Interplay between the ribosomal tunnel, nascent chain, and macrolides influences drug inhibition

Article abstract of DOI:10.1016/j.chembiol.2010.04.008

Accumulating evidence suggests that, during translation, nascent chains can form specific interactions with ribosomal exit tunnel to regulate translation and promote initial folding events. The clinically important macrolide antibiotics bind within the ex

Full text of DOI:10.1016/j.chembiol.2010.04.008

Chemistry & Biology
Article
Interplay between the Ribosomal Tunnel,
Nascent Chain, and Macrolides
Influences Drug Inhibition
Agata L. Starosta,1,2 Viktoriya V. Karpenko,3 Anna V. Shishkina,4 Aleksandra Mikolajka,1,2 Natalia V. Sumbatyan,3
1,2,
Frank Schluenzen,5 Galina A. Korshunova,4 Alexey A. Bogdanov,3,4, and Daniel N. Wilson
*
1Gene Center and Department of Biochemistry
*
2Center for Integrated Protein Science Munich
University of Munich, LMU, Munich D-81377, Germany
3Department of Chemistry
4A.N. Belozersky Institute of Physico-Chemical Biology,
M.V. Lomonosov Moscow State University, Moscow 119992, Russia
5Deutsches Elektronen-Synchrotron, Hamburg D-22603, Germany
*Correspondence: bogdanov@genebee.msu.su (A.A.B.), wilson@lmb.uni-muenchen.de (D.N.W.)
DOI 10.1016/j.chembiol.2010.04.008
SUMMARY
by blocking elongation of the nascent polypeptide chain, which,
in turn, induces peptidyl-tRNA drop-off (Mankin, 2008; Poehls-
Accumulating evidence suggests that, during trans- gaard and Douthwaite, 2003).
Macrolides are a diverse family of antibiotics that contain
lation, nascent chains can form specific interactions
with ribosomal exit tunnel to regulate translation and
promote initial folding events. The clinically impor-
tant macrolide antibiotics bind within the exit tunnel
and inhibit translation by preventing progression
of the nascent chain and inducing peptidyl-tRNA
drop-off. Here, we have synthesized amino acid–
and peptide-containing macrolides, which are used
to demonstrate that distinct amino acids and pep-
tides can establish interaction with components of
a central 12–16-membered lactone ring to which various sugar
substituent’s are attached; for example, erythromycin has a
14-membered ring with C3 cladinose and C5 desosamine
sugars (Figure 1), whereas tylosin has a 16-membered ring with
a C5 mycaminose-mycarose disaccharide as well as a C23
mycinose (Figures 1 and 2). Crystal structures of macrolides
bound to the large ribosomal subunit reveal that the drugs bind
within the ribosomal exit tunnel adjacent to the peptidyltransfer-
ase center (Figures 2A and 2B) (Hansen et al., 2002; Schlu¨ nzen
et al., 2001). Surprisingly, 16-membered macrolides that have a
the ribosomal tunnel and enhance the ribosome- C6 ethyl aldehyde (with the aldehyde functional group at C20),
such as tylosin (Figure 1), form a reversible covalent bond with
the N6 of adenine 2062 (A2062; Escherichia coli numbering
used throughout) of the 23S ribosomal RNA (Figure 2B) (Hansen
binding and inhibitory properties of the macrolide
drugs, consistent with the concept that the exit
tunnel is not simply a Teflon-like channel. Surpris-
ingly, we find that macrolide antibiotics do not inhibit
translation of all nascent chains similarly, but rather
exhibit polypeptide-specific inhibitory effects, pro-
viding a change to our general mechanistic under-
standing of macrolide inhibition.
et al., 2002). The lactone rings of different macrolides are gener-
ally positioned in the tunnel with very similar orientations, such
that the C5 sugars extend up the tunnel toward the PTC (Hansen
et al., 2002; Poulsen et al., 2000; Schlu¨ nzen et al., 2001; Wilson
et al., 2005) (Figure 2B). Consistently, macrolides such as eryth-
romycin with only a single C5 sugar allow synthesis of short
oligopeptides of six to eight amino acids before peptidyl-tRNA
drop-off occurs, whereas macrolides with C5 disaccharides,
INTRODUCTION
such as tylosin, allow only two to four amino acids to be synthe-
sized (Tenson et al., 2003), and can even directly inhibit the pep-
The translational machinery represents one of the major targets tidyltransferase reaction (Karahalios et al., 2006).
within the cell for antibiotics (reviewed by Spahn and Prescott, It should be noted, however, that the length of the pep-
1996; Wilson, 2009). The majority of antibiotics inhibit protein tide synthesized before peptidyl-tRNA drop-off appears to be
synthesis by binding at the active centers on the ribosome; for dependent on the sequence being translated (Tenson et al.,
example, the tetracyclines and aminoglycosides (gentamicin) 2003). Translation of short nascent pentapeptides of defined
bind within the decoding site on the small subunit, whereas the sequence can lead to dissociation of macrolide antibiotics from
phenylpropanoids (chloramphenicol) and oxazolidinones (line- the ribosome and thus confer macrolide resistance (Lovmar
zolid) bind within the A-site of peptidyltransferase center (PTC) et al., 2006; Tenson et al., 1996; Tenson and Mankin, 2001).
to block peptide-bond formation. In contrast, the clinically Moreover, different pentapeptide sequences can confer resis-
important class of macrolide antibiotics bind within the ribo- tance to distinct macrolide members, suggesting a direct and
somal exit tunnel, where they are thought to inhibit translation discriminating interaction between the nascent chain and the
504 Chemistry & Biology 17, 504–514, May 28, 2010 ª2010 Elsevier Ltd All rights reserved

Chemistry & Biology
Macrolide Interaction with the Ribosomal Tunnel
Figure 1. Chemical Structures of Macrolide Antibiotics
Chemical structures of tylosin (Tyl), desmycosin (Des), OMT, erythromycin, and derivatives used in this study (1)–(12) are depicted.
macrolide (Tenson et al., 1997; Tripathi et al., 1998; Vimberg Weisblum, 1989; Ramu et al., 2009; Vazquez-Laslop et al.,
et al., 2004). Inducible expression of some macrolide resistance 2008). Mutations in the ErmC leader peptide as well as within
genes, such as the methyltransferase encoded by the ermC components of the ribosomal exit tunnel, namely A2062 of the
gene, requires macrolide-dependent translational stalling during 23S rRNA, have been identified that relieve the translational stall-
the synthesis of an upstream leader peptide (Mayford and ing (Ramu et al., 2009; Vazquez-Laslop et al., 2008), suggesting
Figure 2. Binding of Macrolide Antibiotics
Within the Ribosomal Tunnel
(A) Transverse section through the large subunit of
the ribosome to visualize the ribosomal exit tunnel.
The position of tylosin (red), P-site tRNA (orange),
and a mock polypeptide chain (yellow) are indi-
cated. CP, central protuberance; rRNA, gray and
ribosomal proteins (dark blue).
(B) Zoom in on the macrolide-binding site within
the ribosomal tunnel. Tylosin (pink) shown with
lactone ring, C23-mycinose, and C5-mycami-
nose-mycarose are labeled. The asterisk indicates
the carbinolamine bond that forms between tylo-
sin and the N6 of 23S rRNA nucleotide A2062,
whereas the arrow indicates the shift in A2062
seen upon ligand binding. The positions of
A- and P-site aminoacyl-tRNAs are shown in green
and orange, respectively.
