind to H50S (25), which possesses the typical eukaryotic
G rather than the common eubacterial A in this position.
An astonishing binding mode has been observed for troleandomycin (TAO), a
semisynthetic derivative of erythromycin (Figure 3a) that has a modest clinical
relevance owing to the toxicity of its metabolites. In TAO all moieties capable
of participating in hydrogen bonding are either methylated or acetylated. Never-
theless, it binds close to the tunnel entrance and exploits the macrolide favorable
binding site, the vicinity of A2058, albeit by making different contacts (8) (Figure
3b). Compared with erythromycin, TAO binds somewhat deeper in the tunnel, and
its lactone ring is almost parallel to the tunnel wall and interacts not only with
23S RNA bases belonging to domain V but also with domain II. Furthermore, in
contrast to other macrolides that interact exclusively with the rRNA, TAO interacts
with the ribosomal protein L32.
At its unique binding orientation, TAO hits an element of the tunnel wall, the
tip of protein L22 beta-hairpin, a tunnel constituent reaching from close to the
PTC to the vicinity of the tunnel opening on the other side of the large subunit
(28, 46). This tip consists of 11 residues and contains at its very end an amino
acid triplet (E. coli residues 88–90) that can act as a “double hook” interacting
31 Aug 2004 11:34 AR AR224-MI58-10.tex AR224-MI58-10.sgm LaTeX2e(2002/01/18)
P1: IKH
246 YONATH
BASHAN
with the RNA, since it is composed of a highly conserved arginine, an invariant
alanine, and a residue that is in most species either an arginine or a lysine. In
the native large subunit structure, most of the double hook is embedded in a
narrow groove so that the space available for its conformational rearrangements is
rather limited. Consequently, TAO binding causes a swing of the entire L22 beta-
hairpin tip across the tunnel (Figure 3b,e). As the native conformation, the resulting
swung conformation is well accommodated on the tunnel walls, stabilized mainly
by electrostatic interactions and hydrogen bonds (8). Importantly, the region of
the tunnel wall that interacts with the L22 hairpin tip at its swung conformation
seems to be designed for these extensive interactions, indicating that swinging L22
beta-hairpin tip may be a general tunnel-shaping mechanism.
Tunnel Mobility and Ribosomal Involvement
in Cellular Regulation
The crystallographically observed alterations in the tunnel conformations could
be correlated not only with the structures of antibiotic-resistant mutations (13, 18)
but also with tunnel participation in gating and sequence discrimination (21, 30,
35, 44, 45, 55, 66, 68, 87). Examples are the SecM (secretion monitor) protein
(45, 55) and the leader peptide of the E. coli tryptophanase (tnaC) operon (21).
Under normal conditions, the SecM biosynthesis rate is similar to that of other
proteins. The SecM protein, produced in conjunction with a protein-export system
that recognizes an export signal, includes a sequence motif that causes elongation
arrest in the absence of a well-functioning protein-export system. This sequence is
located about 150 residues away from the protein N terminus, and within it are the
main features (a proline and a tryptophan, separated by 12 residues) that trigger
elongation arrest.
SecM arrest may be suppressed by mutations that can correlate remarkably
(Figure 3e) with the tunnel regions involved in the interactions with the L22 hairpin
tip at its swung conformation. It is conceivable, therefore, that the L22 hairpin tip
swinging mechanism, revealed by TAO binding, represents the means by which
the ribosome responds to SecM tunnel arrest signals (2, 6, 8). Protein L22 has
an elongated shape and lines the tunnel wall, extending from the tunnel entrance
to the vicinity of the tunnel opening (Figure 3f). Remarkably, this rather unusual
elongated shape is maintained also in the isolated native (70) and mutated protein
(13). Furthermore, in all known L22 structures a flexible region was observed only
in the hairpin tip hinge region (Figure 3f). The efficiency of this swinging motion
as a tunnel gate depends on its accuracy, which is dictated by the positioning of the
L22 hinge region. A pronounced patch of positively charged amino acids, detected
adjacent to the L22 hinge region (70), seems to dominate the positioning of the
hinge region of the L22 hairpin stem, which in turn should enable anchoring of
the double hook to both sides of the tunnel. Thus, the surface properties of L22
support a common mechanism for tunnel blockage accomplished by swinging of
the L22 hairpin tip.
