THE JOURNAL OF BIOLOGICAL CHEMISTRY
© 2003 by The American Society for Biochemistry and Molecular Biology, Inc.
Vol. 278, No. 24, Issue of June 13, pp. 21972–21979, 2003
Printed in U.S.A.
Cross-linking in the Living Cell Locates the Site of Action of
Oxazolidinone Antibiotics*
Received for publication, February 28, 2003, and in revised form, April 8, 2003
Published, JBC Papers in Press, April 10, 2003, DOI 10.1074/jbc.M302109200
Jerry R. Colca‡§, William G. McDonald‡, Daniel J. Waldon‡, Lisa M. Thomasco‡,
Robert C. Gadwood‡, Eric T. Lund‡, Gregory S. Cavey‡, W. Rodney Mathews‡, Lonnie D. Adams‡,
Eric T. Cecil‡, James D. Pearson‡, Jeffrey H. Bock‡, John E. Mott‡, Dean L. Shinabarger‡,
Liqun Xiong ¶, and Alexander S. Mankin¶
From the ‡Pharmacia Corp., Kalamazoo, Michigan 49001 and the ¶Center for Pharmaceutical Biotechnology,
University of Illinois, Chicago, Illinois 60607
Oxazolidinone antibiotics, an important new class of
synthetic antibacterials, inhibit protein synthesis by interfering with ribosomal function. The exact site and
mechanism of oxazolidinone action has not been elucidated. Although genetic data pointed to the ribosomal
peptidyltransferase as the primary site of drug action,
some biochemical studies conducted in vitro suggested
interaction with different regions of the ribosome.
These inconsistent observations obtained in vivo and in
vitro have complicated the understanding of oxazolidinone action. To localize the site of oxazolidinone action
in the living cell, we have cross-linked a photoactive
drug analog to its target in intact, actively growing
Staphylococcus aureus. The oxazolidinone cross-linked
specifically to 23 S rRNA, tRNA, and two polypeptides.
The site of cross-linking to 23 S rRNA was mapped to the
universally conserved A-2602. Polypeptides cross-linked
were the ribosomal protein L27, whose N terminus may
reach the peptidyltransferase center, and LepA, a protein homologous to translation factors. Only ribosomeassociated LepA, but not free protein, was cross-linked,
indicating that LepA was cross-linked by the ribosomebound antibiotic. The evidence suggests that a specific
oxazolidinone binding site is formed in the translating
ribosome in the immediate vicinity of the peptidyltransferase center.
Antibiotics inhibit cell growth by binding to essential molecular components and interfering with their activity. Identification of the specific drug target is a critical starting point for
understanding mechanism of action, possible resistance mechanisms, and for further rational drug refinement. Although one
effective way to identify the target of antibiotics is by mapping
drug resistance mutations, the majority of studies aimed at
understanding the site and the mode of the drug action are
generally carried out in vitro using biochemical techniques.
Such studies, conducted with subcellular fractions, have on
numerous occasions provided critical insights into understanding the mechanism of antibiotic activity. These biochemical
approaches are based on a common assumption that interaction of the drug with its target in vitro closely resembles the one
that takes place in the living cell. This assumption might not be
* 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.
§ To whom correspondence should be addressed: Pharmacia Corp.,
301 Henrietta St., Kalamazoo, MI 49001. Tel.: 269-833-9505; E-mail:
[email protected].
valid for some antibiotics that interact with such a complex and
dynamic macromolecular structure as the ribosome.
A great variety of antibiotics inhibit bacterial growth by
binding to the ribosome and interfering with its functions in
protein synthesis. The ribosome, whose size exceeds that of an
average antibiotic by four orders of magnitude, presents multiple possibilities for the antibiotic binding (1, 2). Furthermore,
in the course of translation, the ribosome cycles through various conformational states and interacts with a number of ligands and effectors, including tRNAs and translation factors,
which affect the spatial structure of the ribosome (3). It is
conceivable that the affinity of antibiotics for specific sites in
the ribosome might depend on the conformational state of the
particle.
Oxazolidinones are the first new class of antibiotics introduced into medical practice in 25 years. This class of protein
synthesis inhibitors is extremely effective against Gram-positive bacteria (4). Several studies suggest that the ribosome is
the primary target of action of these compounds (5–9); however,
interaction of oxazolidinones with the ribosome clearly differs
from other antibiotics because these drugs are active against
isolates that have developed resistance to other ribosomal antibiotics (10). Despite intense interest, the exact site of oxazolidinone action remains to be elucidated. Although the location
of resistance mutations point to the peptidyltransferase center
of the large ribosomal subunit as the possible drug target
(11–14), cross-linking and footprinting of ribosomes in vitro
had suggested interaction of oxazolidinones with different regions of the ribosome including the nucleotides U-2113, A-2114,
U-2118, A-2119, and C-2153 in 23 S rRNA located in the
vicinity of the ribosomal E-site (15) and at a considerable distance from the peptidyltransferase. In addition, these studies
conducted with isolated ribosomes had pointed to possible interactions of oxazolidinones with A-864 in the 16 S rRNA of the
small ribosomal subunit. Lack of a clear understanding of the
oxazolidinone binding site on the ribosome and reliance on
indirect approaches led to the proposal of diverse models of
oxazolidinone action (7, 8, 9, 12, 15–19).
One of the major limitations in studies of oxazolidinone action has been the low binding affinity of these compounds to
isolated 70 S ribosomes or 50 S ribosomal subunits, e.g. 200 ␮M
for eperezolid, one of the reference oxazolidinones (20). The low
affinity binding to isolated ribosomes or ribosomal subunits
limits the utility of standard footprinting techniques that are
routinely used to identify potential direct interactions with
rRNA (21, 22). In contrast to the low affinity of the drug for the
“static” isolated ribosome, these drugs are potent inhibitors of
cell-free translation, where eperezolid exhibits an IC50 of 2.5
21972
This paper is available on line at http://www.jbc.org
Cross-linking in Vivo Identifies Oxazolidinone Binding Site
␮M in an Escherichia coli system (7, 23). Thus, the site of drug
binding in purified non-translating ribosomes may not accurately represent the site responsible for the antibiotic activity.
It is possible that the relevant site may transiently appear only
in the dynamic structure of the ribosome engaged in protein
synthesis.