Chemistry & Biology 17, 504–514, May 28, 2010 ª2010 Elsevier Ltd All rights reserved 505

Chemistry & Biology
Macrolide Interaction with the Ribosomal Tunnel
a complex interplay between the antibiotic, the tunnel, and the agreement with previous findings that modification of the C20
nascent polypeptide chain.
aldehyde group leads to a ꢀ100-fold increase in the MIC value
Although the ribosomal tunnel has for many years been for a variety of gram-positive bacteria (Kirst et al., 1988; Narandja
thought of as a passive conduit for the nascent chain, accumu- et al., 1994; Omura et al., 1982; Omura and Tishler, 1972).
lating evidence indicates the existence of many specific leader
Next, we synthesized a series of macrolide derivatives where
peptides that induce translational stalling by directly interacting amino acids and peptides were attached to the C20 aldehyde
with the ribosomal tunnel, and in doing so regulate translation position to investigate whether amino acids can establish
of a downstream gene (Tenson and Ehrenberg, 2002). Indeed, defined interactions with the ribosomal exit tunnel. We reasoned
recent cryo-EM reconstructions directly visualize interaction that if this is indeed the case, then the peptide-tunnel interac-
between the nascent polypeptide chain and components of tions should improve the binding affinity of the compounds.
the ribosomal tunnel (Becker et al., 2009; Bhushan et al., 2010; As can be seen in Figure 3A, the presence of Ala-Ala at the
Seidelt et al., 2009). Moreover, interaction between the nascent C20 position to generate (2) Tyl-A2 dramatically improves the
chain and tunnel has been proposed to play a more general role ability of the compound to compete with erythromycin for ribo-
in modulating translation rate (Lu and Deutsch, 2008; Seidelt some binding. Similar results were also obtained when Ala-Ala
et al., 2009) as well as influencing initial protein folding events was attached to the C20 position of desmycosin, a degradation
(Bhushan et al., 2010; Kosolapov and Deutsch, 2009; Lu and product of tylosin differing by the absence of a C5 mycarose, to
Deutsch, 2005a, 2005b; Woolhead et al., 2004).
generate (3) Des-A2 (Figure 3A). This finding suggests that,
Here, we demonstrate that abolishing the potential of tylosin to indeed, the Ala-Ala peptides establish interaction with the tunnel
form a covalent bond with A2062 of the ribosome dramatically to improve the binding of the compounds compared with the Tyl-
reduces the binding and inhibitory properties of the drug, but H2; however, it should be noted that the Kd of (2) Tyl-A2 and (3)
that this can be reversed by the addition of amino acid and Des-A2 (ꢀ18 nM for both compounds) are still significantly worse
peptide side chains that enable interaction to be reestablished than their C6 ethyl aldehyde–containing parent compounds tylo-
with the ribosomal exit tunnel. A number of the macrolide- sin (0.65 nM) and desmycosin (0.3 nM) (Table S1; Figure 3A).
peptide compounds are even more potent inhibitors of transla-
To test the inhibitory activity of the compounds, we utilized an
tion than their parent compounds. Surprisingly, we find that E. coli lysate–based, in vitro, coupled transcription-translation
macrolides exhibit nascent chain-specific effects: 5-O-mycami- system (Dinos et al., 2004; Starosta et al., 2009), where the syn-
nosyl-tylonolid (OMT), a precursor of tylosin, has excellent anti- thesis of green fluorescent protein (GFP) was monitored by
microbial activity and is a potent inhibitor of firefly luciferase directly measuring fluorescence (Figure 3B). Consistent with
synthesis. In contrast, OMT is a very poor inhibitor of green the ribosome binding results, tylosin was found to be an excel-
fluorescent protein synthesis and can even protect the transla- lent inhibitor, with an IC50 of 0.25 mM and abolishing translation
tional apparatus from the inhibitory effects of other macrolides, at ꢀ1 mM (Figures 3B and 3C), whereas (1) Tyl-H2 had an IC50
>
such as erythromycin and tylosin. Differential effects are also 103 higher (4 mM) (Table S1) and did not even completely abolish
observed for other macrolide antibiotics but are not seen for non- translation at 25 mM (data not shown). (2) Tyl-A2 exhibited
macrolide antibiotics, such as tetracyclines and chlorampheni- improved inhibitory characteristics, with an IC50 of 0.5 mM—
cols. Collectively, our results demonstrate the direct and specific only slightly worse than that of tylosin (Figure 3C; Table S1).
interaction between peptides and the ribosomal tunnel and This contrasts with the binding assay results where (2) Tyl-A2
suggest that the inhibitory effect of macrolides is dependent was a significantly worse competitor of erythromycin than tylosin
on the sequence of the nascent polypeptide chain being for ribosome binding (Figure 3A). A similar trend was observed
translated.
when comparing desmycosin with (3) Des-A2, in that (3) Des-
A2 was a slightly more effective inhibitor than desmycosin.
However, unlike (2) Tyl-A2, both desmycosin and (3) Des-A2
could not fully abolish translation even at high concentrations
(Figure 3C; Table S1).
RESULTS
Peptide Interaction with the Tunnel Compensates
for Loss of the Covalent Bond
We have previously determined a crystal structure of (2) Tyl-A2
To investigate the effect of the covalent linkage between the C6 bound to the Deinococcus radiodurans large ribosomal subunit
ethyl aldehyde of tylosin and the N6 of A2062 of the 23S rRNA (Wilson et al., 2005), which shows that the lactone ring binds in
(Figure 2B), we synthesized a reduced form of tylosin, (1) a similar position to that seen for tylosin (Hansen et al., 2002).
Tyl-H2, where the C6 ethyl aldehyde is replaced by hydroxyethyl The C20 Ala-Ala peptide protrudes into a small alcove or pocket
group (Figure 1A). Binding of (1) Tyl-H2 to the ribosome was within the tunnel wall, where it stacks against the base of A2062
monitored by competition with radiolabeled [14C]-erythromycin (Figure 3E). The presence of the C20 Ala-Ala peptide on (2) Tyl-
(apparent Kd calculations in Table S1, available online, are based A2 precludes formation of the carbinolamine bond with the N6
on a 2 hr incubation; see Experimental Procedures). In the con- of A2062, and as a result A2062 is in a shifted position compared
trol experiment, we demonstrate that tylosin is an excellent to the tylosin-50S structure (Hansen et al., 2002) (Figure 3D).
competitor, completely abolishing the binding of erythromycin
at ꢀ0.1 mM (Figure 3A). In contrast, (1) Tyl-H2 was totally inactive, Optimized Stacking Interactions with A2062
and no significant loss in erythromycin binding was observed Improves Inhibition
with 1003 higher concentrations (10 mM) of the compound On the basis of the structure of the (2) Tyl-A2 bound to the large
(Figure 3A). These results support the importance of the carbi- subunit, a series of novel C6 tylosin derivatives were synthesized
nolamine covalent linkage for efficient ribosome binding, in that can potentially form better stacking interactions with the
506 Chemistry & Biology 17, 504–514, May 28, 2010 ª2010 Elsevier Ltd All rights reserved

Chemistry & Biology
Macrolide Interaction with the Ribosomal Tunnel
Figure 3. Inhibitory Effect of Tylosin and Desmycosin Ala2 Derivatives
(A) The ability of tylosin (pink), Tyl-H2 (gray), desmycosin (Des, blue), (2) Tyl-A2, and (3) Des-A2 to compete with radiolabeled erythromycin for binding to E. coli
70S ribosomes was determined. The binding of erythromycin in the absence of competing ligand was assigned as 100%.
(B) The amount of GFP produced in an E. coli in vitro transcription-translation assay in the presence or absence of tylosin, (1) Tyl-H2 and (2) Tyl-A2, Des, and (3)
Des-A2 was determined by monitoring fluorescence of GFP. The fluorescence of GFP in the absence of antibiotic was assigned as 100%.
(C) Quantification of (B) with symbols and colors for each compound as in (A).
(D) The binding site of tylosin on the H. marismortui 50S subunit showing the formation of a covalent carbinolamine bond between the C6 ethyl-aldehyde of tylosin
(red) and the N6 of the base of A2062 of the 23S rRNA (Hansen et al., 2002).