31 Aug 2004 11:34 AR AR224-MI58-10.tex AR224-MI58-10.sgm LaTeX2e(2002/01/18)
P1: IKH
PROTEIN MAKING AND ITS OBSTRUCTION
247
Within D50S a single beta-strand chain connects the protein L22 hairpin tip,
and the region of this protein resides over 70
˚
Aaway, in the vicinity of the tunnel
opening on the surface of the ribosome. In this location the C terminus of L22 may
interact with the “pulling protein” and transmit signals according to which the
swinging may be induced or eliminated by a yet unknown mechanism. The SecM
nascent chain can also participate in signal transmission; after the incorporation
of ∼150 amino acids the chain should reach the tunnel opening. The studies
showing that signal transmission through growing chains could be correlated with
structural alterations in the membrane translocon pore that lines up directly with
the exit tunnel (35, 87) support this suggestion.
It seems, therefore, that a swing of the L22 beta-hairpin tip, similar to that in-
duced by TAO binding, is a universal mechanism for tunnel gating and that protein
L22 is a major player in tunnel discrimination. Furthermore, protein L22 as well
as the already formed portion of the nascent chain may be involved in transmitting
signals from the environment into the ribosomal core. Thus, the cellular signals
not only induce tunnel blockage and elongation arrest, they also control and mon-
itor the reverse hinge motions, which are required for alleviation of the arrest by
allowing sufficient space for nascent protein progression.
ACKNOWLEDGMENTS
Thanks are due to J.M. Lehn, P. Ahlberg, M. Lahav, H. Gilon, and to the ribo-
some groups at the Weizmann Institute, MPG for Ribosomal Structure in Ham-
burg and MPI Molecular Genetics in Berlin. X-ray diffraction data were collected
at ID19/SBC/APS/ANL and ID14/ESRF-EMBL. The U.S. National Institutes of
Health (GM34360), the German Ministry for Science & Technology (BMBF05-
641EA), and the Kimmelman Center for Macromolecular Assemblies provided
support. A.Y. holds the Helen and Martin Kimmel Professorial Chair.
The Annual Review of Microbiology is online at http://micro.annualreviews.org
LITERATURE CITED
1. Agmon I, Amit M, Auerbach T, Bashan A,
Baram D, et al. 2004. Ribosomal crystal-
lography: A flexible nucleotide anchoring
tRNA translocation facilitates peptide bond
formation, chirality discrimination and an-
tibiotics synergism. FEBS Lett. 567:20–26
2. Agmon I, Auerbach T, Baram D, Bartels
H, Bashan A, et al. 2003. On peptide bond
formation, translocation, nascent protein
progression and the regulatory properties
of ribosomes. Eur. J. Biochem. 270:2543–
56
3. Auerbach T, Bashan A, Harms J, Schluen-
zen F, Zarivach R, et al. 2002. Antibi-
otics targeting ribosomes: crystallographic
studies. Curr. Drug. Targets Infect. Disord.
2:169–86
4. Ban N, Nissen P, Hansen J, Moore PB,
Steitz TA. 2000. The complete atomic
structure of the large ribosomal subunit at
2.4
˚
A resolution. Science 289:905–20
5. Bashan A, Agmon I, Zarivach R, Schluen-
zen F, Harms J, et al. 2003. Structural ba-
sis for a unified machinery of peptide bond
31 Aug 2004 11:34 AR AR224-MI58-10.tex AR224-MI58-10.sgm LaTeX2e(2002/01/18)
P1: IKH
248 YONATH
BASHAN
formation, translocation and nascent chain
progression. Mol. Cell 11:91–102
6. Bashan A, Zarivach R, Schluenzen F, Ag-
mon I, Harms J, et al. 2003. Ribosomal
crystallography: peptide bond formation
and its inhibition. Biopolymers 70:19–41
7. Berisio R, Harms J, Schluenzen F, Zari-
vach R, Hansen HAS, et al. 2003. Struc-
tural insight into the antibiotic action of
telithromycin on resistant mutants. J. Bac-
teriol. 185:4276–79
8. Berisio R, Schluenzen F, Harms J, Bashan
A, Auerbach T, et al. 2003. Structural in-
sight into the role of the ribosomal tun-
nel in cellular regulation. Nat. Struct. Biol.
