Molecular Cell
Article
Stop Codon Recognition by Release Factors Induces
Structural Rearrangement of the Ribosomal Decoding
Center that Is Productive for Peptide Release
Elaine M. Youngman,1,2 Shan L. He,1,2 Laura J. Nikstad,1,2 and Rachel Green1,2,*
1Howard
Hughes Medical Institute
of Molecular Biology and Genetics
Johns Hopkins University School of Medicine, Baltimore, MD 21205, USA
*Correspondence: [email protected]
DOI 10.1016/j.molcel.2007.09.015
2Department
SUMMARY
Peptide release on the ribosome is catalyzed in
the large subunit peptidyl transferase center by
release factors on recognition of stop codons in
the small subunit decoding center. Here we examine the role of the decoding center in this
process. Mutation of decoding center nucleotides or removal of 20 OH groups from the codon—deleterious in the related process of
tRNA selection—has only mild effects on peptide release. The miscoding antibiotic paromomycin, which binds the decoding center and
promotes the critical steps of tRNA selection,
instead dramatically inhibits peptide release.
Differences in the kinetic mechanism of paromomycin inhibition on stop and sense codons,
paired with correlated structural changes monitored by chemical footprinting, suggest that
recognition of stop codons by release factors
induces specific structural rearrangements in
the small subunit decoding center. We propose
that, like other steps in translation, the specificity of peptide release is achieved through an
induced-fit mechanism.
INTRODUCTION
Translation termination occurs when one of three nearuniversal stop codons is encountered in the A site of the
small ribosomal subunit. Stop codons are recognized by
class I release factor proteins (RFs), which bind to the ribosome and promote hydrolysis of the completed polypeptide from the P site tRNA. In bacteria, the three stop codons are decoded by two class I RFs with overlapping
specificity: RF1 recognizes UAG and UAA while RF2 recognizes UAA and UGA. RF-catalyzed peptide release is
efficiently prevented during elongation, leading to an estimated frequency of inaccurate termination at sense codons in vivo of at most 105 for both RF1 and RF2 (Jorgen-
sen et al., 1993), and discrimination against individual
sense codons in vitro ranging from 103- to over 106-fold
(Freistroffer et al., 2000). This level of specificity is at least
as great as that achieved during EF-Tu-facilitated tRNA
selection (reviewed in Rodnina and Wintermeyer, 2001),
and in this case accuracy is not achieved through a kinetic
proofreading mechanism because GTP hydrolysis-associated steps in termination occur downstream of the committed peptide release step (Freistroffer et al., 1997, 2000).
The accuracy of peptide release arises from the ability
of RFs to promote hydrolysis in the large ribosomal subunit active site specifically at stop codons, which depends
ultimately on discrimination between sense and nonsense
codons by RFs in the distantly located decoding center of
the small subunit. Little is known about the mechanism by
which this discrimination is accomplished. In contrast, the
equivalent discriminatory step during translation elongation, whereby cognate tRNAs are distinguished from
near-cognate tRNAs, has been carefully dissected in
a body of recent structural and kinetic studies (Ogle and
Ramakrishnan, 2005; Rodnina and Wintermeyer, 2001).
Structurally, specific interactions with the three-nucleotide codon:anticodon helix stabilize conformations of
decoding center nucleotides that are not favored in the
absence of bound tRNA: G530 rotates from a syn- to an
anti-conformation, and A1492 and A1493, which normally
stack inside 16S rRNA helix 44 (h44), move to extrahelical
positions (Ogle et al., 2001) (Figure 1A). More global conformational changes in the small subunit collectively referred to as ‘‘domain closure’’ are also observed only in
the presence of a cognate codon:anticodon interaction
(Ogle et al., 2002). Consistent with these structural observations, mutation of A1492, A1493, or G530 leads to kinetic deficiencies in forward steps in the tRNA selection
pathway that can be attributed to defects in closure
(Cochella et al., 2007). The functional relevance of these
structural changes is further supported by the fact that
miscoding antibiotics, which impact the kinetics of multiple steps in the tRNA selection pathway (Gromadski and
Rodnina, 2004b; Pape et al., 2000), promote these same
structural rearrangements even in the absence of bound
tRNA (Carter et al., 2000). The accuracy of tRNA selection
can thus be attributed to ribosome recognition of the
Molecular Cell 28, 533–543, November 30, 2007 ª2007 Elsevier Inc. 533
Molecular Cell
A Role for the Decoding Center in Peptide Release
Figure 1. The Effect of Mutations in the
Ribosomal Decoding Center on Peptide
Release
(A) Structure of the ribosomal decoding center,
showing A1492 and A1493 in their intrahelical
h44 positions in the absence of bound substrate (left) and in their ‘‘induced’’ positions in
the presence of bound mRNA and a cognate
tRNA anticodon stem-loop (ASL) (center), or
in the presence of paromomycin and streptomycin (right). Figure was made using PDB entries 1FJF, 1IBM, and 1FJG (Carter et al.,
2000; Ogle et al., 2001; Wimberly et al., 2000)
and PyMOL (http://pymol.sourceforge.net).
(B–E) Rate constants for catalysis at saturating
release factor concentrations (kcat) for wildtype and mutant ribosomes on (B) stop- or
(C–E) near-stop-programmed ribosomes.
Bars represent the mean ± standard error
from at least two experiments measured at
the following concentrations of RF1: 5 mM for
UAG, 125 mM for CAG, 100 mM for UCG, and
200 mM for UAC. Throughout the figures, black
bars represent release reactions on stop codons while gray bars indicate reactions on
near-stop codons.
geometry of a correct substrate—a codon:anticodon helix
with Watson-Crick geometry—leading to acceptance of
this substrate via a kinetically driven process described
by Rodnina and colleagues (Gromadski and Rodnina,
2004a).
The molecular mechanisms that determine high accuracy recognition of stop codons are not known. Although
an RF ‘‘tripeptide anticodon’’ at least partially responsible
for the observed codon specificities of RF1 and RF2 has
been genetically identified (Ito et al., 2000; Mora et al.,
2003), mutations outside of this region also affect specificity (Ito et al., 1998; Uno et al., 2002). Given that RFs have
low affinity for stop codon-containing RNAs in solution
(Brown and Tate, 1994; Chavatte et al., 2003), it seems
clear that the recognition event must include interactions
with the ribosomal environment in which the codon is presented (i.e., the small subunit decoding center). Low-resolution cryo-EM and medium-resolution crystal structures
of ribosome-bound RFs (Klaholz et al., 2003; Petry et al.,
2005; Rawat et al., 2003) confirm predictions from biochemical and genetic studies that RFs have distinct domains that interact with the peptidyl transferase center
on one hand and the decoding center on the other (Ito
et al., 2000; Scarlett et al., 2003; Wilson et al., 2000), but
no detailed structural view of the interactions between
RFs and stop codons is available. Kinetically, much of
what is known about the accuracy of peptide release
comes from biochemical studies by Ehrenberg and colleagues (Freistroffer et al., 2000), who demonstrated that
RF discrimination against the group of sense codons
that differ from stop codons by a single nucleotide
(‘‘near-stop’’) derives both from reductions in the catalytic
rate constant (kcat) and from increases in the concentration of release factor required to achieve half of kcat (K1/2).
