Search tips
Search criteria 


Logo of nihpaAbout Author manuscriptsSubmit a manuscriptHHS Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
Biochim Biophys Acta. Author manuscript; available in PMC 2009 May 8.
Published in final edited form as:
PMCID: PMC2679260

Biophysical analysis of the interaction of granulysin-derived peptides with enterobacterial endotoxins


To combat infections by Gram-negative bacteria, it is not only necessary to kill the bacteria but also to neutralize pathogenicity factors such as endotoxin (lipopolysaccharide, LPS). The development of antimicrobial peptides based on mammalian endotoxin-binding proteins is a promising tool in the fight against bacterial infections, and septic shock syndrome. Here, synthetic peptides derived from granulysin (Gra-pep) were investigated in microbiological and biophysical assays to understand their interaction with LPS. We analyzed the influence of the binding of Gra-pep on (1) the acyl chain melting of the hydrophobic moiety of LPS, lipid A, by Fourier-transform spectroscopy, (2) the aggregate structure of LPS by small-angle X-ray scattering and cryo-transmission electron microscopy, and 3) the enthalpy change by isothermal titration calorimetry. In addition, the influence of Gra-pep on the incorporation of LPS and LPS-LBP (lipopolysaccharide-binding protein) complexes into negatively charged liposomes was monitored. Our findings demonstrate a characteristic change in the aggregate structure of LPS into multilamellar stacks in the presence of Gra-pep, but little or no change of acyl chain fluidity. Neutralization of LPS by Gra-pep is not due to a scavenging effect in solution, but rather proceeds after incorporation into target membranes, suggesting a requisite membrane-bound step.

Keywords: Endotoxin, Lipopolysaccharide, Antimicrobial peptide, Sepsis, Granulysin

1. Introduction

Gram-negative bacteria such as Escherichia, Salmonella, Yersinia, and Vibrio are responsible for a multitude of infections. In particular, they are inducers of septic shock syndrome that is lethal in more than 50% of cases [1]. It is well accepted that lipopolysaccharide (LPS), the main amphiphilic compound located in the outer leaflet of the outer membrane, is the major pathogenicity factor of Gram-negative bacteria [2]. When LPS is released from bacteria due to cell division or cell death, it may interact with various target cells to induce tumor necrosis factor-á (TNF-α) and interleukins [3]. At low LPS concentrations, these interactions may be beneficial; however, high LPS concentrations activate a cascade of events that can result in multi-organ failure and septic shock for which there is no effective therapy.

The use of antimicrobial peptides based on endotoxin-binding structures of mammalian defense proteins [4-6] represents a new therapeutic approach to septic shock. Granulysin is a cytolytic protein found in the granules of human cytotoxic T lymphocytes and natural killer cells [7, 8]. The crystal structure of granulysin consists of a five-helix bundle suggesting a potential mechanism of action whereby the positive charges of granulysin orient the molecule towards the negatively charged surface of target cells lysing their membrane [9]. Synthetic peptides corresponding to the linear sequence of granulysin kill Gram-positive and Gram-negative bacteria [10]. Furthermore, NK-lysin, the porcine homolog of granulysin, inhibits LPS-induced cytokine production [11], as does a fragment of NK-lysin [12].

An ideal antibiotic should both kill bacteria and neutralize the endotoxins subsequently released. In this study, we performed systematic microbiological, physico-chemical and biophysical studies on the interaction of bacterial endotoxins with four granulysin derivatives (Gra-pep). LPS from Salmonella minnesota Re and Ra, strains R595 and R60, respectively, were used as endotoxins. We monitored the gel to liquid crystalline phase transition behavior of the acyl chains of LPS within its lipid A moiety, the aggregate structure of LPS, their morphology, the intercalation of the peptides into phosphatidylserine (PS) target membranes in the presence of LPS and lipopolysaccharide-binding protein (LBP), and the influence of the peptides on the LPS-induced cytokine production in human mononuclear cells. Our results suggest that neutralization of LPS by Gra-pep involves association of the complex with immune cell membranes.

2. Materials and methods

2.1. Bacterial strains and growth

The bacterial strains used in the study were wild-type or LPS mutant strains of Salmonella enterica (serovar Minnesota) with smooth-form LPS (S form; wild-type), strain R60 with rough LPS (LPS Ra), and strain R595 deep rough LPS (LPS Re). For antibacterial assay, bacteria were grown overnight in N-minimal medium (10.46 g of Bis-Tris, 6.05 g of Tris Base, 0.087 g of K2SO4, 0.136 g of KH2PO4, 0.372 g of KCl, 1.0 g of (NH4)2SO4, 0.1% casamino acids, 38 mM glycerol, 1.0 l distilled water) supplemented with 10 mM MgCl2 (pH 7.4) at 37 °C [13]. Cells were subsequently washed twice with N-minimal medium, inoculated in N-minimal medium+10 μM MgCl2 (pH 7.4) and grown to mid-exponential phase. Cells were washed twice with N-minimal medium +10 μM MgCl2 (pH 7.4) before assay.