(E) The binding site of (2) Tyl-A2 (green) on the D. radiodurans 50S subunit showing the stacking interaction between the Ala-Ala side chain of (2) Tyl-A2 and the
base of A2062 of the 23S rRNA (Wilson et al., 2005).
base of A2062, namely by introducing aromatic or heterocyclic effective than tylosin at competing for ribosome binding with
groups that replace the Ala amino acids. Because base-stacking erythromycin (Figure 3A).
between RNA nucleotides is an efficient and highly ubiquitous
In a separate approach, we also attached carnitine to the C20
interaction mode within RNA particles, the first compound, (4) position to generate (7) Tyl-Car (Figure 1). The positively charged
Tyl-U, was designed with an uridine (U) RNA nucleobase linked side chain improves the aqueous solubility, compared with tylo-
to the lactone C6 carbon by a four-atom linker (Figure 1). In the sin, but was also intended to improve interactions with the highly
translation assay, (4) Tyl-U was still a worse inhibitor than (2) electronegative cleft in the ribosomal tunnel. Indeed, (7) Tyl-Car
Tyl-A2, but was superior to (1) Tyl-H2 (Figure 4A; Table S1). displayed binding (Kd, 0.6 nM) and inhibition (IC50, 0.3) properties
In parallel, we investigated the effect of replacing the Ala-Ala very similar to those of tylosin (Figures 4A and 4B; Table S1), sug-
peptide with the amino acid tyrosine (Tyr) to generate the (5) gesting that, like (2) Tyl-A2, the side chain of (7) Tyl-Car estab-
Tyl=Tyr derivative, which also utilizes
a five-atom linker. lishes interaction with the ribosomal tunnel that compensate
Although this compound had improved binding and inhibitory for the loss of the carbinolamine bond. (7) Tyl-Car was modeled
activity with respect to (1) Tyl-H2, it was still significantly worse in the tunnel with the constraint that the lactone ring and A2062
than (2) Tyl-A2 (Figures 4A and 4B; Table S1). We rationalized were positioned as in (2) Tyl-A2 structure (Wilson et al., 2005).
that the double bond in the linker may restrict the conformational The most energetically favorable solution indicated that the
freedom of the side chain and therefore synthesized a second carnitine side chain adopted a conformation that establishes
compound (6) Tyl-Tyr with a fully saturated linker. In the transla- stacking interactions with A2062 (Figure 4C).
tion assays, (6) Tyl-Tyr was a significantly better inhibitor than (5)
Tyl=Tyr and (2) Tyl-A2 (Figure 4B; Table S1), which was consis- Enhanced Tylosin Antibiotics That Utilize
tent with the improved binding properties of (6) Tyl-Tyr (Fig- Peptide-Tunnel Interaction
ure 4A). What was curious was (6) Tyl-Tyr had an IC50 compa- To further explore the stacking potential of the C20 side chains
rable to that of tylosin (Figure 4B; Table S1), despite being less with the nucleobase of A2062, we decided to introduce
Chemistry & Biology 17, 504–514, May 28, 2010 ª2010 Elsevier Ltd All rights reserved 507

Chemistry & Biology
Macrolide Interaction with the Ribosomal Tunnel
Figure 4. Stacking Interaction with A2062
(A) The ability of tylosin, (1) Tyl-H2, (4) Tyl-U, (5) Tyl=Tyr, (6) Tyl-Tyr, and (7) Tyl-Car to compete with radiolabeled erythromycin for binding to D. radiodurans 70S
ribosomes was determined. The binding of erythromycin in the absence of competing ligand was assigned as 100%.
(B) Quantification of the fluorescence of GFP produced in an E. coli in vitro transcription-translation assay in the presence or absence of tylosin, (1) Tyl-H2, (4)
Tyl-U, (5) Tyl=Tyr, (6) Tyl-Tyr, and (7) Tyl-Car. The fluorescence of GFP in the absence of antibiotic was assigned as 100%.
(C) Model based on Tyl-A2 for the binding site of (7) Tyl-Car (yellow) on the D. radiodurans 50S subunit showing the stacking interaction between the carinitine
sidechain of (7) Tyl-Car and the base of A2062 of the 23S rRNA.
(D) The ability of tylosin, desmycosin (Des), OMT, (8) Tyl-Phe, (9) Des-Phe, and (10) OMT-Phe to compete with radiolabeled erythromycin for binding to
D. radiodurans 70S ribosomes was determined. The binding of erythromycin in the absence of competing ligand was assigned as 100%.
(E) Quantification of the fluorescence of GFP produced in an E. coli in vitro transcription-translation assay in the presence or absence of tylosin, desmycosin (Des),
OMT, (8) Tyl-Phe, (9) Des-Phe, and (10) OMT-Phe. The fluorescence of GFP in the absence of antibiotic was assigned as 100%.
(F) Model based on (2) Tyl-A2 for the binding site of (8) Tyl-Phe (purple) on the D. radiodurans 50S subunit showing the stacking interaction between the linker of (8)
Tyl-Phe and the base of A2062, as well as a between the phenylalanine aromatic residue and the base of C2586 of the 23S rRNA.
a nonpolar phenylalanine (Phe), while extending the linker, but
We also synthesized (9) Des-Phe and (10) OMT-Phe, where the
maintaining its overall flexibility. Binding studies indicate that Phe-containing linker side chain was attached to the C20 posi-
this compound, (8) Tyl-Phe, improved the ribosome bind- tion of desmycosin and OMT, respectively (Figure 1). OMT is
ing affinity (Kd, 3.6 nM) compared to (1) Tyl-H2 (Figure 4D; a precursor of tylosin, which lacks both the C5 mycarose and
Table S1). Consistently, (8) Tyl-Phe was found to be a very the C23 mycinose (Figure 1). Our erythromycin competition
potent inhibitor of translation (Figure 4E), performing slightly studies indicate that both (9) Des-Phe and (10) OMT-Phe were
better than tylosin but significantly better than (2) Tyl-A2 and worse competitors than their respective parent compounds (Fig-
(1) Tyl-H2 (Table S1). The Phe side chain may stabilize the bind- ure 4D). The order of affinities of (8) Tyl-Phe > (9) Des-Phe > (10)
ing of (8) Tyl-Phe by establishing additional stacking interac- OMT-Phe is the same as for tylosin > desmycosin > OMT, and
tions with 23S rRNA nucleotides (Figure 4F). However, the correlates with the predicted surface areas that these
improved ribosome binding affinity cannot be the sole explana- compounds bury on the ribosome; that is, more surface area
tion for the increased translational inhibitory activity relative to buried equals higher binding affinity (Hansen et al., 2002).
tylosin, because (8) Tyl-Phe was significantly less effective However, that each derivative had a lower binding affinity than
compared to tylosin at competing erythromycin from the ribo- the parent compound emphasizes again the important influence
some (Figure 4D).
that the C6-ethyl aldehyde and presumably the carbinolamine
508 Chemistry & Biology 17, 504–514, May 28, 2010 ª2010 Elsevier Ltd All rights reserved

Chemistry & Biology
Macrolide Interaction with the Ribosomal Tunnel
Figure 5. Protective Effects of OMT
Compounds
(A) The ability of tylosin, OMT, (11) BocAAF-OMT,
and (12) fMLF-OMT to compete with radiolabeled
erythromycin for binding to D. radiodurans 70S
ribosomes was determined. The binding of eryth-
romycin in the absence of competing ligand was
assigned as 100%.
(B) Quantification of the fluorescence of GFP
produced in an E. coli in vitro transcription-transla-
tion assay in the presence or absence of tylosin,
OMT, (11) BocAAF-OMT, and (12) fMLF-OMT.
The fluorescence of GFP in the absence of antibi-
otic was assigned as 100%.
(C) Quantification of the fluorescence of GFP
produced in an E. coli in vitro transcription-transla-
tion assay in the presence of increasing concen-
trations (upper x-axis) of erythromycin (Ery) to
5 mM, then with additional presence of increasing
concentrations (lower x-axis) of OMT (cyan), (11)
BocAAF-OMT (magenta), (12) fMLF-OMT (blue),
or (10) OMT-Phe (yellow).