Although differences in the affinity of RFs for stop- and
near-stop-programmed ribosome complexes are expected, the mechanistic underpinning of the observed difference in catalytic rate constant on stop and near-stop
codons is not clear. Are release factors bound at nearstop codons simply positioned inappropriately for catalysis, or does the increased kcat on stop codons reflect
a specific induction of catalysis on stop codon recognition?
Given that tRNAs and release factors must engage the
same two functional centers on the ribosome—the decoding and the peptidyl transferase centers—core strategies
for controlling ribosome activity might be shared. Our initial study indicated that, in fact, tRNA selection and stop
codon recognition rely on different properties of the ribosomal decoding center (Youngman et al., 2006). Here we
534 Molecular Cell 28, 533–543, November 30, 2007 ª2007 Elsevier Inc.
Molecular Cell
A Role for the Decoding Center in Peptide Release
Figure 2. The Effect of Deoxyribonucleotides in the A Site Codon on Peptide Bond Formation and Peptide Release
(A) Rate constants for peptide bond formation on a ribose (filled triangles) or deoxyribose (open triangles) UUU (Phe) codon. On the ribose codon,
fitting the data to a hyperbola yields a rate constant of 8.5 ± 1.3 s1 at saturation. On the deoxyribose codon, the reaction is saturated by 0.5 mM
ribosomes and gives a rate constant of (1.4 ± 0.08) 3 104 s1.
(B) Reaction curves at saturating RF concentrations for peptide release on a ribose (filled triangles) or deoxyribose (open triangles) UAA (stop) codon.
Fitting to a single exponential gives a rate constant of 0.34 ± 0.03 s1 for the ribose codon, and 0.25 ± 0.004 s1 for the deoxyribose codon. RF concentration was 5 mM for ribose, 10 mM for deoxyribose. Error bars represent standard error from at least two experiments.
fully examine the effects of mutations in decoding center
rRNA and mRNA elements and of miscoding aminoglycoside antibiotics on RF-mediated stop codon recognition.
Our results indicate that RF recognition of stop codons induces structural rearrangements in the decoding center
that likely explain the observed differences in rates of catalysis of peptide release on stop and near-stop codons.
Thus, although the specific structural changes involved
are distinct, the specificity of peptide release—like the
specificity of tRNA selection—is achieved through an
induced-fit mechanism.
RESULTS
Decoding Site Perturbations Have Mild Effects
on Peptide Release
To examine the role of the decoding center in peptide release, mutations were incorporated into 16S rRNA nucleotides A1492, A1493, and G530, which play essential roles
in tRNA selection (Cochella et al., 2007). Mutation at these
three positions results in dominant lethality (Cochella et al.,
2007), and so variant ribosomes were inducibly expressed
and purified using a previously described affinity tagging
approach (Youngman and Green, 2005). A pre-steadystate, single-turnover kinetic assay was used to measure
the rate of hydrolysis of f-[35S]-Met from P site-bound
f-[35S]-Met-tRNAfMet on ribosomes programmed with an
mRNA encoding methionine followed by a stop codon
(Youngman et al., 2004). At a saturating RF1 concentration
(5 mM), the decoding center nucleotide substitutions had
no discernible effect on the rate of peptide release on
the stop codon UAG (Figure 1B). Some of these mutations
do appear to have mild effects on release factor binding as
we have reported previously (Youngman et al., 2006).
We next looked at the effect of these mutations on peptide release at near-stop codons. Initiation complexes
were formed on mRNAs that encoded Met followed by
CAG, UCG, or UAC, and rate constants for peptide release
were measured as above. For each message, RF titrations
were performed to establish conditions where the ribosomes (wild-type and variant) were saturated with RF
(data not shown). In these reactions on near-stop codons,
variant ribosomes now all exhibited substantial defects in
catalysis (Figures 1C–1E). These defects range from as little as 2.5-fold in the case of the G530A mutant on a UAC
codon to as much as 115-fold in the case of the A1493C
mutant on UCG. Hence although decoding center mutations had no effect on the rate of peptide release on
stop codons, peptide release catalyzed by these same
variant ribosomes on near-stop codons is characterized
by decreases in the rate constant for peptide release (kcat).
In an effort to perturb the decoding center in a different
manner, we examined the effect of substituting the nucleotides of the A site codon with deoxyribose nucleotides.
The decoding site nucleotides make four hydrogen-bonding interactions with 20 OH groups on the A site codon
during tRNA selection (Ogle et al., 2001). On ribosomes
carrying DNA in the A site, the loss of these stabilizing
interactions between decoding site nucleotides and
20 OHs of the A site codon leads to a 10-fold reduction in
the affinity of aminoacyl-tRNA (Fahlman et al., 2006; Potapov et al., 1995) and is thought to be responsible for failure
to translate a DNA-based message (McCarthy and Holland, 1965). However, the magnitude of the defects in individual steps of tRNA selection had not been measured.
Here, we measure one of these steps, the rate of peptide
bond formation between f-[35S]-Met-tRNAfMet and PhetRNAPhe on ribosomes programmed with a message containing either an all-ribose or all-deoxyribose Phe codon in
the A site. Consistent with the known importance of interactions with the codon:anticodon helix in promoting this
reaction, we observed a 60,000-fold decrease in the rate
constant at saturation on ribosomes carrying the all-deoxyribose Phe codon in the A site (Figure 2A). In contrast,
Molecular Cell 28, 533–543, November 30, 2007 ª2007 Elsevier Inc. 535
Molecular Cell
A Role for the Decoding Center in Peptide Release
Figure 3. The Effect of Aminoglycosides on Peptide Release
Observed rate constants for peptide release on the stop codon UAA or three near-stop codons in the absence or presence of (A) streptomycin or (B)
paromomycin. In every case, rate constants were measured at 53 the K1/2 of RF1 in the absence of antibiotic (Freistroffer et al., 2000). Those concentrations are as follows: 40 nM for UAA, 100 mM for CAA, and 50 mM for UCA and UAC. Streptomycin was present at 20 mM. Paromomycin was used
at 50 mM for UAA and 5 mM for all other codons. Bars represent the mean ± standard error from at least two experiments.
the kcat for peptide release on ribosomes programmed
with a message bearing an all-deoxyribose UAA stop codon was only marginally reduced relative to the all-ribose
message (1.4-fold, Figure 2B), although slightly higher
concentrations of RF1 (10 mM) were required to reach saturation. Peptide release is therefore generally refractory
to changes in the decoding center relative to the larger
effects of these substitutions on tRNA selection.
