Structure and antibacterial activity of a protein from the basal metazoan Hydra

24 Pages·2008·0.83 MB·English

by  Joachim Grötzinger

JBC Papers in Press. Published on November 19, 2008 as Manuscript M804713200
The latest version is at http://www.jbc.org/cgi/doi/10.1074/jbc.M804713200
Hydramacin-1: Structure and activity
1Hydramacin-1: Structure and antibacterial activity of a
protein from the basal metazoan Hydra

Sascha Jung1, Andrew J. Dingley2, René Augustin3, Friederike Anton-Erxleben3,
Mareike Stanisak4, Christoph Gelhaus4, Thomas Gutsmann5, Malte U. Hammer5,
Rainer Podschun6, Alexandre M. J.J. Bonvin7 ,Matthias Leippe4, Thomas C.G. Bosch3
and Joachim Grötzinger1

1Institute of Biochemistry, Christian-Albrechts-University of Kiel, Olshausenstr. 40, 24098 Kiel,
Germany
2Department of Chemistry and School of Biological Sciences, The University of Auckland, Auckland,
New Zealand
3Institute of Zoology, Cell and Developmental Biology, Christian-Albrechts-University of Kiel, Am
Botanischen Garten 1-9, 24118 Kiel, Germany
4Institute of Zoology, Zoophysiology, Christian-Albrechts-University of Kiel, Olshausenstr. 40,
24098 Kiel, Germany
5Research Center Borstel, Leibniz-Center for Medicine and Biosciences, Parkallee 10, 23845 Borstel,
Germany D
6Institute for Infection Medicine, University Hospital Schleswig-Holstein, Campus Kiel, 24105 Kiel, ow
n
Germany lo
a
d
7Bijvoet Center for Biomolecular Research, Science Faculty, Utrecht University, 3584CH, Utrecht, ed
The Netherlands fro
m
h
ttp
Address correspondence to: Joachim Grötzinger, Institute of Biochemistry, Christian-Albrechts- ://w
University of Kiel, Olshausenstr. 40, 24118 Kiel, Germany, Tel. [phone]; Fax. w
w
0049(0)431 880-5007; E-Mail: [email protected] .jb
c
.o
rg
b/
y
Hydramacin-1 is a novel antimicrobial of positively charged residues is sandwiched g
u
e
protein recently discovered during by two hydrophobic areas. Based on this st o
investigations of the epithelial defense of the characteristic surface feature and on n J
a
ancient metazoan Hydra. The amino-acid biophysical analysis of protein-membrane nu
a
sequence of hydramacin-1 shows no sequence interactions we propose a model that ry
1
2
homology to any known antimicrobial describes the aggregation effect exhibited by , 2
0
proteins. Determination of the solution hydramacin-1. 1
9
structure revealed that hydramacin-1
possesses a disulfide-bridge stabilized αβ INTRODUCTION
motif. This motif is the common scaffold of
the knottin protein fold. The structurally The discovery of bacteria resistant to common
closest relatives are the scorpion-toxin like classes of antibiotics is rapidly increasing (1). In
superfamily. Within this superfamily particular, the rise of multidrug-resistant bacteria
hydramacin-1 establishes a new family of is alarming (1). Consequently, there is an urgent
proteins that all share antimicrobial activity. need to discover alternative classes of active
Hydramacin-1 is potently active against antibiotic compounds. One such class, which
Gram-positive and Gram-negative bacteria offers tremendous potential, are antimicrobial
including multi resistant human pathogenic peptides (AMPs) (2). AMPs act predominantly
strains. It leads to aggregation of bacteria as in a mechanical manner by attacking the
an initial step of its bactericidal mechanism. microorganisms membranes (3) in order to form
Aggregated cells are connected via electron- stable or transient pores. Alternatively, AMPs
dense contacts and adopt a thorn apple-like induce micellization in a detergent-like manner
morphology. Analysis of the hydramacin-1 (4). Both mechanisms are effective at destroying
structure revealed an unusual distribution of the target bacteria (5).
amino acid side-chains on the surface. A belt
1
Copyright 2008 by The American Society for Biochemistry and Molecular Biology, Inc.
Hydramacin-1: Structure and activity
Currently, different AMPs have been identified Purification, renaturation and cleavage of the
and isolated from a wide variety of organisms fusion protein
including humans where they mostly participate After expression, bacteria were harvested by
in the first-line of defense against a pathogen (4). centrifugation and resuspended in PBS
AMPs are grouped into superfamilies and containing 0.2 % (v/v) Tween-20. The cell
families based on their three dimensional suspension was sonicated and centrifuged to
structures. isolate the inclusion bodies (IB). The IB pellet
Hydramacin-1 is a cationic antimicrobial peptide was washed twice in PBS/0.2 % Tween-20 and
that was isolated from the basal metazoan Hydra three-times with PBS only. The IB pellet was
and is active against Gram-positive and Gram- dissolved in a denaturing buffer (6 M guanidine
negative bacteria (6). The primary structure of hydrochloride, 50 mM Tris, pH 7.2) and
hydramacin-1, in particular the location of the incubated at 20 °C overnight.