(D) Quantification of the fluorescence of GFP
produced in an E. coli in vitro transcription-transla-
tion assay in the presence of increasing concen-
trations (upper x-axis) of tylosin (Ery) to 5 mM,
then with additional presence of increasing con-
centrations (lower x-axis) of OMT (cyan), (11)
BocAAF-OMT (magenta), (12) fMLF-OMT (blue),
or (10) OMT-Phe (yellow).
bond has on stabilizing the C6-ethyl aldehyde–containing mac- BocAlaAlaPhe and fMetLeuPhe peptides were attached to
rolides on the ribosome. C14(C23) of OMT to generate (11) BocAAF-OMT and (12) fMLF-
In the translation system, (9) Des-Phe and (10) OMT-Phe, like OMT derivatives, respectively. These compounds still retain a C5
(8) Tyl-Phe, were more potent inhibitors than would have been sugar and the C6 ethyl aldehyde, and the placement of the
expected on the basis of their Kd values. (9) Des-Phe and (10) peptide side chain on the C14(C23) position of the lactone ring
OMT-Phe had IC50 values (0.1 mM and 0.3 mM, respectively) orients it such that it would penetrate deeper into the ribosomal
similar to that of tylosin (0.25 mM) (Figure 4E; Table S1), thus indi- tunnel. Like OMT, both (11) BocAAF-OMT and (12) fMLF-OMT
cating that the presence of the peptide and its interaction with were excellent competitors of erythromycin, although the pres-
the tunnel can compensate for loss in binding affinity resulting ence of the peptide did not enhance the affinity of the derivative
from disruption of the covalent linkage with the ribosome. In for the ribosome (Figure 5A). In contrast, the presence of the
contrast, we found unexpectedly that desmycosin and OMT peptide improved the inhibitory effect of the OMT derivatives,
were very poor inhibitors of translation. Even with high concen- particularly, (12) fMLF-OMT, which inhibited the in vitro transla-
trations of antibiotic (50–100 mM), desmycosin never inhibited tion reaction by 50% at ꢀ1 mM (Figure 5B).
translation by more than 50% (Figure 3B), and OMT had little
Because OMT and some of the derivatives were shown to bind
or no significant inhibitory effect (Figure 4E). This finding was efficiently to an empty ribosome, but were poor inhibitors in the
surprising because (1) both OMT and desmycosin bound effi- translation assay, we decided to test whether the compounds
ciently to an empty ribosome (Figure 4D), with Kd values of 3 nM could relieve the translation inhibition effect that results from
and 0.3 nM, respectively (Table S1), similar to that previously other more effective macrolide antibiotics, such as tylosin and
reported (Karahalios et al., 2006) and that is 5- and 50-fold better erythromycin. Figure 5C shows that, at 5 mM erythromycin, the
than that of erythromycin (Kd, 15 nM) (Karahalios et al., 2006); translation of GFP is totally abolished, whereas when OMT,
and (2) both OMT and desmycosin have potent antimicrobial (11) BocAAF-OMT, or (12) fMLF-OMT are additionally titrated
activity (data not shown). Indeed, OMT and desmycosin are into the reaction, the inhibition of translation was gradually alle-
known to have MIC values similar to tylosin for a number of viated. As a control, (10) OMT-Phe, which was shown itself to
gram-positive bacterial strains (e.g., Streptococcus pyogenes, be a potent inhibitor (Figure 4E), could not relieve the inhibitory
Staphlococcus aureus, and Enterococcus faecalis) (Fu et al., effect due to erythromycin (Figure 5C). Indeed, the extent of relief
2006; Kirst et al., 1988; Mutak et al., 2004).
appeared to correlate inversely with the innate inhibitory activity
of the compound; for example, OMT and (11) BocAAF-OMT,
which were the poorest inhibitors, yielded the strongest relief
(up to 40% with 5 mM—that is, equimolar concentration to eryth-
Protective Effects of Peptide-OMT Precursor
Compounds Against Translation Inhibition
To investigate the differential effects of OMT in more detail, romycin), whereas (12) fMLF-OMT, which was a slightly better
we synthesized a series of peptide-OMT compounds, where inhibitor, only restored translation by up to 20% at 5 mM final
Chemistry & Biology 17, 504–514, May 28, 2010 ª2010 Elsevier Ltd All rights reserved 509

Chemistry & Biology
Macrolide Interaction with the Ribosomal Tunnel
Figure 6. Polypeptide Chain-Dependency of Macrolide Antibiotics
(A) Comparison of the effect of desmycosin (Des) and OMT on the in vitro transcription-translation of green fluorescent protein (GFP) and firefly luciferase (Fluc).
(B) Comparison of the effect of tylosin (Tyl) and erythromycin (Ery) on the in vitro transcription-translation of GFP and Fluc.
(C) Comparison of the effect of chloramphenicol (Cam) and tetracycline (Tet) on the in vitro transcription-translation of GFP and Fluc. In (A)–(C), the fluorescence of
GFP or luminescence of Fluc in the absence of antibiotic was assigned as 100%.
concentration (Figure 5C). This latter result was slightly sur- because we do not find such distinct inhibitory effects when
prising, since at 5 mM of (12) fMLF-OMT, the drug itself reduces examining the inhibition of other antibiotic classes, such as the
translation by 60% (to 40%). Similar results were observed when tetracyclines or phenylpropanoids (i.e., chloramphenicol) (Fig-
5 mM tylosin was used instead of erythromycin; however, higher ure 6C). Tetracycline inhibits GFP synthesis with an IC50 of
excesses of OMT and the derivatives were required to obtain the 11 mM and Fluc with an IC50 of 7 mM, and chloramphenicol
equivalent inhibitory relief that was observed with erythromycin inhibits GFP and Fluc with the same IC50 of ꢀ1 mM (Figure 6C).
(Figure 5D). For example, 10 mM (23 tylosin concentration) of
OMT produced only ꢀ30% relief (Figure 5D). This trend is consis-
DISCUSSION
tent with the higher affinity and IC50 of tylosin over erythromycin
(Table S1) and thus supports the idea that the relief is due to
direct competition between the OMT/OMT-derivative and the
inhibitory compound (tylosin or erythromycin). Additional evi-
dence includes the observations that (1) OMT/OMT-derivatives
could not relieve the translation inhibition resulting from nonma-
crolide antibiotics, such as tetracycline or thiostrepton (data not
shown), and (2) compounds that were very poor competitors
with erythromycin, such as (1) Tyl-H2 (Figure 3A), produced
very low inhibition relief with extremely high excesses of com-
pound—for example, 25 mM (53 excess) (1) Tyl-H2 produced
4% relief for 5 mM tylosin (data not shown).