2.2. Synthesis of granulysin peptides

Peptides were synthesized using F-moc chemistry on an Applied Bio-systems (Foster City, CA) automatic peptide synthesizer and purified to >95% homogeneity by reverse-phase HPLC. Peptide composition was confirmed by mass spectrometry and amino acid analysis. Peptide stock solutions (10 mM) were prepared in DMSO and diluted into assay medium at 0.078−10 μM.

The sequences of the Gra-pep are shown in Table 1. These peptides are based on helix 3 and helix 4 of granulysin. This region was previously determined to be important for antibacterial activity [10]. To increase resistance to proteases, all Gra-pep were synthesized with d-amino acids.

Table 1
Amino acid sequences of granulysin-derived peptides (Gra-pep)

2.3. Assay for antibacterial activity

Microtiter plate assay: each peptide was diluted to 20 μM in assay medium and 100 μl of this solution were added to the first well of a 96-well microtiter plate. For twofold serial dilution, 50 μl from each well were transferred to the next well, which contained 50 μl of assay medium. Subsequently, 50 μl of bacteria at 2×105 CFU/ml (1×104 cells) were added to each well. The plates were incubated for 90 min at 37 °C with constant shaking. After 90 min, 100 μl of 2× LB were added to each well, and the plates were further incubated with shaking overnight at 37 °C. Bacterial growth was monitored by measuring the absorbance at 600 nm in a microtiter plate reader (Molecular Devices, California, USA). The MIC (minimal inhibitory concentration) is defined as the lowest peptide concentration at which no bacterial growth was measurable after overnight incubation.

Colony forming unit (CFU) assay: bacteria and Gra-pep were added to microtiter wells as described above and, after a 90 min incubation at 37 °C, bacteria were serially diluted in PBS and plated on LB agar. The plates were incubated overnight at 37 °C, and bacterial colonies were enumerated the following day by automatic colony counter (aCOLyte colony counter, UK). The MBC (minimal bactericidal concentration) is defined as the lowest peptide concentration at which less than 1% of the input bacteria survived.

To test whether pre-incubation of Gra-pep with LPS affects antibacterial activity, 50 μl of different molar ratio of LPS (S-form or LPS Re) and peptide G12.21 were incubated in N-minimal medium+10 μM MgCl2 (pH 7.4) in microtiter wells at room temperature with shaking for 30 min. Then, 50 μl of bacteria (wild-type or strain R595) at 2×105 CFU/ml were added. After an additional 90-min incubation at 37 °C with constant shaking, aliquots were removed, diluted, and plated for CFU assay. The percent survival was determined as CFU (Gra-pep:LPS)/CFU (medium)×100.

2.4. Lipopolysaccharides and lipid:peptide samples

Lipopolysaccharides from the rough mutant Re and Ra S. minnesota (strains R595 and R60, respectively) were extracted by the phenol/chloroform/petrol ether method [14], and S-form LPS was extracted according to the phenol:water procedure [15]. Briefly, an overnight culture of bacteria grown at 37° was purified and lyophilized. The schematic structures of LPS presented in Fig. 1 shows that the two mutant LPS molecules differ essentially in the length of the carbohydrate chain and in the number of negative charges. Lipopolysaccharide-binding protein (LBP) was a kind gift of Russ L. Dedrick (XOMA Co, Berkeley, CA, USA) and was stored at −70 °C at 1 mg/mL stock solution in 10 mM HEPES (150 mM NaCl, 0.002% (v/v) Tween 80, 0.1% F68, pH 7.5). Bovine brain 3-sn-phosphatidylserine (PS) was from Sigma.

Fig. 1
Chemical structures of rough mutant lipopolysaccharides from Salmonella minnesota. LPS Re and LPS Ra were isolated from the strains R595 and R60, respectively. The molecular masses of S-form LPS, LPS Ra, LPS Re are 10, 4.3, 2.65 kDa, respectively. GlcNac, ...

All lipid samples were prepared as aqueous suspensions in 20 mM HEPES (pH 7.0). Briefly, the lipids were suspended directly in buffer and were temperature-cycled 3 times between 5 and 70 °C, interrupted by intensive vortex, and then stored 2 h at 4 °C before measurement. To guarantee physiological conditions, the water content of the samples was ~95%. For preparations of phosphatidylserine liposomes, phosphatidylserine was first solubilized in chloroform, and the solvent was evaporated under a stream of nitrogen. Then, the lipid was resuspended in the appropriate volume of 20 mM HEPES, and treated as described above (temperature-cycling). The resulting liposomes are large and multilamellar as shown by electron microscopy (kindly performed by H. Kühl, Division of Pathology, Forschungszentrum Borstel).

2.5. Cytokine assay

Human mononuclear cells (MNC): heparinized (20 IU/ml) blood obtained from healthy donors was mixed with an equal volume of Hank’s balanced solution, layered over Ficoll, and centrifuged for 40 min (21 °C, 500×g). The interphase layer of MNC was collected and washed twice in Hank’s medium and then resuspended in RPMI 1640 supplemented with 2 mM l-glutamine, 100 U/ml penicillin, 100 μg/ml streptomycin, and 4% heat-inactivated human serum type AB from healthy donors. MNC (200 μl /well; 5×106 cells/ml) were transferred into 96-well culture plates. Twenty microliters of a mixture containing LPS Ra (100 ng/ml or 10 ng/ml) and Gra-pep (10:1 wt.% excess) was added to each well. Supernatants were harvested after 4 h incubation at 37 °C under 5% CO2, and TNFα production was measured in a sandwich-ELISA as described elsewhere [16]. TNFα was determined in duplicate at two different dilutions and the values from two independent experiments were averaged.

