Molecular determinants of microbial resistance to thiopeptide antibiotics

Article abstract of DOI:10.1021/ja909317n

Ribosomally produced thiopeptide antibiotics are highly promising lead compounds targeting the GTPase-associated region (GAR) of the bacterial ribosome. A representative panel of GAR mutants suspected to confer resistance against thiopeptide antibiotics w

Full text of DOI:10.1021/ja909317n

Published on Web 05/04/2010
Molecular Determinants of Microbial Resistance to
Thiopeptide Antibiotics
Sascha Baumann,†,‡ Sebastian Schoof,†,‡ Marcel Bolten,† Claudia Haering,†
Motoki Takagi,§ Kazuo Shin-ya,⊥ and Hans-Dieter Arndt*,†,‡
Fakulta¨t Chemie, Technische UniVersita¨t Dortmund, Otto-Hahn-Strasse 6, D-44221 Dortmund,
Germany, Max-Planck-Institut fu¨r Molekulare Physiologie, Otto-Hahn-Strasse 11, D-44227
Dortmund, Germany, and Biomedicinal Information Research Center (BIRC), 2-4-7 Aomi,
Koto-ku, Tokyo 135-0064, Japan
Received November 3, 2009; E-mail: hans-dieter.arndt@mpi-dortmund.mpg.de
Abstract: Ribosomally produced thiopeptide antibiotics are highly promising lead compounds targeting
the GTPase-associated region (GAR) of the bacterial ribosome. A representative panel of GAR mutants
suspected to confer resistance against thiopeptide antibiotics was reconstituted in vitro and quantitatively
studied with fluorescent probes. It was found that single-site mutations of the ribosomal 23S rRNA binding
site region directly affect thiopeptide affinity. Quantitative equilibrium binding data clearly identified A1067
as the base contributing most strongly to the binding environment. The P25 residue on the ribosomal protein
L11 was essential for binding of the monocyclic thiopeptides micrococcin and promothiocin B, confirming
that the mutation of this residue in the producer organism confers self-resistance. For the bicyclic thiopeptides
thiostrepton and nosiheptide, all studied single-site resistance mutations on the L11 protein were still fully
capable of ligand binding in the upper pM range, both in the RNA-protein complex and in isolated 70S
ribosomes. These single-site mutants were then specifically reconstituted in Bacillus subtilis, confirming
their efficacy as resistance-conferring. It is thus reasoned that, in contrast to modifications of the 23S rRNA
in the GAR, mutations of the L11 protein do not counteract binding of bicyclic thiopeptides, but allow the
ribosome to bypass the protein biosynthesis blockade enforced by these antibiotics in the wild type.
Introduction
astonishing efficacy in Vitro. Very high activity against clinically
relevant methicillin-resistant Staphylococcus aureus (MRSA)
The resistance of pathogenic bacteria to standard antibiotics
is strongly increasing, in both nosocomial and community-
acquired infections.1 To combat the bacterial pathogens in-
volved, new strategies and molecules with unused modes of
action are intensively investigated.2 Recurrent promise in the
area is contributed by thiopeptide antibiotic scaffolds (Figure
1), among them thiostrepton (1) and microccocin (2),3 which
were originally identified more than 50 years ago. Recent
research has conclusively demonstrated that these highly
complex polyheterocyclic molecules are ribosomally synthesized
in Streptomyces, Bacillus, and Micrococcus strains,4 and that
thiostrepton (1) is configurationally labile.5
and Enterococcus faecium (MREF) as well as penicillin-resistant
Streptococcus pneumoniae (PRSP) and vancomycin-resistant
enterococci (VRE) were found for thiopeptide antibiotics or
derivatives.3,6,7 Additionally, in recent screening efforts thiop-
strepton 1 was identified as a promising candidate for further
investigation against infections caused by Mycobacterium
tuberculosis.8 Thiopeptide antibiotics were introduced as feed
(4) A comparative overview: (a) Arndt, H.-D.; Schoof, S.; Lu, J.-Y. Angew.
Chem., Int. Ed. 2009, 48, 6770. Examples: (b) Wieland Brown, L. C.;
Acker, M. G.; Clardy, J.; Walsh, C. T.; Fischbach, M. A. Proc. Natl.
Acad. Sci. U. S. A. 2009, 106, 2549. (c) Liao, R.; Duan, L.; Lei, C.;
Pan, H.; Ding, Y.; Zhang, Q.; Chen, D.; Shen, B.; Yu, Y.; Liu, W.
Chem. Biol. 2009, 16, 141. (d) Kelly, W. L.; Pan, L; Li, C. J. Am.
Chem. Soc. 2009, 131, 4327. (e) Morris, R. P.; Leeds, J. A.; Na¨geli,
H.-U.; Oberer, L.; Memmert, K.; Weber, E.; LaMarche, M. J.; Parker,
C. N.; Burrer, N.; Esterow, S.; Hein, A. E.; Schmitt, E. K.; Krastel,
P. J. Am. Chem. Soc. 2009, 131, 5946. (f) Yu, Y.; Duan, L.; Zhang,
Q.; Liao, R.; Ding, Y.; Pan, H.; Wendt-Pienkowski, E.; Tang, G.; Shen,
B.; Liu, W. ACS Chem. Biol. 2009, 4, 855. First biosynthesis
variation: (g) Acker, M. G.; Bowers, A. A.; Walsh, C. T. J. Am. Chem.
Soc. 2009, 131, 17563.
In bacteria, thiopeptide antibiotics target the ribosome or
ribosome-associated factors.3 By virtue of strongly inhibiting
protein synthesis, they repress Gram-positive species with
† TU Dortmund.
‡ MPI fu¨r Molekulare Physiologie.
§ BIRC, Japan Biological Informatics Consortium (JBIC).
⊥ BIRC, National Institute of Advanced Industrial Science and Technol-
ogy (AIST).
(1) (a) Levy, S. B.; Marshall, B. Nat. Med. 2004, 10, S122. (b) Croft,
A. C.; D’Antoni, A. V.; Terzulli, S. L. Med. Sci. Monit. 2007, 13,
RA103. (c) Nikaido, H. Annu. ReV. Biochem. 2009, 78, 119.
(2) (a) von Nussbaum, F.; Brands, M.; Hinzen, B.; Weigand, S.; Ha¨bich,
D. Angew. Chem., Int. Ed. 2006, 45, 5072. (b) Fischbach, M. A.;
Walsh, C. T. Science 2009, 325, 1089.
(5) Schoof, S.; Arndt, H.-D. Chem. Commun. 2009, 7113.
(6) Schoof, S.; Baumann, S.; Ellinger, B.; Arndt, H.-D. ChemBioChem
2009, 10, 242.
(7) Some examples: (a) Pucci, M. J.; Bronson, J. J.; Barrett, J. F.;
DenBleyker, K. L.; Discotto, L. F.; Fung-Tomc, J. C.; Ueda, Y.
Antimicrob. Agents Chemother. 2004, 48, 3697. (b) Singh, S. B.; Occi,
J.; Jayasuriya, H.; Herath, K.; Motyl, M.; Dorso, K.; Gill, C.; Hickey,
E.; Overbye, K. M.; Barrett, J. F.; Masurekar, P. J. Antibiot. 2007,
60, 565.
(3) (a) Bagley, M. C.; Dale, J. W.; Merritt, E. A.; Xiong, X. Chem. ReV.
2005, 105, 685. (b) Hughes, R. A.; Moody, C. J. Angew. Chem., Int.
Ed. 2007, 46, 7930.
9
10.1021/ja909317n  2010 American Chemical Society
J. AM. CHEM. SOC. 2010, 132, 6973–6981 6973

A R T I C L E S
Baumann et al.