Effects of Aminoglycoside Antibiotics
on Peptide Release
The aminoglycoside antibiotics paromomycin and streptomycin bind to the ribosome in the small subunit decoding
center and lead to structural rearrangements that ultimately
affect the kinetics of tRNA selection and thus decrease
overall accuracy. It has been reported that aminoglycoside
antibiotics of the neomycin class inhibit both RF1- and RF2catalyzed peptide release at stop codons (Brown et al.,
1993) and we have observed paromomycin-mediated inhibition of RF1-catalyzed peptide release at stop codons under conditions of low RF concentration (Youngman et al.,
2006). To examine the effect of miscoding agents on the
accuracy of peptide release, we prebound streptomycin
(20 mM) or paromomycin (5 or 50 mM) to ribosomes programmed with the stop codon UAA or its near-stop neighbors CAA, UCA, or UAC and measured rate constants for
peptide release. In initial experiments with each codon,
the experiment was performed at a concentration of RF1
known to be saturating in the absence of antibiotic (53
the K1/2, Freistroffer et al., 2000). Addition of 20 mM streptomycin had little effect on the rate of peptide release on any
of these codons (Figure 3A). In contrast, paromomycin dramatically inhibited peptide release at all four codons. In this
initial experiment, strong inhibition of peptide release was
seen at 5 mM paromomycin for the near-stop codons but required 50 mM paromomycin for the stop codon (Figure 3B).
This striking difference in responsiveness to paromomycin
led us to postulate that RFs might interact with stop and
near-stop codons in fundamentally different ways. To ex-
amine this possibility, we characterized the kinetic mechanism of paromomycin inhibition. This analysis revealed
a number of critical differences in the signature of paromomycin inhibition on stop and near-stop codons that are
described in the next section.
One intriguing possibility suggested by our results was
that paromomycin functioned in the decoding center independently of the RF to modulate catalytic activity in the
physically remote peptidyl transferase center. This idea
was dismissed after finding that neither the background
rate of peptidyl-tRNA hydrolysis nor the rate of peptidyl
transfer between f-[35S]-Met-tRNAfMet and the A site substrate puromycin was affected by the presence of paromomycin (data not shown).
Paromomycin Inhibits Peptide Release Differently
on Stop and Near-Stop Codons
The decrease in the observed rate constant for peptide release in the presence of paromomycin could be due either
to a decrease in the affinity of RF1 for paromomycinbound ribosomes or to inhibition of catalysis of peptide
release. To distinguish between these possibilities, we
measured the observed rate constant for peptide release
at increasing concentrations of RF1 and fit the resulting
kobs versus [RF1] curves to a hyperbolic equation to determine kcat and K1/2. The values for these constants obtained in the absence and presence of the antibiotic
were then compared. In the case of peptide release on
the stop codon UAA, the inhibitory effect of paromomycin
could be attributed entirely to an increase in the K1/2 of
RF1 in the presence of paromomycin. Reciprocal titrations
of release factor and paromomycin revealed that paromomycin and RF1 behave competitively on stop codon
complexes: the K1/2 of RF1 increased with increasing paromomycin concentration and the IC50 of paromomycin
increased with increasing RF1 concentration (Figures 4A
and 4B). In contrast, addition of up to 150 mM paromomycin had no effect on kcat (i.e., with saturating RF1, Figure 4C).
536 Molecular Cell 28, 533–543, November 30, 2007 ª2007 Elsevier Inc.
Molecular Cell
A Role for the Decoding Center in Peptide Release
accurately determined both in the absence and presence
of paromomycin (UCA and UAC), the magnitude of these
effects was similar: K1/2 was increased by 5- to 10-fold,
while kcat was decreased by about 30-fold (Figures 4D
and 4E). Paromomycin also clearly decreased kcat on the
near-stop codon CAA (from 3.7 3 103 s1 to less than
1 3 104 s1), although we were not able to accurately determine K1/2 on this codon due to the very slow rate constants and the high concentrations of RF1 required in the
presence of paromomycin. These paromomycin inhibition
data for near-stop codon ribosome complexes are thus
distinct from what we observed on stop codon complexes, and are not indicative of simple competitive
binding.
Figure 4. The Kinetic Mechanism of Paromomycin Inhibition
Is Different on Stop and Near-Stop Codons
(A–C) Paromomycin is a competitive inhibitor of RF binding to ribosomes programmed with a stop codon. (A) The K1/2 of RF1 for stop codon-programmed complexes increases with increasing paromomycin
concentration. At each concentration of paromomycin, RF1 titrations
were performed and the resulting kobs versus [RF1] curves were fit to
a hyperbolic equation. The bars represent the best-fit K1/2 values ±
the error in the fit. (B) The IC50 of paromomycin for the release reaction
on stop codon-programmed ribosomes increases with increasing RF1
concentration. IC50 values were determined by performing paromomycin titrations at a given concentration of RF1 and fitting the resulting
kobs versus [paromomycin] curves to a hyperbolic equation. Bars represent the best-fit IC50 values ± the error in the fit. (C) The rate constant
for peptide release at saturating RF1 concentrations does not change
in the presence of increasing concentrations of paromomycin. Bars
represent the mean kcat ± standard error from at least two experiments.
(D and E) Paromomycin affects both K1/2 and kcat values for peptide release on near-stop codon-programmed ribosomes. (D) K1/2 and (E) kcat
values in the absence (open bars) or presence (filled bars) of 5 mM
paromomycin. RF1 titrations were performed, and the resulting kobs
versus [RF1] curves were fit to a hyperbolic equation. Bars represent
the best-fit values ± the standard error in the fit.
The paromomycin inhibition data are markedly different
for the case of the near-stop-codon complexes where the
compound has striking effects on both the RF1 K1/2 and
on kcat. For the two codons where K1/2 and kcat could be
Synthetic Effects on RF Function Are Observed
when Decoding Center Mutations Are Challenged
with Paromomycin
We next examined the combined effects of mutations in
the decoding center and the presence of paromomycin.
In general, the defects associated with paromomycin
addition and mutation of the decoding center appear to
be additive. In release reactions on the near-stop codon
UCG using A1492G or A1493C mutant ribosomes, addition of 5 mM paromomycin reduced the observed rate constant for peptide release to the point that an accurate rate
constant could not be determined (data not shown). A
quantitative evaluation of the effect of paromomycin addition to mutant ribosomes could therefore only be performed on ribosomes programmed with a stop codon
(UAG), and those results are presented below. In the
case of A1492G variant ribosomes, addition of 25 mM paromomycin caused essentially no change in the rate constant for peptide release at a saturating RF1 concentration
(Figure 5A) but resulted in a large (200-fold) increase in K1/2
(Figure 5B). On A1493C mutant ribosomes, however, the
addition of paromomycin resulted in defects in kcat even
on this stop codon, whereas the mutation alone had no
effect on kcat (Figure 1B). In this case, kcat was reduced
by nearly 10-fold (Figure 5A), and the RF1 K1/2 was again
dramatically increased by about 300-fold (Figure 5B).