2.6. Fourier-transform infrared spectroscopy (FTIR)

The infrared spectroscopic measurements were performed on an IFS-55 spectrometer (Bruker, Karlsruhe, Germany). The lipid samples were placed in a CaF2 cuvette with a 12.5 μM teflon spacer. Temperature-scans were performed automatically between 10 and 70 °C with a heating-rate of 0.6 °C /min. Every 3 °C, 50 interferograms were accumulated, apodized, Fourier transformed, and converted to absorbance spectra. For strong absorption bands, the band parameters (peak position, band width, and intensity) were evaluated from the original spectra, or after subtraction of the strong water bands if necessary. The main vibrational band used for the analysis is the symmetrical stretching vibration of themethylene groups ís(CH2) located around 2850 cm−1, a measure of order of the lipid A chains [17].

2.7. Isothermal titration calorimetry (ITC)

Microcalorimetric measurements of peptide binding to the endotoxins were performed on a MCS isothermal titration calorimeter (Microcal Inc., North-ampton, MA, USA) at 37 °C. Briefly, after thorough degassing of the suspensions by ultrasonification, 1.5 ml of endotoxin samples (0.05 mM) were dispensed into the microcalorimetric cell, and 100 μl of peptide solutions (2 mM) were filled into the syringe compartment. After temperature equilibration, the peptides (3 μl) were titrated every 5 min into the lipid-containing cell under constant stirring. The heat of interaction measured by the ITC instrument after each injection was plotted versus time. The titration curves were repeated three times.

2.8. X-ray diffraction

X-ray diffraction measurements were performed at the European Molecular Biology Laboratory (EMBL) outstation at the Hamburg synchrotron radiation facility HASYLAB using the SAXS camera X33 [18]. Diffraction patterns in the range of the scattering vector 0.1<s<1.0 nm −1 (s=2 sin θ/λ, 2θ scattering angle and λ the wavelength=0.15 nm) were recorded at 40 °C with exposure times of 1 min using an image plate detector with online readout (MAR345, MarResearch, Norderstedt/Germany). The s-axis was calibrated with Agbehenate, which has a periodicity of 58.4 nm. The diffraction patterns were evaluated as described previously [19] assigning the spacing ratios of the main scattering maxima to defined three-dimensional structures. The lamellar and cubic structures are the most relevant here. They are characterized by the following features:

  1. Lamellar: the reflections are grouped in equidistant ratios, i.e., 1, 1/2, 1/3, 1/4, etc. of the lamellar repeat distance dl;
  2. Cubic: the different space groups of these non-lamellar three-dimensional structures differ in the ratio of their spacings. The relation between reciprocal spacing shkl =1/dhkl and lattice constant a is
    shkl=[(h2 + k2 + l2)/a
    (hkl=Miller indices of the corresponding set of plane).

2.9. Cryo Transmission Electron Microscopy (Cryo-TEM)

Three microliters of the dispersion were placed on a copper grid with perforated carbon film (Quantifoil R 1.2/1.3, Jena, Germany), and excess liquid was blotted automatically for two s between two strips of filter paper. Subsequently, the samples were rapidly plunged into liquid ethane (cooled to ~ −175 °C) in a cryo-box (Zeiss, Oberkochen). Excess ethane was removed with a piece of filter paper. The sample was transferred with a liquid nitrogen-cooled holder (Gatan 626, München, Germany) into the cryo-TEM (Philips CM 120, Netherlands) and investigated at 120 kV. The micrographs were generated by a Tietz-Fast Scan CCD Camera (TVIPS, Gauting, Germany).

2.10. Fluorescence resonance energy transfer spectroscopy (FRET)

Intercalation of the peptides and of LPS Re into liposomes alone or mediated by lipopolysaccharide-binding protein (LBP) was determined by FRET spectroscopy applied as a probe dilution assay [20]. Liposomes were labelled with the donor dye NBD-phosphatidylethanolamine (NBD-PE) and acceptor dye Rhodamine-PE. Then, the lipids followed by LBP (or vice versa) were added to liposome at a final concentration of 1 μM. Intercalation was monitored as the increase of the ratio of the donor intensity Id at 531 nm to that of the acceptor intensity Ia at 593 nm (FRET signal) in a time-dependent manner.

3. Results

3.1. Effect of different LPS structures on the antimicrobial activity of Gra-pep

It is well known that LPS is a major target for many cationic antimicrobial peptides [12,21,22]. We asked whether LPS structure could affect the antimicrobial activity of a panel of synthetic peptides based on the sequence of helix 3 and helix 4 of granulysin (see Table 1). These peptides are part of a large group of synthetic peptides based on granulysin that exhibit markedly enhanced killing of microbes and minimal toxicity against mammalian cells (Chen, X. and Clayberger, C., unpublished data). Salmonella with distinct LPS sugar structures used as targets were (i) wild-type S. enterica with S-form LPS, which contains a full O-antigenic carbohydrate chain; (ii) mutant strain R60 with LPS Ra, which contains the complete core sugars; and (iii) mutant strain R595 with LPS Re, which lacks the O-antigenic chain and has, besides lipid A, only the 2-keto-3-deoxyoctonate monosaccharides (Fig. 1).