10:366–70
9. Biou V, Shu F, Ramakrishnan V. 1995. X-
ray crystallography shows that translational
initiation factor IF3 consists of two com-
pact alpha/beta domains linked by an alpha-
helix. EMBO J. 14:4056–64
10. Brodersen DE, Clemons WM Jr, Carter
AP, Morgan-Warren RJ, Wimberly BT, Ra-
makrishnan V. 2000. The structural basis
for the action of the antibiotics tetracycline,
pactamycin, and hygromycin B on the 30S
ribosomal subunit. Cell 103:1143–54
11. Carter AP, Clemons WM, Brodersen DE,
Morgan-Warren RJ, Wimberly BT, Ra-
makrishnan V. 2000. Functional insights
from the structure of the 30S ribosomal
subunit and its interactions with antibiotics.
Nature 407:340–48
12. Dallas A, Noller HF. 2001. Interaction of
translation initiation factor 3 with the 30S
ribosomal subunit. Mol. Cell 8:855–64
13. Davydova N, Streltsov V, Wilce M, Liljas
A, Garber M. 2002. L22 ribosomal pro-
tein and effect of its mutation on ribosome
resistance to erythromycin. J. Mol. Biol.
322:635–44
14. Dedkova LM, Fahmi NE, Golovine SY,
Hecht SM. 2003. Enhanced D-amino acid
incorporation into protein by modified ri-
bosomes. J. Am. Chem. Soc. 125:6616–17
15. Douthwaite S, Hansen LH, Mauvais P.
2000. Macrolide-ketolide inhibition of
MLS-resistant ribosomes is improved by
alternative drug interaction with domain II
of 23S rRNA. Mol. Microbiol. 36:183–93
16. Frank J, Zhu J, Penczek P, Li Y, Srivastava
S, et al. 1995. A model of protein synthesis
based on cryo-electron microscopy of the
E. coli ribosome. Nature 376:441–44
17. Gabashvili IS, Agrawal RK, Grassucci R,
Squires CL, Dahlberg AE, Frank J. 1999.
Major rearrangements in the 70S ribosomal
3D structure caused by a conformational
switch in 16S ribosomal RNA. EMBO J.
18:6501–7
18. Gabashvili IS, Gregory ST, Valle M, Gras-
succi R, Worbs M, et al. 2001. The polypep-
tide tunnel system in the ribosome and its
gating in erythromycin resistance mutants
of L4 and L22. Mol. Cell 8:181–88
19. Garcia C, Fortier PL, Blanquet S, Lalle-
mand JY, Dardel F. 1995. 1H and 15N res-
onance assignments and structure of the
N-terminal domain of Escherichia coli ini-
tiation factor 3. Eur. J. Biochem. 228:395–
402
20. Garcia C, Fortier PL, Blanquet S, Lalle-
mand JY, Dardel F. 1995. Solution struc-
ture of the ribosome-binding domain of E.
coli translation initiation factor IF3. Ho-
mology with the U1A protein of the eukary-
otic spliceosome. J. Mol. Biol. 254:247–59
21. Gong F, Yanofsky C. 2002. Instruction of
translating ribosome by nascent peptide.
Science 297:1864–67
22. Green R, Noller HF. 1997. Ribosomes and
translation. Annu. Rev. Biochem. 66:679–
716
23. Gualerzi CO, Pon CL. 1990. Initiation
of mRNA translation in prokaryotes. Bio-
chemistry 29:5881–89
24. Hansen JL, Ippolito JA, Ban N, Nissen P,
Moore PB, Steitz TA. 2002. The structures
of four macrolide antibiotics bound to the
large ribosomal subunit. Mol. Cell 10:117–
28
25. Hansen JL, Moore PB, Steitz TA. 2003.
Structures of five antibiotics bound at the
peptidyl transferase center of the large ribo-
somal subunit. J. Mol. Biol. 330:1061–75
26. Hansen JL, Schmeing TM, Moore PB,
31 Aug 2004 11:34 AR AR224-MI58-10.tex AR224-MI58-10.sgm LaTeX2e(2002/01/18)
P1: IKH
PROTEIN MAKING AND ITS OBSTRUCTION
249
Steitz TA. 2002. Structural insights into
peptide bond formation. Proc. Natl. Acad.
