THE JOURNAL OF BIOLOGICAL CHEMISTRY VOL. 280, NO. 43, pp. 36065–36072, October 28, 2005
© 2005 by The American Society for Biochemistry and Molecular Biology, Inc. Printed in the U.S.A.
Essential Mechanisms in the Catalysis of Peptide Bond
Formation on the Ribosome*
Received for publication, July 21, 2005, and in revised form, August 22, 2005 Published, JBC Papers in Press, August 29, 2005, DOI 10.1074/jbc.M507961200
Malte Beringer‡1, Christian Bruell§1, Liqun Xiong¶, Peter Pfister§, Peter Bieling‡2, Vladimir I. Katunin!,
Alexander S. Mankin¶, Erik C. Böttger§, and Marina V. Rodnina‡3
From the ‡Institute of Physical Biochemistry, University of Witten/Herdecke, Stockumer Strasse 10, 58448 Witten, Germany, the
§
Institute of Medical Microbiology, University of Zurich, Gloriastrasse 30/32, 8006 Zurich, Switzerland, the ¶Center for
Pharmaceutical Biotechnology, University of Illinois, Chicago, Illinois 60607, and the !Petersburg Nuclear Physics Institute,
Russian Academy of Sciences, 188350 Gatchina, Russia
Peptide bond formation is the main catalytic function of the ribosome. The mechanism of catalysis is presumed to be highly conserved in all organisms. We tested the conservation by comparing
mechanistic features of the peptidyl transfer reaction on ribosomes
from Escherichia coli and the Gram-positive bacterium Mycobacterium smegmatis. In both cases, the major contribution to catalysis
was the lowering of the activation entropy. The rate of peptide bond
formation was pH independent with the natural substrate, aminoacyl-tRNA, but was slowed down 200-fold with decreasing pH when
puromycin was used as a substrate analog. Mutation of the conserved base A2451 of 23 S rRNA to U did not abolish the pH dependence of the reaction with puromycin in M. smegmatis, suggesting
that A2451 did not confer the pH dependence. However, the
A2451U mutation alters the structure of the peptidyl transferase
center and changes the pattern of pH-dependent rearrangements,
as probed by chemical modification of 23 S rRNA. A2451 seems to
function as a pivot point in ordering the structure of the peptidyl
transferase center rather than taking part in chemical catalysis.
Ribosomes catalyze peptide bond formation between aminoacyltRNA (aa-tRNA)4 bound to the A site of the ribosome and peptidyltRNA at the P site. The active site for peptide bond formation, the
peptidyl transferase center, is located on the large (50 S) ribosomal
subunit. High-resolution crystal structures of the 50 S subunit have
revealed that the peptidyl transferase center is composed of RNA (23 S
rRNA), with no protein within 15 Å of the active site (1, 2). This implies
that peptide bond formation is catalyzed by RNA and, thus, the ribosome is a ribozyme.
The peptide bond is formed as a result of nucleophilic attack by the
!-amino group of aa-tRNA on the ester carbonyl group of peptidyl-
* The work was supported by grants from the Deutsche Forschungsgemeinschaft (to
M. V. R.), European Union (to M. V. R.), Alfried Krupp von Bohlen und Halbach-Stiftung
(to M. V. R.), Fonds der Chemischen Industrie (to M. V. R.), Swiss National Science
Foundation (to E. C. B.), Bonizzi-Theler Stiftung (to E. C. B.), Russian Foundation for
Basic Research (to V. I. K.), and National Institutes of Health Grant RO1 GM59028 (to
A. S. M.). The costs of publication of this article were defrayed in part by the payment
of page charges. This article must therefore be hereby marked “advertisement” in
accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
1
Both authors contributed equally to this work.
2
Present address: EMBL Heidelberg, Meyerhofstrasse 1, 69117 Heidelberg, Germany.
3
To whom correspondence should be addressed: Inst. of Physical Biochemistry, University of Witten/Herdecke, Stockumer Str. 10, 58448 Witten, Germany. Tel.: 49-2302926205; Fax: 49-2302-926117; E-mail: [email protected].
4
The abbreviations used are: aa-tRNA, aminoacyl-tRNA; DMS, dimethyl sulfate;
CMCT, 1-cyclohexyl-3-(2-morpholinoethyl)-carbodiimide metho-p-toluenesulfonate; Me2SO, dimethyl sulfoxide; EF-Tu, elongation factor Tu; HPLC, high pressure liquid chromatography; Bicine, N,N-bis(2-hydroxyethyl)glycine; BisTris,
2-[bis(2-hydroxyethyl)amino]-2-(hydroxymethyl)propane-1,3-diol; Pipes, 1,4piperazinediethanesulfonic acid; WT, wild type.
OCTOBER 28, 2005 • VOLUME 280 • NUMBER 43
tRNA. The first step is the deprotonation of the !-NH3! group to create
the nucleophilic NH2 group. The pKa of the !-NH3! group in aa-tRNA
is estimated to be around 8, and it is likely that the proton is accepted by
water (3). Subsequent nucleophilic attack of the !-NH2 group on the
electrophilic carbonyl group leads to the formation of the zwitterionic
tetrahedral intermediate, which, by deprotonation, forms the negatively
charged tetrahedral intermediate. The breakdown of the tetrahedral
intermediate is initiated by donating a proton back to the leaving oxygen
to form the products, i.e. P-site deacylated tRNA and A-site
peptidyl-tRNA.
In principle, the ribosome may catalyze the reaction by several mechanisms, such as proper positioning of the peptidyl and aminoacyl ends of
the tRNAs in the active site in a conformation suitable for the spontaneous reaction, general base-acid catalysis during deprotonation and
protonation, or electrostatic stabilization of the transition state(s)
(4 –9). The pH dependence of peptidyl transfer reaction catalyzed by
Escherichia coli ribosomes suggested the importance of an ionizing
group of the ribosome with a pKa of 7.5, which could contribute to
chemical catalysis (10). Based on the crystal structure of the 50 S subunit
from Haloarcula marismortui, a model was proposed suggesting that
one of the RNA bases located within the peptidyl transferase center
takes part in the chemical mechanism of peptide bond formation (2). In
this model, A2451 was proposed to function as a general base that
abstracts a proton from the attacking !-amino group and/or stabilizes
the oxyanion of the tetrahedral intermediate, thereby accelerating the
intermediate formation. As general base catalysis in an aqueous
medium of neutral pH requires that the base is ionizing at that pH, it was
postulated that the pKa of A2451 is shifted toward neutrality by a network of interactions within the peptidyl transferase center (2). Another
likely role of A2451 is to correctly position the substrate in the active
center by forming a hydrogen bond between the attacking !-NH2 group
and the 2" OH group of A2451 (11, 12).