To bridge the information gained from biochemical experiments in vitro and genetic studies in vivo and to identify the
site of oxazolidinone action in intact bacteria, we have used
photo-induced cross-linking, a technique that is commonly used
in vitro. However, in contrast to the previous studies, we have
cross-linked a radioactive probe to its putative target in intact,
actively growing bacteria. We reasoned that this approach applied to living cells would localize drug binding in vivo in the
site relevant for its antibacterial activity. Moreover, we anticipated that this approach would succeed even if the relevant
binding site were formed only transiently in the translating
ribosome.
EXPERIMENTAL PROCEDURES
Reagents and Materials—All of the compounds used in this study
were synthesized at Pharmacia Corp. by previously published methods
(24). Carrier-free 125I-labeled PNU-259621 (abbreviated 125I-XL)1 and
the corresponding unlabeled compound were prepared using chloramine T and sodium iodide (26). Lysostaphin was from Sigma.
Bacterial Strains—The linezolid-sensitive clinical isolate Staphylococcus aureus ATCC 29213 was used for the cross-linking experiments.
S. aureus RN4220 NCTC8325-4 r⫺, m⫹ (restriction, minus; modification, plus; Ref. 27) was used for the generation of insertional mutants.
Cross-linking Experiments—S. aureus cultures were grown overnight with shaking in Mueller Hinton broth (Difco) at 37 °C. Cells were
diluted 1:100 and grown for 3– 4 h at 37 °C. When the optical density
reached an A600 of ⬃0.6, 1–1.4 ml of the resulting culture were centrifuged at 2000 ⫻ g for 2 min. The cells were resuspended in 170 ␮l of
fresh Mueller Hinton medium containing either 20 ␮l of 2% Me2SO or
non-radioactive competitor antibiotics in the same volume of Me2SO
followed by 10 ␮l of the radioactive photoactive oxazolidinone 125I-XL
(⬃2–10 ␮Ci/sample representing 1–2 ␮M final concentration). After a
30-min incubation at 37 °C in the dark, the suspension was exposed in
open tubes to UV light using a Stratalinker 1800 (Stratagene) set to an
energy exposure of 180,000 ␮J (2.1 min, 254 nm). The cells were sedimented in microcentrifuge tubes, washed with Dulbecco’s phosphatebuffered saline (Invitrogen), and resuspended in 180 ␮l of lysis buffer
containing 10 mM Tris-HCl (pH 7.4), 30 mM NH4Cl, 30 mM MgCl2, and
5 ␮g/ml lysostaphin. After a 15-min incubation at 37 °C, the debris of
the lysed cells was sedimented at 18,000 ⫻ g for 15 min at 4 °C. The
resulting supernatant was used for extraction of total RNA or for the
isolation of ribosomes after ultracentrifugation at 450,000 ⫻ g for 30
min in a Beckman TLA-100 ultracentrifuge (4 °C, TLA100 rotor).
Isolation of Total RNA—An aliquot (180 ␮l) of cell lysate supernatant
was diluted with buffer containing 10 mM Tris-HCl (pH 7.6), 6 mM
EDTA, and 0.5% SDS to a volume of 250 ␮l, and RNA was extracted
with phenol. The extracted RNA was precipitated by the addition of 400
␮l of ethanol, and the pellet was washed with 70% ethanol, dried in
vacuo, and either used directly for electrophoresis or subjected to
primer extension or RNase H analysis. Electrophoresis of RNA was
carried out either on 1% agarose gels or on 10% denaturing polyacrylamide gels. Gels were stained with ethidium bromide, photographed,
dried, and exposed to BioMax film (Eastman Kodak Co.).
Identification of 23 S rRNA Bases Cross-linked to Oxazolidinones—
For the preliminary mapping of 23 S rRNA cross-links, total RNA
prepared from the cross-linked sample was subjected to RNase H analysis (28, 29) essentially as described by Dontsova et al. (30). In brief, 4
␮g of total RNA was combined in 20 ␮l of water with 6 pmol of
individual deoxyribooligonucleotides or pairs of deoxyribooligonucleotides complementary to sites in domains V and VI of S. aureus 23 S
rRNA. The solution was incubated for 30 s at 90 °C and then cooled over
10 min to 50 °C. A 2.5-␮l aliquot of 10⫻ reaction buffer (150 mM
HEPES-KOH (pH 7.8), 500 mM NH4Cl, 10 mM MgCl2, 1 mM dithiothreitol) was added followed by the addition of 1 ␮l of RNase H (Seikagaku
1
The abbreviations used are: 125I-XL, 125I-labeled PNU-259621;
nanoLC-MS/MS, nano liquid chromatography tandem mass spectrometry (MS).
21973
America) in 2.5 ␮l of the 1⫻ reaction buffer. The reactions were incubated for 30 min at 55 °C, and RNA was precipitated by the addition of
100 ␮l of 0.3 M sodium acetate (pH 5.5) and 300 ␮l of ethanol. RNA was
recovered by centrifuging for 10 min at 21,000 ⫻ g at 4 °C, resuspending
in 7 ␮l of formamide loading dyes, heating for 30 s at 90 °C, and loading
the sample onto a 6% denaturing polyacrylamide gel. The gels (20 cm ⫻
20 cm ⫻ 1 mm) were electrophoresed at 20 W for 1.5 h, stained with
ethidium bromide, photographed, dried, and exposed to a PhosphorImager screen (Molecular Dynamics).
For precise mapping of the cross-linked nucleotides, the amount of
127
I-XL (in this case non-radioactive) added to S. aureus cells was
increased to 2 mM. Oligonucleotide primers were 5⬘-terminally labeled,
annealed to RNA, and extended with reverse transcriptase as described
(31).
The following oligonucleotides were used for RNase H mapping and
primer extension (numbers in parentheses correspond to the 23 S rRNA
complementary sequence, E. coli numbering): SaL 2550, GCCGACATCGAGGTGCC (2494 –2510); SaL 2720, GTCCATCCCGGTCCTCTCGTAC (2655–2676); SaL 2230, TAGTATCCCACCAGCGTCTC
(2157–2176); SaL2340, ACCTTTGAGCGCCTCCG (2275–2291); EcL
2563, TCGCGTACCACTTTA (2563–2577).