Sci. USA 99:11670–75
27. Harms J, Schlunzen F, Fucini P, Bartels H,
Yonath A. 2004. Alterations at the pep-
tidyl transferase centre of the ribosome
induced by the synergistic action of the
streptogramins dalfopristin and quin-
upristin. BMC Struct. Biol. 2:4
28. Harms J, Schluenzen F, Zarivach R, Bashan
A, Gat S, et al. 2001. High resolution struc-
ture of the large ribosomal subunit from
a mesophilic eubacterium. Cell 107:679–
88
29. Izard T, Ellis J. 2000. The crystal struc-
tures of chloramphenicol phosphotrans-
ferase reveal a novel inactivation mecha-
nism. EMBO J. 19:2690–700
30. Jenni S, Ban N. 2003. The chemistry of
protein synthesis and voyage through the
ribosomal tunnel. Curr. Opin. Struct. Biol.
13:212–19
31. Katunin V, Muth G, Strobel S, Wintermeyer
W, Rodnina M. 2002. Important contribu-
tion to catalysis of peptide bond formation
by a single ionizing group within the ribo-
some. Mol. Cell 10:339–46
32. Kim DF, Green R. 1999. Base-pairing be-
tween 23S rRNA and tRNA in the riboso-
mal A site. Mol. Cell 4:859–64
33. Kurzchalia TV, Wiedmann M, Breter H,
Zimmermann W, Bauschke E, Rapoport
TA. 1988. tRNA-mediated labelling of pro-
teins with biotin. A nonradioactive method
for the detection of cell-free translation
products. Eur. J. Biochem. 172:663–68
34. Lazaro E, Rodriguez-Fonseca C, Porse B,
Urena D, Garrett RA, Ballesta JP. 1996. A
sparsomycin-resistant mutant of Halobac-
terium salinarium lacks a modification at
nucleotide U2603 in the peptidyl trans-
ferase centre of 23 S rRNA. J. Mol. Biol.
261:231–38
35. Liao S, Lin J, Do H, Johnson AE. 1997.
Both lumenal and cytosolic gating of the
aqueous ER translocon pore are regulated
from inside the ribosome during membrane
protein integration. Cell 90:31–41
36. Makowski I, Frolow F, Saper MA, Shoham
M, Wittmann HG, Yonath A. 1987. Single
crystals of large ribosomal particles from
Halobacterium marismortui diffract to 6
˚
A.
J. Mol. Biol. 193:819–22
37. Malkin LI, Rich A. 1967. Partial resistance
of nascent polypeptide chains to proteolytic
digestion due to ribosomal shielding. J.
Mol. Biol. 26:329–46
38. Mankin AS, Garrett RA. 1991. Chloram-
phenicol resistance mutations in the single
23S rRNA gene of the archaeon Halobac-
terium halobium. J. Bacteriol. 173:3559–
63
39. McCutcheon JP, Agrawal RK, Philips SM,
Grassucci RA, Gerchman SE, et al. 1999.
Location of translational initiation factor
IF3 on the small ribosomal subunit. Proc.
Natl. Acad. Sci. USA 96:4301–6
40. Milligan RA, Unwin PN. 1986. Location
of exit channel for nascent protein in 80S
ribosome. Nature 319:693–95
41. Moazed D, Noller HF. 1987. Chloram-
phenicol, erythromycin, carbomycin and
vernamycin B protect overlapping sites in
the peptidyl transferase region of 23S ribo-
somal RNA. Biochemie 69:879–84
42. Moazed D, Noller HF. 1989. Intermediate
states in the movement of transfer RNA in
the ribosome. Nature 342:142–48
43. Moore PB, Steitz TA. 2003. After the ribo-
some structures: How does peptidyl trans-
ferase work? RNA 9:155–59
44. Morris DR, Geballe AP. 2000. Up-
stream open reading frames as regula-
tors of mRNA translation. Mol. Cell Biol.