A number of studies were performed to examine these models. Crystal structures of 50 S subunits with different substrate analogs (11) and
biochemical experiments (13) provided evidence against stabilization of
the transition state oxyanion by a pKa-perturbed RNA base in the peptidyl transferase center but did not exclude other potential roles of
A2451 in catalysis. In E. coli, any mutation of A2451 is lethal (14, 15),
which supports an important functional role of the base. Nevertheless,
ribosomes carrying mutations at A2451 exhibited considerable catalytic
activity in vitro (14, 16). Rapid kinetic analysis has shown that although
the major part of the catalysis is contributed by substrate positioning
(10, 17), the A2451U mutation reduced the rate of puromycin reaction
by about 150-fold (10, 18) and eliminated the inhibition of the reaction
by ionization of a ribosomal group with a pKa of 7.5 (10). These effects
could be explained either by A2451 acting as a general acid/base or by
JOURNAL OF BIOLOGICAL CHEMISTRY
36065
Peptidyl Transferase of M. smegmatis Ribosomes
conformational changes in the peptidyl transferase center caused by the
A2451U mutation. Further biochemical and genetic work has shown
that mutations of G2447, a residue implied in shifting the pKa of A2451
to neutrality, had little effect in vivo and in vitro (14, 16, 19). This suggested that A2451 does not participate in catalysis by the mechanism
suggested (2), but it could not entirely exclude a function of A2451 as a
general acid/base, in particularly in conjunction with other charged
groups (20, 21).
It is reasonable to assume that the mechanism of peptide bond formation is essentially the same on ribosomes from all organisms,
although most of the biochemical data available so far have been
obtained using E. coli ribosomes and little information on the mechanism of reaction in other organisms is available. The Gram-positive
bacterium Mycobacterium smegmatis has significant advantages over
E. coli as a model system for genetic experiments, particularly for introducing mutations in rRNA. Several rRNA mutations that are lethal in
E. coli have viable phenotypes in M. smegmatis (22–24),5 which allows
for the preparation of homogeneous mutant ribosomes without the use
of tag systems (25). These advantages prompted us to use M. smegmatis
ribosomes to analyze the mechanism of peptide bond formation, identify evolutionary conserved features of the mechanism, and further
assess the functional role of A2451. We show that the major conserved
contribution to catalysis comes from the lowering of the activation
entropy of peptide bond formation. The effects of A2451U are different
in E. coli and M. smegmatis, which argues against a mechanism for peptidyl transfer in which A2451 acts as a conserved chemical catalyst;
rather, A2451 has an important function in determining the structure of
the peptidyl transferase center.
MATERIALS AND METHODS
Chemicals and Buffers—Chemicals for ribosome preparation and
activity assays were purchased from Merck (Darmstadt, Germany) or
Sigma (Steinheim, Germany) unless stated otherwise. Chemicals for
RNA probing were from Fisher. Dimethyl sulfate (DMS) and 1-cyclohexyl-3-(2-morpholinoethyl)-carbodiimide metho-p-toluenesulfonate
(CMCT) were from Sigma. MFT-mRNA (5"-GGCAAGGAGGUAAAUAAUGUUCACGAUU-3", underlined sequence coding for fMet-PheThr) was purchased from Dharmacon Research, Inc. (Boulder, CO).
Radioactive amino acids were from MP Biomedicals (Eschwege, Germany) or Hartmann Analytic (Braunschweig, Germany). Radioactive
nucleoside triphosphates used for primer extension were from ICN
(Costa Mesa, CA). Buffer A consisted of: 50 mM Tris-HCl, pH 7.5, 70
mM NH4Cl, 30 mM KCl, 7 mM MgCl2. Buffer B consisted of: 50 mM
Tris-HCl, 20 mM BisTris-HCl (at indicated pH), 70 mM NH4Cl, 30 mM
KCl, 7 mM MgCl2.
M. smegmatis Bacterial Strains and Mutagenesis—The parental
strain used was a single rDNA allelic strain of M. smegmatis, mc2155
SMR5#rrnB, with one remaining functional operon, rrnA, whereas the
rrnB operon was removed completely. No additional resistance determinants were introduced in the genome (23). The integrative vector
pMV361#Kan-Gm, which integrates into the genome of M. smegmatis
at the unique attB site, was constructed by deletion of the aph cassette of
pMV361 with NheI/SpeI and insertion of the gentamicin resistance
cassette of pMV361-Gm carried on a 3-kb PstI fragment. An $1.4-kbp
rRNA gene fragment starting at 23 S RNA position 1929 (E. coli numbering) up to a position 82 bp downstream of the 5 S RNA gene and
containing point mutation A2451U was generated by fusion PCR and
verified by sequencing. The PCR fragments were subcloned into the
5
C. Bruell, P. Pfister, and E. C. Böttger, unpublished observations.
36066 JOURNAL OF BIOLOGICAL CHEMISTRY
pGEMTEasy vector according to the manufacturer’s instructions (Promega, Madison, WI). Subsequently, the PCR fragments were isolated by
EcoRI digestion and integrated into the single EcoRI site of pMV361#KanGm. The resulting plasmid was named pMV361#Kan-Gm-rRNA2451U.
The plasmid pMV361#Kan-Gm-rRNA2451U was transformed into
M. smegmatis mc2155 SMR5#rrnB. Transformants were selected on LB
plates containing 5 "g/ml gentamicin. RecA-mediated homologous
recombination was used to introduce the point mutation A2451U into
the functional rrn operon (rrnA) as described previously (26). Recombinants were colony purified twice; the presence of the mutation was
verified by sequencing the 23 S rRNA gene as well as the 23 S rRNA in
the ribosome preparations. M. smegmatis mc2155 SMR5#rrnB and
A2451U were grown on LB agar plates and plates containing linezolid
(16 "g/ml), respectively, preventing the loss of the point mutation. For
cultivation of bacteria, large flasks with a total volume of 20 liters of LB
medium containing Tween 80 (0.05%) were inoculated with colonies
collected from plates. After 6 –7 generations and an A600 of $0.7, the
cells were harvested and shock-frozen as dry pellets.