Antimicrobial activity was first tested in a microtiter plate assay. As shown in Table 2, all four of these peptides kill wild-type and strain R60 equivalently, and kill strain R595 at lower MICs. Strain R595 has the shortest form of LPS (Fig. 1). Comparison of the growth rates of these three strains in LB medium or N-minimal medium+10 μM MgCl2 (pH 7.4) revealed that strain R60 grows faster, while the growth rates of wild-type and strain R595 are quite similar (data not shown). In addition, strain R60 tends to aggregate even after a short time, while wild-type and strain R595 remain in suspension. This behavior of strain R60 might be due to cell–cell interaction through truncated LPS structures in the cell membrane. Therefore, the antimicrobial activity of Gra-pep on strain R60 could be affected by its faster growth rate and cell clumping. We next tested the effects of peptide G12.21 on these Salmonella strains using the more sensitive Colony Forming Unit (CFU) assay. Consistent with the results from the microtiter plate assay, strain R595 is much more sensitive to treatment with G12.21 than strain R60 and wild-type at all concentrations of peptide tested (Fig. 2), indicating that the loss of the O-antigenic chain and the majority of the core sugar renders the bacteria more sensitive to Gra-peps.

Fig. 2
Activity of G12.21 against S. enterica strains with various LPS carbohydrate structures: wild type (S-form LPS; triangles), strain R60 (LPS Ra; circles), and strain R595 (LPS Re; squares). Percent survival was determined in the CFU assay. Each data point ...
Table 2
Antibacterial activity of Gra-pep against Salmonella with different forms of LPS

3.2. Gra-pep lytic activity is neutralized by preincubation with LPS

Similar to many antimicrobial peptides, Gra-peps have a high net positive charge and can interact with negatively charged amphiphiles such as LPS. We asked whether preincubation of peptide G12.21 with LPS could affect its antibacterial activity. Preincubation of LPS S-form with G12.21 at 1:10 molar ratio totally inhibits the antimicrobial activity of G12.21 against wild-type Salmonella (Fig. 3A), while 10–100 folds more LPS Re is required to neutralize the antimicrobial activity of G12.21 against wild-type Salmonella (Fig. 3B). In contrast, a 10:1 molar ratio of LPS S-form:G12.21 is required to totally inhibit the antimicrobial activity of G12.21 against strain R595 (Fig. 3C), while a 100:1 molar ratio of LPS Re:G12.21 inhibited the activity of G12.21 against strain R595 (Fig. 3D). Interestingly, higher molar ratios of LPS Re:G12.21 are required for total inhibition of antimicrobial activity of G12.21 toward either wild-type or strain R595, indicating that the overall negative charges on LPS play an important role for binding to G12.21 and neutralizing its antimicrobial activity. As a control, the highest concentration of LPS S-form or LPS Re used in these experiments had no effect on cell survival (shown in grey bars). Similar results were obtained when LPS and G12.21 were added together with the bacteria (data not shown), indicating that the interaction between LPS and G12.21 is very rapid. In summary, LPS inhibits the antimicrobial activity of Gra-pep on Salmonella, probably through direct interaction between LPS and Gra-pep.

Fig. 3
LPS inhibits the antimicrobial activity of G12.21 against S. enterica wild-type and strain R595. Different molar ratios of LPS S-form (or LPS Re):G12.21 were pre-incubated for 0.5 h, and the antimicrobial activity was tested on wild-type or strain R595. ...

3.3. Cytokine production of human mononuclear cells

To study the effects of Gra-pep on LPS-induced cytokine production in human mononuclear cells (MNC), Gra-pep:LPS Ra mixtures were added to MNC, and TNFα secretion was monitored in the cell culture supernatant (Fig. 4). There is a clear inhibition of the cytokine production in the presence of Gra-pep. At [LPS Ra] =100 ng/ml, this inhibition is moderate (Fig. 4, left panel), while at 10 ng/ml, it is strong (Fig. 4, right panel). These results suggest that Gra-pep not only kill the bacteria but also are able to neutralize LPS-induced TNFα production.

Fig. 4
Gra-pep inhibit TNFα production induced by LPS Ra. Human mononuclear cells were stimulated with a mixture of LPS Ra (100 or 10 ng/ml) and Gra-pep (10:1 wt.% excess). TNFα level was measured by ELISA. The figure shows one typical experiment. ...

3.4. Gel to liquid crystalline phase transition of the acyl chains of LPS

The phase transition of the acyl chains in the lipid A part of LPS Re can be monitored by infrared spectroscopy. The peak position of the symmetric stretching vibration of the methylene groups around 2850 cm−1 is a measure of lipid order. As shown in Fig. 5, the phase transition temperature Tc and the states of order within the single phases are only marginally influenced by the addition of the Gra-pep. For G12.21, G12.34, and G12.35, the Tc is shifted slightly to higher values and the wavenumbers are reduced, indicating a very small rigidification. For G12.25, the effect is minimal.

