β-turn modified gramicidin S analogues containing arylated sugar amino acids display antimicrobial and hemolytic activity comparable to the natural product

Article abstract of DOI:10.1021/ja0588510

This paper describes the design and synthesis of gramicidin S (GS) analogues 10a-c containing arylated sugar amino acids (SAAs) as a replacement of one of the two ^DPhe-Pro β-turn regions. The cyclic, amphiphilic peptides adopt a β-sheet conform

Full text of DOI:10.1021/ja0588510

Published on Web 05/17/2006
â-Turn Modified Gramicidin S Analogues Containing Arylated
Sugar Amino Acids Display Antimicrobial and Hemolytic
Activity Comparable to the Natural Product
Gijsbert M. Grotenbreg,†, Annelies E. M. Buizert,† Antonio L. Llamas-Saiz,‡
Emile Spalburg,§ Peter A. V. van Hooft,| Albert J. de Neeling,§ Daan Noort,|
Mark J. van Raaij,‡ Gijsbert A. van der Marel,† Herman S. Overkleeft,*,† and
Mark Overhand*,†
Contribution from the Leiden Institute of Chemistry, Gorlaeus Laboratories, P.O. Box 9502,
2300 RA Leiden, The Netherlands, Unidad de Rayos X (RIAIDT), Laboratorio Integral de
Dina´mica y Estructura de Biomole´culas “Jose´ R. Carracido”, Edificio CACTUS, Campus Sur,
UniVersidad de Santiago de Compostela, 15782 Santiago de Compostela, Spain, National
Institute of Public Health and the EnVironment, Research Laboratory for Infectious Diseases,
P.O. Box 1, 3720 BA BilthoVen, The Netherlands, and TNO, Prins Maurits Laboratory,
P.O. Box 45, 2280 AA Rijswijk, The Netherlands
Received December 31, 2005; E-mail: h.s.overkleeft@chem.leidenuniv.nl; overhand@chem.leidenuniv.nl
Abstract: This paper describes the design and synthesis of gramicidin S (GS) analogues 10a-c containing
arylated sugar amino acids (SAAs) as a replacement of one of the two DPhe-Pro â-turn regions. The cyclic,
amphiphilic peptides adopt a â-sheet conformation featuring an unusual reverse turn induced by the SAAs.
The altered turn region induces a slight distortion of the antiparallel â-sheet, as compared to GS; the overall
geometry however closely resembles that of the nonarylated GS analogue 1. GS analogues 10a-c proved
to be as active as the parent GS itself as antibacterial agents and are equally efficient in lysing red blood
cells. These properties are in sharp contrast to the diminished biological activity displayed by 1. We conclude
that the presence of aromaticity in the turn regions of GS derivatives is required for biological activity,
whereas the native conformation of the beta-hairpin is not. Our findings may guide future research toward
efficient and nonhemolytic GS analogues for combating bacterial infections.
Introduction
Gramicidin S (GS, Figure 1) is an amphiphilic cyclic
decapeptide having the C2-symmetrical sequence cyclo(Pro-Val-
Orn-Leu-DPhe)2 that acts as an antibiotic by targeting the
membrane lipid bilayer.1,2 Upon accumulation into the lipid
bilayer, GS induces lysis with bacterial cell death as the ultimate
result. The ability of GS to interact with the bacterial membrane
is attributed to its amphiphilic nature, with the two basic Orn
residues occupying one side of the cyclic peptide and the
aliphatic Leu and Val residues located at the opposite side. The
amphiphilicity of GS is enhanced by its adoption of an
antiparallel â-sheet.3 This conformation is stabilized by four
intramolecular hydrogen bonding interactions at the opposing
† Leiden Institute of Chemistry.
‡ Universidad de Santiago de Compostela.
§ National Institute of Public Health and the Environment.
Figure 1. (A) Structures of gramicidin S and SAA-modified GS analogue
1. (B) The SAA residue in 1 and (C) the â-barrel-like assembly of 1 in the
X-ray structure.
| TNO.
Current address: Whitehead Institute for Biomedical Research, Nine
Cambridge Center, Cambridge, MA 02142, USA.
(1) Gause, G. F.; Brazhnikova, M. G. Nature 1944, 154, 703.
(2) (a) Izumiya, N.; Kato, T.; Aoyagi, H.; Waki, M.; Kondo, M. Synthetic
aspects of biologically actiVe cyclic peptides - gramicidin S and tyro-
cidines; Halstead (Wiley): New York, 1979. (b) Waki, M.; Izumiya, N.