Isolation and Identification of Cross-linked Proteins—Ribosomal pellets were resuspended in 70 ␮l of ribosome buffer (10 mM Tris-HCl, 30
mM NH4Cl, 30 mM MgCl2 (pH 7.4)). Two volumes of glacial acetic acid
were added, and the tubes were mixed for 15 min at 4 °C before the
precipitated RNA was removed by centrifugation at 18,000 ⫻ g for 15
min. The resulting supernatant was brought to 80% acetone and stored
overnight at ⫺20 °C. The acetone pellet was recovered by centrifugation
at 18,000 ⫻ g, washed with cold acetone to remove residual acetic acid,
and lyophilized (32).
Two-dimensional gel electrophoresis was carried out as described by
Adams (33). Lyophilized protein pellets were resuspended in 35 ␮l of
buffer containing 9 M urea, 4% Nonidet P-40, 1% dithiothreitol, and 2%
pH 3.5–10 ampholytes. Samples were applied to isoelectric focusing
tube gels and focused for 23,000 volt-hours. Second-dimension gels were
10 –20% non-linear gradients run in Tris-glycine-SDS buffer (Bio-Rad).
Gels were silver-stained (34), dried, and exposed to Kodak Biomax film.
High performance liquid chromatography separation of the extracted
ribosomal proteins was carried out on a Vydac 218TPT C18 column
according to a modified procedure of Cooperman et al. (35). The elution
(1 ml/min) started with 75% A (water, 0.1% trifluoroacetic acid, v/v),
25% B (acetonitrile, 0.1% trifluoroacetic acid, v/v). Buffer B proportion
was increased from 25 to 80% over a 100-min linear gradient.
Protein Identification—The specifically cross-linked proteins were
excised from the gels, reduced, alkylated, and digested in situ with
modified porcine trypsin (Promega) using a DigestPro robot (LC Packings). Nano liquid chromatography (nanoLC) tandem mass spectrometry analysis (nanoLC-MS/MS) was performed on a Micromass Q-Tof
instrument equipped with a Z-spray ion source. Nanospray MS/MS data
was used to identify proteins by comparing the experimental data with
predicted data derived from protein and DNA databases. Tandem MS
data was searched against the NCBInr and HGS protein databases
using MASCOT (Matrix Science) programs maintained on an in-house
server.
Targeted Disruption of LepA Gene—The LepA knock-out strain of
S. aureus (RN4220) was constructed by insertional mutagenesis using
an E. coli vector construct that is incapable of replication in S. aureus.
A 310-bp region at the 5⬘ end of the S. aureus LepA gene (yqeQ) was
PCR-amplified from the genomic DNA using a pair of primers, CTCGATTATAGCACATATTGAC and TTTGTGCTTCGATACCTTGAG, and
cloned into the pCR2.1 vector (Invitrogen) to create pCR-yqeQ-N310
plasmid. An ermB gene that confers erythromycin resistance in S.
aureus was cloned into the NotI site of the pCR-yqeQ-N310 plasmid,
and the resulting construct was electroporated into S. aureus (36).
Genomic DNA was isolated from several erythromycin-resistant colonies, and yqeQ gene disruption was confirmed by PCR and Southern
blot analysis. The lack of LepA expression in the constructed strain was
confirmed by Western blot analysis using LepA-specific rabbit antibodies (Covance).
RESULTS
In Vivo Cross-linking of Oxazolidinones to Ribosomal RNA—
The structures of the oxazolidinones used in this study are
shown in Fig. 1. In addition to the photosensitive probe (125IXL), the structures are shown for the compounds used for
competition, eperezolid (PNU-100592(S) and its inactive enantiomer, PNU-107112(R) as well as for linezolid (Zyvox威, PNU-
21974
Cross-linking in Vivo Identifies Oxazolidinone Binding Site
FIG. 1. The oxazolidinone photoprobe and competitor compounds. 125I-Labeled PNU-259621 (125I-XL) was the photoreactive oxazolidinone used in these studies. PNU-100592(S), eperezolid, was an
active antibiotic used to compete at the relevant site of action; PNU107112(R) was used as an inactive control. Linezolid, PNU-100766, is
the oxazolidinone currently approved for clinical use.
100766), the first clinically approved oxazolidinone. With the
exception of PNU-107112(R), all compounds were active
against S. aureus (MIC values of 0.7–10 ␮M). All of the compounds contained the central oxazolidinone ring, the main
pharmacophore of this group of drugs (24). 125I-XL, the compound used in cross-linking experiments, contained the photoactive azido group.
For cross-linking in vivo, exponentially growing S. aureus
cells were incubated with the 125I-XL photoprobe with or without competitor compounds for 30 min (optimal under these
conditions) to allow drug penetration into the cell and binding
to the target. To determine the specificity of the interaction,
parallel incubations contained a 20-fold excess of non-photoactive competitors, active PNU-100592(S), or the control, inactive
enantiomer PNU-107112(R) (Fig. 1). After these incubations,
125
I-XL was cross-linked to its target by brief exposure to UV
light (see “Experimental Procedures”).
Extraction and electrophoresis of RNA from S. aureus cells
that had been incubated with 125I-XL demonstrated that 23 S
rRNA and tRNA were labeled selectively by the photoprobe
(Fig. 2). Electrophoresis of the RNA on 1% agarose gels (Fig.
2A) clearly resolved 23 S, 16 S, and 5 S RNA. The 125I probe was
cross-linked only to 23 S rRNA; no radioactive drug was seen
associated with 16 S rRNA even upon prolonged exposure of
the dried gels to film. Cross-linking of 125I-XL to 23 S rRNA
likely occurred from the site of its inhibitory action because the
yield of the cross-link was significantly reduced in the presence
of an active drug PNU-100592(S) but not by its inactive enantiomer PNU-107112(R). (Fig. 2, lanes S and R). Separation of
the extracted RNA by polyacrylamide electrophoresis (Fig. 2B)
demonstrated that although no drug cross-linked to 5S rRNA,
some amount of 125I-XL (much smaller than seen associated
with 23 S rRNA) also cross-linked to RNA molecules co-migrating with tRNA. Specific cross-linking of biologically active
125
I-XL to 23 S rRNA indicated that in their binding site in the
large ribosomal subunit oxazolidinones form close contacts
with certain residues of 23 S rRNA and possibly ribosomebound tRNA.
