JBC Ideal method for primary cell transfection

HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

Originally published In Press as doi:10.1074/jbc.M410116200 on November 30, 2004

J. Biol. Chem., Vol. 280, Issue 7, 5803-5811, February 18, 2005
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow All Versions of this Article:
280/7/5803    most recent
M410116200v1
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Wang, G.
Right arrow Articles by Li, X.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Wang, G.
Right arrow Articles by Li, X.

Correlation of Three-dimensional Structures with the Antibacterial Activity of a Group of Peptides Designed Based on a Nontoxic Bacterial Membrane Anchor*

Guangshun Wang{ddagger}, Yifeng Li, and Xia Li

From the Structure-Fun Laboratory, Eppley Institute for Research in Cancer and Allied Diseases, University of Nebraska Medical Center, Omaha, Nebraska 68198-6805

Received for publication, September 2, 2004 , and in revised form, November 19, 2004.


   ABSTRACT
TOP
 ABSTRACT
INTRODUCTION
 EXPERIMENTAL PROCEDURES
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
To understand the functional differences between a nontoxic membrane anchor corresponding to the N-terminal sequence of the Escherichia coli enzyme IIAGlc and a toxic antimicrobial peptide aurein 1.2 of similar sequence, a series of peptides was designed to bridge the gap between them. An alteration of a single residue of the membrane anchor converted it into an antibacterial peptide. Circular dichroism spectra indicate that all peptides are disordered in water but helical in micelles. Structures of the peptides were determined in membrane-mimetic micelles by solution NMR spectroscopy. The quality of the distance-based structures was improved by including backbone angle restraints derived from a set of chemical shifts (1H{alpha}, 15N, 13C{alpha}, and 13C{beta}) from natural abundance two-dimensional heteronuclear correlated spectroscopy. Different from the membrane anchor, antibacterial peptides possess a broader and longer hydrophobic surface, allowing a deeper penetration into the membrane, as supported by intermolecular nuclear Overhauser effect cross-peaks between the peptide and short chain dioctanoyl phosphatidylglycerol. An attempt was made to correlate the NMR structures of these peptides with their antibacterial activity. The activity of this group of peptides does not correlate exactly with helicity, amphipathicity, charge, the number of charges, the size of the hydrophobic surface, or hydrophobic transfer free energy. However, a correlation is established between the peptide activity and membrane perturbation potential, which is defined by interfacial hydrophobic patches and basic residues in the case of cationic peptides. Indeed, 31P solid state NMR spectroscopy of lipid bilayers showed that the extent of lipid vesicle disruption by these peptides is proportional to their membrane perturbation potential.


   INTRODUCTION
TOP
 ABSTRACT
 INTRODUCTION
EXPERIMENTAL PROCEDURES
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Recent interest in the search for alternative therapeutics is growing because of the drug resistance problem with traditional antibiotics. Antimicrobial peptides have attracted much attention because of their favorable properties, such as rapid killing, wide spectrum, and rare development of drug resistance. It is believed that these properties of antimicrobial peptides can be attributed to their ability to target bacterial membranes (1-4).

Membrane targeting also plays a fundamental role in virus infection and intra- and intercell signal transduction. For example, enzyme IIAGlc1 from Escherichia coli is identified as an amphitropic protein, which can exist either in the cytoplasm or by attaching to the cytoplasmic membrane (5). Both states are essential for the protein cascade to ensure a successful phosphoryl transfer from the high energy molecule phosphoenolpyruvate to the incoming glucose. Membrane association of IIAGlc is achieved through an N-terminal membrane anchor (hereinafter referred to as peptide A1), which, according to two-dimensional NMR characterization, forms a short three-turn amphipathic helical structure in phospholipids (6). Because another membrane-targeting sequence similar to this anchor is conserved in other species (7), the structure of the N-terminal domain of IIAGlc is a useful model for understanding those membrane anchors (6).

To provide an analytical tool for structural and functional studies of antimicrobial peptides, we have created a user-friendly antimicrobial peptide database (aps.unmc.edu/AP/main.html) (8). It is of outstanding interest to note that the N-terminal sequence of the above bacterial membrane anchor, GLFD, is identical to those of 13 antibacterial peptides collected in the database. Because all of these peptides start with the sequence GLFD, we may refer to them as the GLFD family. Such a sequence similarity between an E. coli membrane anchor (6) and a group of antimicrobial peptides from Australian frogs (9) suggests that they might have originated from the same ancestral gene a long time ago. Because both types of peptides act on bacterial membranes, one toxic and the other nontoxic, we are curious to learn what determines their functional differences. Therefore, we have designed a series of peptides (peptides A2, A3, and A4) starting from the bacterial membrane anchor (peptide A1) (6) and ending at a known antibacterial peptide, aurein 1.2 (peptide A5) (9). The sequences and select properties of these peptides are provided in Table I.