Biochemistry of Peptide Antibiotics; Kleinhaug, H., van Doren, H., Eds.;
Walter de Gruyter Company: Berlin, 1990. (c) Kondejewski, L. H.; Farmer,
S. W.; Wishart, D. S.; Hancock, R. E. W.; Hodges, R. S. Int. J. Peptide
Protein Res. 1996, 47, 460-466.
Leu and Val residues and two hydrophobic type II′ â-turns with
the DPhe-Pro dipeptide stretches holding the i + 1 and i + 2
(3) (a) Schmidt, G. M. J.; Hodgkin, D. C.; Oughton, B. M. Biochem. J. 1957,
65, 744-756. (b) De Santis, P.; Liquori, A. M. Biopolymers 1971, 10,
699-710.
9
10.1021/ja0588510 CCC: $33.50 © 2006 American Chemical Society
J. AM. CHEM. SOC. 2006, 128, 7559-7565
7559

A R T I C L E S
Grotenbreg et al.
positions, respectively.4 The observation that GS acts against
the lipid bilayer itself and not against a specific cell membrane
associated biomolecule is underscored by the finding that
enantiomeric GS is equally active.5 GS targets a broad range
of bacteria depending on the type of bioassay, with solution-
based microbiological screening assays showing greater activity
against Gram-negative bacteria than gel-based assays.2c GS
appears rather indiscriminate toward the nature of the lipid
bilayer and kills mammalian cells with equal efficiency. For
instance, GS displays potent hemolytic activity, and it is for
this reason that the use of GS in human medicine is restricted
to topical applications.2,6
those of amino acids.11 SAAs can be readily obtained and
incorporated into oligopeptides as dipeptide isosteres. Our
strategy is to analyze the conformational behavior of SAA
modified GS analogues using both NMR and X-ray analysis
and to correlate the results with toxicity studies toward both
bacterial strains and red blood cells. In the course of these studies
we observed an intriguing conformation in GS analogue 1,
containing a furanoid SAA at one of the turn regions.10b One
of the two hydroxyl functionalities of the SAA moiety in 1 (C3-
OH) is involved in an intraresidue hydrogen bond leading to a
distortion of the â-hairpin structure (Figure 1B), as compared
to GS. Inspection of the molecular packing of 1 in its X-ray
structure revealed a hexameric â-barrel-like structure composed
of six crystallographically equivalent units of 1, with the
positively charged Orn side chains extending into the core and
the Val, Leu, and DPhe residues forming a hydrophobic
periphery (see Figure 1C). This hexameric â-barrel-like pore
structure, the first of its kind, agrees with a molecular arrange-
ment as suggested in a barrel stave type mechanism of
membrane disrupting cationic antibiotic peptides.12
At first, our interest in this X-ray structure was dampened
by the strongly reduced antibacterial activity displayed by 1.10a
We reasoned that the diminished biological activity of 1
compared to GS could be caused by either the altered â-hairpin
structure or by the hydrophilic character of the SAA moiety as
compared to the hydrophobic DPhe-Pro sequence in GS. As one
of the hydroxyl groups of the SAA moiety of 1 is not involved
in hydrogen bonding (C4-OH), it can be modified to introduce
an aromatic functionality, thus more closely resembling the
hydrophobic nature of the original turn sequence. We here
present a synthetic strategy that allows the facile introduction
of aromatic functionality in the SAA turn region of 1 (GS
analogues 10a-c) and provide an in-depth structural analysis,
as well as a biological evaluation. We show that analogues
10a-c adopt the same distorted cyclic â-hairpin secondary
structure as compound 1 but that the biological activity is
comparable to that of the natural product, gramicidin S.
Peptide antibiotics such as GS that target the lipid bilayer as
a whole, and not a specific subcellular target, are of great interest
in the search for new antibiotics.7 In general, bacterial strains
can readily become resistant against compounds that interfere
with a specific metabolic process, or block a specific enzyme
or receptor, by genetically altering the target such that it defies
recognition.8 Gaining resistance against GS would require a
strategy that, for instance, specifically destroys it or blocks its
accumulation in the lipid bilayer. Arguably, such alterations are
less easily attained through genetic mutations, and it is for this
reason that the search for nontoxic GS analogues and related
cationic antimicrobial peptides has found wide attraction.9
Obviously, this search can only be concluded successfully when
compounds are identified that do act upon bacterial lipid bilayers
but leave their mammalian counterparts untouched. For this to
be possible it is of profound importance to understand the exact
mode of action with which GS disrupts lipid bilayers. At present
such detailed information is not available.
We have contributed to the field of GS-based antibiotics by
the synthesis of analogues in which one of the two type II′
â-turns is replaced with selected sugar amino acids (SAA).10
SAAs are sugar peptide hybrids that combine monosaccharide
properties (conformational rigidity, added functionalities) with
(4) Hull, S. E.; Karlsson, R.; Main, P.; Woolfson, M. M.; Dodson, E. J. Nature
1978, 275, 206-207.