Ribosome Preparation—Twenty grams of frozen cells were resuspended in 50 ml of 20 mM Tris-HCl, pH 7.5, 100 mM NH4Cl, 10 mM
MgCl2, 0.5 mM EDTA, and 3 mM 2-mercaptoethanol. Cells were opened
by passing them through a French press (Thermo Spectronic, Rochester, NY) at 18,000 p.s.i. DNase (RNase-free; Sigma; 3 "g/ml) and precooled alumina (20 g) were added to the cell lysate. Alumina and cell
debris were removed by low-speed centrifugation (JA14 rotor; Beckman
Coulter Inc., Fullerton, CA; 9,000 % g, 30 min), and the S-30 fraction was
prepared by centrifugation at 16,000 rpm in a Beckman Ti50.2 rotor for
60 min. 16-ml portions of the S-30 extract were layered on 9-ml sucrose
cushions (1.1 M sucrose in 20 mM Tris-HCl, pH 7.5, 500 mM NH4Cl, 10
mM MgCl2, 0.5 mM EDTA, 3 mM 2-mercaptoethanol). After centrifugation at 33,000 % g in the Ti50.2 rotor for 16 h, ribosome pellets were
washed with 20 mM Tris-HCl, pH 7.5, 350 mM NH4Cl, 10 mM MgCl2, 0.5
mM EDTA, and 3 mM 2-mercaptoethanol, dissolved in the same buffer,
and incubated for 30 min at 4 °C. Ribosomes were pelleted through 1 ml
of sucrose cushion as before (Ti 50.2, 6 h, 50,000 rpm), and washed and
dissolved in the same buffer. Finally, the ribosomes were centrifuged
through 1.5-ml sucrose cushions (Beckman SW28; 28,000 rpm, 13 h).
Ribosome pellets were dissolved in buffer A and stored after shockfreezing in liquid nitrogen. The concentration was determined by
absorption measurements on the basis of 23 pmol/A260 units.
The integrity of 70 S ribosomes was verified by analytical ultracentrifugation. 50 pmol each of 70 S ribosomes from M. smegmatis, E. coli
70 S ribosomes, or E. coli 30 S and 50 S subunits were layered on a
5– 40% sucrose gradient in buffer A (association conditions) or in 50 mM
Tris-HCl, 20 mM BisTris-HCl at pH 7.5, 30 mM KCl, 70 mM NH4Cl, 1
mM MgCl2 (dissociation conditions). The gradients were centrifuged in
a Beckman SW41 rotor at 22,000 % g for 17 h. Gradients were pumped
out from the bottom of the tubes and ribosome fractions were detected
by absorption at 254 nm.
Ribosome Activity and Complex Formation—Purified tRNA and
translation factors from E. coli were prepared as described (27, 28).
Initiation complexes were prepared in buffer A by incubating 70 S ribosomes (1.5 "M), MFT-mRNA (1.7-fold excess over ribosomes), and
[3H]fMet-tRNAfMet (1.7-fold excess over ribosomes) in the presence of
initiation factors 1, 2, and 3 (each in a 1.7-fold excess over ribosomes) for
60 min at 37 °C. To form the ternary complex, EF-Tu!GTP![14C]PhetRNAPhe, EF-Tu (2.4 "M) was incubated with GTP (1 mM), phosphoenolpyruvate (3 mM), and pyruvate kinase (0.1 "g/ml) for 15 min at
37 °C followed by the addition of [14C]Phe-tRNAPhe (1.2 "M). Equal
volumes of initiation and ternary complexes were mixed at 14 mM
VOLUME 280 • NUMBER 43 • OCTOBER 28, 2005
Peptidyl Transferase of M. smegmatis Ribosomes
MgCl2, resulting in pretranslocation complexes. The extent of binding
of [3H]fMet-tRNAfMet to the P site and [14C]Phe-tRNAPhe to the A site
was controlled by nitrocellulose filtration. Filters were dissolved in 10 ml
of scintillation liquid (Quickszint 361; Zinsser Analytic, Frankfurt, Germany) and counted in a TriCarb counter (PerkinElmer Life Sciences).
Post-translocation complexes were prepared by adding 0.1 "M EF-G to
pretranslocation complexes and incubating for 5 min at 20 °C. Posttranslocation complexes were purified and concentrated by centrifugation through 1.1 M sucrose cushions in buffer A at 259,000 % g for 2 h
(M120GX rotor; Sorvall, Asheville, NC). Pellets were dissolved in buffer
B to a concentration of $3 "M, shock-frozen in liquid nitrogen, and
stored at &80 °C.
Kinetic Measurements of Peptide Bond Formation—For the measurement of tripeptide formation, stock solutions of post-translocation
complexes and puromycin were diluted into buffer B; for studying pH
dependence, the buffer was adjusted to various pH values ranging from
6 to 9 (37 °C). Puromycin solutions at concentrations of 15–30 mM (see
Fig. 1) contained 5–10% Me2SO; Me2SO up to 20% had no effect on the
rate of puromycin reaction (data not shown). The final pH of both
ribosome and puromycin solutions was measured directly using a smallsized pH electrode. Quench-flow experiments were carried out at 37 °C
by mixing equal volumes (13 "l) of ribosome and puromycin solutions
in a quench-flow apparatus (KinTek Corp, Austin, TX). After mixing,
the concentration of the post-translocation complex was 0.15 "M. Reactions were quenched with 0.5 M KOH. The [3H]fMet[14C]Phe-puromycin formed in the reaction and unreacted [3H]fMet[14C]Phe, set free
from [3H]fMet[14C]Phe-tRNAPhe by alkaline hydrolysis (30 min, 37 °C),
were separated by reverse phase-HPLC on a RP-8 column and quantified by double label radioactivity counting (10). Alternatively, reactions
were quenched with 25% formic acid, and [3H]fMet[14C]Phe-puromycin was extracted in 750 "l of ethyl acetate in the presence of 500 "l of
1.5 M sodium acetate saturated with MgSO4 at pH 4.5. Double label
radioactivity counting was used to analyze 500 "l of the organic phase.
The rate of dipeptide formation was measured with 0.1 "M initiation
complex and 0.6 "M ternary complex, both in buffer B of the desired pH,
at 37 °C. The reaction products were released from [3H]fMet-PhetRNAPhe by alkaline hydrolysis, separated by reverse phase-HPLC, and
quantified by double label radioactivity counting. Time courses of the
reaction were evaluated by single exponential fitting.
