Pipecolic acid derivatives as small-molecule inhibitors of the legionella MIP protein

Article abstract of DOI:10.1021/jm101156y

The macrophage infectivity potentiator (MIP) protein is a major virulence factor of Legionella pneumophila, the causative agent of Legionnaires disease. MIP belongs to the FK506-binding proteins (FKBP) and is necessary for optimal intracellular survival a

Full text of DOI:10.1021/jm101156y

J. Med. Chem. 2011, 54, 277–283 277
DOI: 10.1021/jm101156y
Pipecolic Acid Derivatives As Small-Molecule Inhibitors of the Legionella MIP Protein
†
†
‡
§
§
€
€
Christina Juli, Martin Sippel, Jens Jager, Alexandra Thiele, Matthias Weiwad, Kristian Schweimer, Paul Rosch,
Michael Steinert,‡ Christoph A. Sotriffer,† and Ulrike Holzgrabe*,†
†Institute of Pharmacy and Food Chemistry, University of Wu€rzburg, Am Hubland, 97074 Wu€rzburg, Germany, ‡Institut fu€r Mikrobiologie,
Technische Universita€t Braunschweig, Spielmannstrasse 7, 38106 Braunschweig, Germany, §Research Center for Enzymology of Proteinfolding,
Max-Planck-Institute Halle, Weinbergweg 22, 06120 Halle, Germany, and Department of Biopolymers, University of Bayreuth,
Universita€tsstrasse 30, 95447 Bayreuth, Germany
Received September 7, 2010
The macrophage infectivity potentiator (MIP) protein is a major virulence factor of Legionella pneumophila,
the causative agent of Legionnaires’ disease. MIP belongs to the FK506-binding proteins (FKBP) and is
necessary for optimal intracellular survival and lung tissue dissemination of L. pneumophila. We aimed to
identify new small-molecule inhibitors of MIP by starting from known FKBP12 ligands. Computational
analysis, synthesis, and biological testing of pipecolic acid derivatives revealed a promising scaffold for new
MIP inhibitors.
Introduction
an extremely stable complex with the immunosuppressive
drugsFK506and rapamycin which both inhibit the enzymatic
function of MIP.6 However, these drugs are contraindicated
for treatment of Legionnaires’ disease because they interact
with human FKBPs exerting immunosuppressive effects.
Recently, Ceymann et al. solved the structure of free MIP
protein and the MIP-rapamycin complex by means of nucle-
ar magnetic resonance (NMR) spectroscopy.7 The complex
structure revealed that binding to MIP is mainly mediated by
rapamycin’s pipecoline moiety, which is anchored in a hydro-
phobic pocket of the binding site. This is similar to FKBP12
(a human PPIase playing a crucial role in immunological
processes), which lends credence to the hypothesis that non-
immunosuppressive small-molecule inhibitors of FKBP12
derived from rapamycin, may constitute a suitable starting
point for obtaining low-molecular-weight inhibitors of the
MIP protein. Accordingly, the goal of this work is the devel-
opment of pipecolic acid derivatives as MIP inhibitors by a
combined approach of structure comparison, computational
analysis, chemical synthesis, and experimental testing, includ-
ing enzymatic methods, cell-based measurements, and NMR
HSQC experiments.
Legionella pneumophila is a Gram-negative aerobic patho-
gen causing two distinct forms of Legionellosis: Legionnaires’
disease, a severe form of pneumonia with a mortality rate
of up to 20%, and Pontiac fever, a milder respiratory disease
with symptoms resembling acute influenza.1 The pathogen
naturally inhabits fresh waters and biofilms, where it para-
sitizes within protozoan hosts. Human infection is by inhaling
aerosols from contaminated man-made hot water systems.
Elderly adults, smokers, and people with impaired immunity
are particularly susceptible. The pathogen can colonize the
lung and replicate intracellularly within macrophages. During
the replicative phase the bacteria reside within a maturation
blocked phagosome. Infection studies in different host cell
models revealed that the Legionella virulence factor MIP protein
contributes to the establishment of infection.2,3 Guinea pig
infections with MIP-positive and MIP-negative L. pneumophila
strains and mutants showed that MIP also contributes to
bacterial dissemination within the lung. Binding assays with
components of the extracellular matrix (ECMa) demonstrated
the interaction of MIP with collagen IV, the prevalent col-
lagen in the human lung.4 Degradation assays with ECM
proteins supported the hypothesis of a concerted action of
MIP and other proteolytic enzyme activities, although the
exact role of MIP in this process remains to be established.4-6
The MIP protein is a peptidyl prolyl cis/trans isomerase
(PPIase) belonging to the FK506 binding protein family. It forms
Results and Discussion
MIP is a homodimeric protein consisting of two 22.8 kDa
monomers. Homodimerization is mediated by the N-terminal
domain, which is connected through a long helix to the
C-terminal domain that resembles FKBP12 and contains the
rapamycin binding site. This C-terminal or FKBP domain
extends from residues 77 to 213 (MIP77-213) of the full-length
protein. In all of the following, “MIP” refers only to this
domain (with the residues numbered consecutively from 1 to
137). Currently, three MIP structures are available in the PDB:
A crystal structure (1FD9)8 and an NMR solution structure
(2UZ5)7 of the unbound protein, as well as an NMR struc-
ture with rapamycin bound to the active site (2VCD).7 The
structures reveal a domain consisting of six β-strands which
*To whom correspondence should be addressed. Phone: þ49-931-
3185460. Fax: þ49-931-3185494. E-mail: u.holzgrabe@pharmazie.uni-
wuerzburg.de.