(5) Shemyakin, M. M.; Ovchinnikov, Y. A.; Ivanov, V. T.; Ryabova, I. D.
Experientia 1967, 23, 326. (b) Goodman, M.; Chorev, M. Acc. Chem. Res.
1979, 12, 1-7.
Materials and Methods
Abbreviations. GS ) gramicidin S, SAA ) sugar amino acid.
Synthesis. Except when indicated otherwise, all reactions were
performed under an inert atmosphere and at ambient temperature. All
reagents were purchased from Aldrich, Acros, or Novabiochem and
were used as supplied. CAUTION: Sodium azide is toxic and can
explode when exposed to heat, pressure, or shock, particularly in
combination with heavy metals or their salts. Organic azides, such as
derivatives 3-6a-c, can in principle also decompose explosively, and
appropriate safety measures must therefore be taken at all times. Full
details, including the characterization of the intermediates, are provided
in the Supporting Information.
(6) Kondejewski, L. H.; Farmer, S. W.; Wishart, D. S.; Hancock, R. E. W.;
Hodges, R. S. Int. J. Peptide Protein Res. 1996, 47, 460-466.
(7) (a) Matsuzaki, K. Biochim. Biophys. Acta 1999, 1462, 1-10. (b) Epand,
R. M.; Vogel, H. J. Biochim. Biophys. Acta 1999, 1462, 11-28. (c) Sitaram,
N.; Nagaraj, R. Biochim. Biophys. Acta 1999, 1462, 29-54. (d) Shai, Y.
Biochim. Biophys. Acta 1999, 1462, 55-70. (d) Dathe M.; Wieprecht, T.
Biochim. Biophys. Acta 1999, 1462, 71-87. (e) Blondelle, S. E.; Lohner,
K.; Aguilar, M.-I. Biochim. Biophys. Acta 1999, 1462, 89-108. (f) Maget-
Dana, R. Biochim. Biophys. Acta 1999, 1462, 109-140. (g) Lohner, K.;
Prenner, E. J. Biochim. Biophys. Acta 1999, 1462, 141-156. (h) Bechinger,
B. Biochim. Biophys. Acta 1999, 1462, 157-183. (i) La Rocca, P.; Biggin,
P. C.; Tieleman D. P.; Sansom, M. S. P. Biochim. Biophys. Acta 1999,
1462, 185-200. (j) Prenner, E. J.; Lewis, R. N. A. H.; McElhaney, R. N.
Biochim. Biophys. Acta 1999, 1462, 201-221. (k) Breukink, E. J.; de
Kruijff, B. Biochim. Biophys. Acta 1999, 1462, 223-234.
(8) Walsh, C. T. Antibiotics: Actions, Origins, Resistance; ASM: Washington,
DC, 2003.
(9) (a) Zasloff, M. Nature 2002, 415, 389-395. (b) Lee, D. L.; Hodges, R. S.
Biopolymers 2003, 71, 28-48. (c) Bulet, P.; Sto¨cklin, R.; Menin, L.
Immunol. ReV. 2004, 198, 169-184. (d) Finlay, B. B.; Hancock, R. E. W.
Nat. ReV. Microbiol. 2004, 2, 497-504. (e) Lehrer, R. I. Nat. ReV.
Microbiol. 2004, 2, 727-738.
Representative Alkylation Procedure. Alcohol 3 (1.96 g, 8.55
mmol) was dissolved in DMF (40 mL) and cooled to 0 °C. To the
solution were added benzyl bromide (1.124 mL, 9.4 mmol, 1.1 equiv)
and sodium hydride (376 mg, 9.4 mmol, 1.1 equiv), and the mixture
(10) (a) Grotenbreg, G. M.; Kronemeijer, M.; Timmer, M. S. M.; El Oualid, F.;
van Well, R. M.; Verdoes, M.; Spalburg, E.; van Hooft, P. A. V.; de
Neeling, A. J.; Noort, D.; van Boom, J. H.; van der Marel, G. A.; Overkleeft,
H. S.; Overhand, M. J. Org. Chem. 2004, 69, 7851-7859. (b) Grotenbreg,
G. M.; Timmer, M. S. M.; Llamas-Saiz, A. L.; Verdoes, M.; van der Marel,
G. A.; van Raaij, M. J.; Overkleeft, H. S.; Overhand, M. J. Am. Chem.
Soc. 2004, 126, 3444-3446. (c) Grotenbreg, G. M.; Christina, A. E.;
Buizert, A. E. M.; van der Marel, G. A.; Overkleeft, H. S.; Overhand, M.