Chemical Probing—RNA probing was carried out generally following
the standard procedure (29, 30) with some modifications. Five "l of
vacant ribosomes, isolated as described earlier and stored at 1 "M concentration in a buffer of 50 mM HEPES (adjusted with KOH to pH 6.5,
7.5, or 8.5), 70 mM NH4Cl, 30 mM KCl, and 7 mM MgCl2 were combined
with 45 "l of a modification buffer with a pH corresponding to that at
which the ribosomes were stored. The following buffers were used for
ribosome dilution and probing: 50 mM PIPES-KOH, pH 6.5, 150 mM
KCl, 7 mM MgCl2; 50 mM HEPES-KOH, pH 7.5, 150 mM KCl, 7 mM
MgCl2; 50 mM Bicine-KOH, pH 8.5, 150 mM KCl, 7 mM MgCl2. After
dilution with modification buffer, ribosome solutions were incubated
for 5 min at 37 °C and stored on ice before use.
For DMS modification, 2 "l of DMS:ethanol mixture (1:5, v/v) was
added (2 "l of ethanol was added to the control samples) and samples
were incubated for 1 h on ice. The reactions were quenched by the
addition of 50 "l of 0.6 M sodium acetate, pH 5.5, 1 M 2-mercaptoethanol, and ribosomes were precipitated by the addition of 300 "l of cold
ethanol and incubation for 10 min at &70 °C. Ribosomes were pelleted
by centrifugation at 20,000 % g for 10 min at 2 °C.
For CMCT modification, 5 "l of 1 "M ribosome solution in storage
buffer, pH 8.5, were combined with 45 "l of Bicine modification buffer,
OCTOBER 28, 2005 • VOLUME 280 • NUMBER 43
pH 8.5 (see earlier). After 5 min of incubation at 37 °C, 50 "l of CMCT
solution (10.5 mg/ml) in Bicine modification buffer, pH 8.5, was added
and samples were incubated for 1 h on ice. Reactions were stopped by
adding 100 "l of 0.6 M sodium acetate, pH 5.5, and 600 "l of cold ethanol
and placing samples at &70 °C for 10 min. Ribosomes were precipitated
by centrifugation at 20,000 % g for 10 min at 2 °C.
Ribosome pellets were dissolved in 200 "l of 0.3 M sodium acetate, pH
5.5, 5 mM EDTA, 0.5% SDS. rRNA was isolated by successive extractions
with equal volumes of phenol, phenol/chloroform, and chloroform, followed by ethanol precipitation of the aqueous phase. RNA pellets were
dissolved in 50 "l of H2O. The distribution of modified nucleotides in 23
S rRNA was assessed by primer extension analysis essentially as
described (29) using the DNA primers MS2720 (GTCCTCTCGTACTAGGGACAG), MS2550 (CCAACCCTTGGGACCTGCT), and
MS2390 (CAGAGTGGTATTTCAACAACGAC), which allowed for
scanning of the entire domain V of M. smegmatis 23 S rRNA.
RESULTS
70 S Ribosomes from M. smegmatis—The 70 S ribosomes were prepared from the M. smegmatis mc2155 SMR5#rrnB strain, which has a
single rRNA operon, rrnA (25). The A2451U mutation was introduced
into the 23 S rRNA gene by RecA-mediated homologous recombination. The resulting cells, which contained a pure population of ribosomes carrying the A2451U mutation, were viable in liquid culture and
formed colonies on agar plates. However, introduction of the A2451U
mutation had a major effect on the cell growth: compared with the
wild-type (WT), the generation time increased from 3– 4 to 15–17 h.
Notably, the A2451U mutation conferred resistance to chloramphenicol and linezolid (to be discussed in detail elsewhere). Wild-type and
mutant M. smegmatis ribosomes were prepared as described earlier for
E. coli ribosomes (27) with some alterations (see “Materials and Methods”). Whereas most of the ribosomal material prepared from WT
M. smegmatis was represented by 70 S particles, 60% of the mutant
ribosome preparation was represented by a 70 S peak and the rest sedimented as free 30 and 50 S ribosomal subunits during gradient centrifugation. The 70 S ribosomes prepared from mc2155 SMR5#rrnB or
mc2155 SMR5#rrnB A2451U cells were 50% active in forming the initiation complex, 70S-mRNA-fMet-tRNAfMet. EF-Tu-dependent binding of Phe-tRNAPhe to the A site and translocation by EF-G resulted in
post-translocation complexes with 40% of peptidyl-tRNA, fMet-PhetRNAPhe, in the P site both for WT and mutant ribosomes.
Peptide Bond Formation on M. smegmatis Ribosomes with Puromycin
as Substrate—The peptidyl transferase activity of WT and A2451U
M. smegmatis ribosomes was tested in a rapid kinetic assay (10) (Fig.
1A). Post-translocation complexes loaded with fMet-Phe-tRNAPhe in
the P site were mixed with puromycin and time courses of fMet-Phepuromycin formation were monitored. Because puromycin binds very
rapidly to the A site, the chemistry step appears to be rate-limiting (10,
17). In the presence of 10 mM puromycin at pH 7.5, the rate of fMetPhe-puromycin formation was 2.3 s&1 on WT ribosomes (Fig. 1B), and
the A2451U mutation reduced the rate $15-fold to 0.15 s&1. This rate
effect could be because of a decreased rate of the chemistry step or to
impaired puromycin binding. To distinguish between these possibilities, the rate of fMet-Phe-puromycin formation was determined at
increasing puromycin concentrations (Fig. 1C). The apparent affinity
for puromycin, Km ' 8 ( 3 mM, was identical for the WT and mutant M.
smegmatis post-translocation complexes and similar to that observed
for E. coli ribosomes, 3– 4 mM (10). The rates of peptide bond formation
at substrate saturation were 4 and 0.25 s&1 for the M. smegmatis WT
and A2451U mutant ribosomes, respectively. The rates on E. coli ribo-
JOURNAL OF BIOLOGICAL CHEMISTRY
36067
Peptidyl Transferase of M. smegmatis Ribosomes
FIGURE 1. Kinetics of peptide bond formation on ribosomes from M. smegmatis with puromycin as substrate. A, preparation of ribosome complexes and rapid kinetic assay of
peptidyl transferase activity. B, time courses of peptide bond formation with WT (closed circles) and A2451U (open circles) M. smegmatis ribosomes at 10 mM puromycin and pH 7.5;
Pmn, puromycin. C, dependence of kpep on puromycin concentration measured with WT (closed circles, left-hand ordinate) and A2451U ribosomes (open circles, right-hand ordinate).