20:8635–42
45. Nakatogawa H, Ito K. 2002. The riboso-
mal exit tunnel functions as a discriminat-
ing gate. Cell 108:629–36
46. Nissen P, Hansen J, Ban N, Moore PB,
Steitz TA. 2000. The structural basis of ri-
bosome activity in peptide bond synthesis.
Science 289:920–30
47. Petrelli D, Garofalo C, Lammi M, Spurio
R, Pon CL, et al. 2003. Mapping the ac-
tive sites of bacterial translation initiation
factor IF3. J. Mol. Biol. 331:541–56
31 Aug 2004 11:34 AR AR224-MI58-10.tex AR224-MI58-10.sgm LaTeX2e(2002/01/18)
P1: IKH
250 YONATH
BASHAN
48. Pioletti M, Schluenzen F, Harms J, Zari-
vach R, Gluehmann M, et al. 2001. Crystal
structures of complexes of the small ribo-
somal subunit with tetracycline, edeine and
IF3. EMBO J. 20:1829–39
49. Polacek N, Gomez MJ, Ito K, Xiong L,
Nakamura Y, Mankin A. 2003. The critical
role of the universally conserved A2602 of
23S ribosomal RNA in the release of the
nascent peptide during translation termina-
tion. Mol. Cell 11:103–12
50. Porse BT, Kirillov SV, Awayez MJ, Ot-
tenheijm HC, Garrett RA. 1999. Direct
crosslinking of the antitumor antibiotic
sparsomycin, and its derivatives, to A2602
in the peptidyl transferase center of 23S-
like rRNA within ribosome-tRNA com-
plexes. Proc. Natl. Acad. Sci. USA 96:
9003–8
51. Ramakrishnan V. 2002. Ribosome struc-
ture and the mechanism of translation. Cell
108:557–72
52. Rodnina MV, Wintermeyer W. 2001. Ribo-
some fidelity: tRNA discrimination, proof-
reading and induced fit. Trends Biochem.
Sci. 26:124–30
53. Rodriguez-Fonseca C, Amils R, Garrett
RA. 1995. Fine structure of the pep-
tidyl transferase centre on 23 S-like
rRNAs deduced from chemical probing
of antibiotic-ribosome complexes. J. Mol.
Biol. 247:224–35
54. Sabatini DD, Blobel G. 1970. Controlled
proteolysis of nascent polypeptides in rat
liver cell fractions. II. Location of the
polypeptides in rough microsomes. J. Cell
Biol. 45:146–57
55. Sarker S, Rudd KE, Oliver D. 2000. Re-
vised translation start site for secM defines
an atypical signal peptide that regulates Es-
cherichia coli secA expression. J. Bacte-
riol. 182:5592–95
56. Schluenzen F, Harms JM, Franceschi F,
Hansen HA, Bartels H, et al. 2003. Struc-
tural basis for the antibiotic activity of ke-
tolides and azalides. Structure 11:329–38
57. Schluenzen F, Tocilj A, Zarivach R, Harms
J, Gluehmann M, et al. 2000. Structure
of functionally activated small ribosomal
subunit at 3.3 angstroms resolution. Cell
102:615–23
58. Schluenzen F, Zarivach R, Harms J, Bashan
A, Tocilj A, et al. 2001. Structural basis for
the interaction of antibiotics with the pep-
tidyl transferase centre in eubacteria. Na-
ture 413:814–21
59. Schmeing TM, Seila AC, Hansen JL,
Freeborn B, Soukup JK, et al. 2002. A
pre-translocational intermediate in protein
synthesis observed in crystals of enzymati-
cally active 50S subunits. Nat. Struct. Biol.