a Abbreviations: MIP, macrophage infectivity potentiator; FKBP,
FK506 binding protein; ECM, extracellular matrix; HSQC, hetero-
nuclear single quantum coherence; PDB, protein database; rmsd, root-
mean-square deviation; cf., confer; DCC, dicyclohexylcarbodiimide;
DMAP, dimethylaminopyridine; PPIase, peptidyl prolyl isomerase;
GA, genetic algorithm; TLC, thin layer chromatography; FT-IR, Fou-
rier transform infrared spectroscopy; HEPES, 2-(4-(2-hydroxyethyl)-1-
piperazinyl)-ethanesulfonic acid.
r
2010 American Chemical Society
Published on Web 12/13/2010
pubs.acs.org/jmc

278 Journal of Medicinal Chemistry, 2011, Vol. 54, No. 1
Juli et al.
Figure 2. FKBP12 inhibitors selected for investigation.
Figure 1. Comparison of the binding pockets of MIP (colored
wheat, with residue numbers) and FKBP12 (colored green) in the
rapamycin complexes (MIP, 2VCD; FKBP12, 1FKB). In MIP,
Tyr109 (top right) protrudes deeper into the pocket.
form an antiparallel sheet and a short helix positioned across
this sheet. The active site is located right between this helix
and the internal side of the sheet and shows a predominantly
hydrophobic character. It is formed by Tyr55, Ala75, Phe77,
Val82, Ile83, and Phe126, as well as Trp86 at the ground of the
pocket. Phe65, Asp66, and Tyr109 are located at the outer rim
of the pocket and are responsible for the main conformational
differences among the three available structures. Because of
the higher relevance of complex structures for the ligand-
bound state of a protein, only structure 2VCD was used for all
subsequent analyses.
Comparison of MIP with FKBP12 reveals highly homo-
logous binding sites formed by identical amino acids, except
for Ala75, which corresponds to a phenylalanine residue in
FKPB12. Nevertheless, superposition of the rapamycin com-
plexes of MIP and FKBP12 (PDB 1FKB) shows significant
differences, in particular with respect to Tyr109, which pro-
trudes much deeper into the binding pocket of MIP, thereby
reducing its overall size (Figure 1). This is reflected by dif-
ferences in the binding mode of the pipecoline moiety of
rapamycin with respect to the orientation of the aliphatic ring
in the hydrophobic pocket, the conformation of the ester, and
the hydrogen-bonding pattern of the carbonyl groups. This
leads to the expectation that also the binding of small mole-
cules may show differences among the two proteins.
Figure 3. Top-ranked binding mode of B (corresponding to S-4c,
cf. below) in MIP. The pipecoline ring is placed in the active-site
pocket, showing good shape complementarity.
the active-site pocket. However, although similarities in the
orientation exist, the placement does not overlap very well
with the rapamycin binding mode in MIP. Furthermore, this
result was obtained only in 2 out of 100 runs, whereas another
orientation (which placed the phenyl ring in the hydrophobic
pocket) was obtained more frequently. More importantly, the
docking score for A in MIP was significantly less favorable
than the top-ranked score of A in FKBP12. Taken together,
these results suggest that although binding of A to MIP may
be possible, it is unlikely to result in a perfectly matching
binding mode leading to high affinity.
Subsequently, docking to MIP was carried out with com-
pound B, for which no crystal structure with FKBP12 is
available. This time, a well-clustered top-ranked result was
obtained, showing a very favorable hydrophobic interaction
score, which was even better than the score obtained upon
docking of B to FKBP12. Indeed, the binding mode of B in
MIP shows a good fit of the pipecoline-sulfonamide anchor
group, which is reflected in the high hydrophobic score
(Figure 3). In contrast to the R-ketoamide moiety of A, the
less bulky sulfonamide of B allows for a better placement
in the narrow MIP pocket. These observations refer to the
S-enantiomer of the ligand (as shown in Figure 3). The
R-enantiomer can also be placed in the pocket, but only in a
somewhat distorted orientation, leading to a less favorable
hydrophobic score.