Fitting the data to a two-step model in which a rapid binding equilibrium is followed by irreversible peptide bond formation yields kpep ' 4 s&1 (WT) and kpep ' 0.25 s&1 (A2451U)
at puromycin saturation and a Km of 8 ( 3 mM (WT and A2451U).
TABLE ONE
Activation parameters for the second-order ribosome-catalyzed peptide bond formation and uncatalyzed ester aminolysis (knon) (25 °C)
Standard deviation of rate constants is 10%, that of activation parameters about (1 kcal/mol.
Ribosomes from
Amine
Ester
Rate constant
M
a
Uncatalyzed
E. colia
M. smegmatis
Data from Ref. 17.
Tris
Puromycin
Puromycin
fMet-tRNAfMet
fMet-Phe-tRNAPhe
fMet-Phe-tRNAPhe
somes were 30 s&1 (on WT ribosomes) and 0.2 s&1 (on A2451U mutant
ribosomes) under the same conditions (pH 7.5, 37 °C) (10).
To obtain further insight into the mechanism of catalysis, activation
parameters of peptide bond formation on M. smegmatis ribosomes were
measured and compared with those of an uncatalyzed model reaction
36068 JOURNAL OF BIOLOGICAL CHEMISTRY
!G ‡
&1 &1
T!S ‡
kcal/mol
s
1 % 10&4
4 % 102
4 % 102
Activation parameters
!H ‡
22.7
14.0
14.0
16.2
16.0
14.5
&6.5
2.0
0.5
using Tris as the reacting nucleophile. The rate of peptide bond formation with fMet-Phe-tRNAPhe and puromycin was measured at different
temperatures; activation parameters (#G ‡, #H ‡, and T#S ‡) were calculated as described (17) (TABLE ONE). The activation parameters of
fMet-Phe-puromycin formation were very similar on M. smegmatis and
VOLUME 280 • NUMBER 43 • OCTOBER 28, 2005
Peptidyl Transferase of M. smegmatis Ribosomes
E. coli ribosomes. The results indicate that the rate enhancement produced by the ribosome is achieved predominantly by lowering the activation entropy, whereas the activation enthalpy is almost the same on
the ribosome and in solution.
pH Dependence of Peptide Bond Formation with Puromycin—Peptidyl transfer reaction on E. coli ribosomes exhibited a strong pH dependence that results from the ionization of both the nucleophilic NH2 group
of puromycin (pKa ' 6.9) and a ribosomal group with a pKa of 7.5 (10,
19). On M. smegmatis ribosomes, the rates of peptide bond formation
increased with pH in a similar fashion to the rates on E. coli ribosomes
(Fig. 2). A maximum rate of 22 ( 2 s&1 was observed with WT ribosomes, which is comparable with the rate of 50 s&1 reported for E. coli
(10, 19) (TABLE TWO). As with E. coli, the kinetic data obtained with
M. smegmatis ribosomes fit a model with 2 ionizing groups involved in
the reaction; in addition to the pKa of puromycin, an ionizing group with
a pKa of 8.0 ( 0.2 was found, which is slightly higher than the pKa of the
ionizing group in E. coli ribosomes (10, 19). Ionization of the group with
pKa of 8.0 ( 0.2 leads to an $200-fold decrease of the reaction rate.
If A2451 were an evolutionarily conserved general acid/base involved
in the peptide bond formation or if A2451 were the base with a pKa of 8
involved in a pH-dependent conformational change, mutations of
A2451 in ribosomes of different species would be expected to change, or
eliminate, the pH dependence of the reaction in a similar manner. However, in contrast to E. coli, the shape of the pH dependence was not
changed by the A2451U mutation in M. smegmatis (Fig. 2). Again, 2
ionizing groups were found, with pKa values of 6.9 (puromycin) and
8.2 ( 0.3, whereas the maximum rate of the reaction was decreased to
2.6 ( 0.4 s&1 (TABLE TWO). Thus, although the A2451U mutation in
M. smegmatis moderately decreased the rate of peptide bond formation
FIGURE 2. Effect of the A2451U mutation on the pH dependence of peptide bond
formation. The rate of fMet-Phe-puromycin formation was measured at 20 mM puromycin on WT (closed circles) and A2451U (open circles) M. smegmatis ribosomes. A model
with 2 ionizing groups influencing the reaction was used for fitting the data, with pKa1 '
6.9 for puromycin (10) (fixed parameter in fitting) and fitting the pKa value of the second
ionizing group; the values for pKa1, pKa2, and kpep at pH 6.5 and 8.5 are given in TABLE
TWO.
(about 10-fold), it did not affect the ionization of group(s) presumably
involved in catalysis.
Peptide Bond Formation with aa-tRNA as Substrate—Peptidyl transfer reaction with aa-tRNA as the substrate was measured on M. smegmatis ribosomes by mixing 70 S ribosomes loaded with mRNA and
fMet-tRNAfMet with the ternary complex, EF-Tu!GTP!Phe-tRNAPhe.
The observed rate of the reaction catalyzed by WT ribosomes, 3 s&1, was
comparable with the respective rate on E. coli ribosomes observed previously, 7 s&1 (31), and was independent of pH (Fig. 3), as also observed
with E. coli ribosomes.6 Introducing the A2451U mutation in M. smegmatis ribosomes had no discernible effect on the rate of peptide bond
formation with aa-tRNA, and the reaction was pH-independent as with
WT ribosomes, suggesting very little or no contribution of A2451 to the
catalysis of peptide bond formation with aa-tRNA within the neutral pH
range.
pH-dependent Conformational Changes of the Peptidyl Transferase
Center—The conformation of 23 S rRNA in the peptidyl transferase
center was probed using DMS and CMCT modification. To monitor
pH-dependent changes in RNA conformation, DMS modification reactions were carried out at 3 pH values. Because RNA modification with
carbodiimide is drastically inhibited at lower pH, CMCT probing was
done only at pH 8.5.
Comparison of the DMS and CMCT modification patterns of WT
and mutant 23 S rRNA at pH 8.5 shows that the A2451U mutation
causes significant alteration in conformation of rRNA in the peptidyl
transferase center. The most prominent differences were observed at
positions U2506 and U2585, which are much more reactive to CMCT in
WT than in mutant, and positions A2060 and A2572, which are more
accessible to DMS modification in mutant than in WT. In addition,
positions U2506 and U2584 showed somewhat enhanced reactivity to
DMS in WT (Fig. 4, TABLE THREE).