J. Org. Chem. 2004, 69, 8331-8339.
(11) (a) Gervay-Hague, J.; Weathers, T. M. J. Carbohydr. Chem. 2002, 21, 867-
910. (b) Schweizer, F. Angew. Chem., Int. Ed. 2002, 41, 230-253. (c)
Gruner, S. A. W.; Locardi, E.; Lohof, E.; Kessler, H. Chem. ReV. 2002,
102, 491-514. (d) Chakraborty, T. K.; Srinivasu, P.; Tapadar, S.; Mohan,
B. K. J. Chem. Sci. 2004, 116, 187-207.
(12) (a) Oren, Z.; Shai, Y. Biopolymers 1998, 47, 451-463. (b) Shai, Y.
Biopolymers 2002, 66, 236-248. (c) Yeaman, M. R.; Yount, N. Y.
Pharmacol. ReV. 2003, 55, 27-55. (d) Brogden, K. A. Nat. ReV. Microbiol.
2005, 3, 238-250.
9
7560 J. AM. CHEM. SOC. VOL. 128, NO. 23, 2006

â
-Turn Modified Gramicidin S Analogues
A R T I C L E S
was stirred for 16 h. The reaction was quenched with methanol (5 mL)
and concentrated in vacuo. The crude mixture was partitioned between
saturated aqueous NaHCO3 and EtOAc, and the aqueous layer was
subsequently extracted with EtOAc twice. The combined organic layers
were dried (MgSO4), filtered, and concentrated. Silica gel column
chromatography (20%f30% EtOAc in light PE) yielded the azide 4a
(2.54 g, 7.95 mmol) in 93% as a transparent oil.
and refined with full-matrix least-squares analysis on F2 using
SHELXL-97.16 Of the reflections, 5% were flagged to compute the
Rfree except for the last refinement cycles where all reflection were
included to calculate the final esd for the refined parameters and
the R1 and wR2 values. Due to the limited resolution of 1.0 Å, the
partial disorder of the peptide, and the presence of very big solvent
channels in the crystal, hydrogen atoms could not be localized for the
ordered water solvent molecules. All non-H atoms were refined
anisotropically (for data and refinement statistics, see Supporting
Information). CCDC 278146 contains the supplementary crystal-
lographic data for this structure. These data can be obtained free of
charge via www.ccdc.cam.ac.uk/conts/retrieving.html (or from the
Cambridge Crystallographic Data Centre, 12 Union Road, Cambridge
CB21EZ, UK).
Representative Methanolysis Procedure. Isopropylidene-protected
4a (1.6 g, 5.0 mmol) was dissolved in methanol (10 mL), and PPTS
(∼50 mg, cat) was added. The mixture was heated to 50 °C and stirred
for 16 h. The reaction was next diluted with EtOAc, extracted with
saturated aqueous NaHCO3, dried (MgSO4), filtered, and evaporated
to dryness. Silica gel column chromatography (50%f70% EtOAc in
light PE) produced diol 5a (0.945 g, 3.39 mmol) in 68% as a white
amorphous solid.
Antibacterial Activity. The following bacterial strains were used:
Staphylococcus aureus (ATCC 29213), Staphylococcus epidermidis
(ATCC 12228), Enterococcus faecalis (ATCC 29212), Bacillus cereus
(ATCC 11778), Escherichia coli (ATCC 25922), and Pseudomonas
aeruginosa (ATCC 27853). Bacteria were stored at -70 °C and grown
at 35 °C on Columbia Agar with sheep blood (Oxoid, Wesel, Germany)
overnight and diluted in 0.9% NaCl. Microtiter plates (96 wells of
100µL) as well as large plates (25 wells of 3 mL) were filled with
Mueller Hinton II Agar (Becton Dickinson, Cockeysvill, USA) contain-
ing serial 2-fold dilutions of peptides 10a-c. To the wells were added
3 µL of bacteria, to give a final inoculum of 104 colony forming units
(CFU) per well. The plates were incubated overnight at 35 °C and the
MIC was determined as the lowest concentration inhibiting bacterial
growth.
Representative Oxidation Procedure. Diol 5a (474 mg, 1.70 mmol)
was dissolved in acetonitrile (10 mL) after which saturated aqueous
NaHCO3 (4 mL) containing KBr (∼20 mg, cat) was added, and the
mixture was cooled to 0 °C. Subsequently, TEMPO catalyst (5 mg) in
acetonitrile (3 mL) was added, and a premixed solution of 15% NaOCl
(7.5 mL, 1.80 mmol), in saturated aqueous NaHCO3 (4.5 mL) and
saturated aqueous NaCl (8.8 mL), was added dropwise to the reaction
mixture, keeping it oscillating between yellow and colorless. The
reaction was then quenched with MeOH, acidified to pH ) 4 with
HCl (1 mL), and extracted with DCM thrice. The combined organic
layers were dried (MgSO4), filtered, and evaporated. Silica gel column
chromatography (20%f30% EtOAc in light PE) quantitatively fur-
nished SAA 6a (507 mg, 1.70 mmol) as a transparent oil.