9:225–30
60. Shaw WV, Leslie AG. 1991. Chloram-
phenicol acetyltransferase. Annu. Rev. Bio-
phys. Biophys. Chem. 20:363–86
61. Shevack A, Gewitz HS, Hennemann B,
Yonath A, Wittmann HG. 1985. Charac-
terization and crystallization of ribosomal
particles from Halobacterium marismor-
tui. FEBS Lett. 184:68–71
62. Spahn CM, Prescott CD. 1996. Throwing
a spanner in the works: antibiotics and the
translation apparatus. J. Mol. Med. 74:423–
39
63. Srivastava S, Verschoor A, Frank J. 1992.
Eukaryotic initiation factor 3 does not pre-
vent association through physical blockage
of the ribosomal subunit-subunit interface.
J. Mol. Biol. 226:301–4
64. Stark H, Mueller F, Orlova EV, Schatz M,
Dube P, et al. 1995. The 70S Escherichia
coli ribosome at 23
˚
A resolution: fitting the
ribosomal RNA. Structure 3:815–21
65. Stark H, Orlova EV, Rinke-Appel J, Junke
N, Mueller F, et al. 1997. Arrangement
of tRNAs in pre- and posttranslocational
ribosomes revealed by electron cryomi-
croscopy. Cell 88:19–28
66. Stroud RM, Walter P. 1999. Signal se-
quence recognition and protein targeting.
Curr. Opin. Struct. Biol. 9:754–59
67. Tan GT, DeBlasio A, Mankin AS. 1996.
Mutations in the peptidyl transferase cen-
ter of 23 S rRNA reveal the site of action of
sparsomycin, a universal inhibitor of trans-
lation. J. Mol. Biol. 261:222–30
31 Aug 2004 11:34 AR AR224-MI58-10.tex AR224-MI58-10.sgm LaTeX2e(2002/01/18)
P1: IKH
PROTEIN MAKING AND ITS OBSTRUCTION
251
68. Tenson T, Ehrenberg M. 2002. Regulatory
nascent peptides in the ribosomal tunnel.
Cell 108:591–94
69. Traut RR, Monro RE. 1964. The puromycin
reaction and its relationship to protein syn-
thesis. J. Mol. Biol. 10:63–72
70. Unge J, Berg A, Al-Kharadaghi S, Nikulin
A, Nikonov S, et al. 1998. The crystal struc-
ture of ribosomal protein L22 from Ther-
mus thermophilus: insights into the mecha-
nism of erythromycin resistance. Structure
6:1577–86
71. Valle M, Zavialov A, Li W, Stagg SM,
Sengupta J, et al. 2003. Incorporation of
aminoacyl-tRNA into the ribosome as seen
by cryo-electron microscopy. Nat. Struct.
Biol. 10:899–906
72. Vester B, Douthwaite S. 2001. Macrolide
resistance conferred by base substitu-
tions in 23S rRNA. Antimicrob. Agents
Chemother. 45:1–12
73. Vester B, Garrett RA. 1988. The impor-
tance of highly conserved nucleotides in the
binding region of chloramphenicol at the
peptidyl transfer centre of Escherichia coli
23S ribosomal RNA. EMBO J. 7:3577–87
74. von Bohlen K, Makowski I, Hansen HA,
Bartels H, Berkovitch-Yellin Z, et al. 1991.
Characterization and preliminary attempts
for derivatization of crystals of large ri-
bosomal subunits from Haloarcula maris-
mortui diffracting to 3
˚
A resolution. J. Mol.
Biol. 222:11–15
75. Weisblum B. 1995. Erythromycin resis-
tance by ribosome modification. Antimi-
crob. Agents Chemother. 39:577–85
76. Deleted in proof
77. Wilson KS, Noller HF. 1998. Molecular
movement inside the translational engine.
Cell 92:337–49
78. Wimberly BT, Brodersen DE, Clemons
WM Jr, Morgan-Warren RJ, Carter AP,
et al. 2000. Structure of the 30S ribosomal
subunit. Nature 407:327–39
79. Yonath A. 2002. The search and its out-
come: high-resolution structures of ribo-
somal particles from mesophilic, ther-
mophilic, and halophilic bacteria at various
functional states. Annu. Rev. Biophys.
Biomol. Struct. 31:257–73
80. Yonath A. 2003. Structural insight into
functional aspects of ribosomal RNA tar-
geting. ChemBioChem 4:1008–17
81. Yonath A. 2003. Ribosomal tolerance
and peptide bond formation. Biol. Chem.