Several groups at the active site of WT ribosomes showed pH-dependent changes of reactivity (Fig. 4). The extent of DMS modification
of U2584 and U2506 increased between pH 7.5 and pH 8.5 (TABLE
THREE), suggesting an apparent pKa close to the kinetic pKa. However,
the increase in reactivity of C2474 occurred between pH 6.5 and 7.5,
which is not reflected in the pH dependence of the peptidyl transferase
rate. Thus, conformation of several nucleotides in the active center may
exhibit pH dependence, only one of which affected the rate of peptide
bond formation.
The pH-dependent changes seemed to be different on WT and
mutant ribosomes. For example, accessibility of A2572 in mutant ribosomes is drastically increased between pH 6.5 and 7.5 but is pH independent in the WT (Fig. 4B, TABLE THREE). Thus, whereas the pHdependent rearrangements were not abolished by the A2451U
mutation, the general structure of the peptidyl transferase center and
structural transitions in rRNA were altered.
6
S. Adio and M. V. Rodnina, unpublished results.
TABLE TWO
pH dependence of peptide bond formation on ribosomes from M. smegmatis and E. coli with puromycin as A-site substrate
Data for E. coli ribosomes were taken from Ref. 10. pKa1 reflects ionization of the !-amino group of puromycin (10); pKa2 reflects apparent pKa of a ribosome
group affecting the peptidyl transfer reaction; kpep (pH 8.5) is the maximum rate of reaction.
Ribosomes
pKa1
pKa2
kpep(pH 8.5)
kpep(pH 6.5)
s
a
M. smegmatis (WT)
M. smegmatis (A2451U)
E. coli (WT)
E. coli (A2451U)
The value was extrapolated from the data of Fig. 2.
6.9 ( 0.2
6.9 ( 0.2
6.9 ( 0.2
6.9 ( 0.2
OCTOBER 28, 2005 • VOLUME 280 • NUMBER 43
8.0 ( 0.2
8.2 ( 0.3
7.5 ( 0.1
0.25 ( 0.02
0.02 ( 0.002a
1.8 ( 0.2
0.06 ( 0.02
&1
22 ( 2
2.6 ( 0.4
50 ( 10
0.3 ( 0.05
JOURNAL OF BIOLOGICAL CHEMISTRY
36069
Peptidyl Transferase of M. smegmatis Ribosomes
DISCUSSION
Conserved Features of the Catalytic Mechanism—The ribosome is an
ancient RNA-based enzyme whose function is to synthesize polypeptides. Despite a high degree of sequence conservation of rRNA, in par-
FIGURE 3. pH dependence of fMet-Phe formation on M. smegmatis ribosomes. Time
courses of dipeptide formation between fMet-tRNAfMet and Phe-tRNAPhe at pH 6.75
(closed circles) and 7.75 (closed triangles) on WT ribosomes and at pH 7.75 on A2451U
(open circles) ribosomes. Dipeptide formation on A2451U ribosomes at pH 6.75 was as
fast as at pH 7.75 (not shown for clarity).
ticular at the peptidyl transferase center (1, 32, 33), the positioning of
groups in the peptidyl transferase active site may differ in detail among
species (1, 34). Interaction of the conserved nucleotides with the less
conserved RNA residues may influence the orientation of the rRNA
bases in the active site of the peptidyl transferase. For instance, a somewhat different neighborhood of the reactive !-amino group of puromycin attached to a short RNA hairpin was reported for the Deinococcus
radiodurans 50 S subunit (35) as compared with a similar substrate
analog on H. marismortui 50 S (11). This can also explain species-specific spectra of mutations conferring antibiotic resistance (24).
The present data provide the first insight into the degree of mechanistic conservation of peptide bond formation among different species.
The reaction rate was slightly lower in the M. smegmatis system as
compared with E. coli, which is compatible with the slower growth rate
of M. smegmatis. On both E. coli and M. smegmatis ribosomes, the catalysis of peptide bond formation was mainly because of a large favorable
contribution to the activation entropy as compared with an uncatalyzed
model reaction in solution. These results further support the view that
the ribosome enhances the rate of peptide bond formation predominantly by positioning the substrates and/or water exclusion within the
FIGURE 4. Chemical probing of 23 S rRNA in the peptidyl transferase center of M. smegmatis ribosomes. C and A, sequencing lanes; K, control without modification; M, after
chemical modification. WT, WT M. smegmatis ribosomes; MUT, A2451U mutant. 6.5, 7.5, 8.5, pH during modification. A, DMS modification, primer MS2390. B, DMS modification, primer
MS2720. C, DMS modification, primer MS2550. D, CMCT modification, primer MS2720.
TABLE THREE
pH-dependent changes in the chemical reactivity of rRNA bases in the peptidyl transferase center of ribosomes from M. smegmatis
a
Position
Reagent
A2060
C2474
A2572
U2506
U2506
U2584
U2585
DMS
DMS
DMS
DMS
CMCT
DMS
CMCT
ND, not determined.
36070 JOURNAL OF BIOLOGICAL CHEMISTRY
6.5
&
!
!
&
NDa
&
ND
WT
7.5
&
!!
!
&
ND
&
ND
8.5
6.5
A2451U
7.5
8.5
&
!!
!
!
!
!
!!!
!
!
!
&
ND
&
ND
!(
!!
!!
&
ND
&
ND
!(
!!
!!
(
(
(
!
VOLUME 280 • NUMBER 43 • OCTOBER 28, 2005
Peptidyl Transferase of M. smegmatis Ribosomes
active site rather than by chemical catalysis (9, 14, 16, 17, 19). The factor
contributed by pH-dependent effects (150 –200-fold for puromycin as
the substrate) was similar with ribosomes from both organisms. However, pKa values determined from reaction rates differed slightly, suggesting that the respective ionizing groups either are not conserved or
are located in different environments. The latter possibility is consistent
with the results of chemical probing that suggested that the structure
and pH-dependent changes of the peptidyl transferase center are different in E. coli and M. smegmatis (see below).