Hemolytic Activity. The hemolytic activity of the peptides was
determined in quadruple. Human blood was collected into EDTA tubes
and centrifuged to remove the buffy coat. The residual erythrocytes
were washed 3 times in 0.85% saline. Serial 2-fold dilutions of the
peptides 10a-c in saline were prepared in sterilized round-bottom 96-
well plates (polystyrene, U-bottom, Costar) using 100 µL volumes
(500-0.5 µM). Red blood cells were diluted with saline to 1/25 packed
volume of cells, and 50 µL of the resulting cell suspension were added
to each well. Plates were incubated while gently shaking at 37 °C for
4 h. Next, the microtiter plate was quickly centrifuged (1000 g, 5 min),
and 50 µL of supernatant of each well were transported into a flat-
bottom 96-well plate (Costar). The absorbance was measured at 405
nm using an mQuant microplate spectrophotometer (Bio-Tek Instru-
ments). The Ablank was measured in the absence of additives and 100%
hemolysis (Atot) in the presence of 1% Triton X-100 in saline. The
percentage hemolysis is determined as (Apep - Ablank)/(Atot - Ablank) ×
100.
Peptide Synthesis. Resin 7 was prepared as reported earlier10a and
was elongated by repetitive removal of the Fmoc protecting group and
coupling of the next amino acids under standard conditions. Final
coupling of 6a, 6b, or 6c (0.2 mmol, 2 equiv), BOP (132 mg, 0.3 mmol,
3 equiv), HOBt (41 mg, 0.3 mmol, 3 equiv), and DiPEA (58 µL, 0.35
mmol, 3.5 equiv) in NMP (2 mL) gave the immobilized azides 8a-c.
The resins were then dispersed in 1,4-dioxane (10 mL) to which
trimethylphosphine (1.6 mL, 1.6 mmol, 1 M in toluene, 16 equiv) and
water (1 mL) were added. After shaking for 4 h, the solid phase was
washed, and the resulting peptides were released from the resin (1%
TFA in DCM) and cyclized by dropwise addition to a vigorously stirred
solution of PyBOP (5 equiv), HOBt (5 equiv), and DiPEA (15 equiv)
to furnish the compound 9a (116 mg, 86 µmol, 86%), 9b (80 mg, 56
µmol, 56%), and 9c (137 mg, 98 µmol, 98%), respectively, as white
amorphous solids, after size exclusion chromatography. The Boc-
protected peptides 9a (58 mg, 43 µmol), 9b (20 mg, 14 µmol), and 9c
(11.6 mg, 8.3 µmol) were dissolved in DCM (2 mL) and cooled to 0
°C. The mixtures were subsequently treated with TFA (2 mL), stirred
for 30 min, and diluted with toluene (10 mL), after which the volatiles
were removed in vacuo. The crude products were purified by semi-
preparative HPLC to give, after lyophilization, 10a-c in 87%, 34%,
and 57% yield, respectively.
Results
Synthesis. The synthesis of GS analogues 10a-c commenced
with the preparation of arylated SAA building blocks SAAs
6a-c as follows. 2,5-Anhydroglucitol 3 (Scheme 1), readily
prepared from D-mannitol 2 following the literature procedure,17
was alkylated with benzyl bromide, 4-biphenylmethyl bromide,
and naphthyl bromide, respectively, under the agency of sodium
hydride to furnish 4a (93%), 4b (89%), and 4c (77%). The
naphthylmethyl- and biphenylmethyl bromides required for the
preparation of 4b and 4c were prepared from their corresponding
X-ray Crystallography. Lyophilized peptide 10c (0.89 mg, 0.74
µmol) was dissolved in 75 µL of MeOH/H2O (2/1, v/v), after which
aliquots of 5 µL of the solution were pipetted into a 96-well microtiter
plate (Terasaki plate) under a layer of n-decane. To the sample, 5 µL
of different crystallization solutions were added, after which the
n-decane layer was replaced with paraffin oil. The microtiter plate was
allowed to stand at 20 °C. Crystals developed at several conditions. A
crystal grown in 15% (v/v) 2-methyl-2,4-pentanediol, 50 mM sodium
acetate, and 10 mM calcium chloride (pH 4.6) was mounted in a
cryoloop and frozen in liquid nitrogen. A complete dataset was collected
at ESRF (Grenoble, France) at beamline ID23-1 from one crystal (0.24
× 0.08 × 0.06 mm3) at 100 K using an oscillation camera equipped
with an MAR Mosaic CCD detector. A total of 329 oscillation images
were taken in three runs. Data were processed using HKL Denzo13
and SORTAV.14 The structure was solved by direct methods (SIR200215)
(13) Otwinowski, Z.; Minor, W. Methods in Enzymology, Volume 276: Mac-
romolecular Crystallography, Part A; Carter, C. W., Jr., Sweet, R. M.,
Eds.; Academic Press: 1997; pp 307-326.