384:1411–19
82. Yonath A, Glotz C, Gewitz HS, Bartels KS,
von Bohlen K, et al. 1988. Characterization
of crystals of small ribosomal subunits. J.
Mol. Biol. 203:831–34
83. Yonath A, Leonard KR, Wittmann HG.
1987. A tunnel in the large ribosomal sub-
unit revealed by three-dimensional image
reconstruction. Science 236:813–16
84. Yonath A, Muessig J, Tesche B, Lorenz S,
Erdmann VA, Wittmann HG. 1980. Crys-
tallization of the large ribosomal subunit
from B. stearothermophilus. Biochem. Int.
1:315–428
85. Yusupov MM, Yusupova GZ, Baucom A,
Lieberman K, Earnest TN, et al. 2001.
Crystal structure of the ribosome at 5.5
˚
A
resolution. Science 292:883–96
86. Zarivach R, Bashan A, Berisio R, Harms J,
Auerbach T, et al. 2004. Functional aspects
of ribosomal architecture: symmetry, chi-
rality and regulation. J. Phys. Org. Chem.
17:1–11
87. Woolhead CA, McCormick PJ, Johnson
AE. 2004. Nascent membrane and se-
cretory proteins differ in FRET-detected
folding far inside the ribosome and in
their exposure to ribosomal proteins. Cell
116:725–36
PROTEIN MAKING AND ITS OBSTRUCTION C-1
See legend on next page
HI-RES-MI58-10-Yonath.qxd 8/31/04 12:51 PM Page 1
C-2 YONATH
■
BASHAN
Figure 1 (a) Left: Space-filling representation of the interface view of T30S (57), with
approximate locations of the mRNA chain (magenta); the three tRNA molecules; the
swung anti-SD region (yellow), serving as an mRNA analog (78); and the IF3C (dark-red
star). Right: Ribbon representation of the solvent side of T30S highlighting the crystallo-
graphically determined (48) positions of IF3C, its neighboring proteins, the anti-SD region,
and the docked IF3N. Middle: The top insert is a cartoon representing the IF3 structure. The
bottom insert is a schematic representation of the tRNA structure and its location within the
ribosome. (b) A table showing the various observed locations of IF3, and an image based
on the cryo-EM structure of T30S in complex with IF3 (39). (c) A close-up view of the
IF3C environment (48), manifesting its superb match. (d) Space-filling representation of
the interface view of T30S with the positions of IF3C, IF3N, IF3 linker, and the SD-region.
The suggested motion of IF3C is represented by the arrow. (e) The action of edeine. Helix
H44 provides the decoding center. The tRNA molecule was docked onto T30S according
to Reference 85. (f) The location of the three tRNA molecules within the 70S ribosome,
docked on the structure of D50S, as in (e). The PTC in the large subunit is colored blue (A-
site) and green (P-site), with tRNA 3-end analogs in both placed according to Reference 5.
Note the strategic position of bridge B2a (yellow), combining the two active sites and inter-
acting with the acceptor stems of both A- and P-site tRNA. The axis of the symmetry-relat-
ed region, discussed below and in References 1 and 5, is shown in red.
HI-RES-MI58-10-Yonath.qxd 8/31/04 12:51 PM Page 2
PROTEIN MAKING AND ITS OBSTRUCTION C-3
See legend on next page
HI-RES-MI58-10-Yonath.qxd 8/31/04 12:51 PM Page 3
C-4 YONATH
■
BASHAN
Figure 2 (a) Top: Ribbon representation of the interface view of D50S (28). The back-
bone of the RNA is shown in blue, and the proteins main chains are in different colors.
Purple regions designate highly flexible, functionally relevant regions that can readily
become disordered, as observed in the crystal structure of H50S (4, 46). The cyan circle
designates the symmetry-related region inner portion. Bottom: The 23S RNA two-dimen-
sional diagram. The colored regions match the colored portion of the three-dimensional
structure shown on top. The symmetry-related region is colored in cyan. (b) A side view of
the PTC as in the complex of D50S with ASM (5) and the docked A-site and P-site tRNAs
(85). Note the immense interaction net of helix H69 and protein L16 with ASM and A-site
tRNA. The chemical compositions of ASM and the puromycin are shown on the left side.