Given these observations, compounds A and B (S-4c) were
approached by synthesis. Instead of compound A, a slightly
modified analogue with an isobutyl chain instead of the 2,2-
dimethylpropyl substituent was more readily accessible and,
thus, synthesized initially. On the basis of the binding model,
the modification was not expected to have any significant
influence on the activity. On the other hand, the benzylsulfon-
amide B (S-4c) was prepared, as well as the corresponding
phenylsulfonamide (i.e., compound 4d below). Docking showed
a convincing binding mode similar to B (S-4c) for this phenyl
derivative, although with a significantly less favorable hydro-
phobic score but granting possibilities for aromatic substitu-
tion in meta position. Accordingly, meta-substituted derivatives
of the phenylsulfonamide were synthesized as well.
Various FKBP12 ligands have been described in the litera-
ture. Among them are compounds which bear a pipecoline
moiety, thus resembling the rapamycin anchoring group.
However, these ligands lack the macrocyclic portion of rapa-
mycin, thereby abolishing the immunosuppressive action.
Given the known binding of rapamycin to MIP, these com-
pounds should constitute the best starting point for obtaining
new MIP inhibitors. Accordingly, the two submicromolar
FKBP12 inhibitors A(Ki =0.010μM)9 and B(Ki =0.230μM)10
were selected for investigation (Figure 2).
For A, a crystal structure in complex with FKBP12 is
available (PDB 1FKG), showing an analogous binding mode
as the pipecoline group of rapamycin. As a reference, ligand A
was separated from the protein and computational redocking
was carried out to reconstitute the complex. This reproduced
the experimental binding mode as top-ranked, well-clustered
˚
result with a root-mean-square deviation (rmsd) of 1.05 A.
Docking of A to MIP with the same procedure yielded a top-
ranked binding mode which also placed the pipecoline ring in

Article
Journal of Medicinal Chemistry, 2011, Vol. 54, No. 1 279
Scheme 1. Synthetic Route to the Pipecolic Acid Derivativesa
Table 1. Synthesized Pipecolic Acid Derivatives and Their IC50 Values
a Reagents and conditions, (a) R1-Cl, NEt3, DMF, RT, 6 h; (b) LiOH,
MeOH, 0 °C, 1 h; (c) R2-OH, DCC, DMAP, p-toluene sulfonic acid,
CH2Cl2, RT, 72 h; (d) SnCl2, EtOH, reflux, 2 h.
The general synthetic route is illustrated in Scheme 1.
Synthesis of the pipecolic acid derivatives started with the con-
version of commercially available ethyl piperidine-2-carboxylate
hydrochloride with 3 equiv of triethylamine and equimolar
amounts of sulfonyl or carbonyl chloride, leading to the
amides 2a-d. Subsequently, the ethyl ester was hydrolyzed
using lithium hydroxide to give the free pipecolic acids 3a-d.
To obtain the corresponding esters 4a-f, the free acids were
converted with equimolar amounts of the corresponding
alcohols, 1.5 equiv of dicyclohexylcarbodiimide, catalytic
amounts of dimethylaminopyridine, and p-toluene sulfonic
acid. The enantiopure compound B (S-4c) was synthesized by
employing the enantiopure pipecolic acid as starting material.
The nitro analogues (4e and 4f) were reduced via SnCl2 to the
corresponding amino compounds 5a and 5b, respectively (see
Scheme 1). The structures of all compounds were confirmed
by NMR spectroscopy.
The inhibition of MIP was determined by a protease-
coupled PPIase assay. This PPIase activity catalyzed by re-
combinant MIP expressed in Escherichia coli was determined
using the synthetic peptide succinyl-Ala-Phe-Pro-Phe-4-
nitroanilide assubstratein the chymotrypsin-coupled reaction
as previously described.11 Because some of the tested com-
pounds resemble common FKBP12 inhibitors,9 the inhibition
of FKBP12 was additionally analyzed by the same method.
Most of the compounds were found to inhibit FKBP12 but
not necessarily the MIP protein. The anti-MIP and anti-
FKBP activity of the compounds is depicted in Table 1. IC50
values >100 μM for MIP and >30 μM for FKBP12 were not
determined. (Note that the IC50 values should only be com-
pared within the inhibitor series of one enzyme, but not
between the two enzymes).
and 4e show substantial activity against MIP. The racemic
compound 4c shows the lowest IC50 of this series, which
indicates that a methylene group next to the sulfonamide is
advantageous for binding, in agreement with the modeling
data. In the case of FKBP, instead, the phenyl and benzyl
sulfonamide derivatives showed equal activity, pointing to a
further difference in comparison to MIP. The introduction of
a nitro group (4f) is detrimental to the MIP activity when
compared with 4d. Some activity can be restored by replacing
the trimethoxyphenyl group with a pyridyl substituent (4e).
The reduction of the nitro group to the corresponding amines
does not lead to active compounds against MIP. Because the
racemic compound 4c was identified as the most promising
MIP inhibitor, the corresponding enantiopure compound B
(S-4c) was synthesized. As anticipated, the S-enantiomer
shows higher activity against MIP (IC50 = 6 μM). Given this
potency, it represents a promising starting structure for the
development of further inhibitors.
With ethyl as R2 substituent attached to the carboxy func-
tion, the compounds show only weak inhibition of MIP (2b)
or no sufficient effect below 100 μM concentration (2a, 2c-d).