Although the error associated with the measurement of
the high K1/2 values seen in the presence of paromomycin
is large, it is clear that paromomycin increases the RF1
K1/2 for these mutants by at least two orders of magnitude.
In all cases, therefore, paromomycin potentiates the
defects of decoding site mutants in peptide release.
Release Factor Binding Displaces Bound
Paromomycin on Stop, but Not Near-Stop, Codons
As discussed above, although paromomycin inhibits peptide release on both stop and near-stop codons, the kinetic mechanism of this inhibition is strikingly different.
On stop codon-programmed ribosomes, the simplest explanation for the observed inhibition is that paromomycin
acts as a competitive inhibitor of RF binding. This predicts
that at the kinetically defined saturating concentration of
Molecular Cell 28, 533–543, November 30, 2007 ª2007 Elsevier Inc. 537
Molecular Cell
A Role for the Decoding Center in Peptide Release
Figure 5. Decoding Site Mutations and
Paromomycin Act Synergistically to Inhibit Peptide Release
(A) kcat and (B) K1/2 values for peptide release
on the stop codon UAG by decoding site variant ribosomes in the absence (open bars) and
presence (closed bars) of 25 mM paromomycin.
kcat values are the mean ± standard error from
2 measurements of the rate constant for peptide release at saturation (5 mM RF1 in the absence or 150 mM RF1 in the presence of paromomycin). K1/2 values were determined as in
Figure 4 and bars represent the best-fit values
± the standard error in the fit.
RF for a given concentration of paromomycin (where kcat
is unaffected), paromomycin is not bound to the decoding
center. To directly assess this possibility, we used chemical footprinting analysis to ask whether paromomycin
binding was affected by the presence of release factor.
This assay takes advantage of the known protection of
16S rRNA nucleotide A1408 from DMS by bound paromomycin. Ribosome complexes were formed containing
tRNAfMet in the P site of ribosomes programmed with
a Met-UAA mRNA. Chemical modification with DMS followed by primer extension analysis resulted in a strong
primer extension stop at position A1408 (Figure 6A). Addition of paromomycin to these same complexes resulted in
strong protection from DMS modification at A1408 as previously described (Moazed and Noller, 1987). Importantly,
addition of RF1 alone to these complexes resulted in discernible but modest protection at this position (Figures 6A
and 6B). When paromomycin was prebound to these ribosome complexes and RF1 then added prior to modification with DMS, paromomycin protection at A1408 was
lost in an RF1-concentration-dependent manner (see
quantitation in Figure 6C). Given that paromomycin makes
direct contacts with A1408 when bound to the ribosome
(Ogle et al., 2001), the observed loss of protection likely reflects paromomycin dissociation. These results are consistent with mutually exclusive binding of paromomycin
and RF1 to stop codon-programmed ribosomes.
On near-stop codons, paromomycin inhibition of peptide release results from both a decrease in kcat and an increase in the K1/2 for RF1. The effects of paromomycin on
kcat even under conditions of saturating RF suggest that
these two agents can bind simultaneously to near-stopprogrammed ribosomes. We again used DMS footprinting
to assess the state of paromomycin binding to ribosomes
programmed with the near-stop codon UCA in the absence and presence of RF1. As on stop codon-programmed ribosomes, paromomycin binding protects
A1408 from modification. In this case, however, addition
of RF1 to complexes containing paromomycin caused little to no change in the protection of A1408 (Figures 6B and
6C). These data indicate that, in contrast to stop codonprogrammed ribosomes, paromomycin remains bound
even at saturating RF1 concentrations when ribosomes
are programmed with a near-stop codon.
DISCUSSION
This quantitative analysis of the effects of decoding center
mutations and aminoglycoside antibiotics on peptide release provides insights into the mechanism of stop codon
recognition by release factors. The most immediate conclusion that emerges from our data is that the mechanism
by which stop codons are recognized on the ribosome
must be quite different than that by which cognate aminoacyl-tRNAs are recognized. Mutation of conserved decoding center nucleotides, demonstrated to cause substantial defects (10- to 50-fold) in the forward rates
associated with tRNA selection (Cochella et al., 2007), in
contrast has very little effect on peptide release at stop codons (Figure 1). Similarly, removal of 20 OH groups from the
A site codon leads to a dramatic (60,000-fold) reduction in
the rate of accommodation of cognate aminoacyl-tRNA
but has almost no effect on peptide release (Figure 2).
Finally, two miscoding antibiotics that bind in or near the
decoding center have distinct effects on the processes of
tRNA selection and peptide release. Paromomycin, which
generally promotes the steps of tRNA selection, instead
acts as a competitive inhibitor of RF binding on stop codons and thus negatively affects RF-mediated stop codon
recognition (Figure 4). Streptomycin, which also has substantial effects on tRNA selection, has no discernible
effects on RF function on stop or near-stop codons
(Figure 3). Preliminary data consistent with the detailed
analysis presented here were recently published in a symposium article (Youngman et al., 2006).
The ribosomal and mRNA elements that are critical for
stop codon recognition by release factors are therefore
distinct from those required for codon recognition during EF-Tu-facilitated tRNA selection. Key nucleotides
A1492, A1493, and G530, which undergo dramatic conformational rearrangements to form critical contacts in
the minor groove of the codon:anticodon helix, are not important for stop codon recognition by release factors, nor
are the 20 OH groups on the mRNA that interact with these
critical rRNA nucleotides. Moreover, the disparate effects
of paromomycin and streptomycin on RF function provide
structural insight into RF-mediated stop codon recognition. Paromomycin binding directly displaces A1492 and
A1493 from h44 into their extrahelical conformations
538 Molecular Cell 28, 533–543, November 30, 2007 ª2007 Elsevier Inc.
Molecular Cell
A Role for the Decoding Center in Peptide Release
Figure 6. RF1 and Paromomycin Binding Are Mutually Exclusive on Ribosomes Programmed with Stop, but Not with
Near-Stop, Codons
(A and B) Primer extension on 16S rRNA from (A) UAA- or (B) UCA-programmed ribosome complexes modified with DMS in the presence or
absence of paromomycin and RF1. G, C, and A are dideoxy sequencing lanes. Arrow, A1408.