Figure 1. Molecular structures of the thiopeptide natural products thiostrepton (1), micrococcin (2), nosiheptide (3), promothiocin B (4), and promoinducin
(5). Crucial residues of 1 and similar rings are labeled.
additives in animal farming early on, and thiostrepton is
currently used in combination therapy ointments to treat
dermatological indications in domestic animals (Panalog, Ani-
max, Solvaderm, Derma-4, etc.). The rather unfavorable phys-
icochemical properties of thiopeptide antibiotics, namely, their
considerable molecular weight and rather low aqueous solubility,
are customarily believed to limit further development as drugs
for systemic application in humans, but for some derivatives
promising activities have been reported in animal models.7
to the rRNA,10 thiopeptides block a cooperative binding region
formed by the rRNA and a ribosomal protein. Over time,
considerable biochemical evidence11,12 and structural studies13
have accumulated, which clarified that most thiopeptides interact
with nucleobases A1067 and A1095 (E. coli numbering) at the
tips of helices 43 and 44 of the 23S rRNA (Figure 2), as well
as a proline-rich helix in the L11 protein N-terminus (21-
PPVGPALQQH-30, Thermus thermophilus sequence and num-
bering). Owing to the switch-like, multiconformational quality
of the GAR13g and the different structure determination tech-
niques and preparation conditions, several models for thiopeptide
binding have been proposed.12d,13c-f Recent X-ray crystal
structures of thiopeptides soaked into the 50S subunit of
Deinococcus radiodurans could be refined to 3.3 Å resolution
for thiostrepton (1) and assigned the binding site of these
The major molecular target of thiopeptide inhibition is the
complex of the 23S rRNA and the ribosomal L11 protein,
situated at the so-called “stalk base” of the GTPase-associated
region (GAR).3 This target remains unused for human therapy
thus far. No cross-resistance with common antibiotics has been
found for it, which should render the thiopeptide binding region
per se highly promising for antibiotic development.4f,6,7,9 Unlike
most other antibiotics targeting the ribosome, which mainly bind
(10) Recent reviews: (a) Poehlsgaard, J.; Douthwaite, S. Nat. ReV.
Microbiol. 2005, 3, 870. (b) Tenson, T.; Mankin, A. Mol. Microbiol.
2006, 59, 1664.
(11) Rosendahl, G.; Douthwaite, S. Nucleic Acids Res. 1994, 22, 357.
(12) (a) Porse, B. T.; Leviev, I.; Mankin, A. S.; Garrett, R. A. J. Mol.
Biol. 1998, 276, 391. (b) Porse, B. T.; Cundliffe, E.; Garrett, R. A. J.
Mol. Biol. 1999, 287, 33. (c) Bowen, W. S.; van Dyke, N.; Murgola,
E. J.; Lodmell, J. S.; Hill, W. E. J. Biol. Chem. 2005, 280, 2934. (d)
Baumann, S.; Schoof, S.; Harkal, S. D.; Arndt, H.-D. J. Am. Chem.
Soc. 2008, 130, 5664.
(8) Lougheed, K. E. A.; Taylor, D. L.; Osborne, S. A.; Bryans, J. S.;
Buxton, R. S. Tuberculosis 2009, 89, 364.
(9) (a) Hrnciar, P.; Ueda, Y.; Huang, S.; Leet, J. E.; Bronson, J. J. J.
Org. Chem. 2002, 67, 8789. (b) Naidu, B. N.; Sorenson, M. E.;
Bronson, J. J.; Pucci, M. J.; Clark, J. M.; Ueda, Y. Bioorg. Med. Chem.
Lett. 2005, 15, 2069.
9
6974 J. AM. CHEM. SOC. VOL. 132, NO. 20, 2010

Thiopeptide Microbial Resistance Determinants
A R T I C L E S
selection.17e,f However, beyond these data the molecular basis
of resistance against thiopeptide antibiotics remains unclear. In
order to fathom the molecular recognition events involved in
these diverging resistance mechanisms, we investigated a
collection of single-site modifications on rRNA and L11 protein
on the molecular level by determining quantitative equilibrium
binding data with different key thiopeptide antibiotics. Critical
cases were reconstituted to study their influence in live cells.
Figure 2 compiles the investigated residues on the ribosomal
crystal structure data obtained for thiostrepton bound to the 50S
subunit of D. radiodurans.13f
Results
In order to quantitatively assess thiopeptide binding to the
mutated GAR, we employed fluorescence anisotropy titrations
using H43/H44 wt RNA fragments (58nt, E. coli sequence
positions 1051-1109) and T. thermophilus L11 (TthL11) wt
protein. Attachment of a fluorescent dye to the terminal
dehydroalanine residue of truncated thiostrepton had been found
to result in high-affinity binders for the complex formed by RNA
and protein.6 We have now investigated other linkers between
the thiostrepton core and the fluoresceine dye and found probe
6 superior to the PEG-linked variant 7 with respect to ease of
synthesis, purification, and long-term stability. Using probe 6,
we found a KD of 0.33 nM to the RNA/protein complex, 2-fold
less than for the previously used probe 7 (0.14 nM) but still
comparable to the parent compound 1 (0.20 nM).6 Displacement
titrations with unlabeled thiopeptides gave similar results for
both probes 6 and 7.
Figure 2. Mutations of the GAR studied in this work mapped on the
thiostrepton crystal structure data at the ribosome (PDB 3cf5).13f Sites
previously reported to lead to reduced sensitivity when mutated are colored
red;13 other residues investigated in this study are colored yellow.
Thiostrepton is labeled similar to Figure 1.
antibiotics to a cleft-like cavity between the L11 protein and
the 23S rRNA,13f in line with the cooperative binding event
and much of the biochemical evidence (Figure 2). Early filter
binding studies and recent investigations with fluorescent labels
on RNA and on thiopeptide ligands have successively detailed
that thiopeptides bind to the respective, reconstituted RNA-
protein complex with affinities in the lower nanomolar to
picomolar range.6,14,15
Resistance development of bacterial pathogens to thiopeptide
antibiotics has not been documented in ViVo. However, for
several bacterial species spontaneous resistance-conferring
mutations have been described to occur at low rates after
challenging the bacteria in Vitro with thiopeptides for selec-
tion.7,16-18 Mutations found on the L11 protein reached from
single amino acid exchange to complete loss of the full protein
or its N-terminal ligand-binding domain.7,16 On the rRNA only
mutations of nucleobases A1067 and A1095 were reported to
contribute to thiopeptide resistance.5,18 Furthermore, thiostrepton
(1)-producing streptomycetes express a specific methyltrans-
ferase to catalyze the 2′O-methylation of A1067, which renders
the organism self-resistant.17 The respective gene is frequently
applied as a tool in Streptomyces biotechnology for strain
(13) (a) Conn, G. L.; Draper, D. E.; Lattman, E. E.; Gittis, A. G. Science
1999, 284, 1171. (b) Wimberly, B. T.; Guymon, R.; McCutcheon,
J. P.; White, S. W.; Ramakrishnan, V. Cell 1999, 97, 491. (c) Lentzen,
G.; Klinck, R.; Matassova, N.; Aboul-ela, F.; Murchie, A. I. H. Chem.
Biol. 2003, 10, 769. (d) Jonker, H. R. A.; Ilin, S.; Grimm, S. K.;
Wo¨hnert, J.; Schwalbe, H. Nucleic Acids Res. 2007, 35, 441. (e) Lee,
D.; Walsh, J. D.; Yu, P.; Markus, M. A.; Choli-Papadopoulou, T.;
Schwieters, C. D.; Krueger, S.; Draper, D. E.; Wang, Y.-X. J. Mol.