This indicates a larger ester moiety is required for proper
binding. However, addition of larger R2 substituents to the
R-ketoamide 2a (resulting in compounds 4a and 4b) did not
lead to any observable gain in activity against MIP. This is in
contrast to FKBP12, where addition of the propyl-trimethox-
yphenyl substituent to 2a provides a clear gain in potency
(compound 4b). Apparently, the dicarbonyl-alkyl group of 2a
(and 4a, 4b) is not well suited to fit the MIP binding pocket.
The lack of activity (below 30 μM) of 4a in FKBP12 despite
the reported submicromolar Ki of A9 should most likely not
only be attributed to the missing methyl group in 4a but rather
to the fact that 4a was obtained and tested as mixture of 8
stereoisomers. In contrast to 4a and 4b, compounds 4c, 4d,
To prove the binding to the MIP protein, the interaction of
compounds 2b, 4c, and B (S-4c) with MIP was determined
by means of NMR-HSQC experiments. The activity of the

280 Journal of Medicinal Chemistry, 2011, Vol. 54, No. 1
Juli et al.
Figure 4. Changes in the chemical shifts of MIP residues as measured by NMR-HSQC experiments with compounds 2b, 4c, and B (S-4c).
ligands was defined by changes in the chemical shifts. As can
be seen in Figure 4, the titration of 15N labeled MIP77-213
(FKBP domain) with compounds 2b, 4c, and B (S-4c) resulted
in significant chemical shift changes of several backbone
amide resonances, indicating a strong binding of these com-
pounds to the MIP-FKBP domain. Because of solubility
problems at ligand:protein ratios of 20:1 (2 mM ligand,
0.1 mM protein), the saturation of MIP with the ligands could
not be achieved; therefore, a determination of the dissociation
constants was not possible.
Figure 5. Structure of the MIP protein with binding regions
The pattern of chemical shift changes is similar for all three
ligands, indicating a similar binding mode at the same region
of the protein. Mapping the chemical shift changes on the
protein surface shows that the affected residues are located in
the hydrophobic binding pocket for rapamycin. It is therefore
concluded that the compounds 2b, 4c, and B (S-4c) indeed
bind to this cavity of MIP (see Figure 5).
To examine the influence of substance B (S-4c) on invasion
and intracellular replication, human macrophage-like U937-
cells were infected with L. pneumophila Philadelphia-1 JR32
(wild type, WT), the MIP-negative mutant L. pneumophila
JR32-2, and the E. coli strain HB101. The results of the
gentamicin infection assays (Figure 6) revealed that substance
B (S-4c) does not influence the replication of wild type
L. pneumophila JR32 in human macrophage-like U937 cells.
(highlighted in orange) of compound B (S-4c) as determined by
NMR. The protein is seen from the same orientation as in Figure 3.
The binding pocket is clearly recognizable in the center.
Incontrast, rapamycintreated wild typeL. pneumophilaJR32,
MIP-negative mutant L. pneumophila JR32-2 and E. coli
strain HB101 were either not able to replicate in U937 cells or
were degraded. These results confirm that the MIP protein,
but not its PPIase activity, which is efficiently inhibited by
substance B (S-4c) (see Protease-Coupled PPIase Assay), is
required for virulence on the cellular level of infection.12
Moreover, these data support the hypothesis that the hydro-
phobic cavity of the PPIase domain rather confers bacterial
adhesion to extracellular targets within lung tissue.4 Thus, it
will be interesting to identify the yet unknown target molecule

Article
Journal of Medicinal Chemistry, 2011, Vol. 54, No. 1 281
1
39.52 ppm, chloroform-d: H 7.26 ppm, 13C 77.16 ppm). The
following abbreviations describing the multiplicity are used:
(s) singlet, (d) doublet, (t) triplet, (q) quartet, (dd) doublet of
doublet, (dt) doublet of triplet, (br) broad signal, (m) multiplet.
Coupling constants are given in hertz. TLC was carried out on
TLC aluminum sheets, silica gel F254 (Merck KGaA, Darmstadt,
Germany). IR spectra were obtained with a Biorad PharmalyzIR
FT-IR spectrometer (Biorad, Cambridge, MA, USA). Melting
points were determined using a capillary melting point appara-
tus (Gallenkamp, Sanyo) and are uncorrected. The purity of the
compounds was confirmed by quantitative microanalyses (C, H,
N); the new compounds were in accordance with the theoretical
value within (0.4%, i.e., purity higher than 95%. All chemicals
were purchased from Sigma-Aldrich Chemicals (Deisenhofen,
Germany), Acros Organics (Geel, Belgium), and VWR Inter-
national (Darmstadt, Germany) and were used without further
purification. The peptide was purchased from Bachem Distri-
bution Services GmbH (Weil am Rhein, Germany).
Figure 6. Influence of B (S-4c) and rapamycin (Rap) on infection of
U937 human macrophage-like cells. Lp = Legionella pneumophila,
MIP- = MIP-deficient L. pneumophila strain.