cysteine residues in the primary structure, Insoluble material was removed by
resembles only two cationic antimicrobial centrifugation and the supernatant was incubated
peptides isolated from the leech: theromacin (7) at 20 °C for 4 h with Ni2+-NTA-agarose beads
and neuromacin (8). The tertiary structures of (Qiagen, Hilden, Germany) that were
theromacin and neuromacin are unknown. equilibrated with the denaturing buffer. After
Furthermore, the mechanism these two peptides incubation, the beads were thoroughly washed
use to exhibit their antimicrobial activity is also with 6 M guanidine hydrochloride, 50 mM Tris,
not entirely understood. pH 8.0 buffer before elution of the target protein D
o
w
In this study, we present the NMR derived with the same buffer containing 250 mM n
lo
solution structure of hydramacin-1, evaluate the imidazole. ad
e
peptide’s antimicrobial spectrum and The eluted fractions were diluted with d fro
characterize its mode of action using various denaturation buffer to a final protein m
h
biophysical techniques. Based on these results a concentration of 0.3 mg/mL. This solution was ttp
model of the interrelation between the unusual diluted rapidly in a ratio of 1:6 with an aqueous ://w
w
structural features of hydramacin-1 and its mode solution containing 4 mM reduced w
of action is presented. glutathione/0.4 mM oxidized glutathione. .jbc
.o
Dilution was achieved by dropping the protein rg
EXPERIMENTAL PROCEDURES solution slowly (~ 1 drop per 2 s) into the by/
g
aqueous glutathione containing redox system. u
e
s
Cloning procedure Throughout this process, the solution was t o
n
N-terminal sequencing of hydramacin-1 which rapidly stirred. After adjustment of the pH to 8.5, Ja
n
u
has been isolated from the pathogen-challenged the solution was incubated at 20 °C for 72 h. a
ry
Hydra magnipapillata enabled the isolation of a After incubation, the protein sample was 1
2
corresponding cDNA1. The nucleic acid concentrated ten-fold with a Vivaspin , 2
0
1
sequence was used for alignment analysis with concentrator (Vivascience AG, Hannover, 9
the Hydra genome. Germany) and dialyzed against 50 mM Tris, pH
The identified cDNA was amplified and cloned 8.0 at 20 °C for 48 h.
into the expression vector pET-32a and The dialyzed protein solution was centrifuged to
transformed into the E. coli strain BL21 (DE3). remove precipitated material. The remaining
soluble protein in the supernatant was cleaved by
Expression the addition of EnterokinaseMaxTM (Invitrogen).
Hydramacin-1 was recombinantly expressed in Cleavage was performed at 37 °C overnight.
the E. coli strain BL21 (DE3) as a fusion protein
comprising a thioredoxin-His -tag fused with an Purification of mature hydramacin-1
6
enterokinase cleavage site N-terminal to the Purification of mature hydramacin-1 was
mature protein. Uniformly 15N- and 15N/13C- performed using reversed-phase high
enriched hydramacin was prepared using M9 performance liquid chromatography (RP-HPLC)
minimal medium (9). When cell cultures reached on a semi-preparative C18-column (Macherey-
an optical density of ~0.2, IPTG was added to Nagel, Düren, Germany). A continuous
the culture to a final concentration of 1 mM to acetonitrile gradient of 1.4 % per min over a
induce protein expression. The cell growth was time range of 13 min starting at a concentration
continued for 3 h. of 25 % (v/v) was used to obtain isolated
hydramacin-1. Fractions containing
2
Hydramacin-1: Structure and activity
hydramacin-1 were pooled and lyophilized. calculations: 2.0 ≤ d(Sγ, Sγ) ≤ 2.1 Å; 3.0 ≤ d(Cβ,
i j i
Purity was confirmed by mass spectrometry in Sγ) ≤ 3.1 Å; 3.0 ≤ d(Sγ, Cβ) ≤ 3.1 Å. In addition,
j i j
linear mode using a 4700 Proteomics Analyzer 20 distance restraints from 10 hydrogen bonds
MALDI TOF/TOF mass spectrometer (Applied were set to a range of 1.8-2.0 Å between the
Biosystems, Framingham, U.S.A.). Size- donor hydrogen atom and the acceptor oxygen
exclusion chromatography on a HiLoad atom and 2.7-3.0 Å between the donor nitrogen
Superdex 75 prep grade (16/60) column atom and the acceptor oxygen atom. Finally, 25
(Amersham Biosciences, Freiburg, Germany) structures with the lowest target functions were
under conditions used for NMR spectroscopy selected from 500 calculated structures. The
confirmed the monomeric state of the peptide average structure was calculated from the
(data not shown). ensemble of these 25 structures and
subsequently energy minimized using the
Sample preparation and NMR experiments GROMOS force field (15). Each of 25 structures
The freeze-dried 15N- or 15N/13C-enriched of the ensemble was further refined in explicit
peptide was dissolved in a buffer consisting of solvent with an 8Å water shell using the
50 mM sodium-phosphate buffer, pH 5.7, HADDOCK program (16).