(14) Blessing, R. H. Acta Crystallogr. 1995, A51, 33-38.
(15) Burla, M. C.; Camalli, M.; Carrozzini, B.; Cascarano, G. L.; Giacovazzo,
C.; Polidoria, G.; Spagna, R. J. Appl. Crystallogr. 2003, 36, 1103.
(16) Sheldrick, G. M. SHELX97; University of Go¨ttingen: Germany, 1997.
(17) Timmer, M. S. M.; Verdoes, M.; Sliedregt, L. A. J. M.; van der Marel, G.
A.; van Boom, J. H.; Overkleeft, H. S. J. Org. Chem. 2003, 68, 9406-
9411.
9
J. AM. CHEM. SOC. VOL. 128, NO. 23, 2006 7561

A R T I C L E S
Grotenbreg et al.
Scheme 1. Synthesis of the Sugar Amino Acidsa
a (i) NaH (1.1 equiv), RBr (1 equiv), DMF, 0 °C, 16 h, 4a, 93%; 4b, 89%; 4c, 77%; (ii) PPTS (cat.), MeOH, 50 °C, 16 h, 5a, 68%; 5b, 58%; 5c, 69%;
(iii) TEMPO (cat.), NaOCl, NaHCO3 (aq), MeCN, 0 °C, 6a, quant; 6b, 52%; 6c, 72%.
Scheme 2. Synthesis of the GS Analoguesa
a (i) Fmoc deprotection: piperidine/NMP (1/4 v/v), condensation: Fmoc-aa-OH (3 equiv) or SAA, 6a, 6b, and 6c (2 equiv), BOP (3 equiv), HOBt (3
equiv), DiPEA (3.3 equiv), NMP, 90 min; (ii) PMe3 (16 equiv), 1,4-dioxane/H2O (10/1 v/v); (iii) TFA/DCM (1/99 v/v), 4 × 10 min; (iV) PyBOP (5 equiv),
HOBt (5 equiv), DiPEA (15 equiv), DMF, 16 h, 9a, 86%; 9b, 56%; 9c, 98%; (V) TFA/DCM (1/1 v/v) 30 min, 10a, 87%; 10b, 34%; 10c, 57%.
alcohols.18 Cleavage of the isopropylidene group using a
catalytic amount of pyridinium p-toluenesulfonate (PPTS) gave
diols 5a-c in their respective yields of 68%, 58%, and 50%.
Oxidation of the primary alcohol in 5a-c, employing a catalytic
amount of TEMPO (2,2,6,6-tetramethyl-piperidinyl-1-oxy) and
NaOCl as co-oxidant, afforded SAAs 6a (quant.), 6b (52%),
and 6c (72%), respectively.
Next, the assembly of 10a-c was undertaken (Scheme 2).
MBHA-resin 7, functionalized with the acid-labile HMPB-linker
system, was condensed with Fmoc-Leu-OH and further elon-
gated using standard Fmoc-based SPPS protocols. The N-
terminal azides of the immobilized nonapeptides 8a-c were
subjected to Staudinger reduction, and the resulting peptides
were released from the resin (1% TFA in DCM) and cyclized
by dropwise addition to a vigorously stirred solution of PyBOP
(5 equiv), HOBt (5 equiv), and DiPEA (15 equiv). Size
exclusion chromatography gave the fully protected 9a (86%),
9b (56%), and 9c (98%), respectively. Removal of the Boc
protective groups and HPLC purification furnished homoge-
neous GS analogues 10a (87%), 10b (34%), and 10c (57%),
respectively.
NMR Analysis. The unambiguous 1H NMR resonance
assignment of each peptide was performed using COSY,
TOCSY, and NOESY or ROESY experiments. The large
resonance distribution, indicative of secondary structure forma-
tion, allowed complete assignment of the data sets of peptides
10a-c. To aid in identifying the presence of secondary structure
elements, the vicinal spin-spin coupling constants19 of the
amide bonds (3JHNR, Figure 2A) and the chemical shift
perturbation20 of the HR of individual amino acid residues
(∆δHR, Figure 2B) were compared to those found in native GS.