Fig. 5
Gel to liquid crystalline phase behaviour of the hydrocarbon chains in the lipid amoiety of LPS Re in the absence and presence of the Gra-pep. The peak position of the symmetric stretching vibration of the methylene groups is plotted versus temperature. ...

3.5. LPS aggregate structures

Aggregates, but not monomers, mediate the biologic function of LPS [23]. Therefore, synchrotron radiation small-angle X-ray scattering (SAXS) was applied to elucidate the aggregate structure of LPS Re in the presence of the Gra-pep. The X-ray diffraction patterns of LPS alone correspond to a unilamellar or mixed unilamellar/inverted cubic structure ([24], data not shown). In the presence of Gra-pep, the patterns correspond to aggregates consisting of multilamellar stacks, deduced from the occurrence of reflections at equidistant ratios (Fig. 6). The values of the periodicities (6.17 nm to 6.30 nm) indicate ’normal’ LPS Re stacks, which have been shown to occur at theses values at high concentrations of divalent cations [25]. This means that there is no increase of the water layer between neighboring stacks, excluding a peripheral binding to the LPS outer surface.

Fig. 6
Small-angle X-ray diffraction patterns of LPS Re in the presence of the Gra-pep at [LPS Re]:[peptides] 6:1 or 4:1 molar. The scattering vector s=1/d=2 sinθ/λ (d=spacing ratio, θ=scattering angle, λ=wavelength=0.15 nm) is ...

3.6. Intercalation into phospholipid liposomal membranes

Since the intercalation of LPS into target cell membranes is mediated by LBP [20, 26], we asked whether Gra-pep are able to inhibit LPS intercalation. Fig. 7A shows that in the presence of G12.21, there is a slight increase of the FRET signal when LPS Re at t=100 s is added to the liposomes, but there is a strong signal when LBP is added at t=150 s, suggesting an LBP-mediated intercalation of LPS into liposomes. LBP alone is able to intercalate into the liposomes, albeit to a lower degree than with LPS (compare the ordinate scales) (Fig. 7D). When G12.21 is added at t=50 s to the liposomes, it is also able to incorporate into the liposomes (Fig. 7B and C). The subsequent addition of LPS leads to a further intercalation, suggesting that the LPS: peptide binding leads to a strong membrane insertion of LPS (Fig. 7B). Lastly, the addition of LBP causes a further increase in intensity. In the absence of LPS (Fig. 7C) the increase in intensity due to LBP is minimal, indicating that LPS is necessary for effective incorporation into liposomes. The results for each Gra-pep are shown in Fig. 8, G12.34 and G12.35 seem to better enhance the membrane insertion of LPS, while the effect of G12.21 and G12.25 is stronger in the presence of LBP.

Fig. 7
Intercalation of G12.21/LPS/LBP samples and HEPES buffer control into target cell liposomes measured by fluorescence resonance energy transfer spectroscopic (FRET). Presented is the FRET signal Id/Ia versus time. The samples were added subsequently at ...
Fig. 8
Intercalation of Gra-pep/LPS/LBP samples and HEPES buffer control into target cell liposomes measured by fluorescence resonance energy transfer spectroscopic (FRET). Presented is the FRET signal Id/Ia versus time. The samples were added subsequently at ...

3.7. Morphology of the LPS: Gra-pep complexes

Cryo-transmission electron microscopy was used to investigate the morphology of the LPS:peptide complexes. Pure LPS Ra exhibits fibrillary LPS with cylindrical forms (Fig. 9A). The diameter of the fibrilles is in the range of 12 nm, which would correspond to the lamellar repeat distance of hydrated LPS Ra, i.e. the thickness of one bilayer of LPS Ra [24]. The picture completely changes when peptide G12.21 is added (Fig. 9B). Regularly stacked ribbons can be seen, the repeat distance of the stacks lies approximately at 9 nm, which corresponds to tightly packed multilamellar structures of LPS Ra [24].

Fig. 9
Cryo-transmission electron micrographs of LPS from Salmonella minnesota Ra (A) and in the presence of granulysin peptide G12.21 ([LPS]: [G12.21]=5:1 molar) (B).

3.8. Binding of peptides to LPS

Isothermal titration calorimetry (ITC) was used to characterize the binding of LPS to the peptides. Equal volumes (3 μl) of G12.21 (2 mM) and LPS Ra (0.05 mM) were combined and the resulting thermal reactions were monitored at 37 °C. The ITC curve of enthalpy change versus [G12.21]: [LPS Ra] molar ratio is plotted in Fig. 10. These data show that the binding is exothermic, and that binding saturation occurs around [G12.21]:[LPS Ra] =1:1 mol/mol. In addition, a similar basic course of the reaction was observed with the other peptides (data not shown).

Fig. 10
Isothermal calorimetric titration of 0.05 mM LPS Ra with 2 mM G12.21 (30×3 μl every 5 min). Presented are the integrated peak areas of the single titrations versus G12.21:LPS molar ratio.