0.05 % NaN and 93 % H O/7 % D O v/v and
3 2 2
placed into a symmetrical matched microtube Graphical representations
(Shigemi, Inc., U.S.A.). All molecular graphical representations were
NMR experiments were recorded at 298 K on a generated using the programs Ribbons (17) and D
o
w
Bruker DRX600 spectrometer equipped with 5- GRASP (18). n
mm z-gradient 1H/15N/13C cryoprobe optimized load
e
for 1H detection. Sequence-specific backbone Determination of antimicrobial activity d fro
and side-chain resonance assignments of The antimicrobial activity of hydramacin-1 was m
h
hydramacin-1 were achieved using the following investigated as previously described (19). ttp
three-dimensional spectra: HNCA, HNCO, Briefly, test strains were grown in brain heart ://w
w
H(CCO)NH, C(CO)NH, HCACO, infusion broth at 37 °C for 2 to 3 h. The cells w
CBCA(CO)NH, CBCANH, 15N-edited TOCSY were washed three times in 10 mM sodium .jbc
and 13C-edited TOCSY. Distance restraints were phosphate buffer, pH 7.4, supplemented with .org
obtained from 3D 15N-edited and 13C-edited 1 % tryptic soy broth (TSB), and the cell number by/
NOESY-HSQC experiments recorded with adjusted to 104 to 105 cells/mL. From the gu
e
s
mixing times of 100 and 120 ms, respectively. prepared cell suspension, 100 μL were mixed t o
n
The long-range quantitative-JNC' 2D H(N)CO with 10 μL of hydramacin-1 dissolved in 10 mM Jan
u
experiment was used to identify H-bond scalar sodium phosphate buffer, pH 7.4, and incubated a
ry
couplings (10). Proton chemical shifts were at 37 °C. The final peptide concentrations tested 1
2
referenced to TSP, while 15N and 13C chemical were between 0.0125 and 100 μg/mL. After 2 h , 2
0
1
shifts were indirectly referenced (11). of incubation, the colony forming units (CFU) 9
All spectra were processed using the program were determined. For the negative control,
NMRPipe (12). Analysis of the processed bacteria suspensions were supplemented with
spectra was performed with the program 10 μl of phosphate buffer, 1 % TSB.
NMRview (13). Antimicrobial activity was tested for all strains
by determination of LD (90 % lethal dose) or
90
Structure calculations minimal bactericidal concentrations (MBCs) (≥
Structure calculations were performed using the 99.9 % killing).
program CYANA (14). The calculations were
based on 1046 inter-proton distance constraints Preparation of liposomes used for determination
extracted from the 15N- and 13C-edited NOESY of membrane integration by fluorescence
spectra. Distances were calibrated so that the spectroscopy
median NOE intensity corresponded to a Liposomes were prepared essentially as
distance of 2.6 Å with a tolerance of 0.125 times described by Pick (20). Liposomes composed of
the bound squared. Methyl intensities have been defined phospholipids were resuspended in
divided by 2.0. All lower bounds were set to 10 mM HEPES, 1 mM EDTA, 150 mM NaCl,
1.8 Å. Furthermore, 24 distance restraints pH 7.4 by passing them over a NAP-5 column
between eight cysteine residues generating four (Amersham Biosciences). This served as a stock
disulfide bridges were introduced into the suspension for use in measurements. Except for
3
Hydramacin-1: Structure and activity
one the liposomes used have been generated Germany) was used. Measurements were
from lipids purchased from Avanti Polar Lipids performed using a continuous buffer flow of
Inc. (Alabaster, AL, U.S.A.). The lipids are: L-α- 20 µL/min at 22 °C. By injection of 100 µL
phosphatidylethanolamine (PE), L-α- poly-L-lysine (PLL, 60 µg/mL, Fluka) and the
phosphatidyl-DL-glycerol (PG), L-α- subsequent injection of liposomes or LPS
phosphatidylinositol (PI), L-α- (100 µM) the bilayer was prepared and finally
phosphatidylserine (PS), L-α- 100 µl hydramacin-1 (100 µg/mL in 0.01°%
phosphatidylcholine (PC), sphingomyeline or trifluoric acid (TFA)) was added. All
asolectin (Fluka, Buchs, Switzerland). experiments were performed at 22 °C.

Preparation of liposomes used for SAW- and FRET-based fusion assay
FRET-assays To determine the peptide-induced fusion of lipid
Rough mutant lipopolysacchride (LPS) from liposomes/aggregates of various compositions
E. coli strain WBB01 was extracted by the we used a FRET-based fusion assay. One half of
phenol/chloroform/petroleum ether method (21), the respective liposomes were doped with the
purified, lyophylized and transformed into the donor dye NBP-PE and the other half with the
triethylamine salt form. acceptor dye Rh-PE. Mixtures in a ratio of 1:1 of
Phospholipid liposomes and LPS aggregates the differently labeled liposomes were prepared
were prepared as 1 mM aqueous dispersions of with a final concentration of 10 µM. In each
the phospholipids or LPS in three different experiment, 5 µl of hydramacin-1 (1 mg/mL in D
o
buffers as follows. The lipids were dissolved in 0.01°% TFA) was added after 50 s. A Fluorolog wn
lo
chloroform and the LPS in chloroform:methanol F1 T11 (Spex Instruments, Edison, NJ, U.S.A.) ad
e
(F1R0E:1T) etox pae rcimonecnetsn t1ra %tio n(m oofl /1m molg) /omfL t.h eF odro nthoer wfluaos resucseendc e tion tenssiimtyu lotaf ntehoeu dsolyn orm aenads uarcec eptthoer d from
dye NBD-phosphatidylethanolamine (NBD-PE) dyes. All experiments were performed at 37 °C. http
or of the acceptor dye rhodamine-PE (Rh-PE) ://w
w
(both were obtained from Molecular Probes, Assay for pore-forming activity w
Eugene, OR, U.S.A.) was added. The solvent Pore-forming activity of hydramacin-1 was .jbc
.o
was evaporated under a stream of nitrogen; the determined by measuring fluorimetrically the rg
lipids were resuspended in buffer, mixed dissipation of a valinomycin-induced membrane by/
g
thoroughly and sonicated for 1 min (1 mL potential in liposomes, prepared from soy bean u
e
s
solution). Subsequently, the preparation was asolectin, as previously described (22). t o
n
cooled at 4 °C for 30 min, heated at 56 °C for Fluorescence was measured by a fluorescence Ja
n
30 min, and cooled to 4 °C. Preparations were spectrophotometer (model LS 50B; Perkin ua
ry
stored at 4 °C overnight prior to measurements. Elmer) using excitation and emission 1
2
wavelengths of 620 nm and 670 nm, , 2
0
1
Fluorescence spectroscopy respectively. Pore-forming activity was 9
Fluorescence experiments were carried out using measured as the initial change in fluorescence
a F-2500 fluorescence spectrophotometer intensity over time after addition of the sample.