(C) Quantitation of the data in (A) and (B). A1408 band intensity was first
normalized to a constant band in each lane (arrowhead). The data are
presented as normalized A1408 intensity in the presence of paromomycin, with 1.0 defined as the A1408 band intensity under the same
conditions in the absence of paromomycin.
(Ogle et al., 2001); the extrahelical positioning of these nucleotides is thus thought to represent a productive state
for tRNA selection. The competitive nature of paromomycin inhibition of RF-mediated stop codon recognition
strongly argues that the productive state for peptide
release places A1492 and A1493 in their h44-intrahelical
positions. Streptomycin binds nearby at the junction of
the small subunit shoulder with the h44 region and directly
promotes the closure interactions that are required for
tRNA selection (Carter et al., 2000; Ogle et al., 2002).
The failure of streptomycin to affect peptide release at
any codon tested argues that small subunit closure is
not critical for peptide release. Thus, the molecular deter-
minants for RF-mediated stop codon recognition are distinct from those utilized during tRNA selection.
Most importantly, the data presented here provide
structural insight into the mechanism by which the specificity of peptide release is achieved. These insights come
from the observation that certain perturbations of the decoding center have strikingly different effects on peptide
release at stop and near-stop (sense) codons. First, while
decoding center mutations (A1492, A1493, and G530 variants) have no effect on peptide release at stop codons,
they result in substantial rate defects in peptide release
at near-stop codons (generally 10- to 100-fold, with mutations at A1493 being the most deleterious) (Figure 1).
In addition, although paromomycin negatively affects
peptide release on both stop and near-stop codons, the
kinetic mechanisms of this inhibition are distinct. The K1/2
of RF1 for both stop and near-stop complexes is increased in the presence of paromomycin—the simplest
explanation for this observation is competitive binding of
the two molecules in the decoding center. However, on
near-stop codons, addition of paromomycin results not
only in substantial effects on the K1/2 for RF binding, but
also in substantial decreases in the kcat for peptide release
(Figure 4).
These data are most easily explained by a model
wherein stop codon recognition by RFs induces conformational rearrangements in the termination complex that
are incompatible with continued binding to paromomycin.
The correlated prediction would be that binding of RF to
near-stop-codon complexes fails to induce these same
conformational rearrangements, thus allowing for simultaneous binding of RF and paromomycin and ultimately resulting in effects on catalysis (kcat) by the aminoglycoside
(Figure 7). Chemical modification analysis was used to test
these predictions by examining binding of paromomycin
in the decoding center of stop and near-stop complexes
in the presence of increasing amounts of RF. Indeed,
paromomycin apparently remained bound to near-stop
complexes even in the presence of kinetically saturating
concentrations of RF, while addition of RF to authentic
stop codon complexes resulted in the loss of paromomycin binding (Figure 6). A second binding site for paromomycin, located in the large ribosomal subunit near helix
69, was recently reported in a crystallographic study by
Doudna Cate and colleagues (Borovinskaya et al., 2007).
That study argued that paromomycin binding to this helix
69 site is responsible for paromomycin-mediated inhibition of ribosome recycling through prevention of ribosome
conformational rearrangements in intersubunit bridges
(including helix 69) normally promoted by binding of ribosome recycling factor to posttermination complexes. Although we have not directly assessed the role of this binding site in paromomycin-mediated inhibition of peptide
release, the inhibition we observe can readily be accounted for by the presence or absence of paromomycin
in the decoding center.
We argue here that the structural rearrangement that
occurs on stop codon recognition (which here results in
Molecular Cell 28, 533–543, November 30, 2007 ª2007 Elsevier Inc. 539
Molecular Cell
A Role for the Decoding Center in Peptide Release
Figure 7. Model for the Accuracy of Stop
Codon Recognition
(A) When RF1 binds to ribosomes with a stop
codon in the A site, a change in the conformation of the termination complex is induced that
results ultimately in an ‘‘induced’’ ribosome
state that is fully active for peptide release.
This state includes a conformation of the decoding center that is not compatible with paromomycin binding.
(B) When RF1 binds to ribosomes with a nearstop codon in the A site, it is unable to induce
the conformational change that leads to fast
peptide release. In the context of this ‘‘uninduced’’ decoding site conformation, paromomycin remains bound and further inhibits
peptide release.
paromomycin dissociation) is normally required for fast induction of catalysis of peptide release in the large subunit
active site, and that the rate of peptide release on nearstop codons is impaired because the ribosome:RF complex does not efficiently rearrange to this maximally active
state (Figure 7). Earlier studies in both bacteria and eukaryotes have provided indirect evidence that peptide release involves an initial codon-independent binding step
followed by a rearrangement to a stop codon-specific
state that is fully active for peptide release (Chavatte
et al., 2003; Freistroffer et al., 2000; Yoshimura et al.,
1999). Our data are consistent with these ideas and provide insights into the nature of the rearrangements that
may be required to reach the state maximally active for
peptide release.
The increased sensitivity of peptide release to decoding
center mutations at near-stop codons relative to stop codons (Figure 5) is also consistent with this view. These
data indicate that the mutations in the decoding center further impair the already hindered rearrangement to the active state on near-stop complexes. However, rearrangement upon RF recognition of stop codons is sufficiently
robust to be insensitive to these mutations. The simultaneous presence of paromomycin and at least some decoding center mutations is then sufficiently inhibitory to
cause kcat effects even on stop codons. The texture of
these data is reminiscent of synthetic approaches used
in a wide variety of systems to establish that events are associated with a common pathway (Duncan et al., 2007;
Hartman et al., 2001).
At a practical level, the data presented here provide evidence that the in vivo effects of aminoglycosides are likely
to be manifested both on the process of tRNA selection
(stimulatory) and on the process of peptide release (inhibitory). The in vivo efficacy of aminoglycoside therapy used
to stimulate nonsense codon readthrough in prematurely
truncated genes (Kaufman, 1999) likely benefits from
both facets of drug action. While these ideas have previously been explored with reporter constructs in yeast
(Salas-Marco and Bedwell, 2005), the competitive nature
of termination and near-cognate tRNA incorporation at
nonsense codons makes the relative contribution of these
two effects difficult to evaluate. The in vitro biochemistry
performed here allows us to address these effects independently. Further experiments will be required to determine whether aminoglycosides also inhibit peptide release
in eukaryotes, where the decoding center of the ribosome
is conserved but release factor structure is distinct (Frolova
et al., 1999).