(c–h) The unified rotatory mechanism. The RNA backbone is purple, and blue and green
represent the A-site and P-site regions, respectively, as well as the ASM and the derived P-
site mimic. (c) The RNA backbone of the symmetry-related region as observed in all
known structures, with the center of symmetry region in red. The core of the twofold sym-
metry-related region, including the PTC (pale cyan), is circled in red. The PTC and its
immediate environment are circled in red. (d) The region circled in red in (c) showing
nucleotides A2602 and U2585 and the approximate positions of the A-site and P-site with-
in the PTC. (e) The symmetry-related region in the large subunit and its functional exten-
sions bridging remote ribosomal functional sites (gold and pink). Bridge B2a, connecting
the PTC with the decoding site in the small subunit, is shown in yellow. (f–h) The A-site to
P-site passage, shown as an assembly of snapshots of the spiral motion, obtained by suc-
cessive rotations from A-site to P-site of the rotating moiety shown as a gradual transition
from blue to green. The blue-green turning arrows show the rotation direction. (f) Left: The
initial and final points of tRNA translocation on the large subunit, showing the stem shift
and the 3 end (shaded in blue) rotating moiety. Right: The simulated rotatory motion seen
perpendicular to the rotation axis. (g) Same as in (f) right, but as seen from the tunnel
toward the PTC. (h) A representation of the formation of the peptide bond.
HI-RES-MI58-10-Yonath.qxd 8/31/04 12:51 PM Page 4
PROTEIN MAKING AND ITS OBSTRUCTION C-5
See legend on next page
HI-RES-MI58-10-Yonath.qxd 8/31/04 12:51 PM Page 5
C-6 YONATH
■
BASHAN
Figure 3 Large subunit antibiotics and ribosomal involvement in cellular regulation.
(a) Chemical formulas of the macrolides and ketolides shown in (a–f). (b) Various
macrolide binding modes to D50S. Top: A section of the D50S ribosomal tunnel at the
vicinity of macrolides-ketolides binding height. rRNA is purple, and various macrolides
and ketolides bound to it (3, 7, 8, 27, 56, 58) are different colors. Nucleotide 2058 is high-
lighted. Bottom: The additional contacts made by telithromycin (left) and the special bind-
ing mode of troleandomycin (right). (c) The structural basis to macrolide selectivity based
on the identity of the nucleotide in position 2058. The region leading to short contacts is
highlighted in pink. (d) The difference in azithromycin binding to eubacterial (D50S) and
archaeon (H50S) ribosomes. Note that the latter provides only limited tunnel blockage.
(e) L22 swinging induced by troleandomycin (gold) binding together with a modeled
nascent poly-alanine (blue), on which the positions of the two key residues for SecM arrest,
Pro and Trp, are in red. The native L22 conformation is cyan, the swung conformation is
magenta, and the ribosome RNA components are purple. The cyan- and purple-shaded
areas correspond to the regions containing the mutations that bypass the elongation arrest
of SecM protein (45) in L22 and in the 23S RNA, respectively. (f) Front and top views of
D50S with P-site tRNA (green), the native (cyan) and the swung (magenta) L22 confor-
mations, a modeled poly-alanine (gold), and the crystallographically determined position
of troleandomycin (red). Middle: All known conformations of protein L22. Note the simi-
larity of these structures, except for the tip of the hairpin region. (g) U2585 native confor-
mation in D50S (U2585n) and its alterations induced by binding the synergetic strep-
togramin Synercid
©
(27), called U2582f. For comparison, the minor swing triggered by
virginamycin M (the S
A
streptogramin component) to H50S (25), U2585h, is shown. Left:
The relevant PTC region. Middle: The relevant PTC region together with the A-site and P-
site 3 ends. Right: Same as middle panel but with the entire simulated A-site to P-site pas-
sage (see Figure 2f ) and both Synercid components.
HI-RES-MI58-10-Yonath.qxd 8/31/04 12:51 PM Page 6