When aminoacyl-tRNA, rather than puromycin, was used as the
A-site substrate, the rate of peptide bond formation on M. smegmatis
ribosomes was independent of pH. The rate of the reaction on E. coli
ribosomes was higher but was also independent of pH over 2 pH units.6
At high pH, the reaction rate was limited by the accommodation of
aa-tRNA in the A site, precluding rate measurements of the chemistry
step with aa-tRNA substrates (36). Notably, mutations of a number of
residues at the peptidyl transferase center of E. coli ribosomes impaired
the reaction with puromycin but had no effect when aa-tRNA was used
(18). These observations may indicate that groups ionizing at neutral pH
are not critical for the reaction with the natural substrate.
Structural Rearrangements in the Active Site—The structure of the
peptidyl transferase center has long been known to be sensitive to pH
and monovalent cation concentrations (37). Chemical probing experiments with E. coli ribosomes showed that the DMS reactivities of several bases, including A2451, A2453, A2572, and U2584 in 23 S rRNA, are
affected by pH transition in the 6.5– 8.5 range (30). However, the pHdependent structural changes differ in ribosomes from various species
despite the high degree of structural conservation of the peptidyl transferase center. The pH-dependent reactivity of A2451 varies in ribosomes of different organisms (38). C2474 shows increased reactivity at
pH 7.5 and 8.5 in ribosomes of M. smegmatis and Thermus aquaticus
(Ref. 39, and the current study) but not in E. coli (30). These data suggest
that ribosomes may undergo conformational changes that involve several ionizing groups that are not necessarily conserved in the peptidyl
transferase center with apparent pKa values between 7 and 8. The chemical nature of these groups does not have to be limited to rRNA residues,
as ionization of ribosomal proteins may also affect the structure of
rRNA in the peptidyl transferase center. Because the structure of the
peptidyl transferase center changes considerably in response to pH,
the effect of pH on the peptidyl transferase rate (Refs. 10 and 19, and the
current study) per se does not provide support for the catalytic models
involving general acid/base catalysis.
The A2451U mutation in M. smegmatis ribosomes altered the intensity and pH dependence of chemical modification of several bases in the
peptidyl transferase center, i.e. A2060, U2506, A2572, U2584, and
U2585. Interestingly, 2 of the effects that we observed in the M. smegmatis mutant, namely, increased accessibility of A2572 and decreased
accessibility of U2584 to DMS modification, parallel those observed in
the inactive conformation of the E. coli ribosome (30). This may suggest
that the A2451U mutation alters the structure of the peptidyl transferase center toward the inactive conformation. However, as the functional
effect of the mutation is rather moderate compared with the peptidyl
transferase inactivation by monovalent cation depletion, additional differences must be present between the structures induced by the A2451
mutation versus the absence of monovalent cations (30, 37).
Role of A2451 in the Peptidyl Transfer Reaction—Whereas mutations
of A2451 to 3 other bases are lethal in E. coli (14, 15), M. smegmatis cells
bearing the A2451U mutation are viable, although they grow slowly. A
similar mutation confers resistance to chloramphenicol in mouse mitochondria, indicating that this mutation is not lethal in mitochondrial
OCTOBER 28, 2005 • VOLUME 280 • NUMBER 43
ribosomes either (41). In reconstituted large ribosomal subunits, mutations of A2451 as well as incorporation of several nonnatural nucleobases instead of adenine or even removal of the adenine base preserved
considerable peptidyl transferase activity (12, 14, 16, 19).
In E. coli, the A2451U mutation decreased the rate of puromycin
reaction by 150-fold. It resulted in the loss of the corresponding ionization (pKa 7.5) and reduced the reaction rate by the same factor as protonation did with WT ribosomes (10). Kinetic analysis of the A2451U
M. smegmatis mutant demonstrated that the rate of peptidyl transfer to
puromycin was decreased by a factor of 10. However, the A2451U mutation in M. smegmatis ribosomes did not eliminate pH dependence of the
puromycin reaction and, thus, did not affect the ionization of a ribosomal group with a pKa of 8.0, which directly or indirectly affects the
catalysis. Therefore, in M. smegmatis, A2451 is not responsible for the
pKa of around 8, which then must arise from another ionizing group
within the peptidyl transferase center.
Two nonconventional pairs, A2453–C2499 and A2450 –C2063, have
been proposed as candidates that may convey pH-dependent conformational changes to the peptidyl transferase center (10). Each is presumed to possess a near-neutral pKa and both are located in the vicinity
of A2451. The A2453–C2499 base pair, but not the A2450 –C2063 base
pair, was found to contribute to the pH-dependent structural rearrangement of A2451 in E. coli (31, 40). However, the lack of nucleotide conservation at the 2453 and 2499 positions suggests that the A2453–
C2499 base pair, and consequently its presumed near-neutral pKa, is
unlikely to be a part of the fundamental, evolutionarily conserved mechanism of peptide bond formation (40). Rather, deprotonation of
group(s) in the peptidyl transferase center may accelerate peptide bond
formation by stabilizing a favorable conformation of the active site,
thereby contributing to catalysis on both M. smegmatis and E. coli ribosomes with puromycin as substrate (10).
Replacements of rRNA base A2451 in the “inner shell” of the peptidyl
transferase center (18) has a moderate to no effect on the rate of peptide
bond formation depending on the source of the ribosomes and the
nature of the A-site substrate (puromycin or aminoacyl-tRNA) (Refs. 10
and 18, and the current study). Mutations of the neighboring G2447
have very little effect on the rate of peptide bond formation (19). Both
replacements do not cause lethality in M. smegmatis (Ref. 24, and the
current study). This strongly argues against an involvement of the base
moiety of A2451 in the universally conserved mechanism of peptide
bond formation. Rather, A2451 may act as one of the pivots for assembling the structure of the peptidyl transferase center required for rapid
and proper positioning of substrates, shielding charges, and/or organizing water molecules.
Acknowledgments—We thank Wolfgang Wintermeyer for valuable comments
on the manuscript and Petra Striebeck, Astrid Böhm, Carmen Schillings,
Simone Möbitz, and Tanja Janusic for expert technical assistance.