Biol. 2007, 367, 1007. (f) Harms, J. M.; Wilson, D. N.; Schlu¨nzen,
F.; Connell, S. R.; Stachelhaus, T.; Zaborowska, Z.; Spahn, C. M. T.;
Fucini, P. Mol. Cell 2008, 30, 26. (g) Li, W.; Sengupta, J.; Rath, B. K.;
Frank, J. RNA 2006, 12, 1240.
(17) (a) Cundliffe, E. Nature 1978, 272, 792. (b) Thompson, J.; Schmidt,
F.; Cundliffe, E. J. Biol. Chem. 1982, 257, 7915. (c) Bechthold, A.;
Floss, H. G. Eur. J. Biochem. 1994, 224, 431. (d) Dunstan, M. S.;
Hang, P. C.; Zelinskaya, N. V.; Honek, J. F.; Conn, G. L. J. Biol.
Chem. 2009, 284, 17013. (e) Thompson, C. J.; Ward, J. M.; Hopwood,
D. A. J. Bacteriol. 1982, 151, 668. (f) Kieser, T.; Bibb, M. J.; Buttner,
M. J.; Chater, K. F.; Hopwood, D. A. Practical Streptomyces Genetics;
John-Innes-Foundation: Norwich (U.K.), 2000.
(14) Thompson, J.; Cundliffe, E.; Stark, M. Eur. J. Biochem. 1979, 98,
261.
(15) Bausch, S. L.; Poliakova, E.; Draper, D. E. J. Biol. Chem. 2005, 280,
29956.
(16) (a) Pestka, S.; Weiss, D.; Vince, R.; Wienen, B.; Sto¨ffler, G.; Smith,
I. Mol. Gen. Genet. 1976, 144, 235. (b) Cameron, D. M.; Thompson,
J.; Gregory, S. T.; March, P. E.; Dahlberg, A. E. Nucleic Acids Res.
2004, 32, 3220.
(18) (a) Hummel, H.; Bo¨ck, A. Biochimie 1987, 69, 857. (b) Thompson,
J.; Cundliffe, E.; Dahlberg, A. E. J. Mol. Biol. 1988, 203, 457.
9
J. AM. CHEM. SOC. VOL. 132, NO. 20, 2010 6975

A R T I C L E S
Baumann et al.
The unperturbed binding and moderate influence of linker
type and length or the negatively charged dye on the probe’s
properties indicated that the linker attachment point is benign
and exposed to the aqueous solvent, in line with previous
studies.12d,13 However, the presence of the L11 protein was
absolutely indispensable for observing stable binding events of
the labeled probes under these homogeneous equilibrium binding
conditions.6,12c,15 We could not observe any binding of the
labeled probe 6 or 7 to the respective RNA fragment up to
concentrations of 10 µM in the absence of L11 (see Supporting
Information). The independence of the binding data from linker
length and the displacement titration data (Vide infra) suggests
that the unlabeled ligands perform similarly. It cannot be
excluded that the wt RNA fragment or its frequently used mutant
U1061A19 with its artificially stabilized secondary structure
displays weak residual affinity to unlabeled thiopeptides.20
However, in either case ligand affinity weakened by at least
3-4 orders of magnitude is highly unlikely to have conse-
quences in cells when considering the modest solubility of
thiopeptide antibiotics in aqueous medium (low µM range).6
The H43/H44 wt RNA fragment had been reported before to
be dependent on Mg2+ for reaching a fold competent for ligand
binding.21 In line with these reports, we detected no ternary
complex formation of probe 6, L11, and wt RNA in the absence
of Mg2+ at neutral pH (7.4). An apparent EC50(Mg2+) ) 0.05
mM and complete saturation at 1 mM Mg2+ were determined
for this interaction by titration (Supporting Information), nicely
corresponding to physiological Mg2+ levels. Under Mg2+-free/
high K+ conditions and slightly acidic pH (6.1), which had been
applied for solution structure determination of the Thermotoga
maritima L11 protein in complex with RNA,13d probe binding
was apparent, but we found the affinity 100-fold reduced to 34
( 3 nM (Supporting Information).
We then screened for the influence of single-point RNA
mutants on the binding affinity of the fluorescent probes.
Fragments of the E. coli 23S rRNA (58nt, positions 1051-1109)
carrying all possible single-site mutations of A1067 and A1095
were synthesized by T7-mediated in Vitro transcription. Muta-
tions proximal (U1066, G1068) and distal (A1089) were
investigated as well, together with the 58 nt fragment incorpo-
rating a 2′O-methylated nucleotide at A1067, which was
obtained by total chemical synthesis.
The 2′O-methylated A1067 RNA derivative showed a drastic
affinity decrease (∼3000-fold) to probe 6, reporting the ef-
fectiveness of the resistance mechanism of the thiostrepton
producer Streptomyces azureus.16 High affinity losses were also
found for the transversion mutants A1067C and A1067U
(∼1000-fold), but not for the transition A1067G, which
performed comparably to the wt RNA. This indicated that
thiopeptide affinity to residue 1067 is mainly triggered by the
purine or pyrimidine nature of the base heterocycle. In fact,
previous studies11 on the A1067G mutation showed that this
modification had only very weak impact on thiostrepton activity
in bacteria.16,18,20
Figure 3. Summary of quantitative binding data of fluorescently labeled
thiostrepton to TthL11 protein and 23S rRNA mutants. Note the logarithmic
scale.
Table 1. Affinities of RNA Single-Point Mutant/wt TthL11
Complexes to Thiostrepton Probe 6
entry
RNA
KD/nM
1
2
3
4
5
6
7
8
wild type (wt)
A1067C
A1067U
A1067G
A1095C
A1095U
A1095G
U1066C
G1068U
0.33 ( 0.17
410 ( 168
460 ( 90
0.34 ( 0.07
5.2 ( 2.8
2.3 ( 1.2
4.5 ( 1.9
0.28 ( 0.15
0.29 ( 0.09
0.36 ( 0.17
970 ( 20
9
10
11
A1089G
2′-O-methyl A1067
each mutation led to a fair decrease of the probe affinity (7- to
15-fold), in line with the increased thiopeptide tolerance
described. Taken together, these results show that resistance-
conferring mutations on the RNA level implicate a reduction
of thiopeptide affinity dependent on RNA structure. The impact
of mutations at the A1095 site was much less pronounced than
at A1067, indicating this nucleobase as the main element of
RNA recognition by 1 (Figure 3).
We then analyzed the impact of 14 different point mutations
in the L11 N-terminussincluding six thiopeptide-resistance
mutations previously described16 (Figure 2)son their impact
on the binding of fluorescently labeled 1. All proteins were
generated by site-directed mutagenesis of the TthL11 wt protein
rplK gene. Notably, none of them showed significantly altered
binding affinities to the probe (Figure 3, Table 2), which varied
only between 0.27 and 0.69 nM (Table 2). Alkylating cysteine
residues in the binding region with iodoacetamide (where
possible) did not influence the probe binding to a major extent
either (data not shown).
To rule out that our results were skewed by the fluorescently
labeled probe, we performed competition titrations with 1. The
apparent affinities thus determined were comparable (∼0.2 nM,
see Figure 5 and Supporting Information) for the three proteins
tested (L11 wt, G24C, and P25C), demonstrating that the
fluorescent label itself is benign and that 1 binds to these mutants
with similar affinity. Similar results were obtained for the closely
related bicyclic thiopeptide nosiheptide (3) (Figure 5 and Table
3). Hence binding of 1 and 3 to TthL11/RNA complexes seemed
almost unperturbed by mutations of the protein.
Interestingly, mutations of residues directly adjacent to A1067
had no effect on probe binding and are hence not expected to
be involved in thiopeptide recognition (Figure 3, Table 1). The
same was true for A1089G far away from the proposed
thiopeptide binding site. Analysis of position 1095 revealed that
(19) Lu, M.; Draper, D. E. J. Mol. Biol. 1994, 244, 572.