The distinctive trends found in these data were largely compa-
rable to GS and 1 and consistent with a â-sheet structure.10
The preservation of the â-sheet structure in benzyl-derivative
10a was further corroborated by the observation of interstrand
NH-NH NOEs, such as Val2-Leu9 (NOE b, Figure 3) and Val7-
(19) Wu¨thrich, K. NMR of Proteins and Nucleic Acids, John Wiley & Sons:
New York, 1986.
(18) Brun, K. A.; Linden, A.; Heimgartner, H. HelV. Chim. Acta 2002, 85, 3422-
(20) Wishart, D. S.; Sykes, B. D.; Richards, F. M. Biochemistry 1992, 31, 1647-
3443.
1651.
9
7562 J. AM. CHEM. SOC. VOL. 128, NO. 23, 2006

â
-Turn Modified Gramicidin S Analogues
A R T I C L E S
analogue 10a. Similar observations through spectroscopic
comparison were made for biphenyl- and naphthyl-functional-
ized 10b and 10c, respectively. Thus, the monoarylated SAAs
6a-c appear to induce the same reverse turn conformation in
GS analogues 10a-c in solution as the fully hydroxylated SAA
moiety does in GS analogue 1.
X-ray Analysis. Buffered aqueous solutions of 10c were
allowed to evaporate slowly under oil. After 4 months, crystals
of 10c developed, and the crystal structure of one of them was
solved by X-ray diffraction (Figure 4A). This provided final
confirmation, in line with our spectroscopic data, that the peptide
indeed adopts a â-sheet structure with a distinctive reverse turn
that is identical to 1 (Figure 1B). Inspection of the molecular
packing of a single unit cell revealed a porelike assembly
comprised of 12 individual cyclic peptides (Figure 4B). Peptide
monomers related by a binary axis perpendicular to the pore
channel form dimers (Figure 4C). Six of these dimers, arranged
in a helix, give rise to a complete turn of the helical pore. This
pore subsequently forms an infinite helical tube displaying 65
screw axis symmetry. The pore assemblies of 10c, those
observed in the molecular packing of a hydrated GS-urea
complex (Figure 4D),21 and GS-trifluoroacetic acid-HCl
(Figure 4E) are all helical motifs and are distinct to the stacking
of hexameric barrels displayed by 1 (Figure 4F). For example,
the â-strands of the cyclic peptide units 10c are almost
perpendicular with the axis of the pore, whereas the â-strands
of 1 are closer to a parallel orientation with respect to the pore
axis. Moreover, the assembly of 10c (Figure 4B) forms a larger
pore than the assembly of 1 (Figure 1C). However, all
assemblies are similar with respect to their hydrophilic interior
and hydrophobic exterior.
Figure 2. NMR analysis of SAA-modified GS analogues 10a-c. (A)
Coupling constants (3JHNR). (B) The chemical shift perturbations (∆δHR )
observed δHR - random coil δHR).
Antibacterial and Hemolytic Activity. The bactericidal
activity of GS analogues 10a-c against a number of Gram-
positive and -negative strains was evaluated (Table 1). The
Gram-positive strains S. aureus, S. epidermidis, E. faecalis, and
B. cereus proved to be highly sensitive toward the GS analogues
10a-c, with MIC values comparable to those observed for
native GS. As we had established earlier,10a the nonarylated GS
analogue 1 was largely inactive against all bacterial strains
included in the assay. The Gram-negative E. coli and P.
aeruginosa strains proved resistant to the action of 10a-c, as
well as to GS.2c The nature of the aromatic appendage from
the SAA does not influence the biological profile, as the benzyl,
biphenyl, and naphthyl derivatives proved to be equally active.
The hemolytic properties of peptides 10a-c toward human
erythrocytes were similarly assessed and compared to those of
native GS (Figure 5).
Peptides 10a-c show a similar toxicity compared to GS that
is not influenced by the difference in aromatic moieties, as the
benzyl, biphenyl, and naphthyl derivatives display comparable
hemolytic profiles. These data demonstrate that the advantageous
effect of aromatic decoration of the SAAs on the capacity of
peptides 10a-c to arrest proliferation of various strains of
bacteria concurrently increases the toxicity toward human
erythrocytes.22
Figure 3. Amide region of the ROESY spectrum (600 MHz, CD3OH) of
peptide 10a.
Leu4 (NOE d), as well as the sequential NH-NH cross-peak
of SAA1-Leu9 (NOE c). Moreover, the characteristic NOE
contact between SAA1-NH and its neighboring Val2-NH was
discerned (NOE a, Figure 3), providing evidence that the
structure adopted by 1, as observed through 1H NMR and single-
crystal X-ray analysis, was again assumed by aromatic GS
(21) Tishchenko, G. N.; Andrianov, V. I.; Vainstein, B. K.; Woolfson, M. M.;
Dodson, E. Acta Crystallogr. 1997, D53, 151-159.