Below, the general synthesis procedures are reported, as well
as experimental and spectroscopic details for the compounds
S-1, S-2b, S-3b, and S-4c. Details for all other compounds can be
found in the Supporting Information.
of the MIP mediated PPIase activity and to analyze the effect
of substance B (S-4c) during guinea pig infections.
Experimental Section
Synthesis of (S)-Ethyl piperidine-2-carboxylate S-1. (S)-Piper-
idine-2-carboxylic acid (1 equiv) was dissolved in 40 mL of
anhydrous ethanol, and thionyl chloride (3 equiv) was added
dropwise. The mixture was refluxed for 4 h under stirring. After
the reaction had ceased (TLC control), the solvent was removed
in vacuo. The residue was suspended in 20 mL of NaHCO3 and
washed with 4 ꢀ 30 mL CHCl3. The organic layers were dried
over anhydrous Na2SO4, and the solvent was removed in vacuo
to yield compound (S)-1 as a yellow waxy substance (0.90 g,
90%). 1H NMR (CDCl3, δ [ppm], J [Hz]): 4.15 (q, 2H, 3J = 7.1),
3.52 (dd, 1H, 3J = 3.5, 3J = 9.8), 3.22-3.14 (m, 1H); 2.79-2.69
(m, 1H), 2.03-1.95 (m, 1H), 1.92 (s, 1H), 1.77-1.40 (m, 5H),
1.22 (t, 3H, 3J = 7.1). IR [cm-1]: 1046, 1205, 1370, 1456, 1740,
2695, 2906.
General Procedure for the Synthesis of Compounds 2a-d and
S-2b. Ethyl piperidine-2-carboxylate hydrochloride 1 and S-1,
respectively (1 equiv), were dissolved in 40 mL of anhydrous
DMF, and triethylamine (3 equiv) was added dropwise followed
by carbonyl and sulfonyl chloride (1 equiv), respectively. The
mixture was stirred for 6 h at RT and subsequently the solvent
removed in vacuo. The residue was suspended in 30 mL CHCl3
and washed with 4 ꢀ 30 mL 2 M HCl and H2O. The organic
layer was dried over anhydrous Na2SO4, and the solvent was
removed in vacuo. If necessary, a subsequent purification by
column chromatography was performed to obtain compounds
2a-d and S-2b.
Computational Docking. The MIP NMR structure 2VCD7
was used for structural analyses and computational docking
studies. Because 2VCD consists of an ensemble of 16 structures,
the least restrictive conformation with respect to pocket size for
ligand accommodation was selected. On the basis of the mutual
distance between Asp66, Phe77, and Tyr109, 2VCD-4 (confor-
mer 4) was used for this purpose. Focusing on the least restric-
tive conformation was based on the reasoning that upon dock-
ing at this exploratory stage, any ligand should be granted
as much space as the protein is able to provide without any
significant energy cost. Although this might not be the best
conformation for all ligands, it should at least be the one which
best prevents that true ligands are inadvertently excluded due to
unsuccessful docking.
Docking was carried out with AutoDock3.013 and GOLD4.0.14
For docking with AutoDock, the protein was setup by removing
ligand and water molecules, as well as all hydrogen atoms. Polar
hydrogens were then readded with the protonate tool, and
Kollman united-atom partial charges and solvation parameters
were assigned to all atoms.15 The active site was defined by a box
of 60 ꢀ 60 ꢀ 60 grid points, centered between Phe77 and Tyr109.
˚
The grid spacing was set to 0.375 A. Subsequently, the interaction
maps were calculated with the autogrid module. Gasteiger-
Marsili charges16 were added to the ligand using Sybyl8.0.17 The
conversion to the pdbq format as well as the definition of
rotatable bonds was done with the autotors module of Auto-
Dock. The docking protocol used an initial population of 50
individuals, a maximum number of 3.0 ꢀ 106 energy evalua-
tions, a mutation rate of 0.02, a crossover rate of 0.80, and an
elitism value of 1. The Solis and Wets local search parameters
were set as follows: The probability of performing a local search
on an individual was set to 0.06, a maximum number of 300
iterations was performed, and the number of consecutive suc-
cesses or failures before changing the size of the local search
space was 4. 100 GA runs were carried out.
(S)-Ethyl 1-(Benzylsulfonyl)piperidine-2-carboxylate (S-2b).
S-2b was obtained as a yellow oil (0.63 g; 70%). 1H NMR
(CDCl3, δ [ppm], J [Hz]): 7.49-7.33 (m, 5H), 4.55-4.47 (m, 1H),
4.27 (s, 2H), 4.24-4.13 (m, 2H), 3.49-3.41 (m, 1H), 3.15 (dt, 1H,
3
3J = 12.4, J = 3.1), 2.20-2.12 (m, 1H), 1.73-1.52 (m, 3H),
1.51-1.37 (m, 1H), 1.33-1.20 (m, 4H). IR [cm-1]: 1148, 1200,
1335, 1455, 1732, 2860, 2940. Anal. Calcd for C15H21NO4S
(MW: 311.40 g/mol): C, 57.86; H, 6.80; N, 4.50; S, 10.30. Found:
C, 58.17; H, 6.54; N, 4.78; S, 10.36.