The surface of the exit tunnel is largely hydrophilic and has no
patches of hydrophobic surface large enough to form a sig-
nificant binding site for hydrophobic sequences in nascent
polypeptide chains (Nissen et al., 2000). Although this general
‘‘nonstick’’ characteristic of the exit tunnel may hold for the
majority of polypeptides being synthesized by the ribosome,
growing evidence indicates that particular nascent chain
sequences interact with the ribosomal tunnel during their egres-
sion (Becker et al., 2009; Bhushan et al., 2010; Cruz-Vera et al.,
2005; Cruz-Vera and Yanofsky, 2008; Gong and Yanofsky, 2002;
Kosolapov and Deutsch, 2009; Lawrence et al., 2008; Lu and
Deutsch, 2005a, 2005b, 2008; Nakatogawa and Ito, 2002; Seid-
elt et al., 2009; Tenson and Ehrenberg, 2002; Vazquez-Laslop
et al., 2008; Woolhead et al., 2004). Here, we have utilized mac-
Macrolide Antibiotics Exhibit Polypeptide-Specific
Inhibitory Effects
Because of the inconsistency between the potent antimicrobial rolide antibiotics to direct amino acid and peptide sequences to
activity of OMT and the poor inhibitory activity in the in vitro bind and interact with the ribosomal tunnel. Our results indicate
translation assay, we decided to test whether OMT displayed that specific amino acid and peptide sequences can indeed
some template-specific effects. To do this, we substituted the establish distinct interactions with components of the ribosomal
GFP template with firefly luciferase (Fluc) and monitored trans- tunnel, such as stacking interactions between the backbone and
lation of Fluc using luminescence. Figure 6A shows that OMT aromatic moieties with nucleotide A2062 of the 23S rRNA. A2062
is an excellent inhibitor of translation of Fluc, with an IC50 of is a universally conserved nucleotide that appears to adopt dif-
0.15 mM, similar to that observed for the inhibition of GFP syn- ferent conformations within the ribosomal tunnel dependent
thesis by tylosin (0.25 mM). Similarly, we found that Des, which we upon the functional state of the ribosome. In the crystal structure
had shown to be a poor inhibitor of GFP (Figure 3C), was a potent of the apo Haloarcula marismortui 50S subunit, A2062 lies flat
inhibitor of Fluc synthesis (Figure 6A). Less dramatic differences against the tunnel wall, whereas when P-site ligands are bound
were observed when comparing the effect of tylosin and erythro- A2062 shifts position to extend into the lumen of the ribosomal
mycin on GFP and Fluc synthesis (Figure 6B). The effect of tylo- tunnel (Figure 2B). It appears that A2062 also establishes inter-
sin on GFP and Fluc synthesis was similar, whereas the corre- action with the nascent polypeptide chains, as observed by
sponding values for erythromycin were 1 mM (GFP) and 0.75 mM cryo-EM of stalled translating ribosomes (Bhushan et al., 2010;
(Fluc) (Table S1). We believe that this template dependency in Seidelt et al., 2009). Mutations of A2062 relieve the translational
inhibition is specific for the macrolide class of antibiotics, arrest resulting during synthesis of the ErmC leader peptide,
510 Chemistry & Biology 17, 504–514, May 28, 2010 ª2010 Elsevier Ltd All rights reserved

Chemistry & Biology
Macrolide Interaction with the Ribosomal Tunnel
Figure 7. Factors Influencing Macrolide Inhibition
(A) Schematic view of a transverse section through the large ribosomal subunit revealing a P-site tRNA (yellow) attached to a polypeptide chain (orange) in the
tunnel exit. The nucleotide A2062 (blue) has been proposed to interact with some polypeptide chains (Ramu et al., 2009; Vazquez-Laslop et al., 2008).
(B) Tylosin (red) binds within the tunnel where it forms a covalent bond with A2062 of the 23S rRNA and allows only translation of short peptides.
(C) Reducing the C6 ethylaldehyde of tylosin to produce (1) Tyl-H2 prevents the covalent interaction with A2062, reducing the binding affinity and allowing trans-
lation in the presence of the drug.
(D) Addition of amino acids or peptides to the C20 position prevents the covalent interaction with A2062, but establishes new compensatory interaction with the
tunnel that increases the binding affinity and thus allows only translation of short peptides.
(E) OMT (cyan) binds within the tunnel where it forms a covalent bond with A2062 of the 23S rRNA and allows only translation of short peptides of the Fluc protein.
(F) Translation of short pentapeptide sequences confers resistance to macrolide antibiotics, such as erythromycin (Tenson et al., 1996). The peptide is thought to
interact with the drug and remove it from the binding site when the peptide is released from the tRNA (Tenson and Mankin, 2001).
(G) Translation of GFP is not inhibited by OMT (cyan) either because the polypeptide chain removes the drug from its binding site or can pass the drug
undisturbed.
suggesting that this nucleotide might be involved in monitoring modes, kinetics, and formation of the carbinolamine interaction
the nascent chain as it passes through the tunnel (Figure 7A) with A2062. Nevertheless, we could show that the loss of the
(Vazquez-Laslop et al., 2008).
covalent bond can be compensated for by addition of amino
A2062 is also involved in the interaction of 16-membered mac- acid or peptide side chains to the C6(C23) position of the drug
rolides, such as tylosin, with the ribosome because it forms (Figure 7D). On the basis of the observation that (2) Tyl-A2 stacks
a covalent bond with the drug. Here we show that reducing the upon the nucleobase of A2062 (Figures 3E and 7D), a series of
C6-ethyl aldehyde of tylosin (Tyl-H2), which abolishes the poten- compounds was designed to explore different modes of interac-
tial to form a covalent linkage with the N6 of A2062, dramatically tions with this region of the ribosome, culminating in a series of
reduces the Kd value by >100-fold (Table S1). The corresponding compounds—(6) Tyl-Tyr, (7) Tyl-Car, (8) Tyl-Phe, (9) Des-Phe,
loss in effectiveness of this antibiotic in translation inhibition is and (10) OMT-Phe—that were similar or better inhibitors of trans-
seen by a reduction in IC50 by >10-fold (Table S1), suggesting lation compared to tylosin (Figure 4; Table S1). These results also
that the nascent polypeptide chain can now expel the drug demonstrate the potential to use the ribosome antibiotic struc-
from the ribosome (Figure 7C). It should be noted, however, tures as a basis for rational design of novel derivatives of antimi-
that the presence of a C6-ethyl aldehyde is not strictly necessary crobial agents.
for efficient binding and potent inhibitory activity because eryth-
One observation with respect to many of the macrolide-
romycin has a C6-hydroxyl and thus cannot and does not form peptide derivatives is the lack of correlation between the Kd
the carbinolamine bond (Tu et al., 2005). The differences in orien- and IC50 values, namely that all of the macrolide-peptide deriva-
tation (Hansen et al., 2002; Tu et al., 2005; Wilson et al., 2005) tives are less effective at competing erythromycin from the ribo-
and kinetics (Petropoulos et al., 2009) of binding between some than are tylosin, but are relatively better inhibitors of trans-
14-membered macrolides, such as erythromycin, and 16- lation. This suggests that the binding affinity per se does not
membered macrolides, such as tylosin, seems to obviate the solely determine the effectiveness of the compound and that
C6-ethyl aldehyde in the former case. It would be interesting in the C20-side chains contribute an additional factor, perhaps by
the future to analyze the characteristics of C6-ethyl aldehyde sterically blocking the path of the nascent polypeptide chain
containing 14-membered macrolides, with respect to binding (Figure 7D). The most dramatic example of this is seen for
Chemistry & Biology 17, 504–514, May 28, 2010 ª2010 Elsevier Ltd All rights reserved 511

Chemistry & Biology
Macrolide Interaction with the Ribosomal Tunnel
OMT and (10) OMT-Phe, because OMT has a Kd value ꢀ93 The proposed model suggests that the translation of a particular
lower than that of (10) OMT-Phe (Table S1), yet (10) OMT-Phe subset of proteins in vivo may not be inhibited by macrolide anti-
is a vastly superior inhibitor of GFP synthesis (Figure 4E). The biotics, which could potentially provide the cell an additional
ability of OMT and the peptide-OMT derivatives, (11) BocAAF- regulatory mechanism, for example, to up-regulate relevant
OMT and (12) fMLF-OMT to restore translation in the presence resistance genes. Future works needs to address the ubiquity
of inhibitory concentrations of erythromycin or tylosin (Figures of this ‘‘translating nascent chain-mediated macrolide resis-
5C and 5D) is consistent with the binding data (Figures 4D and tance’’ and to determine whether specific consensus sequences
5A), supporting the conclusion that these compounds do bind exist for distinct macrolides.
to the ribosome, and thus are likely to reverse the effect of eryth-
romycin and tylosin by competing the drugs from their binding
site in tunnel.