4. Discussion

The development of new antimicrobial peptides to treat infections and systemic septic shock is an important goal. A number of groups are evaluating synthetic peptides based on natural endotoxin-binding structures, including lactoferrin-based peptides [22], LALF (Limulus-anti-LPS factor)-peptides [27], or NK-lysin derived peptides [12]. Using a panel of synthetic peptides based on Helix 3/4 of granulysin (Table 1), we conducted a biophysical study of their interaction with S. minnesota strains R595 and R60 and with the corresponding chemically well-characterized endotoxins LPS Re and LPS Ra (structures Fig. 1). The former LPS was taken in various measurements since the interpretation of biophysical techniques such as SAXS is facilitated due to the short sugar part, whereas for biological experiments the choice of LPS Ra is more adequate because it has been shown that it represents the biologically active principle within the heterogeneous smooth form LPS [28]. However, the bioactivities of Re and Ra LPS are similar in most systems [29].

Our results indicate that G12.21 is much more effective at killing bacteria expressing the truncated LPS Re than the longer forms of LPS (Table 2 and Fig. 2) but that G12.21 binds more tightly to the long form of LPS than to the short form (Fig. 3). This is indicative of a process that inhibits peptide-mediated antimicrobial activity (Fig. 2) as well as cytokine induction by LPS (Fig. 4). In order to understand these effects, biophysical parameters were investigated. We show that Gra-pep interaction with LPS causes only minimal changes in the fluidity of the lipid A acyl chain, indicating that binding of Gra-pep does not lead to a significant change of acyl chain order (Fig. 5). This is notable because the whole LPS assembly is completely rearranged into a multilamellar form by the addition of Gra-pep (Fig. 6), but apparently without changing the order of the hydrophobic moiety. Previously it was shown that pure LPS or lipid A in the biologically active form adopt unilamellar and/or an inverted, mostly cubic structures. In contrast, some biologically inactive non-enterobacterial LPS or lipid A such as lipid A from the phototropic strain Rhodobacter capsulatus adopt multilamellar structures [30]. Moreover, for proteins or peptides that suppress LPS-induced cytokine production, e.g., Limulus-anti-LPS-factor (LALF, in recombinant form called endotoxin-binding protein ENP) and peptides of their binding domains, the binding to lipid A and LPS leads to a conversion into a multilamellar aggregate [27,31]. Thus, the conversion into a multilamellar structure is a pre-requisite for inactivation of biologically active LPS/lipid A.

The electron micrographs of LPS in the absence and presence of Gra-pep (Fig. 9A, B) are in complete accordance with this mechanism. With the exception of older papers on a poorly characterized smooth form LPS (see, for example, [32]) EM pictures of LPS have not been reported. Pure LPS Ra exists as elongated fibrillary cylinders (Fig. 9A) corresponding in diameter to the bilayer thickness, which would be in accordance with a former SAXS study which showed only broad diffraction patterns consistent with unilamellar bilayer structures [24]. In the presence of G12.21, the LPS aggregation changes completely into domains with parallel lying tightly packed stacks with a repeat distance of about 9 nm (Fig. 9B), corresponding to tightly arranged mutilamellar structures. At this place it should be emphasized that for the first time the detailed morphology of LPS assemblies in parallel SAXS and EM investigations has been elucidated. Both methods indicate the formation of multilamellar aggregates by binding of LPS with G12.21.

An important step in cell activation is the binding of LPS to LBP with the subsequent intercalation into target membranes [20,33] (see Fig 7A). Although it is possible that Gra-pep binding to endotoxin inhibits its membrane interaction and/or impedes the interaction of LPS with important binding proteins such as LBP or CD14 (LPS receptor expressed in monocytes), the data in Figs. 7 and and88 show that the situation is more complex. The Gra-pep not only intercalate by themselves into target liposomes, but they also considerably enhance the intercalation of LPS (Figs. 7B and and8).8). These data confirm and extend our earlier model in which we proposed a membrane step of cell activation by LPS [34]. Once LPS has incorporated into the target membrane of immune cells, it may diffuse within subdomains to associate with relevant signaling proteins. Due to the conical shape of its lipid A part, LPS represents a strong sterical disturbance within the membrane and thus may lead to a conformational change of proteins involved in cell signaling. In the present investigation, the peptides themselves and the LPS: peptide complexes incorporate into the membrane. The peptide causes a conversion into a lamellar structure, abrogating the ability of the lipid A part to present a sterical disturbance. The binding reaction of the peptide to LPS remains exothermic over the whole concentration range tested until saturation occurs at approximately [G12.21:[LPS] =1 mol/mol (Fig. 10). The exothermic nature of this reaction implies that the dominant reaction comes from the electrostatic attraction between the positive charges of the peptide and the negative charge of the endotoxin. An additional, entropically-governed reaction between LPS and Gra-pep, due to the removal of the highly ordered water layer in the backbone region of LPS [35], is apparently much less important than the former effect. Further investigations are in progress to elucidate details of these processes.