General Procedure for the Synthesis of Compounds 3a-d and
S-3b. The compounds 3a-d and S-3b, respectively, were synthe-
sized using a procedure described by Holt et al.9 for analogous
substances. Compounds 2 (1 equiv) were dissolved in 30 mL of
methanol and cooled to 0 °C with an ice bath. To this mixture a
solution of lithium hydroxide (10 mL, 1 M) was added dropwise,
and then it was stirred for 1 h. After the reaction had ceased
(TLC control), 10% HCl was added to adopt a pH value of 1.
The free carboxylic acid was extracted with 4 ꢀ 30 mL CH2Cl2.
The combined organic layers were washed with 4 ꢀ 50 mL H2O
and saturated NaCl and then dried over anhydrous Na2SO4.
The solvent was removed in vacuo to give compounds 3a-d
and S-3b.
For docking with GOLD, the binding site was defined by a
˚
sphere of 8 A radius, centered between Tyr55 and Tyr109 of the
fully protonated 2VCD-4 structure. GOLDscore was used as
scoring function. 50 GA runs were performed. Genetic algorithm
parameters were set as follows: Population size 100, number of
islands 5, selection pressure 1.1, number of operations 89500,
niche size 2, migrate 10, mutate 95, and crossover 95.
Chemical Synthesis. 1H and 13C NMR spectra were recorded
on an Avance 400 nuclear magnetic resonance spectrometer,
Bruker Biospin GmbH, Rheinstetten, Germany (1H 400.132
MHz, 13C 100.613 MHz). As internal standard the signals of the
1
deuterated solvents were used (DMSO-d6: H 2.50 ppm, 13C

282 Journal of Medicinal Chemistry, 2011, Vol. 54, No. 1
Juli et al.
(S)-1-(Benzylsulfonyl)piperidine-2-carboxylic Acid (S-3b).18
S-3b was obtained as a yellow oil (0.57 g, 90%). 1H NMR
(CDCl3, δ [ppm], J [Hz]): 7.49-7.33 (m, 5H), 4.60-4.55 (m, 1H),
4.27 (s, 2H), 3.49-3.41 (m, 1H), 3.15 (dt, 1H, 3J = 12.4, 3J =
3.1); 2.20-2.12 (m, 1H), 1.73-1.52 (m, 3H), 1.51-1.18 (m, 2H).
IR [cm-1]: 1060, 1118, 1317, 1455, 1728, 2854, 2952, 3192.
General Procedure for the Synthesis of Compounds 4a-f and
S-4c. The syntheses of compounds 4a-f and S-4c were per-
formed using a procedure described by Holt et al.9 for analogous
substances. Compounds 3 (1 equiv), the corresponding alcohol
(1.1 equiv), DCC (1.5 equiv), DMAP (0.3 equiv), and p-toluene
sulfonic acid (0.3 equiv) were dissolved in 50 mL of anhydrous
CH2Cl2. This mixture was stirred for 72 h at RT. After the
reaction had ceased (TLC control), the precipitate was filtered
off and the solvent was removed in vacuo. The residue was
suspended in ethyl acetate, and the resulting precipitate was filtered
off again. This procedure was repeated as long as a precipitate was
formed. Afterward, the filtrate was dried over Na2SO4 and subse-
quently the solvent removed in vacuo. If it was necessary, a
subsequent purification by silica gel column chromatography
was performed to obtain the compounds 4a-f and S-4c.
(S)-3-(3,4,5-Trimethoxyphenyl)propyl 1-(benzylsulfonyl)-
piperidine-2-carboxylate (S-4c). After purification by column chro-
matography (aluminum oxide; chloroform: ethyl acetate = 1:1)
S-4c was obtained as a yellow waxy substance (0.12 g, 21%). 1H
NMR (CDCl3, δ [ppm], J [Hz]): 7.49-7.33 (m, 5H), 6.41 (s, 2H),
4.55-4.50 (m, 1H), 4.27 (s, 2H), 4.24-4.15 (m, 2H), 3.84 (s, 6H),
3.82 (s, 3H), 3.46-3.40 (m, 1H), 3.17 (dt, 1H, 3J = 12.4, 3J = 3.1),
2.68-2.62 (m, 2H), 2.20-2.12 (m, 1H), 2.03-1.95 (m, 2H),
1.75-1.50 (m, 3H), 1.51-1.18 (m, 2H). IR [cm-1]: 1124, 1335,
1455, 1589, 1733, 2857, 2938. Anal. Calcd for C25H33NO7S (MW:
491.60 g/mol): C, 61.08; H, 6.77; N, 2.85; S, 6.52. Found: C, 60.86;
H, 6.87; N, 3.06; S, 6.13.