View this table:
[in this window]
[in a new window]
  		TABLE I
Primary sequences and properties of group A peptides

 
To understand the structure and activity relationship of this group of peptides, it is essential to elucidate their three-dimensional structures at high resolution. Antimicrobial peptides are usually cationic with less than 50 residues (1-4, 8). Therefore, they are very suitable for NMR studies. In the antimicrobial peptide database (8), 68 peptides (13%) have been investigated by traditional two-dimensional homonuclear NMR spectroscopy in lipid-mimetic environments such as organic solvents or detergent micelles (9-11). In micelles, structures of antimicrobial peptides are determined based primarily on distance restraints derived from NOEs (10-12). This is because scalar coupling data, which also contain valuable structural information (12), are not amenable to measure by homonuclear NMR methods as a result of line broadening of peptide signals from micelle binding. In contrast, chemical shifts are easy to measure and provide an alternative approach to obtaining angle restraints by using the NMR program TALOS (13). Although TALOS-derived angle restraints have been included in the structural refinement of isotope-labeled proteins, the use of such restraints in the structural refinement of micelle-bound peptides without isotope labeling has not been demonstrated. We show in this study the improvement of the structural quality of these peptides using backbone angle restraints predicted by TALOS based on a set of heteronuclear chemical shifts. This work benefitted from the recent installation of a cool probe to the 600-MHz NMR spectrometer, which allows rapid data collection in hours, even for natural abundance peptides. High quality structures allow a better correlation with the activity of these peptides. Although no good correlation was found between the peptide activity and numerous structural parameters, the membrane perturbation potential as we defined shows a nice correlation.


   EXPERIMENTAL PROCEDURES
TOP
 ABSTRACT
 INTRODUCTION
 EXPERIMENTAL PROCEDURES
RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 
Materials—All antibacterial peptides (>95%) investigated here are synthesized and purified by Genemed Synthesis, Inc. (San Francisco). The primary sequences and related information for these peptides can be found in Table I. Deuterated SDS (>99%) was obtained from Cambridge Isotope Laboratories, Inc. (Andover, MA). Protonated DOPG (>98%), dioleoyl phosphatidylglycerol, and dioleoyl phosphatidylethanolamine (>99%) were purchased from Avanti%20Polar%20Lipids">Avanti Polar Lipids (Alabaster, AL). Chloroform was removed from phospholipids under a stream of nitrogen gas followed by evaporation under vacuum overnight. The lipids and SDS were used without further purification.

Lipid Vesicles—The lipid vesicles (4 mM) of dioleoyl phosphatidylglycerol (25%) and dioleoyl phosphatidylethanolamine (75%) (molar ratio) in 25 mM Tris buffer in D2O, pD 7, were made as described previously (51) by repeated cycles of vortexing on a mixer and 30-s heating in a water bath at 57 °C.

Antibacterial Assay—The E. coli strain K12 3000 was a gift from Dr. Alan Peterkofsky (National Institutes of Health). The antibacterial activity of the peptides was analyzed using the standard approach of microdilution (14). In brief, a small culture was grown overnight. A fresh culture was inoculated with a small aliquot of the overnight culture and incubated at 37 °C until the optical density reached the logarithm stage. The culture was then diluted and partitioned into a 96-well plate with ~106 cells/well (90 µl each). The cells were then treated with 10 µl of the peptide at a series of concentrations, allowing the minimum inhibition concentration measurement for each. The plate was then further incubated at the same temperature overnight (~16 h) and read on an Ultra Microplate Reader at 620 nm (Bio-TEK Instruments).

CD Spectroscopy—CD spectra were collected at 25 °C on a Jasco J-810 spectropolarimeter from 185 to 250 nm using a 0.1-mm path length cell, with a scan rate of 100 nm/min, a time constant 1.0 s, a band width of 1 nm, and a sensitivity of 100 millidegrees. Each spectrum is the average of 20 scans. After background subtraction, the spectrum was expressed in molar ellipticity.

NMR Spectroscopy—For NMR measurements, the peptide concentration is typically ~2mM in 0.6 ml of aqueous solution of 90% H2O and 10% D2O at pH 5.4. The pH of each sample was adjusted by using microliter aliquots of HCl or NaOH solution and measured directly in the 5-mm NMR tube with a micro-pH electrode (Wilmad-Labglass). Based on detergent titrations, the peptide/SDS molar ratio was 1:40 and the peptide/DOPG ratio was 1:5. All proton NMR data were collected at 25 °C on a Varian INOVA 600 MHz NMR spectrometer equipped with a triple-resonance cryoprobe. A set of NMR spectra was collected for each peptide using States-TPPI (15). NOESY spectra (16) were acquired at mixing times of 50, 100, and 150 ms for peptide-micelle complexes. For peptides in water, NOESY spectra were collected at 200 ms. TOCSY experiments were performed with a mixing time of 75 ms using a clean MLEV-17 pulse sequence (17-19). ROSEY spectra (20) were collected at a mixing time of 35 ms. Typically, two-dimensional homonuclear spectra were collected with 512 increments (16-32 scans each) in t1 and 2,048 data points in t2 time domains using a spectral width of 8,510.6 Hz in both dimensions with the 1H carrier on the water resonance. The water signal was suppressed by low power presaturation during both the relaxation delay and the mixing period in NOESY experiments and during relaxation delay only for TOCSY (18) and DQF-COSY (21) experiments.