EXPERIMENTAL PROCEDURES
The lack of effect of OMT on GFP synthesis (Figures 5E and
6A) was surprising because OMT is known to be a potent antimi-
crobial agent that targets the translation machinery (Fu et al.,
2006; Kirst et al., 1988; Mutak et al., 2004). This finding prompted
us to test the effect of macrolides on translation of Fluc, which
we could monitor using luminescence. OMT was found to be a
potent inhibitor of Fluc synthesis (Figure 7E), as was tylosin
(Figure 6A). Similarly, we found differential effects for desmyco-
sin, which was a poor inhibitor of GFP but an excellent inhibitor of
Fluc synthesis (Figure 6A). In contrast, few significant differences
were observed between the inhibition of GFP and Fluc synthesis
by other macrolides, such as tylosin (Figure 6B), erythromycin,
(Figure 6C), nor for other classes of antibiotics, such as chloram-
phenicols and tetracyclines (Figure 6C). We also generated
chimeric Fluc-GFP templates where the first 5–15 amino acids
of Fluc were placed at the N terminus of GFP and vice versa.
Unfortunately, the efficiency of translation of these proteins ap-
peared to be dramatically affected, such that the interpretation
of the results was inconclusive (data not shown).
Synthesis of Peptide Derivatives of Macrolides
(1) Tyl-H2 was prepared by reduction of Tyl with sodium borohydride (Kirst
et al., 1988). (2) Tyl-A2 and (3) Des-A2 were synthesized as described else-
where (Sumbatyan et al., 2003). (8) Tyl-Phe, (9) Des-Phe, and (10) OMT-Phe
were prepared by analogy with (2) Tyl-A2 and (3) Des-A2 (Sumbatyan et al.,
2003) by the condensation of 2-aminooxyacetylphenylalanyl ethyl ester with
Tyl (Des or OMT) in 0.4 M Na-acetate buffer (pH 4.7) at 50ꢁC (Sumbatyan
et al., 2010). (4) Tyl-U and (5) Tyl = Tyr were prepared by coupling of Tyl with
3-(uracyl-1-yl)alanine (or tyrosine) in methanol in the presence of sodium meth-
ylate at room temperature. (5) Tyl = Tyr was reduced with sodium borohydride
to give (6) Tyl-Tyr (Sumbatyan et al., 2010). The compounds were purified by
the column chromatography on silica gel, and their homogeneity and structure
were confirmed by TLC, HPLC, NMR, and mass spectra. BocAAF-OMT and
fMLF-OMT were synthesized as described elsewhere (Korshunova et al.,
2007). (7) Tyl-Car was synthesized starting from tylosin (24 mg, 0.026 mmol,
1 eq.) and carnitine hydrazide (Oka et al., 1963) (8 mg, 0.045 mmol, 1.7 eq.),
which were dissolved in Na-acetate buffer (pH 4.7) (1.5 ml), and the mixture
was stirred for 12 hr at 50ꢁC. The crude product was extracted with chloroform
and purified by silica gel column chromatography in chloroform-methanol (3:1,
v/v) system to give (7) Tyl-Car as white crystals: the yield, 42% (11.1 mg, 0.011
mmol); TLC: Rf (CHCl3 - CH3OH, 3:1) 0;7, Rf (CHCl3 - CH3OH – CH3COOH,
6:1:0.1) 0;25, Rf (CHCl3 - CH3OH, 6:1) 0;28, HPLC: tR = 19.7 min (gradient of
CH3CN in 0,1% CF3COOH 0%–60%); MALDI MS: m/z calculated. for
C53H93N4O18 1073.7; found 1070.6; NMR 1H (600 MHz, 303K, DMSO): 0.86
(3H, t, H17), 0.94 (3H, d, H18), 1.08 (3H, d, H5’’CH3), 1.15 (3H, s, H3’’CH3),
1.23 (3H, d, H21), 1.25 (3H, d, H50CH3), 1.28 (3H, d, H50’’CH3), 1.58 (1H, m,
H16), 1.65 (2H, m, H4, H2’’), 1.80 (3H, s, H22), 1.82 (4H, m, H7, H16, H2,’’
Ha Car), 1.90 (3H, m, H2, H7, H16), 1.95 (1H, m, H2), 2.48 (6H, 1H, s, m,
N(CH3)3, H30), 2.77 (2H, m, H6, H19), 2.88 (1H, m, H14), 2.98 (1H, dd, H20’’),
3.13 (m, 4H, H8, H40, H50), 3.38 (1H, m, H20), 3.39 (9H, s, N+(CH3)3Car), 3.41
(2H, m, H5,’’ H3), 3.44 (3H, s, H20’’OCH3), 3.50 (3H, s, H30’’OCH3), 3.52 (3H,
m, Hg Car, H23), 3.61 (3H, m, H30’’, H5, H50’’), 3.80 (1H, m, H23), 4.12 (3H, m,
H10, H4,’’ Hb Car), 4.38 (1H, m, H40’’), 4.47 (1H, d, H10’’), 4.95 (3H, m, H15,
H1,’’ NHCar), 5.82 (1H, d, H13), 6.49 (1H, d, H10), 7.10 (1H, d, H11), 8.16
(1H, s, H20).
Collectively, the data suggest that the sequence of the
nascent chain may influence the inhibitory ability of particular
macrolide antibiotics. Precedents for this include the observa-
tion that translation of specific pentapeptides can confer resis-
tance to macrolide antibiotics by chasing the drug from the ribo-
some (Figure 7F). Distinct pentapeptides sequences are specific
for particular macrolide or ketolide members, leading to the
proposal of a direct interaction between the nascent chain and
the drug in the ribosomal tunnel (Lovmar et al., 2006; Tenson
et al., 1996; Tenson and Mankin, 2001; Tenson et al., 1997;
Tripathi et al., 1998; Vimberg et al., 2004). By analogy, we
propose that the GFP nascent polypeptide chain establishes
specific interactions with OMT, leading to its removal from the
ribosome (Figure 7G).
In Vitro Transcription-Translation Assay
All coupled transcription-translation experiments were performed using an
E. coli lysate–based system in the presence and absence of antibiotics, as
described elsewhere (Dinos et al., 2004; Starosta et al., 2009; Szaflarski
et al., 2008). Two microliters of each reaction was diluted with 50 mL of buffer
A (10 mM HEPES/KOH [pH 7.8], 10 mM MgCl2, 60 mM NH4Cl, and 4 mM
b-mercaptoethanol), mixed, and then transferred into black 96-well chimney
flat bottom microtiter plates. The GFP fluorescence was either on a Tecan
Infiniteꢀ M1000 with an excitation wavelength of 395 nm and emission
509 nm, or measured at 520 nm (filter cutoff) with a Typhoon Scanner 9400
(Amersham Bioscience) using a Typhoon blue laser module (excited at
488 nm). Images were then quantified using ImageQuantTL (GE Healthcare)
and were represented graphical using SigmaPlot (Systat Software, Inc.).
Synthesis of firefly luciferase (Fluc) was performed as described above for
GFP, but using pIVEX-2.3MCS with Fluc cloned into the NdeI and SacI restric-
tion sites as a template. After incubation at 30ꢁC with shaking for 4–5 hr, 2 mL of
each reaction was added directly to white 96-well chimney flat bottom micro-
titer plates, after which 50 mL of luminol substrate (Promega) was quickly
added, shaken for 3 s, and the luminescence was immediately detected using
Because macrolide antibiotics cannot bind to and inhibit
translating ribosomes (Contreras and Vazquez, 1977; Tai et al.,
1974), we do not believe that the differential inhibitory effects
of OMT on GFP and Fluc synthesis result from the ability of the
GFP, but not Fluc, nascent polypeptide chain to pass the drug
in the exit tunnel, but we cannot totally exclude this possibility.
Macrolide antibiotics have been shown to allow synthesis of up
to six to eight amino acids before peptidyl-tRNA drop-off occurs
(Tenson et al., 2003). This may hint that the N-terminal six to eight
residues of translating nascent chains, in this case GFP, inter-
acts with drug (OMT) and removes it from the tunnel by con-
tinued translation of the nascent chain (Figure 7G) as suggested
previously (Mankin, 2008), in contrast to drug removal via pep-
tidyl-tRNA hydrolysis, as occurs in the case of the macrolide-
resistance peptides (Figure 7F) (Tenson and Ehrenberg, 2002).