(Hitachi Ltd., Tokyo, Japan) in the 300 to The pore-forming peptide alamethicin (Sigma–
400 nm spectral region at 20 °C. Tryptophan Aldrich, Germany) served as a positive control
side chains were excited at λ = 280 nm. The to calibrate the assay.
interaction between hydramacin-1 and the
different liposome types was measured using Assay for permeabilization of bacterial
620 µL of a hydramacin-1 solution (20 µg/mL) membranes
in 50 mM sodium phosphate with either an Bacteria with compromised membranes were
acidic pH of 5.7 or a pH of 7.4 and 15 µL of a detected by monitoring the fluorescence of the
liposome stock suspension. Spectra were DNA-binding dye SYTOX Green (Invitrogen,
recorded immediately after adding the Molecular Probes) as previously described (23).
liposomes. All spectra were measured in B. megaterium ATCC 14581 in mid-logarithmic
triplicate. phase were washed twice and resuspended in
10 mM HEPES, pH 7.4 containing 25 mM NaCl.
SAW-biosensor assay A flat bottom 96-well microtiter plate (Sarstedt,
To determine the peptide-membrane interaction Germany) was coated with 0.1 % bovine serum
the SAW-biosensor system (Biosensor GmbH, albumin (A2153, Sigma-Aldrich, Munich) for
4
Hydramacin-1: Structure and activity
15 min prior to its use in the assay. Hydramacin- incubated with 1,2-propylendioxide for at least
1 was two-fold serially diluted in 10 mM 5 min before this compound was replaced
HEPES, 25 mM NaCl, pH 7.4. Bacteria (1 x 105 stepwise by an epoxy resin (agar 100 resin kit,
CFU/50 µL) were incubated with the diluted hard, Plano GmbH, Germany). The first step
peptides (25 µL) and 2 µM of the fluorescent ratio of propylendioxide to epoxy resin was 2:1
dye SYTOX Green (25 µL in 10 mM HEPES, (v/v). In the subsequent steps the ratio was 1:1
25 mM NaCl, pH 7.4) at 20 °C for 1 h. and 1:2 before the bacteria samples were finally
Permeabilization of the bacterial cytoplasmic suspended in epoxy resin only. Suspended
membrane allows the dye to cross the membrane samples were incubated overnight. Samples
and to intercalate with the DNA. Excitation of were embedded in fresh epoxy resin and were
the DNA-bound dye at 495 nm resulted in an incubated at 60 °C overnight. From the hardened
increase of emitted fluorescence at 538 nm. epoxy resin blocks ultrathin sections with a
Measurements were made in a microtiter plate thickness of 60 nm were prepared using an
reader (Fluoroskan II; Labsystems). Membrane- ultramicrotome (Ultracut S, Leica Microsystems
permeabilizing activity of the peptide was GmbH, Germany) equipped with a diamond
expressed as a percentage of permeabilized knife (histo, DiS-Galetzka, Germany). These
bacteria. For monitoring the activity at pH 5.2, a sections were transferred onto copper grids
20 mM MES and 25 mM NaCl buffer was used. (G2500C, Plano GmbH, Germany) covered with
As a control, the antimicrobial peptide cecropin a plastic film (Pioloform-F, Plano GmbH,
P1 (Sigma-Aldrich, Germany) was used. For Germany). Finally, samples were stained with D
o
w
maximum permeabilization of the bacteria uranyl acetate and lead citrate according to n
lo
(100 % value), cells were incubated with 70 % Reynolds (24). Electron micrographs where ad
e
ethanol for 10 min. The values were expressed taken with a TEM 208 S (Philips, Hamburg). d fro
as the mean of at least two independent m
h
experiments, each performed in duplicates. RESULTS ttp
://w
w
Electron microscopy Expression and purification w
Bacteria (E. coli DH5λ and Staphylococcus Hydramacin-1 was recombinantly expressed as .jbc
.o
aureus ATCC 12600) grown to exponential an insoluble fusion protein. The inclusion bodies rg
phase were received using the method of Sahly were dissolved under denaturating conditions, by/
g
et al. (19). Bacteria have been washed twice in the fusion protein purified by immobilized-metal u
e
s
10 mM sodium-phosphate buffer (pH 5.7) and affinity chromatography and renaturated. After t o
n
were concentrated to an optical density of 4.6. proteolytic cleavage to remove the thioredoxine- Ja
n
u
All further procedures has been carried out at His -tag and carrier protein, mature hydramacin- a
6 ry
20 °C. The suspension has been incubated with 1 was further purified by RP-HPLC and its 1
2
hydramacin-1 at final concentrations of 5 µM or average mass was confirmed to be 7009 Da by , 2
0
1
100 µM peptide. Incubations were performed for MALDI-TOF mass spectrometry. 9
30 min, 1 h or 2 h. As a negative control bacteria
were incubated with buffer only. Tertiary structure of hydramacin-1
Subsequently, 10 % glutaraldehyde (Fluka) in Structure calculations were performed using