(20) Thompson, J.; Cundliffe, E. Biochimie 1991, 73, 1131.
(21) Blyn, L. B.; Risen, L. M.; Griffey, R. H.; Draper, D. E. Nucleic Acids
Res. 2000, 28, 1778.
In order to find out if a different binding profile resulted from
the GAR in the native ribosomal environment, we investigated
intact 70S ribosomes from E. coli. In the past, fluorescently
9
6976 J. AM. CHEM. SOC. VOL. 132, NO. 20, 2010

Thiopeptide Microbial Resistance Determinants
A R T I C L E S
Table 2. Affinities of TthL11 Single-Point Mutant/wt RNA
Complexes to Thiostrepton Probe 6
(wt or mutant) were not affected by detergent, and fluorescence
lifetime measurements showed a similar trend when excess wt
ribosomes (100 nM, τ ) 3.72 ( 0.11 ns) or L11-deficient
ribosomes (τ ) 3.87 ( 0.12 ns) were added to free probe 6 (τ
) 3.34 ( 0.08 ns). We conclude that L11 is indispensable for
high-affinity thiopeptide binding on fully assembled ribosomes,
just as on the isolated RNA/protein complex, and that binding
to ribosomes has to be expected for all the L11 mutants studied
at least in the low nM range.
These results prompted us to analyze if these L11 point
mutations could be solely responsible for rendering bacterial
cells refractory to thiopeptide action, as surmised earlier.16 In
fact, in the previous studies the respective protein mutations
have been identified by sequencing of the resistant organism’s
rplK genes coding for the L11 protein, but had not been
specifically reintroduced into cells. As thiopeptide antibiotics
do not penetrate the outer membrane of Gram-negative bacteria
(such as T. thermophilus or E. coli),25 we conducted studies in
Gram-positive B. subtilis where an L11-deficient knockout strain
has been described (Bsu rplK-).26 We found this strain heavily
impaired compared to the wild type, as indicated by strongly
reduced growth. Under our conditions we observed a doubling
time of 20 ( 1 min for Bsu wt, whereas for Bsu rplK- we found
a 6-fold increase (116 ( 17 min).26 This underlines the
importance of the L11 protein for bacterial viability27 and
suggests that resistance by loss of L11 should strongly suspend
bacterial pathogenicity.
TthL11 wt protein as well as the single-point exchange
mutants P21S, P22S, G24C, and P25C were then transiently
overexpressed in Bsu rplK- and compared to the empty
expression vector28 as control. The resulting cultures were
screened for thiopeptide resistance. With the application of
thiostrepton (1, 0.5 µM, Figure 5), B. subtilis wt showed no
growth at all, whereas the L11-deficient strain was tolerant to
1, as expected.26 Importantly, the transient expression of TthL11
wt protein in the L11-deficient strain restored the sensitivity to
1, as evidenced by a significantly reduced growth rate (-60%).
All other L11 mutants tested (P22S, G24C, P25C) showed no
response to treatment with 1 in this assay, indicating that they
were fully insensitive to the application of thiostrepton within
the limits of this experiment. Taken together, these results
confirmed the relevance of the protein mutations9 in cells and
showed that inhibitor binding does not significantly impact cell
viability in the mutant cells studied.
entry
L11
KD/nM
1
2
3
4
5
6
7
8
wild type (wt)
Q11C
A20C
P21S
P22S
G24C
P25C
Q29C
H30C
G31C
A32C
I34C
M35C
V38C
P55H
0.33 ( 0.05
0.48 ( 0.04
0.35 ( 0.01
0.34 ( 0.01
0.66 ( 0.06
0.49 ( 0.09
0.35 ( 0.06
0.41 ( 0.03
0.37 ( 0.06
0.69 ( 0.05
0.27 ( 0.02
0.56 ( 0.03
0.36 ( 0.03
0.44 ( 0.03
0.65 ( 0.07
9
10
11
12
13
14
15
labeled erythromycin derivatives have been successfully used
for investigation and screening of the macrolide binding site
on E. coli ribosomes.22 To make the GAR amenable for such a
study, ribosome chimeras featuring the TthL11 protein and its
respective mutants were generated using a described L11
knockout strain23 with mild overexpression of His6-tagged
TthL11 protein mutants. To isolate the ribosomes efficiently,
we utilized a mild affinity-based chromatography protocol.24
The 23S rRNA remains highly intact using this procedure,24b
which is crucial considering the exposed positioning of the GAR
on the 70S ribosome’s rim. Correct incorporation of the tagged
L11 in the isolated ribosome fractions was confirmed by SDS-
PAGE analysis and immunoblotting using anti-His6 antibodies
(Figure 4).
Titration of probe 6 with the purified ribosome chimeras in
Vitro led to sigmoidal-shaped binding curves, which showed
apparent affinities in the 0.5-2 nM range (Figure 4) irrespective
of the protein mutant investigated. Given the finite homogeneity
of any ribosome preparation and the biophysical limitations of
this 1:1 binding study (nM concentrations of labeled probe 6
were necessary), we consider these data only approximating the
upper limit for the binding constants of probe 6 to full-size
ribosomes, and in extension thiostrepton (1) itself. The “true”
affinities may be much higher. However, if the mutations
strongly reduced ligand binding to a level affecting function in
cells (higher nM to µM range), they should become detected in
this assay.
In order to further clarify the interplay of binding and
resistance, we investigated the monocyclic thiopeptide natural
L11-deficient ribosomes did not show stable probe binding
(Figure 4) in the presence of detergent (NP-40). In the absence
of detergent, an ambiguous increase of anisotropy (Figure 4)
and brightness (not shown) was observed for L11-deficient
ribosomes, likely indicating nonspecific binding or aggregation.
Indeed, anisotropy measurements with L11-carrying ribosomes
(25) (a) Kelly, J; Schraft, W. C.; Kutscher, A. H.; Tuoti, F. Antibiot.
Chemother. 1959, 9, 87. (b) Kelly, J.; Kutscher, A. H.; Tuoti, F. Oral
Surg. Oral Med. Oral Pathol. 1959, 12, 1334.
(26) Zhang, S.; Scott, J. M.; Haldenwang, W. G. J. Bacteriol. 2001, 183,
2316.
(22) (a) Langlois, R.; Cantor, C. R.; Vince, R.; Pestka, S. Biochemistry
1977, 16, 2349. (b) Turconi, S.; Shea, K.; Ashman, S.; Fantom, K.;
Earnshaw, D. L.; Bingham, R. P.; Haupts, U. M.; Brown, M. J. B.;
Pope, A. J. J. Biomol. Screen. 2001, 6, 275. (c) Yan, K.; Hunt, E.;
Berge, J.; May, E.; Copeland, R. A.; Gontarek, R. R. Antimicrob.
Agents Chemother. 2005, 49, 3367. (d) Li, J.; Kim, I. H.; Roche, E. D.;
Beeman, D.; Lynch, A. S.; Ding, C. Z.; Ma, Z. Bioorg. Med. Chem.
Lett. 2006, 16, 794.
(27) In experiments with E. coli L11 knockout strains we observed a weaker
influence on the growth rate and protein overexpression capability.
The magnitude of the L11 influencing bacterial viability may therefore
depend on species and/or strain, but is frequently observed. Examples:
(a) Cundliffe, E.; Dixon, P.; Stark, M.; Stoffler, G.; Ehrlich, R.;
Stoffler-Meilicke, M.; Cannon, M. J. Mol. Biol. 1979, 132, 235. (b)
Stoffler, G.; Cundliffe, E.; Stoffler-Meilicke, M.; Dabbs, E. R. J. Biol.