512 Chemistry & Biology 17, 504–514, May 28, 2010 ª2010 Elsevier Ltd All rights reserved

Chemistry & Biology
Macrolide Interaction with the Ribosomal Tunnel
a Tecan Infiniteꢀ M1000. All measurements were repeated in triplicate and
had a standard deviation of less than 10%.
Contreras, A., and Vazquez, D. (1977). Cooperative and antagonistic interac-
tions of peptidyl-tRNA and antibiotics with bacterial ribosomes. Eur. J. Bio-
chem. 74, 539–547.
Cruz-Vera, L.R., and Yanofsky, C. (2008). Conserved residues Asp16 and
Pro24 of TnaC-tRNAPro participate in tryptophan induction of Tna operon
expression. J. Bacteriol. 190, 4791–4797.
Binding Assay
Binding of all compounds to empty ribosomes was examined using a compe-
tition assay with radiolabeled [14C]Erythromycin (Perkin Elmer), as described
elsewhere (Karahalios et al., 2006; Petropoulos et al., 2009). Briefly, all
reactions contained 0.25 mM D. radiodurans 70S ribosomes and 1.25 mM
Cruz-Vera, L.R., Rajagopal, S., Squires, C., and Yanofsky, C. (2005). Features
of ribosome-peptidyl-tRNA interactions essential for tryptophan induction of
tna operon expression. Mol. Cell 19, 333–343.
[
14C]erythromycin in binding buffer (10 mM HEPES/KOH [pH 7.8], 30 mM
MgCl2, 150 mM NH4Cl, and 6 mM b-mercaptoethanol), which equated with
60% binding from the saturation curve (data not shown). The saturation curve
indicated that 80% of ribosomes were active (data not shown). To measure the
apparent Kd value for each of the compound, reactions were performed in the
absence or presence of increasing concentrations of the competing com-
pounds. After incubation at room temperature for 2 hr, reactions were passed
through nitrocellulose filters, type HA, 0.45 mm pore size (Millipore). Filters were
washed three times with binding buffer and then scintillation counted in the
presence of Filtersafe (Zinsser Analytic) scintillant. All measurements were
repeated in duplicate and had a standard deviation of 5%–10%.
Dinos, G., Wilson, D.N., Teraoka, Y., Szaflarski, W., Fucini, P., Kalpaxis, D., and
Nierhaus, K.H. (2004). Dissecting the ribosomal inhibition mechanisms of
edeine and pactamycin: the universally conserved residues G693 and C795
regulate P-site tRNA binding. Mol. Cell 13, 113–124.
Fu, H., Marquez, S., Gu, X., Katz, L., and Myles, D.C. (2006). Synthesis and
in vitro antibiotic activity of 16-membered 9-O-arylalkyloxime macrolides.
Bioorg. Med. Chem. Lett. 16, 1259–1266.
Gong, F., and Yanofsky, C. (2002). Instruction of translating ribosome by
nascent peptide. Science 297, 1864–1867.
Hansen, J.L., Ippolito, J.A., Ban, N., Nissen, P., Moore, P.B., and Steitz, T.A.
(2002). The structures of four macrolide antibiotics bound to the large ribo-
somal subunit. Mol. Cell 10, 117–128.
Modeling and Figure Preparation
All chemical structures were drawn with ChemSketch and then were modified
with Adobe Illustrator. Small molecule structures for macrolide-peptide deriv-
atives were generated using ChemSketch. The lactone rings were then
replaced with the lactone ring from D. radiodurans 50S subunit in complex
with (2) Tyl-A2 (Wilson et al., 2005), and the derivative sidechains were
modeled on the basis of the position of the Ala-Ala sidechain of (2) Tyl-A2.
The models of the macrolide-peptide compounds in complex with the large
subunit were then minimized using CNS (Bru¨ nger et al., 1998). All 3D structural
figures were produced using PyMol (http://www.pymol.org).
Karahalios, P., Kalpaxis, D.L., Fu, H., Katz, L., Wilson, D.N., and Dinos, G.P.
(2006). On the mechanism of action of 9-O-arylalkyloxime derivatives of
6-O-mycaminosyltylonolide, a new class of 16-membered macrolide antibi-
otics. Mol. Pharmacol. 70, 1271–1280.
Kirst, H.A., Toth, J.E., Debono, M., Willard, K.E., Truedell, B.A., Ott, J.L.,
Counter, F.T., Felty-Duckworth, A.M., and Pekarek, R.S. (1988). Synthesis
and evaluation of tylosin-related macrolides modified at the aldehyde function:
a new series of orally effective antibiotics. J. Med. Chem. 31, 1631–1641.
Korshunova, G.A., Sumbatyan, N.V., Fedorova, N.V., Kuznetsova, I.V., Shish-
kina, A.V., and Bogdanov, A.A. (2007). Peptide derivatives of tylosin-related
macrolides [in Russian]. Bioorg. Khim. 33, 235–244.
SUPPLEMENTAL INFORMATION
Supplemental information includes one table and can be found with this article
online at doi:10.1016/j.chembiol.2010.04.008.
Kosolapov, A., and Deutsch, C. (2009). Tertiary interactions within the ribo-
somal exit tunnel. Nat. Struct. Mol. Biol. 16, 405–411.
Lawrence, M.G., Lindahl, L., and Zengel, J.M. (2008). Effects on translation
pausing of alterations in protein and RNA components of the ribosome exit
tunnel. J. Bacteriol. 190, 5862–5869.
ACKNOWLEDGMENTS
We thank Shura Mankin and Roland Beckmann for helpful discussions, and
the Beckmann laboratory for generous support and creating a stimulating
work environment. This study was financed by the Deutsche Forschungsge-
meinschaft (grant WI3285/1-1 to D.N.W.) and the Russian Foundation for Basic
Research (grant 07-04-00902a to G.A.K.). V.K. would like to acknowledge the
FEBS Fellowship Committee for granting her a Collaborative Experimental
Scholarship for Central and Eastern Europe. A.A.B. would like to thank the
Alexander von Humboldt Foundation for a very generous support.
Lovmar, M., Nilsson, K., Vimberg, V., Tenson, T., Nervall, M., and Ehrenberg,
M. (2006). The molecular mechanism of peptide-mediated erythromycin resis-
tance. J. Biol. Chem. 281, 6742–6750.
Lu, J., and Deutsch, C. (2005a). Folding zones inside the ribosomal exit tunnel.
Nat. Struct. Mol. Biol. 12, 1123–1129.
Lu, J., and Deutsch, C. (2005b). Secondary structure formation of a transmem-
brane segment in Kv channels. Biochemistry 44, 8230–8243.
Lu, J., and Deutsch, C. (2008). Electrostatics in the ribosomal tunnel modulate
Received: February 23, 2010
Revised: March 22, 2010
Accepted: April 2, 2010
Published: May 27, 2010
chain elongation rates. J. Mol. Biol. 384, 73–86.
Mankin, A.S. (2008). Macrolide myths. Curr. Opin. Microbiol. 11, 414–421.
Mayford, M., and Weisblum, B. (1989). ermC leader peptide: amino acid
sequence critical for induction by translational attenuation. J. Mol. Biol. 206,
69–79.
REFERENCES
Mutak, S., Marsic, N., Kramaric, M.D., and Pavlovic, D. (2004). Semisynthetic
macrolide antibacterials derived from tylosin: synthesis and structure-activity
relationships of novel desmycosin analogues. J. Med. Chem. 47, 411–431.
Becker, T., Bhushan, S., Jarasch, A., Armache, J.P., Funes, S., Jossinet, F.,
Gumbart, J., Mielke, T., Berninghausen, O., Schulten, K., et al. (2009). Struc-
ture of monomeric yeast and mammalian Sec61 complexes interacting with
the translating ribosome. Science 326, 1369–1373.