10 mM sodium-phosphate buffer (pH 7.8) has 1090 distance restraints (see Table 1). From the
been added in a 1:1 ratio to the bacteria 500 calculated structures the 25 lowest target
suspension and incubated for a further 2 h. function structures were selected and subjected
Subsequently, samples were washed three times to a final water refinement using the
in 10 mM sodium-phosphate buffer (pH 7.8) and HADDOCK software package. An overlay of
incubated with 3 % (w/v) osmium tetroxide in a this ensemble is depicted in Figure 1A. No
10 mM sodium-phosphate buffer (pH 7.8) for distance restraint was violated by more than
1 h. This was followed by threefold washing (see 0.5 Å in any of the 25 structures. The backbone
above). Afterwards samples were treated with exhibits a root mean square deviation (rmsd) of
non-denatured EtOH at increasing 1.13 ± 0.15 Å, whereas the secondary structure
concentrations with an initial concentration of elements showed a smaller rmsd of 0.58 ±
60 % EtOH. Concentrations were subsequently 0.12 Å (Table 1). This reflects the well defined
increased in 10 % steps until a pure EtOH secondary structure elements held together by an
solution was used. Each incubation step took arrangement of four disulfide bonds and
15 min. After the EtOH gradient samples were highlights that hydramacin-1 contains unordered
5
Hydramacin-1: Structure and activity
regions. The two large loops represent the positions of the cysteine residues constituting the
majority of the unordered regions. The ensemble knottin fold (Fig. 2D).
of 25 structures was used to calculate an average In four out of the five scorpion-toxin like
structure (Fig. 1B). The molecule possesses two superfamily members six cysteines are spatially
short α-helices (residues 10 to 14 and 27 to 33) identical and involved in identical cysteine
at the N-terminus which are separated by a long pairings. In the one exception, heliomicin, a
flexible loop. The C-terminal region contains single cysteine is located at a different position
two β-strands (residues 38 to 42 and 56 to 60) in in space. Nevertheless, we placed the amino-acid
an antiparallel arrangement separated by a long residues of heliomicin flanking this cysteine
flexible loop. according to the conservation of that residue in
Inspection of amino-acid side chains the primary structures of both peptides. The
contributing to the surface of hydramacin-1 sequential alignment of hydramacin-1 and its
revealed an unusual distribution of arginines and homologues theromacin and neuromacin (Fig.
lysines. These residues form a positively charged 2D) clearly showed that these proteins constitute
belt thereby dividing the molecular surface into a new family within the scorpion-toxin like
two large hydrophobic hemispheres (Fig. 1C). superfamily, which we have named the macin
family.

Structural alignment Antimicrobial activity of hydramacin-1
Using the solved solution structure of To test the antimicrobial activity of hydramacin- D
o
w
hydramacin-1, a structural similarity search was 1 against microbes we used a wide range of n
lo
performed using the Dali server (25). The search human Gram-negative and Gram-positive ad
e
led to the identification of only two structures: pathogens in a broth-microdilution assay. d fro
the Petunia hybrida defensin 1 (PhD1, pdb Hydramacin-1 revealed highest activities against m
h
accession code: 1n4n) (26) and the Nicotiana the Gram-negative species Citrobacter freundii ttp
alata defensin 1 (NaD1, pdb accession code: (C7), Enterobacter cloaceae (Va12270/03), ://w
w
1mr4) (27). Both proteins belong to the family Escherichia coli, Klebsiella pneumoniae and K. w
of plant defensins, which share the knottin fold oxytoca, Salmonella typhimurium and Yersinia .jbc
.o
comprising a disulfide-bridge stabilized αβ enterolitica (Table 2). In each case, very low rg
motif. The striking characteristics of the knottin peptide concentrations of less than 1 µM were by/
g
fold are a knot formed by three disulfide bonds sufficient to kill all (99.9 %) of these bacteria. ue
s
and part of the protein backbone in which the The highest killing activity against Gram- t o
n
disulfide bridge formed by the 3rd and 6th positive pathogens was observed for Ja
n
u
cysteine threads through a macrocycle formed Staphylococcus hemolyticus, for which the a
ry
by the 1st, 2nd, 4th and 5th cysteine and part of the peptide concentration needed for total killing 1
2
protein backbone (28,29). The family of plant was 1.8 µM. Hydramacin-1 was effective at , 2
0
1
defensins belongs to the scorpion-toxin like killing most of the tested multi-resistant strains, 9
superfamily in the class of small proteins as albeit at higher concentrations.