To map the approximate site(s) of 125I-XL cross-linking to 23
FIG. 2. Analysis of RNA isolated from intact S. aureus crosslinked by [125]I-XL by agarose gel electrophoresis (A) or polyacrylamide gel electrophoresis (B). Exponentially growing S. aureus cells were cross-linked with 2 ␮M [125]I-XL alone (lane C) or in the
presence of a 20-fold excess of biologically active non-cross-linkable
competitor (lane S) or its inactive enantiomer (lane R). Panel A shows a
representative ethidium bromide-stained 1% agarose gel (left) and its
autoradiogram (right) of extracted cellular RNA. Panel B is a similar
representation of the separation on denaturing 10% polyacrylamide
gels.
S rRNA, RNase H mapping was utilized (28, 29, 37). Total RNA
prepared from cross-linked S. aureus cells was incubated with
oligodeoxyribonucleotides complementary to specific regions of
23 S rRNA (Fig. 3A). The RNA in the formed DNA/RNA heteroduplex was cleaved with RNase H, and the resulting fragments were resolved by denaturing gel electrophoresis and
identified by autoradiography. As shown in Fig. 3B, when
cross-linked 23 S rRNA was cleaved with RNase H in the
presence of oligonucleotide SaL 2550, complementary to the
rRNA segment 2494 –2510 (E. coli numbering), an ⬃400-nucleotide-long radiolabeled 23 S rRNA fragment representing the
3⬘ end of 23 S rRNA was released. No label was found associated with the longer 5⬘ segment. Digestion of 23 S rRNA with
RNase H in the presence of two oligonucleotides, SaL 2550 and
SaL 2720 (23 S rRNA positions 2655–2676), released a single
radiolabeled rRNA fragment ⬃170 nucleotides long corresponding to the rRNA segment between the two oligonucleotides. Thus, 125I-XL was cross-linked exclusively to the 2494 –
2676 segment of 23 S rRNA.
The precise location of the cross-link was mapped by primer
extension. A single additional band at position 2603 was seen
when the rRNA sample prepared from cells incubated with
non-radioactive 127I-XL was used as a template (Fig. 3C), corresponding to the drug cross-link at position A-2602. As expected from the results with the radioactive probe, the intensity of this band decreased when the cross-linking was
prevented by incubation in the presence of a 20-fold excess of
PNU-100592(S) in vivo. Thus, the single rRNA base crosslinked by the photoprobe in actively growing bacteria was
A-2602.
A-2602 is one of the central nucleotides in the ribosomal
peptidyltransferase center (38). Thus, cross-linking of the oxazolidinone photoprobe to A-2602 places the site of the drug
action in the peptidyltransferase center.
Oxazolidinone Cross-linking to Ribosome-associated Proteins
in Vivo—To investigate possible in vivo cross-linking of 125I-XL
Cross-linking in Vivo Identifies Oxazolidinone Binding Site
21975
FIG. 4. Ribosome-associated proteins cross-linked in vivo by
I-XL. A, autoradiogram of a SDS-PAGE separation of proteins extracted from the ribosome pellet prepared from cells cross-linked by
125
I-XL in the absence (⫺) or in the presence (⫹) of a 20-fold excess of
competing PNU-100592(S). B, resolution of the 125I-XL-cross-linked
11-kDa ribosome-associated protein on BASO two-dimensional gels.
The left panel is a representative silver-stained gel. The center panel
and the right panel are gel autoradiograms from incubations of cells
without and with competing PNU-100592(S), respectively. The crosslinked protein is shown by the arrow. C, resolution of the 125I-XL-crosslinked 64-kDa ribosome-associated protein on Iso-dalt two-dimensional
gels. The left panel is a silver-stained gel. The center panel and the right
panel are gel autoradiograms from incubations without and with excess
PNU-100592(S), respectively. The specifically cross-linked protein is
shown by the arrow. D, left panel, Western blot of LepA in ribosomal
pellets (Pellet) and post-ribosomal supernatant (Super) from cells crosslinked with 125I-XL in the presence (⫹) or in the absence (⫺) of competing PNU-100592(S). Right panel, the autoradiogram of the gel.
125
FIG. 3. Identification of the site of in vivo 23 S RNA crosslinking by [125I]-XL. A, the relative location of the primers (thin bars)
complementary to 23 S rRNA (thick bar) used in RNase H mapping. The
approximate sizes of the RNA fragments produced by RNase H cleavage
are indicated. B, the autoradiogram of the denaturing 6% polyacrylamide gel resolving the products of 23 S rRNA hydrolysis by RNase H.
The location of 16 S and 23 S rRNA as observed in ethidium bromidestained gels are shown as well as the sizes of the 23 S RNA fragments
released after RNase H treatment, as determined by comparison with
the ethidium bromide-stained size markers. nt, nucleotides. C, the
results of primer extension mapping of the site of cross-link. C, U, A,
and G were sequencing lanes. Lane K, control RNA prepared from
untreated cells. Lane UV, RNA prepared from cells irradiated with UV
light in the absence of antibiotic. Lane X, RNA extracted from cells that
were incubated with the photoactive oxazolidinone 127I-XL before UV
irradiation. Lane S, RNA from cells incubated with 127I-XL in the
presence of PNU-100592(S), the oxazolidinone competitor, before UV
irradiation. An arrow indicates the nucleotide residue corresponding to
the cross-linked position.
to ribosome-associated proteins, ribosomes were pelleted from
the lysates of S. aureus cells cross-linked with the drug, and
proteins were fractionated by SDS electrophoresis. SDS-PAGE
analysis of the proteins in the ribosome pellet demonstrated
two radiolabeled protein bands of ⬃64 and 11 kDa that were
selectively cross-linked (Fig. 4A). The 125I-XL cross-linking to
these the proteins showed the same specificity as was seen for
the 23 S RNA in these samples; cross-linking was reduced in
the presence of the biologically active competitor eperezolid,
PNU-100592(S), but not by its inactive enantiomer. The specifically cross-linked proteins were purified, and their identity
was determined by nanoLC-MS/MS analysis as described
below.
The 11-kDa cross-linked protein resolved as a single image
on two-dimensional electrophoresis with an isoelectric point
close to 11 (Fig. 4B). The radioactive spot corresponding to the
specifically cross-linked 11 kDa protein did not precisely overlay a silver stained protein on the respective gels, most likely
due to the effect of the cross-linked drug on protein mobility.