While we have little structural information about the rearrangement promoted during RF-mediated stop codon
recognition, because it is incompatible with paromomycin
binding it likely ends in a state where A1492 and A1493
adopt their intrahelical h44 positions. Given that conformational changes in the large ribosomal subunit active
site must also occur for the hydrolysis reaction to proceed
(Bieling et al., 2006; Brunelle et al., 2006; Schmeing et al.,
2005), and that the active site and decoding center are
spanned by the release factor, one view is that stop codon
recognition promotes structural rearrangements in the RF
that then activate the large subunit peptidyl transferase
center for catalysis. In fact, RFs can undergo large conformational rearrangements, although the physiological relevance of such large conformational changes is currently
debated (Graille et al., 2005; Klaholz et al., 2003; Petry
et al., 2005; Rawat et al., 2003; Vestergaard et al., 2005;
Zoldak et al., 2007). The rearrangement of the ribosomal
decoding center seen here may indicate a role for the ribosome in promoting rearrangements in the release factor,
or may simply be required for accommodating structural
rearrangements in the RF. High-resolution structures of
RFs bound to ribosome complexes containing stop and
near-stop codons will be required to elucidate the specific
interactions that facilitate this rearrangement, as has been
done so beautifully for the case of tRNA selection (Ogle
and Ramakrishnan, 2005). Although such structural detail
is currently lacking, our data argue that conformational rearrangements occur in response to RF recognition of stop
codons, and further, that the accuracy of stop codon recognition, like the accuracy of tRNA selection (Gromadski
and Rodnina, 2004a; Ogle and Ramakrishnan, 2005), is
driven by an induced-fit mechanism. The repeated use
540 Molecular Cell 28, 533–543, November 30, 2007 ª2007 Elsevier Inc.
Molecular Cell
A Role for the Decoding Center in Peptide Release
of induced fit by the ribosome to control catalysis suggests a broad mechanism for regulation by a variety of extraribosomal factors.
EXPERIMENTAL PROCEDURES
Strains and Reagents
Tight-couple 70S ribosomes from strain MRE600 were isolated from
cells in exponential growth as previously described (Moazed and Noller, 1986) and stored in HiFi buffer (50 mM Tris-HCl [pH 7.5], 70 mM
NH4Cl, 30 mM KCl, 3.5 mM MgCl2, 0.5 mM spermidine, 8 mM putrescine, and 2 mM DTT). Native IFs 1 and 3 and His-tagged IF2 were overexpressed and purified as described (Brunelle et al., 2006) and stored
in IF storage buffer (20 mM Tris-HCl [pH 7.5], 70 mM NH4Cl, 30 mM
KCl, 7 mM MgCl2, 1 mM DTT, and 20% glycerol) at 80 C. N-terminally His-tagged RF1 was expressed and purified as previously described (Shimizu et al., 2001) and stored in RF storage buffer (30 mM
Tris-HCl [pH 7.5], 70 mM NH4Cl, 30 mM KCl, 7 mM MgCl2, 5 mM
b-mercaptoethanol, 50% glycerol) at 20 C. mRNAs were made by
PCR and T7 transcription to produce the sequence 50 -GGGUUAA
CUUUAGAAGGAGGUAAAAAAAAUGNNNUUUUUCUUU-30 with an
appropriate second codon as specified in the Results. mRNAs with
20 deoxynucleotides were purchased from Dharmacon and were identical to the sequence above except with coding sequence 50 -AUGdU
dAdAUUUUUCUUU-30 for peptide release and 50 -AUGdUdUdU
AAAUUCUUU-30 for peptidyl transfer. tRNAs were purchased from
Sigma-Aldrich.
DMS Footprinting and Primer Extension
Complexes for DMS modification were formed in DMS buffer (80 mM
HEPES-KOH [pH 7.8], 10 mM MgCl2, and 100 mM NH4Cl) by incubating 1 mM 70S ribosomes, 1.5 mM tRNAfMet, and 3 mM mRNA for 30 min
at 37 C. Twenty picomoles of ribosome complex were modified in
each reaction. Stop codon complexes were diluted to 50 nM for modification; near-stop complexes were modified at 1 mM. Antibiotics were
added for 2 min prior to incubation with RF1 for 5 min. DMS (SigmaAldrich) was diluted 1:12 in ethanol and added to complexes at a final
concentration of 44 mM. Modification was carried out for 10 min at
37 C and stopped with 1 volume of DMS stop buffer (1 M Tris-HCl
[pH 7.5] and 1 M bME). Modified complexes were ethanol precipitated
and resuspended in RNAqueous lysis buffer, and RNA was extracted
using the RNAqueous kit (Ambion). Primer extension was then carried
out as described (Merryman and Noller, 1998) using a primer that begins extension at 16S rRNA nucleotide 1490. Quantitation was done
using ImageQuant software. Normalization was performed by dividing
A1408 intensity by the intensity of a constant band (arrowhead, Figures
6A and 6B). To correct for the modest protection by RF1 seen in the
absence of paromomycin, modification in the presence of paromomycin was then expressed as a fraction of the modification seen at the
same concentration of RF1 in the absence of paromomycin.
The RF1 concentration ranges used for both the stop and near-stop
complexes were chosen to represent 0.625, 6.25, and 31.25 times the
measured K1/2 value at the chosen paromomycin concentration. Thus,
the RF1 concentrations used in these experiments are similarly saturating for both codons tested. Additionally, extended incubation of
near-stop paromomycin complexes with a high concentration of release factor gave no further loss of A1408 protection (data not shown).
ACKNOWLEDGMENTS
Kinetic Assays
Wild-type ribosome initiation complexes were formed in buffer A (50
mM Tris-HCl [pH 7.5], 70 mM NH4Cl, 30 mM KCl, 7 mM MgCl2, and
1 mM DTT) by incubating 2 mM 70S ribosomes, 2.6 mM f-[35S]-MettRNAfMet, 6 mM mRNA, 3 mM each IFs 1, 2, and 3, and 1 mM GTP at
37 C for 30 min. Complexes were then pelleted through a 1 ml sucrose
cushion (1.1 M sucrose in buffer A) in a Beckman TLA 100.3 rotor at
69,000 rpm for 2 hr at 4 C. Pelleted complexes were resuspended in
HiFi buffer and stored in aliquots at 80 C. Concentration was determined by scintillation counting of [35S]-Met. Ribosomes with mutations
in decoding center nucleotides were expressed in strain DH5a from
plasmid pSpurMS2 and isolated as described (Youngman and Green,
2005). Initiation complexes containing mutant ribosomes were formed
as above except that 1 mM purified 50S subunits from strain MRE600
(isolated as in Cochella and Green, [2005]) were included during complex formation to replace 50S subunits lost during purification of mutant ribosomes. Release assays were performed in HiFi buffer at
37 C by mixing equal volumes of initiation complex and RF1. At appropriate time points, reactions were quenched with 25% formic acid and
released f-[35S]-Met was resolved from unreacted f-[35S]-Met-tRNAfMet
by electrophoretic TLC as described (Youngman et al., 2004). Generally, initiation complexes were present at 0.25 mM in reactions. In cases
where K1/2 or IC50 values less than 1 mM were being measured, initiation
complexes were used at 10 nM in all reactions. For K1/2 and IC50 measurements, rate constants were measured at least in duplicate at a minimum of 8 concentrations of the reagent being titrated (either RF1 or
paromomycin). Release reactions were carried out on ribosomes programmed with the stop codon UAA except in the case of assays involving tagged ribosomes, where UAG-programmed ribosomes were used
to eliminate background from low levels of RF2 contamination in ribosome preparations. Peptidyl transfer reactions were performed as described (Cochella and Green, 2005). For reactions containing antibiotics, initiation complexes formed as above were preincubated with
paromomycin or streptomycin (Sigma-Aldrich) at twice the desired final
concentration at 37 C for 2 min before being mixed with an equal volume of release factor or ternary complex.