specified in the Structural Classification of
Proteins (SCOP) database (30). Membrane permeabilization
Comparison of the hydramacin-1, PhD1 and Membrane permeabilization assays were
NaD1 structures is depicted in Fig. 2A-C. Six of performed to determine the mechanism by which
possible eight cysteine residues are in identical hydramacin-1 kills microbes. The results showed
spatial positions. that hydramacin-1 is able to permeabilize
The scorpion-toxin like superfamily consists of membranes of viable bacteria at low and neutral
five families, namely the long- and short-chain pH values (Fig 3). To further characterize the
scorpion toxins, the MDG-1 defensin, the insect peptide’s mode of action, its potential pore-
defensins and the plant defensins. From each forming activity was investigated by using a
family the first listed member has been chosen minimalistic membrane system consisting of
for a structural superposition with hydramacin-1 multilamellar liposomes prepared from asolectin,
in order to determine which family hydramacin- a crude phospholipids mixture of soy bean, and
1 most closely resembles. To aid this by monitoring the liposome depolarization
superposition we performed a sequential induced by hydramacin-1. Under the conditions
alignment based on the conserved spatial used, pore-forming activity was not detectable
(insert Fig 3).
6
Hydramacin-1: Structure and activity
exists between LPS and hydramacin-1 compared
Fluorescence spectroscopy to the interaction between PE + PG and
The five tryptophans of hydramacin-1 are all hydramacin-1. Using the same buffer system, a
solvent exposed (Fig. 1D) indicating that these FRET-based fusion assay was performed. The
residues are likely to play a key role in the results show that hydramacin-1 does not induce
interaction with and insertion into the target the fusion of LPS aggregates or PE + PG
membrane. To examine whether hydramacin-1 is liposomes (Fig. 4A, lower trace).
able to insert into the hydrophobic environment To further investigate the influence of the lipid
of liposomes via the tryptophan residues, matrix and pH of the buffer on the fusion of
fluorescence spectroscopy was used. We tested liposomes and LPS aggregates, we used buffers
negatively and neutrally charged liposomes containing 50 mM sodium phosphate, pH 5.7
prepared from various types of phospholipids at and 7.4. At both pH values no interaction
pH values of 5.7 and 7.4. Table 3 summarizes between hydramacin-1 and PC liposomes was
the results obtained by fluorescence observed (Fig. 4B, upper trace). In the case of
spectroscopy. pure PG liposomes, the FRET signal increased at
A shift of wavelength of the fluorescence both pH values. This is indicative for a peptide-
emission maximum was observed with induced fusion (Fig. 4B, middle trace). The
negatively charged liposomes only. The interaction was stronger and in particular faster
strongest shift to a shorter wavelength (9.5 nm) at the lower pH. Hydramacin-1 induced no
was observed with PG liposomes at pH 5.7. This fusion of LPS aggregates but the peptide D
o
w
is indicative of tryptophan residues interacting intercalates into the LPS membranes at both pH n
lo
with the membrane bilayer. Liposomes values; however, the interaction at pH 7.4 was ad
e
consisting of asolectin exhibited a slightly significant weaker (Fig. 4B, lower trace). The d fro
smaller shift (7.0 nm). PS and PI liposomes FRET assay does not allow a distinction between m
h
resulted in a shift of 5.0 nm. Zwitterionic an intercalation of the peptide without any ttp
liposomes prepared from PC, PE and further effects and an intercalation with ://w
w
sphingomyelin did not induce a significant shift subsequent aggregation of the LPS aggregates. w
of the fluorescence emission maximum Therefore, we used the SAW-biosensor .jbc
.o
indicating that the tryptophan residues were not experiments at pH 5.7 and 7.4 in which we first rg
in contact with the lipid bilayer. A similar injected two times LPS WBB01 aggregates, in a by/
g
situation was observed at pH 7.4 but with second step hydramacin-1 and in a third step u
e
s
smaller shifts when PS and PI were used. LPS WBB01 aggregates once again. Figure 4C t o
n
This indicates that hydramacin-1 preferentially shows that at low pH hydramacin-1 bound Ja
n
u
interacts with negatively charged membrane stronger to the LPS membranes (Fig. 4C, upper a
ry
constituents, such as PG. trace) and that the subsequent injection of LPS 1
2
During experiments we observed the formation aggregates led to a change of the amplitude (Fig. , 2
0
1
of precipitates consisting of hydramacin-1-lipid 4C, lower trace). This result is indicative for a 9
complexes. In order to assess whether the hydramacin-1-induced interaction between two
precipitates are formed due to fusion of LPS matrices at pH 5.7 but not at pH 7.4.
liposomes SAW- and FRET-assays were carried
out. Electron microscopy of hydramacin-1 treated
bacteria.
SAW-biosensor experiments and FRET assays Incubation of E. coli DH5λ with hydramacin-1
The phase/time as well as the amplitude/phase at 5 µM led to the formation of electron-dense
diagram (Fig. 4A, upper and middle trace) is in contacts between the cells that were not
agreement with the results from the tryptophan- observed with control bacteria treated with
fluorescence experiments: there is no interaction buffer only (Fig. 5A and B). The peptide
between hydramacin-1 and PC. In contrast PE + concentration tested was 103 to 104 times less
PG and LPS matrices led to similar adsorption of than the MBCs determined. The incubation of
the peptide as observed by the increase of the the peptide at 20-times higher concentration
phase. The amplitude, which depends to a first (which is closer to the MBC values) showed that
approximation on the viscoelastic properties of the bacteria retained the cell-cell contacts but in
the material adsorbed on the sensor surface, was addition the cells had changed their morphology
only influenced in the case of LPS membranes. to a thorn apple-like shape (Fig. 5C).