Chem. 1980, 255, 10517. (c) Schnier, J.; Gewitz, H. S.; Behrens, S. E.;
Lee, A.; Ginther, C.; Leighton, T. J. Bacteriol. 1990, 172, 7306. (d)
McElwain, K. B.; Boynton, J. E.; Gillham, N. W. Mol. Gen. Genet.
1993, 241, 564. (e) Ochi, K.; Zhang, D.; Kawamoto, S.; Hesketh, A.
Mol. Gen. Genet. 1997, 256, 488. (f) Kawamoto, S.; Zhang, D.; Ochi,
K. Mol. Gen. Genet. 1997, 255, 549.
(23) Baba, T.; Ara, T.; Hasegawa, M.; Takai, Y.; Okumura, Y.; Baba, M.;
Datsenko, K. A.; Tomita, M.; Wanner, B. L.; Mori, H. Mol. Syst. Biol.
2006, 2, 2006:0008.
(24) (a) Maguire, B. A.; Wondrack, L. M.; Contillo, L. G.; Xu, Z. RNA
2008, 14, 188. (b) McLellan, T. J.; Marr, E. S.; Wondrack, L. M.;
Subashi, T. A.; Aeed, P. A.; Han, S.; Xu, Z.; Wang, I. K.; Maguire,
B. A. Acta Crystallogr. D 2009, 75, 1270.
(28) Nguyen, H. D.; Nguyen, Q. A.; Ferreira, R. C.; Ferreira, L. C. S.;
Tran, T. L.; Schumann, W. Plasmid 2005, 54, 241.
9
J. AM. CHEM. SOC. VOL. 132, NO. 20, 2010 6977

A R T I C L E S
Baumann et al.
Figure 4. Incorporation of mutant L11 protein in ribosome chimeras and binding study. (A) SDS-PAGE analysis of the ribosomal proteins from isolated
ribosomes. M ) molecular weight marker (Precision Plus, Bio-Rad, Mu¨nchen, Germany); arrows indicate the L11 protein band. (B) Western blot analysis
of His6-tagged proteins from isolated 70S ribosomes. (C) Isothermal titration binding curves of probe 6 to isolated ribosome chimeras. Only on L11-deficient
ribosomes (rplK-) anisotropy measurements were strongly influenced by detergent.
Figure 5. Reconstitution of L11 mutations in B. subtilis cells by
complementing an L11-deficient strain with transiently expressed mutant
TthL11 proteins and screening for thiostrepton (TS) resistance.
Figure 6. Displacement titrations of 1-5 on mutant L11/rRNA complexes.
Table 3. Affinities of Thiopeptide Natural Products 1-5 to
Stoichiometric amounts of RNA and fluorescent probe (5 nM each) were
complexed with an excess of L11 (0.6 µM) and titrated with increasing
amounts of the thiopeptide ligands.
Selected L11 Mutant/wt RNA Complexes
KD/nM
TthL11
1
2
3
4
5
thiopeptides with an enlarged A-ringssuch as GE2270A or
thiomuracinsdirectly target EF-TU.4e,34
wild type 0.23 ( 0.02 1.46 ( 0.11 0.11 ( 0.01 1040 ( 125 113 ( 13
G24C
P25C
0.24 ( 0.01 1.64 ( 0.20 0.11 ( 0.01 1304 ( 171 195 ( 21
0.30 ( 0.02 >10 000 0.07 ( 0.01 >10 000 139 ( 19
Interestingly, when performing competition titrations using
micrococcin (2), probe 6 could not be displaced from complexes
of wt RNA and the TthL11 mutant P25C (residue 26 in E. coli
and B. subtilis numbering, Figure 6). Apparently, 2 is not able
to bind to complexes carrying a mutation at P25 any more (Table
3), confirming indications from Bacillus megaterium.12b Notably,
in the micrococcin producer Bacillus cereus additional copies
of the rplK gene exist that code for an exchange ProfThr at
this position.4b We thus confirm that mutation of this particular
proline residue selectively abolishes micrococcin binding to the
GAR and can hence be regarded as the natural resistance
mechanism of the micrococcin producer.
products micrococcin (2), promothiocin B (4), and promoinducin
(5) in comparison to the bicyclic thiopeptides 1 and 3.
Micrococcin (2) binds to the GAR, but was reported to affect
the GTPase activity of the elongation factors in a different
fashion from the bicyclic compounds 1 and 3.29-33 Monocyclic
(29) Cundliffe, E.; Thompson, J. Eur. J. Biochem. 1981, 118, 47.
(30) Cameron, D. M.; Thompson, J.; March, P. E.; Dahlberg, A. E. J. Mol.
Biol. 2002, 319, 27.
Similar results were obtained for the structurally correspond-
ing promothiocin B (4). In this case reduced affinity to P25
mutants was observed likewise, suggesting a binding mode
similar to micrococcin. The monocyclic thiopeptide promoin-
(31) Rodnina, M. V.; Savelsbergh, A.; Matassova, N. B.; Katunin, V. I.;
Semenkov, Y. P.; Wintermeyer, W. Proc. Natl. Acad. Sci. U. S. A.
1999, 96, 9585.
(32) (a) Weisblum, B.; Demohn, V. FEBS Lett. 1970, 11, 149. (b) Pestka,
S. Biochem. Biophys. Res. Commun. 1970, 40, 667.
(33) Starosta, A. L.; Qin, H.; Mikolajka, A.; Leung, G. Y. C.; Schwing-
hammer, K.; Nicolaou, K. C.; Chen, D. Y.-K.; Cooperman, B. S.;
Wilson, D. N. Chem. Biol. 2009, 16, 1087.
(34) (a) Anborgh, P. H.; Parmeggiani, A. EMBO J. 1991, 10, 779. (b)
Heffron, S. E.; Jurnak, F. Biochemistry 2000, 39, 37.
9
6978 J. AM. CHEM. SOC. VOL. 132, NO. 20, 2010

Thiopeptide Microbial Resistance Determinants
A R T I C L E S
ducin (5), with a much larger A-ring (35 vs 26 atoms), did not
discriminate L11 wt, G24C, and P25C, indicating an indifferent
binding mode not crucially involving P25 (Table 3). These data
imply that the repulsive interaction with P25 mutants depends
on the size of the main macrocycle (A-ring) and can be masked
by a constraining second macrocycle (B-ring) in the thiopeptide
scaffold, which can be assumed to rigidify and tune the binding
conformation.
fashion. Most of the relevant mutations on L11 involve Pro
residues. An increase in conformational flexibility of the proline-
rich helix in the L11 N-terminal domain can hence be anticipated
in the resistance mutants. This might allow adaptation of this
ribosomal region to the obvious blockade occurring when an
inhibitor is bound to the wt GAR. The well-studied dissimilar
action of micrococcin (2) and thiostrepton (1) at the GAR of
wt ribosomes provides circumstantial support for this notion. 2
was reported to promote GTP hydrolysis from EF-G in Vitro,29,30
but 1 inhibits GTP turnover.30-33 Such differences imply that
binding of the different ligands is functionally distinct and that
it is not mutually exclusive with elongation factor (EF) binding
and hence translational processing. Furthermore, recent crystal
structure data indicate conformational changes in this region
upon binding of EF-G and EF-TU to the GAR, in particular
movements of the L11 N-terminal domain with respect to the
23S rRNA.37 Resistance mutations of crucial residues of L11
could allow overcoming the conformational blockade exerted
by the molecular “glue” of a bound ligand, which apparently
rigidifies the respective orientations of 23S rRNA and L11
protein by tightly inserting into the dynamic cleft separating
the two biomacromolecules.13d,e,g,37,38 This added flexibility
might then rescue translation by re-enabling EF-G catalysis,
which stalls in the wt,38 although alternative explanations for
co-occupancy of inhibitor and EFs cannot be excluded at this
stage.