However, when protein was eluted from the gel beneath the
radioactive spot and digested with trypsin, five peptides were
identified by nanoLC-MS/MS as corresponding to the tryptic
fragments of the 50 S subunit protein L27. Fig. 5A shows the
S. aureus L27 sequence; the tryptic peptides identified are in
bold. No other proteins were identified in this portion of the
two-dimensional gel. To further confirm that the probe had
cross-linked L27, ribosomal proteins prepared from cells crosslinked with 125I-XL were subjected to high performance liquid
chromatography purification. The fraction containing the 11kDa radioactive protein was collected and analyzed by MS/MS.
The identity of L27 was again confirmed by the detection of five
tryptic fragments that covered 39% of the total sequence. L27 is
the only ribosomal protein that is suspected to reach very close
to the active site of the peptidyltransferase center of the bac-
21976
Cross-linking in Vivo Identifies Oxazolidinone Binding Site
FIG. 5. Identification of the proteins cross-linked in vivo by
I-XL. The complete amino acid sequences of the S. aureus L27 (A)
and LepA (B) are shown with the tryptic peptides identified by MS/MS
shown in bold.
125
terial ribosome (see “Discussion”). Therefore, cross-linking of
this protein by the oxazolidinone probe is compatible with the
23 S rRNA cross-link and with the general location of the site
of the drug action within the ribosomal peptidyltransferase
center.
Analysis of the 64-kDa cross-linked protein by two-dimensional electrophoresis is shown in Fig. 4C. The gel beneath the
radioactive spot was excised, digested with trypsin, and analyzed by nanoLC-MS/MS as described under “Experimental
Procedures.” Eleven tryptic fragments were identified (shown
in bold in Fig. 5B) that matched the S. aureus protein LepA. No
other proteins were identified in this region of the gel. LepA
belongs to the GTPase family of and exhibits significant homology to the translation factors EF-G and EF-Tu, indicating its
possible involvement in translation and association with the
ribosome. Antibodies generated against LepA showed the presence of the protein in both the ribosomal pellet and the postribosomal supernatant (Fig. 4D, left panel). Notably, however,
the protein cross-linked by 125I-XL was detected only in the
ribosomal pellet (Fig. 4D, right panel), suggesting that association of LepA with the ribosome was required for cross-linking
by the drug.
LepA Is Not the Target of Oxazolidinone Action—To investigate whether LepA represented a target of oxazolidinone action, we constructed a S. aureus strain in which the LepA gene
(yqeQ) was disrupted by insertional mutagenesis (see “Experimental Procedures”). LepA gene disruption was confirmed by
PCR and Southern blotting; the lack of LepA expression in the
engineered strain was also verified by Western blotting using
LepA-specific antibodies. The LepA knockout strain exhibited
growth characteristics comparable with those for the wild type.
When the 125I-XL was cross-linked to its target in the LepA
knockout strain, the 64-kDa cross-linked band was absent,
confirming that LepA was a target for oxazolidinone crosslinking. However, the 23 S rRNA, tRNA, and protein L27
cross-links remained unaffected as did the minimum inhibitory
concentration (MIC) for the active oxazolidinones used in this
study (data not shown). Thus, although LepA bound to the
ribosome was cross-linked specifically by 125I-XL, 125LepA is
not required for oxazolidinone action and does not appear to be
a direct target of the drug action.
DISCUSSION
In this study we describe the first use of cross-linking to
identify the site of antibiotic action in vivo in actively growing
cells. We have applied this technique to mapping the site of
action of the clinically important oxazolidinones. The data demonstrate that photoactive oxazolidinones cross-link to the pep-
tidyltransferase center in the exponentially growing Grampositive pathogen, S. aureus.
Only specific RNA and protein molecules were cross-linked
when the biologically active oxazolidinone azido-derivative, XL,
was activated by UV irradiation after its accumulation in the
bacterial cells. The cellular macromolecules contacted by the
drug in the putative site of action included 23 S rRNA (A-2602),
tRNA, and the two proteins L27 and LepA. The relevance of
these contacts is supported by the high selectivity of the interaction; only one of the three ribosomal RNAs (23 S rRNA) and
only one of more than 50 different ribosomal proteins (L27) was
cross-linked by the drug in the living cell. It is highly likely that
125
I-XL was cross-linked to these macromolecules from the site
of oxazolidinone action since the biologically active oxazolidinone, eperezolid, was able to compete with 125I-XL for the
binding site, whereas its inactive isomer was without effect.
This conclusion is further supported by the ⬃75% reduction of
all cross-links in two strains of S. aureus carrying oxazolidinone-resistant mutations G2447U and G2576U in 23 S rRNA
(not shown).
Cross-linking of the photoactive oxazolidinone in vivo to
A-2602, one of the central nucleotides in the ribosomal peptidyltransferase center, is the key result that defines ribosomal
peptidyltransferase as the main site of drug action. Fig. 6
shows the position of A-2602 in relationship to the known
linezolid resistance mutations identified in Gram-positive
(green), Gram-negative (gold), or archaeal (blue) ribosomes
(11–14). The site of the drug cross-link and the resistance
mutations are located in close vicinity to each other and are all
clustered within or nearby the central loop of domain V of 23 S
rRNA, the core of the catalytic peptidyltransferase center of the
ribosome. This proximity of this site of cross-linking of the drug
in vivo to the mutations producing oxazolidinone resistance is
particularly evident in the crystallographic structure of the
large ribosomal subunit (Fig. 6B). The spatial arrangement of
the resistance mutations versus A-2602 cross-linked by the
azido group attached from the C-ring side of the oxazolidinone
indicates a possible general orientation of the drug molecule
within its binding site, with the pharmacophoric oxazolidinone
group protruding toward the active site of the peptidyltransferase center and the extended (C-ring) appendage reaching
toward A-2602.
The site of oxazolidinone action in the ribosomal peptidyltransferase center is also supported by the polypeptides crosslinked by 125I-XL in the living cell. The data conclusively show
an interaction of 125I-XL with the large ribosomal subunit
protein, L27. Existing biochemical evidence suggests a close
association of protein L27 with the peptidyltransferase center.