Normalized chemical changes were expressed as weighted
geometric average of chemical shift changes (1H and 15N) by
normshift = (δ(1H)2 þ (0.1δ(15N))2)1/2. Normalized chemical
shift changes larger than 0.03 ppm were considered as signifi-
cant.
Infection Assay of U937 Cells. The human macrophage-like
cell line U937 was infected as described previously.20 In brief,
cells were infected with a multiplicity of infection (MOI) of 10
with or without 50 μM of substance B (S-4c) or rapamycin. The
used bacterial strains were L. pneumophila Philadelphia-1 JR32
(wild type, WT), the MIP-negative mutant L. pneumophila
Philadelphia-1 JR32-23, and the E. coli strain HB101. Two
hours after infection, extracellular bacteria were killed by gen-
tamicin and infection samples were plated on BCYE agar to
determine CFU counts immediately and after 24 and 48 h.
During incubation of the respective samples, 50 μM of rapamy-
cin or substance B (S-4c) were always present. Control experi-
ments showed that rapamycin and substance B (S-4c) do not
alter replication rates of bacterial cultures.
Acknowledgment. Financial support by the Deutsche
Forschungsgemeinschaft (SFB 630) is gratefully acknowledged.
Supporting Information Available: Experimental details of the
synthesis, analytical and spectral data of all synthesized com-
pounds. This material is available free of charge via Internet at
http://pubs.acs.org.
References
(1) Steinert, M.; Hentschel, U.; Hacker, J. Legionella pneumophila: an
aquatic microbe goes astray. FEMS Microbiol. Rev. 2002, 26 (2),
149–162.
(2) Cianciotto, N. P.; Stamos, J. V.; Kamp, D. W. Infectivity of
Legionella pneumophila MIP mutant for alveolar epithelial cells.
Curr. Microbiol. 1995, 30 (4), 247–250.
(3) Wintermeyer, E.; Ludwig, B.; Steinert, M.; Schmidt, B.; Fischer,
G.; Hacker, J. Influence of site specifically altered MIP proteins on
intracellular survival of Legionella pneumophila in eucaryotic cells.
Infect. Immun. 1995, 63 (12), 4576–4583.
(4) Wagner, C.; Khan, S. A.; Kamphausen, T.; Schmausser, B.; Unal,
C.; Lorenz, U.; Fischer, G.; Hacker, J.; Steinert, M. Collagen
binding protein MIP enables Legionella pneumophila to transmi-
grate through a barrier of NCI-H292 lung epithelial cells and
extracellular matrix. Cell. Microbiol. 2007, 9, 450–462.
General Procedure for the Synthesis of Compounds 5a-b. The
reduction of 4e and 4f was performed according to Bellamy
et al.19 Compound 4e and 4f (1 equiv), respectively, was sus-
pended in 50 mL of anhydrous ethanol and SnCl2 (5 equiv)
added in portions. This reaction was refluxed for 2 h. The
reaction was monitored by TLC (silica gel; triethylamine:ethyl-
acetate = 5:95). Thirty mL of water and NaHCO3 were added to
neutralize the mixture. The aqueous phase was washed with 4 ꢀ
30 mL ethyl acetate. The combined organic layers were dried
over anhydrous Na2SO4, and the solvent was removed in vacuo.
If necessary, a subsequent purification by silica gel column
chromatography was performed to obtain compounds 5a-b.
Protease-Coupled PPIase Assay. The PPIase activity of MIP
(and FKBP12, respectively) was determined by using the pep-
tide substrate Succinyl-Ala-Phe-Pro-Phe-4-nitroanilide in the
protease-coupled PPIase assay as described by Fischer et al.6 To
study the effect of compounds 2a-d, 4a-f, S-4c, and 5a-b on
PPIase activity, 40 nM MIP (20 nM FKBP12), and 40 μM
peptide substrate were preincubated in the presence or absence
of various inhibitor concentrations for 4 min at 10 °C in 35 mM
HEPES/NaOH buffer (pH 7.8). Then reaction was started by
addition of 100 μg/mL of the isomer specific protease chymo-
trypsin, and the release of 4-nitroaniline was monitored by
measuring the absorbance at 390 nm.
€
(5) Gelse, K.; Poschl, E.; Aigner, T. Collagens structure, function
;
and biosynthesis. Adv. Drug Delivery Rev. 2003, 12, 1531–1546.
(6) Fischer, G.; Bang, H.; Ludwig, B.; Mann, K.; Hacker, J. MIP
protein of Legionella pneumophila exhibits peptidyl-prolyl-cis/trans
isomerase (PPIase) activity. Mol. Microbiol. 1992, 6 (10), 1375–
1383.
(7) Ceymann, A.; Horstmann, M.; Ehses, P.; Schweimer, K.; Paschke,
A.; Steinert, M.; Faber, C. Solution structure of the Legionella
pneumophila MIP-rapamycin complex. BMC Struct. Biol. 2008,
8, 17.