We thank M. Rodnina, J. Lorsch, and members of the Green lab for
valuable discussions; S. Urban, J. Shaw, and L. Cochella for critical
review of the manuscript; NIH for funding the project; and HHMI for salary support to R.G.
Received: July 11, 2007
Revised: August 28, 2007
Accepted: September 7, 2007
Published: November 29, 2007
REFERENCES
Bieling, P., Beringer, M., Adio, S., and Rodnina, M.V. (2006). Peptide
bond formation does not involve acid-base catalysis by ribosomal residues. Nat. Struct. Mol. Biol. 13, 423–428.
Borovinskaya, M.A., Pai, R.D., Zhang, W., Schuwirth, B.S., Holton,
J.M., Hirokawa, G., Kaji, H., Kaji, A., and Cate, J.H. (2007). Structural
basis for aminoglycoside inhibition of bacterial ribosome recycling.
Nat. Struct. Mol. Biol. 14, 727–732.
Brown, C.M., and Tate, W.P. (1994). Direct recognition of mRNA stop
signals by Escherichia coli polypeptide chain release factor two.
J. Biol. Chem. 269, 33164–33170.
Brown, C.M., McCaughan, K.K., and Tate, W.P. (1993). Two regions of
the Escherichia coli 16S ribosomal RNA are important for decoding
stop signals in polypeptide chain termination. Nucleic Acids Res. 21,
2109–2115.
Brunelle, J.L., Youngman, E.M., Sharma, D., and Green, R. (2006). The
interaction between C75 of tRNA and the A loop of the ribosome stimulates peptidyl transferase activity. RNA 12, 33–39.
Carter, A.P., Clemons, W.M., Brodersen, D.E., Morgan-Warren, R.J.,
Wimberly, B.T., and Ramakrishnan, V. (2000). Functional insights
from the structure of the 30S ribosomal subunit and its interactions
with antibiotics. Nature 407, 340–348.
Molecular Cell 28, 533–543, November 30, 2007 ª2007 Elsevier Inc. 541
Molecular Cell
A Role for the Decoding Center in Peptide Release
Chavatte, L., Frolova, L., Laugaa, P., Kisselev, L., and Favre, A. (2003).
Stop codons and UGG promote efficient binding of the polypeptide
release factor eRF1 to the ribosomal A site. J. Mol. Biol. 331, 745–758.
Cochella, L., and Green, R. (2005). An active role for tRNA in decoding
beyond codon:anticodon pairing. Science 308, 1178–1180.
Cochella, L., Brunelle, J.L., and Green, R. (2007). Mutational analysis
reveals two independent molecular requirements during transfer
RNA selection on the ribosome. Nat. Struct. Mol. Biol. 14, 30–36.
Moazed, D., and Noller, H.F. (1987). Interaction of antibiotics with functional sites in 16S ribosomal RNA. Nature 327, 389–394.
Mora, L., Zavialov, A., Ehrenberg, M., and Buckingham, R.H. (2003).
Stop codon recognition and interactions with peptide release factor
RF3 of truncated and chimeric RF1 and RF2 from Escherichia coli.
Mol. Microbiol. 50, 1467–1476.
Ogle, J.M., and Ramakrishnan, V. (2005). Structural insights into translational fidelity. Annu. Rev. Biochem. 74, 129–177.
Duncan, M.C., Ho, D.G., Huang, J., Jung, M.E., and Payne, G.S.
(2007). Composite synthetic lethal identification of membrane traffic
inhibitors. Proc. Natl. Acad. Sci. USA 104, 6235–6240.
Ogle, J.M., Brodersen, D.E., Clemons, W.M., Jr., Tarry, M.J., Carter,
A.P., and Ramakrishnan, V. (2001). Recognition of cognate transfer
RNA by the 30S ribosomal subunit. Science 292, 897–902.
Fahlman, R.P., Olejniczak, M., and Uhlenbeck, O.C. (2006). Quantitative analysis of deoxynucleotide substitutions in the codon-anticodon
helix. J. Mol. Biol. 355, 887–892.
Ogle, J.M., Murphy, F.V., Tarry, M.J., and Ramakrishnan, V. (2002).
Selection of tRNA by the ribosome requires a transition from an
open to a closed form. Cell 111, 721–732.
Freistroffer, D.V., Pavlov, M.Y., MacDougall, J., Buckingham, R.H.,
and Ehrenberg, M. (1997). Release factor RF3 in E.coli accelerates
the dissociation of release factors RF1 and RF2 from the ribosome in
a GTP-dependent manner. EMBO J. 16, 4126–4133.
Pape, T., Wintermeyer, W., and Rodnina, M.V. (2000). Conformational
switch in the decoding region of 16S rRNA during aminoacyl-tRNA
selection on the ribosome. Nat. Struct. Biol. 7, 104–107.
Freistroffer, D.V., Kwiatkowski, M., Buckingham, R.H., and Ehrenberg,
M. (2000). The accuracy of codon recognition by polypeptide release
factors. Proc. Natl. Acad. Sci. USA 97, 2046–2051.
Frolova, L.Y., Tsivkovskii, R.Y., Sivolobova, G.F., Oparina, N.Y., Serpinsky, O.I., Blinov, V.M., Tatkov, S.I., and Kisselev, L.L. (1999). Mutations in the highly conserved GGQ motif of class 1 polypeptide release
factors abolish ability of human eRF1 to trigger peptidyl-tRNA hydrolysis. RNA 5, 1014–1020.
Graille, M., Heurgue-Hamard, V., Champ, S., Mora, L., Scrima, N.,
Ulryck, N., van Tilbeurgh, H., and Buckingham, R.H. (2005). Molecular
basis for bacterial class I release factor methylation by PrmC. Mol. Cell
20, 917–927.
Gromadski, K.B., and Rodnina, M.V. (2004a). Kinetic determinants of
high-fidelity tRNA discrimination on the ribosome. Mol. Cell 13, 191–
200.
Gromadski, K.B., and Rodnina, M.V. (2004b). Streptomycin interferes
with conformational coupling between codon recognition and GTPase
activation on the ribosome. Nat. Struct. Mol. Biol. 11, 316–322.