Some experimental data on thiostrepton competition with
EF-G and/or stabilization of a ternary ribosome/thiostrepton/
EF-G complex have been reported.30,31,39,40 Additionally, IF-2
was shown to be affected by 1 in a similar fashion to EF-G.40
Hence, the resistance mutations studied on the L11 protein might
also facilitate intermittent displacement of the strongly bound
bicyclic thiopeptide inhibitors by ribosomal factors, potentially
promoted by phosphate release from their GTP-loaded states.
Further substantiation of this property must await investigations
of structure and dynamics of this ribosomal region when the
L11 protein is mutated. Nonetheless, on the 23S rRNA resistance
mutations or modifications of crucial residues in the GAR were
found to directly compromise molecular recognition properties
for thiopeptide antibiotics, which reduce ligand affinity to levels
below physiological consequence.
Crystal structure data for 213f indicated a binding pose with
direct involvement of the acyclic appendage (“tail”) with the
binding site, which was not present for 1. Our data for 2 and
the highly related promothiocin B (4) support this view.
Compared to 2, the dehydrobutyrine residue in the A-ring is
missing in 4, and the characteristically different tail structure
of 4 features three dehydroalanines instead of catenated
heterocycles. This results in a 5000-fold loss in affinity for 4.
The thiazole rings in the tail of 2 hence might indeed reinforce
ligand binding of monocyclic thiopeptide antibiotics, potentially
by a productive hydrophobic stacking interaction with A1095.13f
The dehydrobutyrine residue is key for tight binding as well,
as evidenced by a 10-fold affinity increase from 4 to promoin-
ducin (5), which even features an unfavorably enlarged A-ring.
Overall, the GAR seems to display affinity to a certain range
of hydrophobic ligands, but crucial residues and conformational
constraints select between at least two different high-affinity
binding modes. These are nicely reported by the distinct
resistance mechanisms for thiostrepton (1) and micrococcin (2).
Discussion
On the molecular level, the binding of the L11/RNA complex
in the GAR by bicyclic thiopeptides carries a bipartite character.
The molecular recognition of the composite RNA/protein target
complex by thiopeptide antibiotics seems to be highly dependent
on local RNA structure, in particular on the two crucial residues
A1067 and A1095. Interestingly, the adaptability of the GAR
and especially of residue A1067 have previously been identified
as key factors in the processing of aminoacyl-tRNA during
decoding and tRNA accommodation,13g which suggests that this
region is a hot spot for binding and recognition events.
Resistance mutations on the rRNA lead to reduced affinity for
thiopeptides, and the S. azureus resistance gene responsible for
2′O-methylation of A1067 capitalizes on this property as well.
Hence bicyclic thiopeptide antibiotics such as 1 primarily
recognize the rRNA portion of the GAR, although the L11
protein is absolutely indispensable for high affinity binding.
Resistance to the monocyclic thiopeptide antibiotic micro-
coccin (2) can arise from a point mutation in the L11 protein,
which does abrogate inhibitor affinity, presumably by affecting
a key binding interaction with P25. Mutations on the L11 protein
likewise confer indifference to the bicyclic thiopeptide antibiot-
ics 1 and 3, as clearly shown by specific reconstitution of
individual mutants in B. subtilis. However, thiostrepton (1) and
nosiheptide (3) bind tightly to the target structure, both in the
isolated L11/RNA complex and in full-size 70S ribosomes.
Hence resistance does not go in line with reduced binding in
these cases.
Conclusion
The GTPase-associated region (GAR) is a highly promising
target on the bacterial ribosome for the development of new
anti-infective lead compounds. We have characterized the
individual molecular contributions of thiopeptide binding to the
reconstituted complex from T. thermophilus L11 protein and
rRNA in a quantitative fashion. We found that in the cooperative
binding event the molecular recognition of rRNA by thiostrepton
(35) Das, K.; Ding, J. P.; Hsiou, Y.; Clark, A. D.; Moereels, H.; Koymans,
L.; Andries, K.; Pauwels, R.; Janssen, P. A. J.; Boyer, P. L.; Clark,
P.; Smith, R. H.; Smith, M. B. K.; Michejda, C. J.; Hughes, S. H.;
Arnold, E. J. Mol. Biol. 1996, 264, 1085.
(36) Tu, D.; Blaha, G.; Moore, P. B.; Steitz, T. A. Cell 2005, 121, 257.
(37) (a) Schmeing, T. M.; Voorhees, R. M.; Kelley, A. C.; Gao, Y. G.;
Murphy, F. V.; Weir, J. R.; Ramakrishnan, V. Science 2009, 326, 688.
(b) Gao, Y. G.; Selmer, M.; Dunham, C. M.; Weixlbaumer, A.; Kelley,
A. C.; Ramakrishnan, V. Science 2009, 326, 694.
Resistance mutations that do bypass stable inhibitor binding
are not very common but have been described, for example for
HIV-1 reverse transcriptase inhibitors35 or for macrolide
antibiotics binding the 50S ribosomal subunit.36 At the GAR,
this functional dichotomy cannot easily be explained by changes
in the topography of a static binding site environment, which
would accommodate the binding inhibitors in a “lock-and-key”
(38) Ilin, S.; Hoskin, A.; Ohlenschla¨ger, O.; Jonker, H. R. A.; Schwalbe,
H. ChemBioChem 2005, 6, 1611.
(39) Seo, H.-S.; Abedin, S.; Kamp, D.; Wilson, D. N.; Nierhaus, K. H.;
Cooperman, B. S. Biochemistry 2006, 45, 2504.
(40) Brandi, L.; Marzi, S.; Fabbretti, A.; Fleischer, C.; Hill, W. E.; Gualerzi,
C. O.; Lodmell, J. S. J. Mol. Biol. 2004, 335, 881.
9
J. AM. CHEM. SOC. VOL. 132, NO. 20, 2010 6979

A R T I C L E S
Baumann et al.
isolated plasmids. Primer sequences can be found in the Supporting
Information. The wild type and mutant L11 proteins were expressed
from the pQE-30 Xa vectors as N-terminally His6-tagged fusion
proteins in E. coli XL1 blue cells and purified by Ni2+-NTA-affinity
chromatography.12d Homogeneity was assessed by SDS-PAGE
(>95% pure in all cases), and MALDI-TOF mass spectrometry was
used to confirm the molecular weight of the mutant proteins. Correct
and conforming folding was verified by CD spectroscopy, as
described.12d Extinction coefficients wt: ε ) 8250 cm-1 M-1; P55H:
and nosiheptide conforms to a lock-and-key pattern, whereas
binding is much less influenced by resistance mutations of the
L11 protein, as conclusively shown by full-size ribosome studies
and in-cell reconstitution. Notably, micrococcin was found to
sense a resistance mutation in the L11 protein, in line with
genetic data from B. cereus.
The L11 protein was indispensable for high affinity binding
in all cases, suggesting that it acts as a hydrophobic complement
of the RNA binding/recognition event of thiopeptide ligands,
which conformationally lock the orientation of 23S rRNA and
the L11 protein. Resistance mutations on the L11 protein enable
overcoming this conformational blockade, likely by increasing
conformational flexibility and restoring EF-G function. These
insights are expected to open up new avenues for making
thiopeptides and the GAR applicable for combating bacterial
pathogens and should stimulate new compound designs and
screening formats for research on anti-infectives.
ε ) 8250 cm-1 M-1, all other mutants: ε ) 8370 cm-1 M-1
.
Thiopeptide Ligands. Thiostrepton (1) was obtained from
Calbiochem. Micrococcin (2), promothiocin B (4), and promoin-
ducin (5) were isolated from natural sources as described.43-45
Nosiheptide was a gift from H. G. Floss (U Washington) and T.
Mahmud (Oregon State U). All thiopeptide ligands were more than
1
95% pure by H NMR, HPLC, MALDI-TOF, and LC-ESI MS.