The 3⬘ end of azido derivatives of aminoacyl- and peptidyltRNA bound in the E. coli ribosomal peptidyltransferase active
site cross-link to L27 in vitro (39 – 41). Biochemical and genetic
studies indicate that this interaction may engage the N terminus of the protein2; by inference, one would expect that the
oxazolidinone cross-link also engages the N-terminal part of
L27. Indeed, cleavage of the cross-linked L27 at Lys 57 with
endoproteinase Glu-C demonstrated that the 125I cross-link
was associated with the N terminus of the L27 protein (data
not shown). L27 was previously shown to be cross-linked in
vitro by several other antibiotics that affect ribosomal peptidyltransferase activity (42– 46). This observation suggests that
the oxazolidinone binding site is located in the vicinity of the
sites impacted by other peptidyltransferase-targeted drugs.
However, the mode of oxazolidinone binding must be unique
2
B. A. Maguire, A. D. Beniaminov, L. Xiong, A. S. Mankin, and R. A.
Zimmermann, submitted for publication.
Cross-linking in Vivo Identifies Oxazolidinone Binding Site
21977
FIG. 6. Location of the site of the 125I-XL cross-link and sites of oxazolidinone resistance mutations in the 23 S rRNA. Secondary (A)
and three-dimensional (B) structures of S. aureus 23 S rRNA. A-2602 cross-linked by photoactive oxazolidinone is shown in red and indicated by
an arrow in the secondary structure diagram of the central loop of domain V (25). Nucleotides whose mutations confer resistance to oxazolidinones
in Gram-positive bacteria are indicated by green dots, whereas the resistance mutations identified in the Archaea Halobacterium halobium are
shown as blue dots. The single resistance mutation identified in E. coli is shown by a gold dot on the secondary structure diagram. B, the orientation
of corresponding nucleotide positions are shown in the crystallographic structure of the H. marismortui large ribosomal subunit (47) where relevant
nucleotides are colored to match the secondary structure diagram. Numbering corresponds to the E. coli 23 S rRNA.
since the majority of ribosomal mutations conferring resistance
to other ribosomal antibiotics does not affect cell sensitivity to
oxazolidinones (10).
Although first high resolution structure of the 50 S ribosomal
subunit of Haloarcula marismortui x-ray studies of showed no
proteins in the peptidyltransferase active site of the archaeal
ribosome (47), Archaea lack the L27 homolog. The crystal
structure of the large ribosomal subunit from the eubacterium
21978
Cross-linking in Vivo Identifies Oxazolidinone Binding Site
Deinococcus radiodurans shows a close proximity of L27 to the
peptidyltransferase center (48). The globular C-terminal part
of L27 is “buried” under the central protuberance of the large
ribosomal subunit, whereas the unstructured N-terminal region emerges at the subunit interface and may protrude in the
direction of the peptidyltransferase center. Though the precise
location of the N terminus of L27 in the solved structure is
ambiguous, the highly flexible nature of the N-terminal protein
“tail” is consistent with the notion that it might reach close to
the drug binding site in the peptidyltransferase center (48). An
interference with L27 function might also be expected to affect
ribosome assembly and the fidelity of translation (49), functions that are also inhibited by oxazolidinones (19, 50). We
were not able to test, however, whether L27 is required for drug
interaction since the L27 knockout appears to be lethal in
S. aureus.
Another protein cross-linked by 125I-XL in vivo is LepA. The
function of LepA in the cell is unclear. This protein, which was
originally found associated with the cell membrane fraction
(39), exhibits considerable homology to the translation factor
GTPases, and thus, its interaction with the ribosome is likely
(51). The protein is distributed between the ribosomal pellet
and free (supernatant) fractions (Fig. 4D). Importantly, only
the ribosome-associated fraction of LepA was cross-linked by
125
I-XL, suggesting that cross-linking occurs from the ribosome-associated drug. This is supported by the observation
that rRNA mutations conferring resistance to oxazolidinones
significantly decrease 125I-XL cross-linking to LepA in parallel
to reductions in the other cross-links (not shown). Despite
cross-linking of photoactive oxazolidinone to the LepA protein,
inactivation of the LepA gene did not confer any appreciable
resistance to oxazolidinones, suggesting that LepA was crosslinked essentially as a bystander rather than as the target of
drug action. It is also possible that the function of LepA may be
replaced by a redundant translation factor in the knockout
mutant. However, an important functional implication of this
cross-linking result is that this presents the first biochemical
indication of LepA involvement in protein synthesis. The results demonstrate that a portion of this protein, while bound to
the ribosome in vivo, must extend to the site of oxazolidinone
binding in the ribosomal peptidyltransferase center.
Cross-linking of photoactive oxazolidinone to a critical base,
A-2602, in the peptidyltransferase center as well as clustering
of oxazolidinone resistance mutations, all point to the peptidyltransferase as the primary site of the drug action. Although
binding of oxazolidinones to the ribosome and 50 S subunit can
be observed in vitro (52), oxazolidinone binding in vitro demonstrates a 50 –100-fold weaker affinity than would be predicted from their in vivo inhibitory concentrations (20, 52).
Because there is no evidence that oxazolidinones are concentrated within bacteria, these results suggest that the ribosome
must adopt a specific conformation to create a high affinity
oxazolidinone binding site. We were unable to show any specific
(competed by eperezolid) cross-linking in isolated ribosomes
(not shown).
“Opening” of a oxazolidinone binding site in the translating
ribosome may occur after interaction of the ribosome with one
of its ligands. Binding of several different ribosome-targeted
antibiotics is known to be stimulated by simultaneous binding
of co-ligands. For example, the affinity of sparsomycin significantly increases in the presence of peptidyl-tRNA (53), enhancing its ability to also cross-link to A-2602 (54). Moreover, erythromycin binding may be stimulated by specific nascent
peptides (55). An additional ligand might either allosterically
affect the conformation of nucleotides involved in oxazolidinone
binding or directly participate in formation of the drug binding
site. Consistent with this model, cross-linking of tRNA was
observed in our studies, suggesting that occupation of one of
the tRNA sites on the ribosome may facilitate drug binding.
Several studies demonstrate that oxazolidinones may interfere
with positioning of fMet-tRNA in the P-site of 70 S ribosomes,
thereby preventing the initiation of translation (8, 16, 18). In
agreement with this conclusion, the A-2602 base that is exclusively cross-linked by 125I-XL in the living cell has been implicated in the interaction of the 3⬘ end of transfer RNA with the
ribosome (56). Although the identity of the tRNA species crosslinked by oxazolidinones is not known, the inability of this
antibiotic class to inhibit polysome elongation (7) suggests that
fMet-tRNA is a likely candidate. Initial interaction of fMettRNA with the ribosome could facilitate drug binding, leading
to interference with accommodation of fMet-tRNA in the ribosomal P-site.