(22) Several less toxic GS analogues have been reported: (a) Ando, S.;
Nishikawa, H.; Takiguchi, H.; Lee, S.; Sugihara, G. Biochim. Biophys. Acta
1993, 1147, 42-49. (b) Lee, D. L.; Hodges, R. S. Biopolymers 2003, 71,
28-48. (c) Kawai, M.; Yamamura, H.; Tanaka, R.; Umemoto, H.; Ohmizo,
C.; Higuchi, S.; Katsu, T. J. Pept. Res. 2005, 65, 98-104.
9
J. AM. CHEM. SOC. VOL. 128, NO. 23, 2006 7563

A R T I C L E S
Grotenbreg et al.
Figure 4. (A) Monomeric X-ray structure of 10c. (B) The molecular packing of 10c along the pore axis. (C) The dodecameric helical assembly viewed
perpendicular to the pore. (D) The hexameric helical assembly of the GS-urea complex. (E) The nonameric helical assembly of the GS-TFA-HCl complex.
(F) The hexameric barrel of 1. Solvent and water molecules have been omitted for clarity.
Table 1. Antimicrobial Activity (MIC in µg/mL) of Peptides 10a-c in Comparison to GSa
S. aureusb
S. epidermidisb
E. faecalisb
B. cereusb
E. colic
P. aeruginosac
peptide
25Wd
MTe
25Wd
MTe
25Wd
MTe
25Wd
MTe
25Wd
MTe
25Wd
MTe
GS
4
8
8
4-8
4
2
8
4
2
2
2
2
2
2
4
4
8
8
16
8-16
8
4
4
8
4
4
4
8
4
>64
64
>64
>64
64
64
>64
>64
>64
>64
>64
>64
>64
>64
>64
>64
10a
10b
10c
16-32
16
8
a Measurements were executed using standard agar 2-fold dilution techniques. b Gram-positive. c Gram-negative. d 3 mL/25 well plates. e 100 µL/96 microtiter
plates.
Figure 5. Hemolytic activity of GS analogues 10a-c in comparison with GS.
Discussion
irrespective of the aromatic group appended. Examination of
the biological activity revealed that the antibacterial properties
of the aromatic SAA-containing GS analogues was completely
restored to the level of the parent compound GS, with a
concomitant increase in hemolytic activity. We conclude that
the presence of aromatic functionality in the turn region of these
GS derivatives is a requirement for antibacterial activity.
GS analogue 1, the intriguing structure of which we have
reported previously and that exhibits limited antibacterial
activity, formed the basis for the design of a new series of
aromatic SAA-containing GS analogues 10a-c. All three cyclic
peptides 10a-c adopt a â-sheet structure with the SAAs
occupying a distinctive reverse turn analogous to peptide 1,
9
7564 J. AM. CHEM. SOC. VOL. 128, NO. 23, 2006

â-Turn Modified Gramicidin S Analogues
A R T I C L E S
Alteration of the conformation of one of the turns of GS is
allowed with respect to the biological activity as long as this
modified turn contains an aromatic functionality. This finding
strongly suggests an important role of the aromatic function-
alities in the mechanism of action of membrane disruption by
GS.
At this point we cannot correlate conclusively the molecular
packing of SAA analogues of GS, as observed in the X-ray
structure, with their biological properties. However, we do
believe it is striking that the active compounds so far analyzed
(10c and native GS) form helical pores in crystals, while the
less active compound 1 forms stacks of â-barrels. The work
presented here encourages further efforts to design and synthe-
size alternate compounds to deepen our understanding of the
structural requirements for biological activity of GS and related
cationic antibiotic peptides, with the ultimate goal to arrive at
GS analogues that target bacterial strains but are inactive against
human blood cells.
Acknowledgment. This work was financially supported by
the Council for Chemical Sciences of The Netherlands Organi-
zation for Scientific Research (CW-NWO), The Netherlands
Technology Foundation (STW), and DSM Research. Kees
Erkelens and Fons Lefeber are gratefully acknowledged
for assistance with NMR experiments. We thank Nico
Meeuwenoord and Hans van den Elst for their technical
assistance. We thank Dr. David Hall for help with crystal-
lographic data collection. A.L.L.-S. and M.J.v.R. also thank the
Spanish Ministry of Education and Science for research funding
and a Ramo´n y Cajal research fellowship to M.J.v.R.
Supporting Information Available: Detailed experimental
procedures, compound characterization data, and experimental
data for crystal structure determination of GS analogue 10c.
This material is available free of charge via the Internet at
http://pubs.acs.org.
JA0588510
9
J. AM. CHEM. SOC. VOL. 128, NO. 23, 2006 7565