Stock solutions were prepared in trifluoroethanol (TFE) and
quantified by UV spectroscopy. Thiostrepton (1):6 ε280 ) 0.027 cm-1
µM-1; micrococcin (2):6 ε280 ) 0.034 cm-1 µM-1; nosiheptide (3):6
ε280 ) 0.039 cm-1 µM-1; promothiocin B (4):45 ε313 ) 0.010 cm-1
Experimental Section
µM-1; promoinducin (5):44 ε315 ) 0.015 cm-1 µM-1
.
Binding Experiments Using Fluorescence Anisotropy. Fluo-
rescence anisotropy measurements were performed using a Tecan
Safire II plate reader in 384-well plates (Optiplate-384 F, Perki-
nElmer) using an excitation wavelength of 470 nm, emission
wavelength of 520 nm, emission bandwidth of 5 nm, and 50 scans
per well at 22 °C. Affinity determination experiments were
performed and evaluated following the protocol described.6 The
fluoresceine-tagged thiostrepton derivative 6 was used for all
measurements. For the analysis of probe binding to wt/mutant RNA
complexes, constant concentrations of the L11 wt protein (0.6 µM)
and fluorescent probe 6 (5 nM) were titrated with varying
concentrations of the respective wt/mutant RNA in measurement
buffer (50 mM MOPS, 5 mM MgCl2, 50 mM KCl, 2 mM TCEP,
5% (v/v) TFE, pH 7.5) in 50 µL final analysis volume. For the
analysis of direct probe binding to wt/mutant protein complexes,
constant concentrations of the protein L11 wt/mutant protein (0.6
µM) and fluorescent probe (5 nM) were titrated with varying
concentrations of wt RNA. All measurements were conducted after
equilibration at 4 °C for 15 h. All data were analyzed using Origin
7.5 (OriginLab Corporation). Sigmoidal-shaped curves were fitted
using the equations derived earlier,6 yielding the apparent KD values
for the binding of the thiostrepton probe to the respective protein/
RNA complex.
Displacement Titrations. For displacement titrations 5 nM
fluorescent probe 6, 5 nM RNA, and 0.6 µM L11 protein were
titrated with varying concentrations of the analyte compounds. In
the wells of a 384-well plate, to measurement buffer (see above)
were added aliquots of protein, fluorescent probe, displacement
analyte, and RNA, to give a final volume of 50 µL, and equilibrated
for 15 h at 4 °C. Measurements were taken as described above.
After plotting against the thiopeptide concentration, the data were
fitted to Hills equation (Origin 7.5, OriginLab Corp.) for analysis
to yield apparent IC50 values, which were converted into apparent
KD data using eq 1.
RNA Synthesis and Purification. Canonical RNAs were
synthesized by T7-polymerase-mediated in Vitro transcription from
cDNA oligonucleotides, coding for the sequence of a wild type or
mutant 58nt fragment of Escherichia coli 23s rRNA (1051-1109)
flanked on the 5′-end by the sequence of the T7 promotor as well
as EcoRI and XbaI restriction sites.12d A list of the sequences used
can be found in the Supporting Information. To ensure homogeneity
and to allow sequence validation, the complementary oligonucle-
otides were annealed and ligated into an EcoRI/XbaI predigested
pUC19 vector. Each resulting plasmid was transformed into E. coli
XL1 blue cells by electroporation, amplified, isolated, and inde-
pendently sequenced. The purified plasmids were linearized with
XbaI to determine the 3′-ends of RNA transcripts and then used as
templates for in Vitro transcription (MEGAscript T7 High Yield
Transcription Kit, Ambion). The RNA transcripts were directly
purified by preparative native 10% PAGE (10 mM Tris, 10 mM
boric acid, 2.5 mM EDTA, pH 8.3), eluted with H2O, desalted with
the aid of NAP5 columns (GE Healthcare), and stored in H2O at
-20 °C. RNA purity was confirmed by native 10% PAGE, UV
spectroscopy, and CD spectroscopy (see Supporting Information).
The 58nt 2′OMe-A1067 RNA derivative was obtained from
Dharmacon (ABgene Limited, Epsom, UK) in fully 2′OH-ACE
protected form.41 For deprotection, the protected RNA was dis-
solved in 400 µL of 2′-deprotection buffer (100 mM acetic acid,
adjusted to pH 3.8 with TEMED) and heated to 60 °C for 30 min.
The deprotected RNA was freeze-dried, redissolved in H2O, desalted
with the aid of NAP5 columns, and stored at -20 °C. The 2′OMe-
RNA was assessed by native 10% PAGE and UV and CD
spectroscopy and found to be of similar purity compared to the
samples synthesized by in Vitro transcription. All RNA was
quantified by UV spectroscopy, extinction coefficients wt: ε )
640 400 cm-1 M-1; 2′O-Me-A1067: ε ) 640 400 cm-1 M-1
;
A1067U: ε ) 635 900 cm-1 M-1; A1067G: ε ) 636 300 cm-1 M-1
A1067C: ε ) 634 000 cm-1 M-1; A1095U: ε ) 638 100 cm-1 M-1
A1095G: ε ) 637 500 cm-1 M-1; A1095C: ε ) 634 800 cm-1 M-1
U1066C: ε ) 637 100 cm-1 M-1; G1068U: ε ) 640 400 cm-1 M-1
;
;
;
;
IC50
Kapp
)
× k2
(1)
TS*
A1089G: ε ) 639 700 cm-1 M-1
.
Protein Synthesis and Purification. The T. thermophilus (Tth)
rplK gene coding for the L11 protein was cloned from T.
thermophilus HB8 genomic DNA and ligated into the pQE-30 Xa
vector (Qiagen) using a StuI/BamHI strategy.12d L11 mutants were
produced using the “site-directed mutagenesis by whole plasmid
synthesis” methodology42 and confirmed by sequencing of the
where Kapp is the apparent KD, TS* is the concentration of the probe
(5 nM), and k2 is the KD of the probe (0.33 nM for 6).
Heterologous Expression of TthL11 in B. subtilis. T. thermo-
philus rplK wt and mutant genes were amplified by PCR using the
(43) (a) Su, T. L. Br. J. Exp. Pathol. 1948, 29, 473. (b) Heatley, N. G.;
Doery, H. M. Biochem. J. 1951, 50, 247. (c) Fuller, A. T. Nature
1955, 175, 722.
(44) Yun, B. S.; Seto, H. Biosci. Biotechnol. Biochem. 1995, 59, 876.
(45) Yun, B. S.; Hidaka, T.; Furihata, K.; Seto, H. J. Antibiot. 1994, 47,
510.
(41) Scaringe, S. A.; Wincott, F. E.; Caruthers, M. H. J. Am. Chem. Soc.
1998, 120, 11820.
(42) Weiner, M. P.; Costa, G. L.; Schoettlin, W.; Cline, J.; Mathur, E.;
Bauer, J. C. Gene 1994, 151, 119.
9
6980 J. AM. CHEM. SOC. VOL. 132, NO. 20, 2010

Thiopeptide Microbial Resistance Determinants
A R T I C L E S
¨
pQE30 Xa constructs as template and ligated (BamHI, XbaI) into
the E. coli/B. subtilis shuttle vector pHT08 (MoBiTec, Go¨ttingen,
Germany)28 to allow a chloramphenicol-dependent selection of
positive B. subtilis transformants as well as an IPTG-triggered
expression of L11 as His8-tagged fusion protein in B. subtilis.
pHT08-TthL11 forward primer sequence: TAGGATCCATGAA-
GAAAGTCGTTGCGGTGGTC; reverse primer sequence: TATCTA-
GATTAGGCATCCTTCACCTCCGGAG. The plasmids were trans-
formed into E. coli XL1 blue cells, which were grown on selection
medium (ampr) for amplification, and the plasmids were isolated
and sequenced to confirm identity.
to an automated AKTA FPLC system (GE Healthcare) or operated
semimanually using a peristaltic pump (P1, Pharmacia).