In the “high affinity” conformation of the ribosome, induced
either by binding of one of the ribosomal ligands (tRNA, translation factors) or by ribosome cycling through different conformational states during translation, the rRNA nucleotides involved in drug binding might adopt an orientation favoring
their interaction with the drug. Binding of the antibiotic would
then “lock” the ribosome in this conformation, thereby preventing subsequent structural transitions of rRNA required for
protein synthesis. Such a mode of action is reminiscent of that
proposed for pactamycin and several other ribosome-targeted
antibiotics (1, 57). In this respect it is worth noting that the
nucleotide cross-linked by 125I-XL in vivo, A-2602, appears to
be the most flexible in the peptidyltransferase center, assuming a different conformation depending upon the state of the
ribosome (47, 48, 58). Specific orientations of A-2602 were
proposed to be critical for some activities of the peptidyltransferase center (58). One of the conformations of this nucleotide
may favor oxazolidinone binding. A similar dependence of sparsomycin binding on A-2602 orientation was recently proposed
(59).
Overall, the in vivo cross-linking approach used in this study
provides strong evidence that a binding site for the oxazolidinone antibiotics is formed at or near the peptidyltransferase
center in actively translating bacterial ribosomes. This is consistent with the location of the resistance mutations and with
the biochemical studies that suggest that oxazolidinones interfere with first peptide bond formation (5, 7, 8, 16, 18). In its
binding site in the ribosome, the oxazolidinones appear to be
able to interact with ribosomal protein L27, ribosome-associated protein LepA, and tRNA. Such interactions could explain
the additional reported effects of the oxazolidinones on ribosome formation and fidelity of translation. The information
gained from these studies in actively growing bacteria should
help unravel the mechanism of oxazolidinone action, whereas
the approach itself can be used for studies of other currently
used and newly developed anti-ribosomal drugs.
Acknowledgments—We gratefully acknowledge the help of Sara Morin
and Gary Zurenko for susceptibility testing and Mark McCroskey for
purification of LepA. We also acknowledge John Chosay and James Blinn
for assistance with interpreting the three-dimensional representation
shown in Fig. 6.
REFERENCES
1. Brodersen, D. E., Clemons, W. M., Jr., Carter, A. P., Morgan-Warren, R. J.,
Wimberly, B. T., and Ramakrishnan, V. (2000) Cell 103, 1143–1154
2. Schlunzen, F., Zarivach, R., Harms, J., Bashan, A., Tocilj, A., Albrecht, R.,
Yonath, A., and Franceschi, F. (2001) Nature 413, 814 – 821
3. Ramakrishnan, V. (2002) Cell 108, 557–572
4. Ford, C. W., Hamel, J. C., Wilson, D. M., Moerman, J K., Stapert, D., Yancey,
R. J., Jr., Hutchinson, D. K., Barbachyn, M. R., and Brickner, S. J. (1996)
Antimicrob. Agents Chemother. 40, 1508 –1513
5. Eustice, D. C., Feldman, P. A., and Slee, A. M. (1988) Biochem. Biophys. Res.
Commun. 150, 965–971
6. Eustice, D. C., Feldman, P. A., Zajac, I., and Slee, A. M. (1988) Antimicrob.
Cross-linking in Vivo Identifies Oxazolidinone Binding Site
Agents Chemother. 32, 1218 –1222
7. Shinabarger, D. L., Marotti, K. R., Murray, R. W., Lin, A. H., Melchior, E. P.,
Swaney, S. M., Dunyak, D. S., Demyan, W. F., and Buysse, J. M. (1997)
Antimicrob. Agents Chemother. 41, 2132–2136
8. Swaney, S. M., Aoki, H., Ganoza, M. C., and Shinabarger, D. L. (1998) Antimicrob. Agents Chemother. 42, 3251–3255
9. Patel, U., Yan, Y. P., Hobbs, F. W., Jr., Kaczmarczyk, J., Slee, A. M., Pompliano, D. L., Kurilla, M. G., and Bobkova, E. V. (2001) J. Biol. Chem. 276,
37199 –37205
10. Stevens, D. L., Herr, D, Lampiris, H. Hunt, J. L., Batts, D. H., and Hafkin, B.
(2002) Clin. Infect. Dis. 34, 1481–1490
11. Swaney, S. M., Shinabarger, D. L., Schaadt, R. D., Bock, J. H., Slightom, J. L.,
and Zurenko, G. E. (1998) Abstracts of the 38th Interscience Conference on
Antimicrobial Agents and Chemotherapy, San Diego, CA, September 24 –27,
1998, Abst. C-104, American Society for Microbiology, Washington, D. C.
12. Kloss, P., Xiong, L., Shinabarger, D. L., and Mankin, A. S. (1999) J. Mol. Biol.
294, 93–101
13. Xiong, L., Kloss, P., Douthwaite, S., Andersen, N. M., Swaney, S., Shinabarger,
D. L., and Mankin, A. S. (2000) J. Bacteriol. 182, 5325–5331
14. Prystowsky, J., Siddiqui, F., Chosay, J., Shinabarger, D. L., Millichap, J.,
Peterson, L. R., and Noskin, G. A. (2001) Antimicrob. Agents Chemother. 45,
2154 –2156
15. Matassova, N. B., Rodnina, M. V., Endermann, R., Kroll, H. P., Pleiss, U.,
Wild, H., and Wintermeyer, W. (1999) RNA (N. Y.) 5, 939 –946
16. Aoki, H., Ke, L., Poppe, S. M., Poel, T. J., Weaver, E. A., Gadwood, R. C.,
Thomas, R. C., Shinabarger, D. L., and Ganoza, M. C. (2002) Antimicrob.