5. Conclusions

In this study, multiple biophysical analyses of the interaction of cationic granulysin-based peptides with bacteria and with isolated LPS provide a mechanistic basis for the ability of the Gra-pep to kill bacteria as well as to neutralize LPS. Both antimicrobial activity and neutralization of LPS are desirable characteristics of anti-septic agents. Decrease in the antimicrobial activity of Gra-pep by LPS could be compensated by combination with other antibiotics, whereas a lack in LPS neutralization cannot be compensated by current antibiotics. Peptide neutralization of LPS involves conversion of the LPS aggregate structure into a multilamellar form and exothermic binding to LPS. This contrasts with fluidization of the LPS acyl chains by agents such as polymyxin [36]. Perhaps most interestingly, the final inactivation process of LPS by Gra-pep involves interactions with membranes.


We are indebted to G. von Busse, C. Hamann, and K. Stephan for technical assistance in the IR spectroscopic, FRET, and TNFα measurements, respectively. This work has been carried out with financial support from the Deutsche Forschungsge-meinschaft (SFB 617, project A17) and the Commission of the European Communities, specific RTD programme “Quality of Life and Management of Living Resources”, QLK-CT-2002-01001, ‘Antimicrobial endotoxin neutralizing peptides to combat infectious diseases’. This work was supported by an NIH Postdoctoral Fellowship (to X.C.) and NIH grant U19 AI056548 (to C. C.). A.M.K is the Shelagh Galligan Professor of Pediatrics.