These results indicate that a stronger interaction Staphylococcus aureus bacteria that have been
7
Hydramacin-1: Structure and activity
insensitive to hydramacin-1 under the conditions cloacae and multiresistant Klebsiella oxytoca.
tested to determine the MBC were used as the This makes hydramacin-1 a promising template
negative control. These bacteria did not show for a new class of antibiotics.
cell-cell contacts at a concentration of 5 µM Hydramacin-1 promotes aggregation of bacteria
hydramacin-1 (Fig. 5D). No differences were at significantly lower concentrations relative to
observed between the three incubation periods the amount of bacteria present as compared to
tested. the MBCs found. Therefore this cell aggregation
effect occurs prior to the extensive killing effects
DISCUSSION of the peptide and may represent the initial step
of the peptide’s killing mechanism. Bacteria
The solution structure of the antibacterial protein exposed to hydramacin-1 form electron-dense
hydramacin-1 from the basal metazoan Hydra contacts and change their cell morphology to a
magnipapillata reveals a disulfide-bridge thorn apple-like shape. Due to the double-
stabilized αβ motive which is the common amphipathic character of hydramacin-1
scaffold of the knottin-protein fold that was first represented by the two hydrophobic hemispheres
described by Rees and Lipscomb (28). SCOP sandwiched by a belt of positive charges, we
(30) places hydramacin-1 as a member of the suggest a model that explains the observed
scorpion-toxin like superfamily which is further aggregation of bacteria in the presence of
subdivided into five families. Comparison of hydramacin-1 (Fig. 6). The two hydrophobic
several members of all five families of the patches of the peptide immerse into the outer D
o
w
scorpion-toxin like superfamily suggests that leaflets of the membranes of two individual n
lo
hydramacin-1 represents the first member of a bacterial cells. Thereby, negative charges of the ad
e
new family within the scorpion-toxin like phospholipids of the membrane surfaces, that d fro
superfamily. Comparison of the hydramacin-1 would probably decrease the possible m
h
structure with all 103 structures of the members spontaneous aggregation of individual bacterial ttp
of the scorpion-toxin like superfamily that are cells, are compensated by the band of positive ://w
w
listed in the SCOP data base (30) revealed that, charges surrounding the molecule at the w
except for one, none of these structures contains immersion point. Therefore the peptide-lipid .jbc
.o
an additional α-helix at the N-terminus. The one complex is stabilized by hydrophobic as well as rg
exception is the scorpion toxin Bj-xtrIT (PDB electrostatic forces. However, peptide-lipid by/
g
ID: 1bcg) which belongs to the long-chain interactions might be initiated mainly by u
e
s
scorpion toxins. This additional N-terminal helix attraction of opposite charges because the t o
n
can also be described as one long α-helix (Fig. immersion of hydramacin-1 into the membrane Ja
n
u
1B and Fig. 2A) that is interrupted by a loop as well as the fusion or precipitation effect of a
ry
containing three glycine residues. The glycine liposomes and LPS aggregates is highly affected 1
2
residues confer flexibility to this region. As this by pH. , 2
0
1
loop consists of primarily hydrophobic residues, The phenomenon of aggregation was observed 9
this region may represent an important with liposomes as well. As these model
interaction site with target membrane moieties. membranes consist of only lipids, the
Compared to hydramacin-1 none of the listed aggregation must be a result exclusively of
proteins have long flexible loops separating the peptide-lipid and peptide-charge interactions.
typical secondary structure elements of the Consequently, one can exclude receptor
knottin fold. In addition, hydramacin-1 is mediation.
derived from Hydra, which is an organism of Enforced proximity of cells might lead to fusion
ancient origin compared to scorpions, spiders, as observed with model membranes from
insects, plants or mussels. Moreover, liposomes. However, fused cells have not been
hydramacin-1 exhibits a high degree of sequence observed and the absence of such an observation
identity to two other antimicrobial peptides may be a consequence of the higher complexity
isolated from leech: theromacin and neuromacin of the bacterial membrane compared with
(7,8). This observation points to two additional liposomes.
members of this new family, which we have Explanations for the observed change in cell
termed the macins. morphology are currently suggestive. The results
Here, we demonstrate that hydramacin-1 is clearly showed that hydramacin-1 permeabilizes
active against a broad range of microbes the bacterial membrane. Consequently,
including some clinical isolates of Enterobacter hydramacin-1 may invade the cells and affect
8
Hydramacin-1: Structure and activity
intracellular secondary targets like the details of these secondary events after the initial
cytoskeleton (31), leading to the observed aggregation/permeabilization of cells.
change in morphology. Interference with the
polymerization or depolymerization of filaments ACKNOWLEDGMENTS
might lead to invaginations or formation of The authors would like to thank Sonja Hollmer,
protuberances of the bacterial membrane as Christine Hamann, Antje Thomas and Silvia
observed. However, further investigations are Voss for excellent technical assistance and
required to better understand the molecular Matthias Michalek for preparing the liposomes.