Agents Chemother. 46, 1080 –1085
17. Shinabarger, D. L. (1999) Exp. Opin. Investig. Drugs 8, 1195–1202
18. Burghardt, H., Schimz, K. L., and Muller, M. (1998) FEBS Lett. 425, 40 – 44
19. Thompson, J., O’Connor, M., Mills, J. A., and Dahlberg, A. E. (2002) J. Mol.
Biol. 322, 273–279
20. Zhou, C. C., Swaney, S. M., Shinabarger, D. L., and Stockman, B. J. (2002)
Antimicrob. Agents Chemother. 46, 625– 629
21. Moazed, D., and Noller, H. F. (1987) Biochimie (Paris), 69, 879 – 884
22. Moazed, D., and Noller, H. F. (1987) Nature 327, 389 –394
23. Brickner, S. J., Hutchinson, D. K., Barbachyn, M. R., Manninen, P. R., Ulanowicz, D. A., Garmon, S. A., Grega, K. C., Hendges, S. K., Toops, D. S.,
Ford, C. W., and Zurenko, G. E. (1996) J. Med. Chem. 39, 673– 679
24. Barbachyn, M. R., Brickner, S. J., Gadwood, R. C., Garmon, S. A., Grega, K. C.,
Hutchinson, D. K., Munesada, K., Reischer, R. J., Taniguchi, M., Thomasco,
L. M., Toops, D. S., Yamada, H., Ford, C. W., and Zurenko, G. E. (1998) Adv.
Exp. Med. Biol. 456, 219 –238
25. Cannone, J. J., Subramanian, S., Schnare, M. N., Collett, J. R., D’Souza,
L. M., Du, Y., Feng, B., Lin, N., Madabusi, L. V., Muller, K. M., Pande, N.,
Shang, Z., Yu, N., and Gutell, R. R. BMC Bioinformatics (2002) 3,
http://www.biomedcentral.com/content/pdf/1471-2105-3-2.pdf
26. Greenwood, F. C., Hunter, W. M., and Glover, J. S. (1963) Biochem. J. 89,
114 –163
27. Kreiswirth, B. N., Lofdahl, S., Betley, M. J., O’Reilly, M., Schlievert, P. M.,
Bergdoll, M. S., and Novick, R. P. (1983) Nature 305, 709 –712
28. Donis-Keller, H. (1979) Nucleic Acids Res. 7, 179 –192
29. Mankin, A. S., Skripkin, E. A., Chichkova, N. V., Kopylov, A. M., and Bogdanov, A. A. (1981) FEBS Lett. 131, 253–256
30. Dontsova, O., Kopylov, A., and Brimacombe, R. (1991) EMBO J. 10, 2613–2620
21979
31. 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
32. Barritault, D., Expert-Bezancon, A., Guerin, M. F., and Hayes, D. (1976) Eur.
J. Biochem. 63, 131–135
33. Adams, L. (1995) in Current Protocols in Molecular Biology (Ausubel, F.,
Brent, R., and Kingston, R., eds) pp. 10.3.1–10.3.12, John Wiley & Sons,
Inc., New York
34. Shevchenko, A., Wilm, M., Vorm, O., and Mann, M. (1996) Anal. Chem. 68,
850 – 858
35. Cooperman, B. S., Weitzmann, C. J., and Buck, M. A. (1988) Methods Enzymol.
164, 523–532
36. Schenk, S., and Laddaga, R. A. (1992) FEMS Microbiol. Lett. 94, 133–138
37. Rinke-Appel, J., Junke, N., Stade, K., and Brimacombe, R. (1991) EMBO J. 10,
2195–2202
38. Nissen, P., Hansen, J., Ban, N., Moore, P. B., and Steitz, T. A. (2000) Science
289, 920 –930
39. Wower, J., Hixson, S. S., and Zimmermann, R. A. (1989) Proc. Natl. Acad. Sci.
U. S. A. 86, 5232–5236
40. Wower, J., Wower I. K., Kirillov, S. V., Rosen, K. V., Hixson, S. S., and
Zimmermann, R. A. (1995) Biochem. Cell Biol. 73, 1041–1047
41. Wower, J., Kirillov, S. V., Wower, I. K., Guven, S., Hixson, S. S., and Zimmermann, R. A. (2000) J. Biol. Chem. 275, 37887–37894
42. Sonenberg, N., Wilchek, M., and Zamir, A. (1973) Proc. Natl. Acad. Sci.
U. S. A. 70, 1423–1426
43. Tejedor, F., and Ballesta, J. P. (1985) J. Antimicrob. Chemother. 16, Suppl. A,
53– 62
44. Arévalo, M. A., Tejedor, F., Polo, F., and Ballesta, J. P. (1989) J. Med. Chem.
32, 2200 –2204
45. Bischof, O., Kruft, V., and Wittmann-Liebold, B. (1994) J. Biol. Chem. 269,
18315–18319
46. Nicholson, A. W., Hall, C. C., Strycharz, W. A., and Cooperman, B. S. (1982)
Biochemistry 21, 3797–3808
47. Ban, N., Nissen, P., Hansen, J., Moore, P. B., and Steitz, T. A. (2000) Science
289, 905–920
48. Harms, J., Schluenzen, F., Zarivach, R., Bashan, A., Gat, S., Agmon, I.,
Bartels, H., Franceschi, F., and Yonath, A. (2001) Cell 107, 679 – 688
49. Wower, I. K., Wower, J., and Zimmermann, R. A. (1998) J. Biol. Chem. 273,
19847–19852
50. Champney, W. S., and Miller, M. (2002) Curr. Microbiol. 44, 350 –356
51. Caldon, C. E., Yoong, P., and March, P. E. (2001) Mol. Microbiol. 41, 289 –297
52. Lin, A. H., Murray, R. W., Vidmar, T. J., and Marotti, K. R. (1997) Antimicrob.
Agents Chemother. 41, 2127–2131
53. Monro, R. E., Celma, M. L., and Vazquez, D. (1969) Nature 222, 356 –358
54. Porse, B. T., Kirillov, S. V., Awayez, M. J., Ottenheijm, H. C. J., and Garrett,
R. A. (1999) Proc. Natl. Acad. Sci. U. S. A. 96, 9003–9008
55. Weisblum, B. (1995) Antimicrob. Agents Chemother. 39, 797– 805
56. Moazed, D., and Noller, H. F. (1989) Cell 57, 585–597
57. Mankin, A. S. (1997) J. Mol. Biol. 274, 8 –15
58. Polacek, N., Gomez, M. G., Ito, K., Nakamura, Y., and Mankin, A. S. (2003)
Mol. Cell 11, 103–112
59. Hansen, J. L., Schmeing, T. M., Moore, P. B., and Steitz, T. A. (2002) Proc.
Natl. Acad. Sci. U. S. A. 99, 11670 –11675