1. Cohen J, Guyatt G, Bernard GR, Calandra T, Cook D, Elbourne D, Marshall J, Nunn A, Opal S. New strategies for clinical trials in patients with sepsis and septic shock. Crit Care Med. 2001;29:880–886. [PubMed]
2. Rietschel ET, Brade H, Holst O, Brade L, Muller-Loennies S, Mamat U, Zahringer U, Beckmann F, Seydel U, Brandenburg K, Ulmer AJ, Mattern T, Heine H, Schletter J, Loppnow H, Schonbeck U, Flad HD, Hauschildt S, Schade UF, Di Padova F, Kusumoto S, Schumann RR. Bacterial endotoxin: Chemical constitution, biological recognition, host response, and immunological detoxification. Curr Top Microbiol Immunol. 1996;216:39–81. [PubMed]
3. Rietschel ET, Brade L, Schade UF, Seydel U, Zähringer U, Loppnow H, Flad H-D, Brade H. Bacterial Endotoxins: Relationships between chemical structure and biological activity. In: Gregoriadis G, Allison AC, Poste G, editors. Immunological Adjuvants and Vaccines Plenum Corporation. 1989. pp. 61–74.
4. Boman HG. Peptide antibiotics and their role in innate immunity. Annu Rev Immunol. 1995;13:61–92. [PubMed]
5. Hancock RE. Cationic antimicrobial peptides: towards clinical applications. Expert Opin Investig Drugs. 2000;9:1723–1729. [PubMed]
6. Andrä J, Gutsmann T, Garidel P, Brandenburg K. Mechanisms of endotoxin neutralization by synthetic cationic compounds. J Endotoxin Res. 2006;12:261–277. [PubMed]
7. Pena SV, Krensky AM. Granulysin, a new human cytolytic granule-associated protein with possible involvement in cell-mediated cytotoxicity. Semin Immunol. 1997;9:117–125. [PubMed]
8. Krensky AM, Clayberger C. Granulysin: a novel host defense molecule. Am J Transp. 2005;5:1789–1792. [PubMed]
9. Anderson DH, Sawaya MR, Cascio D, Ernst W, Modlin R, Krensky A, Eisenberg D. Granulysin crystal structure and a structure-derived lytic mechanism. J Mol Biol. 2003;325:355–365. [PubMed]
10. Wang Z, Choice E, Kaspar A, Hanson D, Okada S, Lyu SC, Krensky AM, Clayberger C. Bactericidal and tumoricidal activities of synthetic peptides derived from granulysin. J Immunol. 2000;165:1486–1490. [PubMed]
11. Andersson M, Girard R, Cazenave P. Interaction of NK lysin, a peptide produced by cytolytic lymphocytes, with endotoxin. Infect Immun. 1999;67:201–205. [PMC free article] [PubMed]
12. Andrä J, Koch MH, Bartels R, Brandenburg K. Biophysical characterization of endotoxin inactivation by NK-2, an antimicrobial peptide derived from mammalian NK-lysin. Antimicrob Agents Chemother. 2004;48:1593–1599. [PMC free article] [PubMed]
13. Tu X, Latifi T, Bougdour A, Gottesman S, Groisman EA. The PhoP/PhoQ two-component system stabilizes the alternative sigma factor RpoS in Salmonella enterica. Proc Natl Acad Sci U S A. 2006;103:13503–13508. [PubMed]
14. Galanos C, Lüderitz O, Westphal O. A new method for the extraction of R lipopolysaccharides. Eur J Biochem. 1969;9:245–249. [PubMed]
15. Westphal O, Lüderitz O, Bister F. Uber die Extraktion von Bakterien mit Phenol/Wasser. Z Naturforsch. 1952;7:148–155.
16. Jürgens G, Müller M, Koch MH, Brandenburg K. Interaction of hemoglobin with enterobacterial lipopolysaccharide and lipid A. Physico-chemical characterization and biological activity. Eur J Biochem. 2001;268:4233–4242. [PubMed]
17. Brandenburg K, Seydel U. Infrared spectroscopy of glycolipids. Chem Phys Lipids. 1998;96:23–40. [PubMed]
18. Koch MHJ. Instruments and methods for small-angle scattering with synchrotron radiation. Makromol Chem Macromol Symp. 1988;15:79–90.
19. Brandenburg K, Jürgens G, Andrä J, Lindner B, Koch MH, Blume A, Garidel P. Biophysical characterization of the interaction of high-density lipoprotein (HDL) with endotoxins. Eur J Biochem. 2002;269:5972–5981. [PubMed]
20. Schromm AB, Brandenburg K, Rietschel ET, Flad HD, Carroll SF, Seydel U. Lipopolysaccharide-binding protein mediates CD14-independent intercalation of lipopolysaccharide into phospholipid membranes. FEBS Lett. 1996;399:267–271. [PubMed]
21. Hamamoto K, Shimizu T, Kida Y, Kuwano K. Interactions of a small linear cationic peptide with lipopolysaccharide and lipoteichoic acid. Kurume Med J. 2003;50:99–107. [PubMed]
22. Andrä J, Lohner K, Blondelle SE, Jerala R, Moriyon I, Koch MH, Garidel P, Brandenburg K. Enhancement of endotoxin neutralization by coupling of a C12-alkyl chain to a lactoferricin-derived peptide. Biochem J. 2005;385:135–143. [PubMed]
23. Mueller M, Lindner B, Kusumoto S, Fukase K, Schromm AB, Seydel U. Aggregates are the biologically active units of endotoxin. J Biol Chem. 2004;279:26307–26313. [PubMed]
24. Seydel U, Koch MH, Brandenburg K. Structural polymorphisms of rough mutant lipopolysaccharides Rd to Ra from Salmonella minnesota. J Struct Biol. 1993;110:232–243. [PubMed]
25. Brandenburg K, Koch MH, Seydel U. Phase diagram of deep rough mutant lipopolysaccharide from Salmonella minnesota R595. J Struct Biol. 1992;108:93–106. [PubMed]
26. Gutsmann T, Schromm AB, Koch MHJ, Kusumoto S, Fukase K, Oikawa M, Seydel U, Brandenburg K. Lipopolysaccharide-binding protein-mediated interaction of lipid A from different origin with phospholipid membanes. Phys Chem Chem Phys. 2000;2:4521–4528.
27. Andrä J, Garidel P, Majerle A, Jerala R, Ridge R, Paus E, Novitsky T, Koch MH, Brandenburg K. Biophysical characterization of the interaction of Limulus polyphemus endotoxin neutralizing protein with lipopolysaccharide. Eur J Biochem. 2004;271:2037–2046. [PubMed]
28. Jiao BH, Freudenberg M, Galanos C. Characterization of the lipid A component of genuine smooth-form lipopolysaccharide. Eur J Biochem. 1989;180:515–518. [PubMed]
29. Riestchel ET, Kirikae T, Schade FU, Mamat U, Schmidt G, Loppnow H, Ulmer AJ, Zahringer U, Seydel U, Di Padova F, Schreier M, Brade H. Bacterial endotoxin: molecular relationships of structure to activity and function. FASEB J. 1994;8:217–225. [PubMed]
30. Brandenburg K, Wiese A. Endotoxins: relationships between structure, function, and activity. Curr Top Med Chem. 2004;4:1127–1146. [PubMed]
31. Andrä J, Lamata M, Martinez de Tejada G, Bartels R, Koch MH, Brandenburg K. Cyclic antimicrobial peptides based on Limulus anti-lipopolysaccharide factor for neutralization of lipopolysaccharide. Biochem Pharmacol. 2004;68:1297–1307. [PubMed]
32. Shands JW. The Physical Structure of Bacterial Lipopolysacchrides. In: Weinbaum G, Kadis S, Ajl J, editors. Microbial Toxins. Academic Press; New York and London: 1971. pp. 127–144.
33. Hailman E, Lichenstein HS, Wurfel MM, Miller DS, Johnson DA, Kelley M, Busse LA, Zukowski MM, Wright SD. Lipopolysaccharide (LPS)-binding protein accelerates the binding of LPS to CD14. J Exp Med. 1994;179:269–277. [PMC free article] [PubMed]
34. Seydel U, Wiese A, Schromm AB, Brandenburg K. A biophysical view on the function and activity of endotoxins. In: Brade H, Opal SNSM, Vogel DC, editors. Endotoxin in Health and Disease. New York: 1999. pp. 195–220.
35. Brandenburg K, David A, Howe J, Koch MH, Andrä J, Garidel P. Temperature dependence of the binding of endotoxins to the polycationic peptides polymyxin B and its nonapeptide. Biophys J. 2005;88:1845–1858. [PubMed]
36. Brandenburg K, Moriyon I, Arraiza MD, Lehwark-Yvetot G, Koch MHJ, Seydel U. Biophysical investigations into the interaction of lipopolysaccharide with polymyxins. Thermochim Acta. 2002;382:189–198.