Nakatogawa, H., and Ito, K. (2002). The ribosomal exit tunnel functions as
a discriminating gate. Cell 108, 629–636.
Narandja, A., Suskovic, B., Kelneric, Z., and Djokic, S. (1994). Structure-
activity relationship among polyhydro derivatives of tylosin. J. Antibiot. (Tokyo)
47, 581–587.
Bhushan, S., Gartmann, M., Halic, M., Armache, J.P., Jarasch, A., Mielke, T.,
Berninghausen, O., Wilson, D.N., and Beckmann, R. (2010). a-helical nascent
polypeptide chains visualized within distinct regions of the ribosomal exit
tunnel. Nat. Struct. Mol. Biol. 17, 313–317.
Nissen, P., Hansen, J., Ban, N., Moore, P.B., and Steitz, T.A. (2000). The struc-
tural basis of ribosome activity in peptide bond synthesis. Science 289,
920–930.
Bru¨ nger, A.T., Adams, P.D., Clore, G.M., DeLano, W.L., Gros, P., Grosse-
Kunstleve, R.W., Jiang, J.S., Kuszewski, J., Nilges, M., Pannu, N.S., et al.
(1998). Crystallography & NMR system: a new software suite for macromolec-
ular structure determination. Acta Crystallogr. D Biol. Crystallogr. 54, 905–921.
Oka, Y., Yonemoto, H., and Yurugi, S. (1963). Carnitine derivatives. Takeda
Kenkyusho Nenpo 22, 13–18.
Chemistry & Biology 17, 504–514, May 28, 2010 ª2010 Elsevier Ltd All rights reserved 513

Chemistry & Biology
Macrolide Interaction with the Ribosomal Tunnel
Omura, S., and Tishler, M. (1972). Relationship of structures and microbiolog-
Szaflarski, W., Vesper, O., Teraoka, Y., Plitta, B., Wilson, D.N., and Nierhaus,
K.H. (2008). New features of the ribosome and ribosomal inhibitors: non-enzy-
matic recycling, misreading and back-translocation. J. Mol. Biol. 380,
193–205.
ical activities of the 16-membered macrolides. J. Med. Chem. 15, 1011–1015.
Omura, S., Miyano, K., Matsubara, H., and Nakagawa, A. (1982). Novel dimeric
derivatives of leucomycins and tylosin, sixteen-membered macrolides.
J. Med. Chem. 25, 271–275.
Tai, P.-C., Wallace, B.J., and Davis, B.D. (1974). Selective action of erythro-
mycin on initiating ribosomes. Biochemistry 13, 4653–4659.
Petropoulos, A.D., Kouvela, E.C., Starosta, A.L., Wilson, D.N., Dinos, G.P., and
Kalpaxis, D.L. (2009). Time-resolved binding of azithromycin to Escherichia
coli ribosomes. J. Mol. Biol. 385, 1179–1192.
Tenson, T., and Ehrenberg, M. (2002). Regulatory nascent peptides in the ribo-
somal tunnel. Cell 108, 591–594.
Poehlsgaard, J., and Douthwaite, S. (2003). Macrolide antibiotic interaction
and resistance on the bacterial ribosome. Curr. Opin. Investig. Drugs 4,
140–148.
Tenson, T., and Mankin, A. (2001). Short peptides conferring resistance to
macrolide antibiotics. Peptides 22, 1661–1668.
Tenson, T., Deblasio, A., and Mankin, A. (1996). A functional peptide encoded
Poulsen, S.M., Kofoed, C., and Vester, B. (2000). Inhibition of the ribosomal
peptidyl transferase reaction by the mycarose moiety of the antibiotics carbo-
mycin, spiramycin and tylosin. J. Mol. Biol. 304, 471–481.
in the Escherichia coli 23S rRNA. Proc. Natl. Acad. Sci. USA 93, 5641–5646.
Tenson, T., Xiong, L.Q., Kloss, P., and Mankin, A.S. (1997). Erythromycin resis-
tance peptides selected from random peptide libraries. J. Biol. Chem. 272,
17425–17430.
Ramu, H., Mankin, A., and Vazquez-Laslop, N. (2009). Programmed drug-
dependent ribosome stalling. Mol. Microbiol. 71, 811–824.
Tenson, T., Lovmar, M., and Ehrenberg, M. (2003). The mechanism of action of
macrolides, lincosamides and streptogramin B reveals the nascent peptide
exit path in the ribosome. J. Mol. Biol. 330, 1005–1014.
Schlu¨ nzen, F., Zarivach, R., Harms, J., Bashan, A., Tocilj, A., Albrecht, R.,
Yonath, A., and Franceschi, F. (2001). Structural basis for the interaction of
antibiotics with the peptidyl transferase centre in eubacteria. Nature 413,
814–821.
Tripathi, S., Kloss, P.S., and Mankin, A.S. (1998). Ketolide resistance conferred
by short peptides. J. Biol. Chem. 273, 20073–20077.
Seidelt, B., Innis, C.A., Wilson, D.N., Gartmann, M., Armache, J.P., Villa, E.,
Trabuco, L.G., Becker, T., Mielke, T., Schulten, K., et al. (2009). Structural
insight into nascent polypeptide chain-mediated translational stalling. Science
326, 1412–1415.
Tu, D., Blaha, G., Moore, P., and Steitz, T. (2005). Structures of MLSBK antibi-
otics bound to mutated large ribosomal subunits provide a structural explana-
tion for resistance. Cell 121, 257–270.
Vazquez-Laslop, N., Thum, C., and Mankin, A.S. (2008). Molecular mechanism
Spahn, C.M.T., and Prescott, C.D. (1996). Throwing a spanner in the works:
of drug-dependent ribosome stalling. Mol. Cell 30, 190–202.
antibiotics and the translational apparatus. J. Mol. Med. 74, 423–439.
Vimberg, V., Xiong, L., Bailey, M., Tenson, T., and Mankin, A. (2004). Peptide-
mediated macrolide resistance reveals possible specific interactions in the
nascent peptide exit tunnel. Mol. Microbiol. 54, 376–385.
Starosta, A.L., Qin, H., Mikolajka, A., Leung, G.Y., Schwinghammer, K., Nico-
laou, K.C., Chen, D.Y., Cooperman, B.S., and Wilson, D.N. (2009). Identifica-
tion of distinct thiopeptide-antibiotic precursor lead compounds using transla-
tion machinery assays. Chem. Biol. 16, 1087–1096.
Wilson, D.N. (2009). The A-Z of bacterial translation inhibitors. Crit. Rev. Bio-
chem. Mol. Biol. 44, 393–433.
Sumbatyan, N.V., Korshunova, G.A., and Bogdanov, A.A. (2003). Peptide
derivatives of antibiotics tylosin and desmycosin, protein synthesis inhibitors.
Biochemistry (Mosc.) 68, 1156–1158.
Wilson, D.N., Harms, J.M., Nierhaus, K.H., Schlu¨ nzen, F., and Fucini, P. (2005).
Species-specific antibiotic-ribosome interactions: Implications for drug devel-
opment. Biol. Chem. 386, 1239–1252.
Sumbatyan, N.V., Kuznetsova, I.V., Karpenko, V.V., Federova, N.V., Chertkov,
V.A., Korshunova, G.V., and Bogdanov, A.A. (2010). Amino acid and peptide
derivatives of the tylosin family of antibiotics modified by aldehyde function.
Russ. J. Bioorganic Chem. 36, 245–256.
Woolhead, C.A., McCormick, P.J., and Johnson, A.E. (2004). Nascent
membrane and secretory proteins differ in FRET-detected folding far inside
the ribosome and in their exposure to ribosomal proteins. Cell 116, 725–736.
514 Chemistry & Biology 17, 504–514, May 28, 2010 ª2010 Elsevier Ltd All rights reserved