REFERENCES
1. Ban, N., Nissen, P., Hansen, J., Moore, P. B., and Steitz, T. A. (2000) Science 289,
905–920
2. Nissen, P., Hansen, J., Ban, N., Moore, P. B., and Steitz, T. A. (2000) Science 289,
920 –930
3. Wolfenden, R. (1963) Biochemistry 338, 1090 –1092
4. Barta, A., and Halama, I. (1996) in Ribosomal RNA and Group I Introns (Green, R., and
Schroeder, R., eds) pp. 35–54, R. G. Landes Company, Austin, TX
5. Krayevsky, A. A., and Kukhanova, M. K. (1979) Prog. Nucleic Acids Res. Mol. Biol. 23,
1–51
6. Lieberman, K. R., and Dahlberg, A. E. (1995) Prog. Nucleic Acids Res. Mol. Biol. 50,
1–23
7. Maden, B. E., and Monro, R. E. (1968) Eur. J. Biochem. 6, 309 –316
JOURNAL OF BIOLOGICAL CHEMISTRY
36071
Peptidyl Transferase of M. smegmatis Ribosomes
8. Moore, P. B., and Steitz, T. A. (2003) Annu. Rev. Biochem. 72, 813– 850
9. Nierhaus, K. H., Schulze, H., and Cooperman, B. S. (1980) Biochem. Int. 1, 185–192
10. Katunin, V. I., Muth, G. W., Strobel, S. A., Wintermeyer, W., and Rodnina, M. V.
(2002) Mol. Cell 10, 339 –346
11. Hansen, J. L., Schmeing, T. M., Moore, P. B., and Steitz, T. A. (2002) Proc. Natl. Acad.
Sci. U. S. A. 99, 11670 –11675
12. Erlacher, M. D., Lang, K., Shankaran, N., Wotzel, B., Huttenhofer, A., Micura, R.,
Mankin, A. S., and Polacek, N. (2005) Nucleic Acids Res. 33, 1618 –1627
13. Parnell, K. M., Seila, A. C., and Strobel, S. A. (2002) Proc. Natl. Acad. Sci. U. S. A. 99,
11658 –11663
14. Thompson, J., Kim, D. F., O’Connor, M., Lieberman, K. R., Bayfield, M. A., Gregory,
S. T., Green, R., Noller, H. F., and Dahlberg, A. E. (2001) Proc. Natl. Acad. Sci. U. S. A.
98, 9002–9007
15. Muth, G. W., Ortoleva-Donnelly, L., and Strobel, S. A. (2000) Science 289, 947–950
16. Polacek, N., Gaynor, M., Yassin, A., and Mankin, A. S. (2001) Nature 411, 498 –501
17. Sievers, A., Beringer, M., Rodnina, M. V., and Wolfenden, R. (2004) Proc. Natl. Acad.
Sci. U. S. A. 101, 7897–7901
18. Youngman, E. M., Brunelle, J. L., Kochaniak, A. B., and Green, R. (2004) Cell 117,
589 –599
19. Beringer, M., Adio, S., Wintermeyer, W., and Rodnina, M. V. (2003) RNA (N.Y.) 9,
919 –922
20. Bevilacqua, P. C., Brown, T. S., Nakano, S., and Yajima, R. (2004) Biopolymers 73,
90 –109
21. Weinger, J. S., Parnell, K. M., Dorner, S., Green, R., and Strobel, S. A. (2004) Nat.
Struct. Mol. Biol. 11, 1101–1106
22. Pfister, P., Hobbie, S., Brull, C., Corti, N., Vasella, A., Westhof, E., and Bottger, E. C.
(2005) J. Mol. Biol. 346, 467– 475
23. Pfister, P., Hobbie, S., Vicens, Q., Bottger, E. C., and Westhof, E. (2003) Chembiochem.
4, 1078 –1088
24. Sander, P., Belova, L., Kidan, Y. G., Pfister, P., Mankin, A. S., and Bottger, E. C. (2002)
Mol. Microbiol. 46, 1295–1304
36072 JOURNAL OF BIOLOGICAL CHEMISTRY
25. Sander, P., Prammananan, T., and Bottger, E. C. (1996) Mol. Microbiol. 22, 841– 848
26. Prammananan, T., Sander, P., Springer, B., and Bottger, E. C. (1999) Antimicrob.
Agents Chemother. 43, 447– 453
27. Rodnina, M. V., and Wintermeyer, W. (1995) Proc. Natl. Acad. Sci. U. S. A. 92,
1945–1949
28. Rodnina, M. V., Savelsbergh, A., Matassova, N. B., Katunin, V. I., Semenkov, Y. P., and
Wintermeyer, W. (1999) Proc. Natl. Acad. Sci. U. S. A. 96, 9586 –9590
29. Merryman, C., and Noller, H. F. (1998) in RNA:Protein Interactions, A Practical Approach (Smith, C. W. J., ed) pp. 237–253, Oxford University Press, Oxford
30. Bayfield, M. A., Dahlberg, A. E., Schulmeister, U., Dorner, S., and Barta, A. (2001)
Proc. Natl. Acad. Sci. U. S. A. 98, 10096 –10101
31. Hesslein, A. E., Katunin, V. I., Beringer, M., Kosek, A. B., Rodnina, M. V., and Strobel,
S. A. (2004) Nucleic Acids Res. 32, 3760 –3770
32. Gutell, R. R., Weiser, B., Woese, C. R., and Noller, H. F. (1985) Prog. Nucleic Acids Res.
Mol. Biol. 32, 155–216
33. Noller, H. F., and Woese, C. R. (1981) Science 212, 403– 411
34. Harms, J., Schluenzen, F., Zarivach, R., Bashan, A., Gat, S., Agmon, I., Bartels, H.,
Franceschi, F., and Yonath, A. (2001) Cell 107, 679 – 688
35. Bashan, A., Agmon, I., Zarivach, R., Schluenzen, F., Harms, J., Berisio, R., Bartels, H.,
Franceschi, F., Auerbach, T., Hansen, H. A., Kossoy, E., Kessler, M., and Yonath, A.
(2003) Mol. Cell 11, 91–102
36. Pape, T., Wintermeyer, W., and Rodnina, M. V. (1998) EMBO J. 17, 7490 –7497
37. Miskin, R., Zamir, A., and Elson, D. (1968) Biochem. Biophys. Res. Commun. 33,
551–557
38. Muth, G. W., Chen, L., Kosek, A. B., and Strobel, S. A. (2001) RNA (N.Y.) 7,
1403–1415
39. Xiong, L., Polacek, N., Sander, P., Bottger, E. C., and Mankin, A. (2001) RNA (N.Y.) 7,
1365–1369
40. Bayfield, M. A., Thompson, J., and Dahlberg, A. E. (2004) Nucleic Acids Res. 32,
5512–5518
41. Kearsey, S. E., and Craig, I. W. (1981) Nature 290, 607– 608
VOLUME 280 • NUMBER 43 • OCTOBER 28, 2005