For the purification of ribosomes, bacterial cells were harvested
and lysed by ultrasonication as suspensions (0.5 g wet cells/mL)
in ice-cold lysis buffer (20 mM Tris, 10.5 mM Mg(OAc)2, 100
mM NH4Cl, 0.5 mM Na2-EDTA, pH 7.5). Cell debris was removed
by centrifugation, and the lysate was cleared by filtration through
a 0.45 µm filter. The affinity column was equilibrated with 10
volumes of lysis buffer directly before use. The lysate was loaded
with a rate of ∼1 mL/min, and the column was eluted with 10
volumes of lysis buffer. Pure ribosomes were eluted with elution
buffer (20 mM Tris, 10.5 mM Mg(OAc)2, 300 mM NH4Cl, 0.5
mM Na2-EDTA, pH 7.5) in a gradient fashion. Ribosome concen-
tration of appropriate fractions was determined by measuring the
absorption at 260 nm and assuming an extinction coefficient of
For heterologus expression, B. subtilis wt (strain 1012) and B.
subtilis rplK- were used.26 Accounting for its slow growth, B.
subtilis rplK- was routinely preselected by seeded growth on agar
plates (37 °C overnight) containing 0.5 µM thiostrepton as selection
marker to ensure strain homogeneity and the thiopeptide-resistance
phenotype. Single thiostrepton-resistant colonies were picked from
these selection plates and used to prepare transformation-competent
cells of B. subtilis rplK-, following a modified Spizizen’s
protocol.26,46 Competent cells were transiently transformed with a
set of the pHT08-TthL11 wt/mutant constructs (1-2 µg of DNA,
camr) following Spizizen’s protocol.46 For the selection of positive
transformants, the cells were plated on chloramphenicol-containing
selection plates and grown overnight at 37 °C. Transformants
growing on these plates were used to inoculate individual cultures
(4 mL of LB medium) for the thiopeptide-resistance scan.
B. subtilis wt (strain 1012) and B. subtilis rplK- expressing the
His8-tagged T. thermophilus L11 wt or mutant proteins were grown
overnight at 37 °C in an orbital shaker (180 rpm) in 4 mL of LB
medium supplemented with 1 mM IPTG (and chloramphenicol in
the case of the transformants) to induce protein expression. On the
next day, the cultures were diluted to an OD600 of 0.1 and used to
inoculate triplicates of 2 × 4 mL of fresh LB medium (1:10) + 1
mM IPTG. The cells were grown for 2 h at 37 °C (180 rpm) and
then supplemented with 0.5 µM thiostrepton (5 µL of a 100 µM
solution in TFE) or 5 µL of TFE as control. The cells were further
grown overnight at 37 °C in an orbital shaker (180 rpm), and the
OD600 of each culture was determined after 24 h. The obtained
OD600 data were normalized to the growth of the TFE controls per
strain/mutant () 100%).
ε260 ) 4.2 × 107 M-1 cm-1
.
47 The purity of ribosomes was checked
by PAGE gels (for ribosomal proteins) and agarose gel electro-
phoresis (for rRNA) in comparison to a sample prepared by an
ultracentrifugation protocol (courtesy of Dr. F. Seebeck, MPI
Dortmund).
Immunoblotting for His6-Tagged Ribosomal Protein Incorpora-
tion. Purified chimeric E. coli ribosomes were thermally denatured
for 10 min at 95 °C in SDS-PAGE loading buffer and separated
on an analytical 15% SDS-PAGE gel. The protein bands were
electrophoretically transferred onto a PVDF membrane. After the
transfer the blot membranes were blocked (3.8% (w/v) milk powder
in TBS-T) for 1 h. For the detection of the His-tagged TthL11
protein a directly HRP (horse radish peroxidase)-coupled R-His5-
antibody (Qiagen, Hilden, Germany) was used (1:1000 in 3.8%
(w/v) milk powder in TBS-T; overnight, rt). The membranes were
washed twice with TBS-T (10 min at rt), and the R-His5-HRP
conjugate was visualized using the SuperSignal West Pico luminol
reaction kit (Pierce Protein Research Products, Rockford, IL).
Affinity Determination on Isolated Ribosomes. Apparent
dissociation constants of probe 5 and purified ribosomes were
determined by titrating 5 nM of the fluorescent probe with
increasing concentrations of ribosome chimeras, incubating for 24 h
to equilibrate, and monitoring for fluorescence anisotropy as
described above. The influence of detergent was tested by adding
0.001 to 1% w/v of NP-40 to the titration buffer. Plotting of the
resulting data against the ribosome concentration gave sigmoid
curves, which were fitted using Hill’s equation.
Design and Synthesis of Ribosome Chimeras. B. subtilis
ribosome chimeras were obtained from B. subtilis rplK- expressing
the His8-tagged T. thermophilus L11 wt as described above. For
the production of E. coli ribosome chimeras the E. coli K12 W3110
rplK knockout strain (Keio collection, kanr)23 was transiently
transformed with the previously constructed pQE30 Xa-derived
TthL11 mutant plasmids (ampr). Positive clones were identified on
kanamycin-ampicillin selection medium. For the ribosome prepa-
ration, transformants were grown in 100-300 mL of LB cultures
(+ampicillin, +kanamycin). Instead of forcing strong overexpres-
sion of the His-tagged L11 fusion protein by the addition of IPTG,
basal expression resulting from the leaky pQE Xa T5 promotor
system was fully sufficient and simplified the purification process.
Purification of Bacterial Ribosomes. 70S ribosomes were
purified following the affinity-based method of Maguire et al.24a
For the preparation of the affinity matrix, 5 mL of SulfoLink matrix
slurry (Pierce Protein Research Products, Rockford, IL) was
equilibrated with 3 × 1 vol. of coupling buffer (50 mM Tris, 5
mM Na2-EDTA, pH 8.5) and reacted with 1 vol. of 50 mM cysteine
hydrochloride (in coupling buffer) for 1 h at room temperature.
Acknowledgment. This work was supported by the DFG
(Emmy-Noether young investigator grant to H.-D.A.), the Fonds
der Chemischen Industrie, and the International Max Planck
Research School Chemical Biology (predoctoral fellowship to S.S.).
We thank Prof. W. G. Haldenwang (UT San Antonio) and the
“National BioResource Project (NIG, Japan): E. coli” for contribut-
ing L11-deficient B. subtilis and E. coli strains, Dr. H. D. Ngyuen
(CGC Dortmund) for sharing B. subtilis transfection vectors, Dr.
F. Seebeck (MPI Dortmund) for samples of E. coli 70S ribosomes
for comparison, Profs. Dr. H. G. Floss (U Washington) and T.
Mahmud (Oregon SU) for a sample of nosiheptide, M. Sc. J. Koo
and Dr. C. Czeslik (TU Dortmund) for support in fluorescence
spectroscopy, and Prof. Dr. M. Helm (now U Mainz) for discussions.
Supporting Information Available: Oligonucleotide se-
quences, additional supporting figures, preparation of the labeled
probe, and Mg2+-dependency of the interaction. This information
is available free of charge via the Internet at http://pubs.acs.org.
¨
The matrix was loaded into an AKTA column (GE Healthcare)
and washed with 10 volumes of H2O. The column was either fitted
(46) Anagnostopoulos, C.; Spizizen, J. J. Bacteriol. 1961, 81, 741.
(47) Arfvidsson, C.; Wahlund, K. G. Anal. Biochem. 2003, 313, 76.
JA909317N
9
J. AM. CHEM. SOC. VOL. 132, NO. 20, 2010 6981