Hartman, J.L., IV, Garvik, B., and Hartwell, L. (2001). Principles for the
buffering of genetic variation. Science 291, 1001–1004.
Ito, K., Uno, M., and Nakamura, Y. (1998). Single amino acid substitution in prokaryote polypeptide release factor 2 permits it to terminate
translation at all three stop codons. Proc. Natl. Acad. Sci. USA 95,
8165–8169.
Ito, K., Uno, M., and Nakamura, Y. (2000). A tripeptide ‘anticodon’
deciphers stop codons in messenger RNA. Nature 403, 680–684.
Jorgensen, F., Adamski, F.M., Tate, W.P., and Kurland, C.G. (1993).
Release factor-dependent false stops are infrequent in Escherichia
coli. J. Mol. Biol. 230, 41–50.
Kaufman, R.J. (1999). Correction of genetic disease by making sense
from nonsense. J. Clin. Invest. 104, 367–368.
Klaholz, B.P., Pape, T., Zavialov, A.V., Myasnikov, A.G., Orlova, E.V.,
Vestergaard, B., Ehrenberg, M., and van Heel, M. (2003). Structure
of the Escherichia coli ribosomal termination complex with release factor 2. Nature 421, 90–94.
McCarthy, B.J., and Holland, J.J. (1965). Denatured DNA as a direct
template for in vitro protein synthesis. Proc. Natl. Acad. Sci. USA 54,
880–886.
Merryman, C., and Noller, H.F. (1998). Footprinting and modificationinterference analysis of binding sites on RNA. In RNA-Protein Interactions: A Practical Approach, C.W.J. Smith, ed. (Oxford: Oxford University Press), pp. 237–253.
Moazed, D., and Noller, H.F. (1986). Transfer RNA shields specific
nucleotides in 16S ribosomal RNA from attack by chemical probes.
Cell 47, 985–994.
Petry, S., Brodersen, D.E., Murphy, F.V., IV, Dunham, C.M., Selmer,
M., Tarry, M.J., Kelley, A.C., and Ramakrishnan, V. (2005). Crystal
structures of the ribosome in complex with release factors RF1 and
RF2 bound to a cognate stop codon. Cell 123, 1255–1266.
Potapov, A.P., Triana-Alonso, F.J., and Nierhaus, K.H. (1995). Ribosomal decoding processes at codons in the A or P sites depend differently on 20 -OH groups. J. Biol. Chem. 270, 17680–17684.
Rawat, U.B., Zavialov, A.V., Sengupta, J., Valle, M., Grassucci, R.A.,
Linde, J., Vestergaard, B., Ehrenberg, M., and Frank, J. (2003).
A cryo-electron microscopic study of ribosome-bound termination
factor RF2. Nature 421, 87–90.
Rodnina, M.V., and Wintermeyer, W. (2001). Fidelity of aminoacyltRNA selection on the ribosome: kinetic and structural mechanisms.
Annu. Rev. Biochem. 70, 415–435.
Salas-Marco, J., and Bedwell, D.M. (2005). Discrimination between
defects in elongation fidelity and termination efficiency provides mechanistic insights into translational readthrough. J. Mol. Biol. 348, 801–
815.
Scarlett, D.J., McCaughan, K.K., Wilson, D.N., and Tate, W.P. (2003).
Mapping functionally important motifs SPF and GGQ of the decoding
release factor RF2 to the Escherichia coli ribosome by hydroxyl radical
footprinting. Implications for macromolecular mimicry and structural
changes in RF2. J. Biol. Chem. 278, 15095–15104.
Schmeing, T.M., Huang, K.S., Strobel, S.A., and Steitz, T.A. (2005). An
induced-fit mechanism to promote peptide bond formation and
exclude hydrolysis of peptidyl-tRNA. Nature 438, 520–524.
Shimizu, Y., Inoue, A., Tomari, Y., Suzuki, T., Yokogawa, T., Nishikawa,
K., and Ueda, T. (2001). Cell-free translation reconstituted with purified
components. Nat. Biotechnol. 19, 751–755.
Uno, M., Ito, K., and Nakamura, Y. (2002). Polypeptide release at sense
and noncognate stop codons by localized charge-exchange alterations in translational release factors. Proc. Natl. Acad. Sci. USA 99,
1819–1824.
Vestergaard, B., Sanyal, S., Roessle, M., Mora, L., Buckingham, R.H.,
Kastrup, J.S., Gajhede, M., Svergun, D.I., and Ehrenberg, M. (2005).
The SAXS solution structure of RF1 differs from its crystal structure
and is similar to its ribosome bound cryo-EM structure. Mol. Cell 20,
929–938.
Wilson, K.S., Ito, K., Noller, H.F., and Nakamura, Y. (2000). Functional
sites of interaction between release factor RF1 and the ribosome. Nat.
Struct. Biol. 7, 866–870.
Wimberly, B.T., Brodersen, D.E., Clemons, W.M., Jr., Morgan-Warren,
R.J., Carter, A.P., Vonrhein, C., Hartsch, T., and Ramakrishnan, V.
(2000). Structure of the 30S ribosomal subunit. Nature 407, 327–339.
Yoshimura, K., Ito, K., and Nakamura, Y. (1999). Amber (UAG) suppressors affected in UGA/UAA-specific polypeptide release factor 2
542 Molecular Cell 28, 533–543, November 30, 2007 ª2007 Elsevier Inc.
Molecular Cell
A Role for the Decoding Center in Peptide Release
of bacteria: genetic prediction of initial binding to ribosome preceding
stop codon recognition. Genes Cells 4, 253–266.
Youngman, E.M., and Green, R. (2005). Affinity purification of in vivoassembled ribosomes for in vitro biochemical analysis. Methods 36,
305–312.
Youngman, E.M., Brunelle, J.L., Kochaniak, A.B., and Green, R. (2004).
The active site of the ribosome is composed of two layers of conserved
nucleotides with distinct roles in peptide bond formation and peptide
release. Cell 117, 589–599.
Youngman, E.M., Cochella, L., Brunelle, J.L., He, S., and Green, R.
(2006). Two distinct conformations of the conserved RNA-rich decoding center of the small ribosomal subunit are recognized by tRNAs and
release factors. Cold Spring Harb. Symp. Quant. Biol. 71, 545–549.
Zoldak, G., Redecke, L., Svergun, D.I., Konarev, P.V., Voertler, C.S.,
Dobbek, H., Sedlak, E., and Sprinzl, M. (2007). Release factors 2
from Escherichia coli and Thermus thermophilus: structural, spectroscopic and microcalorimetric studies. Nucleic Acids Res. 35, 1343–
1353.
Molecular Cell 28, 533–543, November 30, 2007 ª2007 Elsevier Inc. 543