€
(8) Riboldi-Tunnicliffe, A.; Konig, B.; Jessen, S.; Weiss, M. S.;
Rahfeld, J.; Hacker, J.; Fischer, G.; Hilgenfeld, R. Crystal struc-
ture of MIP, a prolylisomerase from Legionella pneumophila.
Nature Struct. Biol. 2001, 8, 779–783.
(9) Holt, D. A.; Luengo, J. I.; Yamashita, D. S.; Oh, H.; Konialian,
A. L.; Yen, H.; Rozamus, L. W.; Brandt, M.; Bossard, M. J.; Levy,
M. A.; Eggleston, D. S.; Liang, J.; Schultz, L. W.; Stout, T. J.;
Clardy, J. Design, Synthesis, and Kinetic Evaluation of High-
Affinity FKBP Ligands and the X-ray Crystal Structures of Their
Complexes with FKBP12. J. Am. Chem. Soc. 1993, 115, 9925–9938.
(10) Holt, D. A.; Konialian-Beck, A. L.; Oh, H.; Yen, H.; Rozamus,
L. W.; Krog, A. J.; Erhard, K. F.; Ortiz, E.; Levy, M. A.; Brandt,
M.; Bossard, M. J.; Luengo, J. Structure-activity studies of
synthetic FKBP Ligands as peptidyl-prolyl isomerase inhibitors.
Bioorg. Med. Chem. Lett. 1994, 4, 315–320.
NMR Spectroscopy: HSQC Experiments. 15N labeled MIP77-213
was expressed and purified as described previously.7 NMR
samples contained 0.1 mM 15N labeled MIP77-213 in 20 mM
potassium phosphate buffer, pH 6.5, and 10% D2O for field/
1
frequency lock. Standard H, 15N FHSQC experiments were
recorded at 298 K with a Bruker Avance 800 MHz NMR
spectrometer equipped with a cryogenically cooled probe.
Ligands were dissolved in deuterated acetonitrile to a concen-
tration of 23 mM. For each ligand, HSQC experiments were
recorded with protein:ligand ratios of 1:5, 1:10, and 1:20. A set
of reference experiments was recorded by adding the same
amount of deuterated acetonitrile without any ligand to the
protein solution.
€
(11) Fischer, G.; Mech, C.; Bang, H. Nachweis einer Enzymkatalyse fur
die cis/trans Isomerisierung der Peptidbindung in prolinhaltigen
Peptiden. Biomed. Biochim. Acta. 1984, 43, 1101–1111.
€
€
€
€
(12) Kohler, R.; Fanghanel, J.; Konig, B.; Luneberg, E.; Frosch, M.;
Rahfeld, J. U.; Hilgenfeld, R.; Fischer, G.; Hacker, J.; Steinert, M.
Biochemical and functional analyses of the MIP protein: influence
of the N-terminal half and of peptidylprolyl isomerase activity on

Article
the virulenceof Legionella pneumophila. Infect. Immun. 2003, 71 (8),
Journal of Medicinal Chemistry, 2011, Vol. 54, No. 1 283
Singh, U. C.; Weiner, P. K.; Kollman, P. A. AMBER 6; University
of California: San Francisco, 1999.
(16) Gasteiger, J.; Marsili, M. Iterative partial equilization of orbital
4389–4397.
(13) Morris, G. M.; Goodsell, D. S.; Halliday, R. S.; Huey, R.; Hart,
W. E.; Belew, R. K.; Olson, A. J. Automated docking using
a Lamarckian genetic algorithm and an empirical binding
free energy function. J. Comput. Chem. 1998, 19 (14), 1639–
1662.
(14) Jones, G.; Willett, P.; Glen, R. C.; Leach, A. R.; Taylor, R.
Development and validation of a genetic algorithm for flexible
docking. J. Mol. Biol. 1997, 267 (3), 727–748.
(15) Case, D. A.; Pearlman, D. A.; Caldwell, J. W.; Cheatham, T. E., III;
Ross, W. S.; Simmerling, C. L.; Darden, T. A.; Merz, K. M.;
Stanton, R. V.; Cheng, A. L.; Vincent, J. J.; Crowley, M.; Tsui, V.;
Radmer, R. J.; Duan, Y.; Pitera, J.; Massova, I.; Seibel, G. L.;
electronegativity a rapid access to atomic charges. Tetrahedron
;
1980, 36, 3219–3228.
(17) SYBYL Molecular Modeling Software; Tripos Inc.: St. Louis, MO, 2007.
(18) Hamilton, G. S.; Li, J.; Steiner, J. P. Method of using neurotrophic
sulfonamide compounds. U.S. Patent US 5721256 A 19980224, 1998.
(19) Bellamy, F. D.; Ou, K. Selective reduction of aromatic nitro
compounds with stannous chloride in nonacidic and nonaqueous
medium. Tetrahedron Lett. 1984, 25, 839–842.
(20) Hacker, J.; Ott, M.; Wintermeyer, E.; Ludwig, B.; Fischer, G.
Analysis of virulence factors of Legionella pneumophila. Zentralbl
Bakteriol. 1993, 278 (2-3), 348–358.