D
o
w
n
lo
a
d
e
d
fro
m
h
ttp
://w
w
w
.jb
c
.o
rg
b/
y
g
u
e
s
t o
n
J
a
n
u
a
ry
1
2
, 2
0
1
9
9
Hydramacin-1: Structure and activity
REFERENCES

1. Alanis, A. J. (2005) Arch Med Res 36(6), 697-705
2. Marshall, S. H., and Arenas, G. (2003) Electronic Journal of Biotechnology [online] 6(3),
271-284
3. Zasloff, M. (2002) Nature 415(6870), 389-395
4. Jenssen, H., Hamill, P., and Hancock, R. E. (2006) Clin Microbiol Rev 19(3), 491-511
5. Gallo, R. L., Murakami, M., Ohtake, T., and Zaiou, M. (2002) J Allergy Clin Immunol 110(6),
823-831
6. Bosch, T. C. G., Augustin, R., Anton-Erxleben, F., Fraune, S., Hemmrich, G., Zill, H.,
Rosenstiel, P., Jacobs, G., Schreiber, S., Leippe, M., Stanisak, M., Grötzinger, J., Jung, S.,
Podschun, R., Bartels, J., Harder, J., and Schröder, J. M. (2008) Dev Comp Immunol (in press)
7. Tasiemski, A., Vandenbulcke, F., Mitta, G., Lemoine, J., Lefebvre, C., Sautiere, P. E., and
Salzet, M. (2004) J Biol Chem 279(30), 30973-30982
8. Schikorski, D., Cuvillier-Hot, V., Leippe, M., Boidin-Wichlacz, C., Slomianny, C., Macagno,
E., Salzet, M., and Tasiemski, A. (2008) J Immunol 181(2), 1083-1095
9. Dingley, A. J., Lorenzen, I., and Grotzinger, J. (2008) Methods Mol Biol 451, 441-462
10. Cordier, F., and Grzesiek, S. (1999) J. Am. Chem. Soc. 121(7), 1601-1602
11. Markley, J. L., Bax, A., Arata, Y., Hilbers, C. W., Kaptein, R., Sykes, B. D., Wright, P. E.,
and Wuthrich, K. (1998) J Biomol NMR 12(1), 1-23 D
o
w
12. Delaglio, F., Grzesiek, S., Vuister, G. W., Zhu, G., Pfeifer, J., and Bax, A. (1995) J Biomol n
lo
NMR 6(3), 277-293 ad
e
d
13. Johnson, B. A. (2004) Methods Mol Biol 278, 313-352 fro
14. Guntert, P. (2004) Methods Mol Biol 278, 353-378 m
h
15. van Gunsteren, W. F., Billeter, S. R., Eising, A. A., Hünenberger, P. H., Krüger, P., Mark, A. ttp
E., Scott, W. R. P., and Tironi, I. G. (1996) Biomolecular Simulation: The GROMOS96 ://w
w
Manual and User Guide, vdf Hochschulverlag AG, Zürich w
16. Dominguez, C., Boelens, R., and Bonvin , A. M. J. J. (2003) J. Am. Chem. Soc 125, 1731- .jbc
.o
1737 rg
b/
17. Carson, M. (1991) J. Appl. Crystallogr. 24, 946-950 y
g
18. Nicholls, A., Sharp, K. A., and Honig, B. (1991) Proteins 11(4), 281-296 ue
s
19. Sahly, H., Schubert, S., Harder, J., Rautenberg, P., Ullmann, U., Schroder, J., and Podschun, t o
n
R. (2003) Antimicrob Agents Chemother 47(5), 1739-1741 Ja
n
u
20. Pick, U. (1981) Arch Biochem Biophys 212(1), 186-194 a
ry
21. Galanos, C., Luderitz, O., and Westphal, O. (1969) European journal of biochemistry / FEBS 1
2
9(2), 245-249 , 20
1
22. Leippe, M., Ebel, S., Schoenberger, O. L., Horstmann, R. D., and Muller-Eberhard, H. J. 9
(1991) Proc Natl Acad Sci U S A 88(17), 7659-7663
23. Herbst, R., Ott, C., Jacobs, T., Marti, T., Marciano-Cabral, F., and Leippe, M. (2002) J Biol
Chem 277(25), 22353-22360
24. Reynolds, E. S. (1963) The Journal of cell biology 17, 208-212
25. Holm, L., and Sander, C. (1996) Science 273(5275), 595-603
26. Janssen, B. J., Schirra, H. J., Lay, F. T., Anderson, M. A., and Craik, D. J. (2003)
Biochemistry 42(27), 8214-8222
27. Lay, F. T., Schirra, H. J., Scanlon, M. J., Anderson, M. A., and Craik, D. J. (2003) J Mol Biol
325(1), 175-188
28. Rees, D. C., and Lipscomb, W. N. (1982) J Mol Biol 160(3), 475-498
29. Le Nguyen, D., Heitz, A., Chiche, L., Castro, B., Boigegrain, R. A., Favel, A., and Coletti-
Previero, M. A. (1990) Biochimie 72(6-7), 431-435
30. Murzin, A. G., Brenner, S. E., Hubbard, T., and Chothia, C. (1995) J Mol Biol 247(4), 536-
540
31. Errington, J. (2003) Nature cell biology 5(3), 175-178
32. Yang, Y. S., Mitta, G., Chavanieu, A., Calas, B., Sanchez, J. F., Roch, P., and Aumelas, A.
(2000) Biochemistry 39(47), 14436-14447
10