To obtain backbone 15N, 13C{alpha}, and 13C{beta} chemical shifts, gradient-enhanced HSQC spectra (22), between 1H and 15N as well as between 1H and 13C, were collected at the natural abundance on the Varian 600 MHz NMR spectrometer. The 1H, 15N, and 13C carriers were set at 4.77, 118.27, and 36.37 ppm, respectively. Typically 30 increments (128 scans) and 80 increments (256 scans) were collected for the 15N (spectral width 2,200 Hz) and aliphatic 13C (spectral width 12,000 Hz) dimensions, respectively.

31P NMR spectra for vesicles, consisting of a mixture of the E. coli lipids, were recorded on a 500 MHz Varian NMR spectrometer (202.2 MHz 31P frequency) at 25 °C. The 90° pulse for 31P on this direct detected probe is 9 µs. 160,000 complex data points were collected at a spectral width of 50 kHz and a relaxation delay of 1.5 s with proton decoupling. The 31P chemical shift of phosphoric acid (85%) was set at 0.0 ppm.

All NMR data were processed on an Octane work station (SGI) using the NMRPipe software (23). The data points in the indirect dimension were doubled by liner prediction (24). NMR data were apodized by a 63° shifted squared sine-bell window function in both dimensions, zero-filled prior to Fourier transformation to yield a data matrix of 2,048 x 1,024. Because anionic DSS interacts with cationic peptides (6), it was not used as an internal chemical shift standard to prevent the detergent additive effect. Instead, the proton chemical shifts of the peptide were referenced to the water signal, which in turn was referenced to internal DSS at 0.00 ppm (25). 15N and 13C chemical shifts were referenced based on the ratios recommended by IUPAC (26). NMR data were analyzed with the program PIPP (27).

The peptide proton signals were assigned using the standard procedure (12) based on two-dimensional TOCSY, DQF-COSY, and NOESY spectra. 15N, 13C{alpha}, and 13C{beta} chemical shifts of the peptides were assigned on the basis of the known proton chemical shifts. Occasional ambiguity in assigning heteronuclear chemical shifts as a result of overlap is removed by comparison with a spectrum collected at a slightly different temperature or by comparison with standard chemical shifts as well as with the assignments of other similar peptides investigated here.

Structure Calculations—Three-dimensional structures of the peptides in SDS-d25 at pH 5.4 and 25 °C were calculated based on both distance and angle restraints by using the simulated annealing protocol (28) in the National Institutes of Health version of X-PLOR (29, 30). To see the impact of TALOS-derived backbone angles on the structural quality, a separate calculation was performed for each peptide by omitting those angle restraints.

The distance restraints were obtained by classifying the NOE cross-peak volumes into strong (1.8-2.8 Å), medium (1.8-3.8 Å), weak (1.8-5.0 Å), and very weak (1.8-6.0 Å) ranges (31). The distance was calibrated on the basis of the typical NOE patterns in an {alpha} helix (12). Peptide backbone restraints were obtained from the TALOS (13) analysis of a set of heteronuclear chemical shifts, including 1H{alpha}, 13C{alpha}, 13C{beta}, and 15N. A broader range (±20°) than predicted was allowed for each angle in the structural calculations. The side chain {chi}1 angles were derived from a combined analysis of a short mixing time NOESY and ROESY spectra (32). A list of the number of NMR distance and angle restraints used for structural calculations of each peptide is given in Table II. A covalent peptide structure with random {varphi}, {Psi}, and {chi} angles but trans-planar peptide bonds was used as a starting structure. All peptide structural templates were also amidated at the C terminus using X-PLOR (29). In total, 100 structures were calculated. An ensemble of 50 structures with the lowest total energy was chosen for structural analysis. This final ensemble of accepted structures satisfies the following criteria: no NOE violations greater than 0.50 Å, r.m.s.d. for bond deviations from the ideal less than 0.01 Å, and r.m.s.d. for angle deviations from the ideal less than 5°.


View this table:
[in this window]
[in a new window]
  		TABLE II
Structural statistics of group A peptides bound to SDS micelles at a peptide/SDS molar ratio of 1:40, pH 5.4, and 25 °C

 

   RESULTS AND DISCUSSION
TOP
 ABSTRACT
 INTRODUCTION
 EXPERIMENTAL PROCEDURES
 RESULTS AND DISCUSSION
CONCLUSIONS
 REFERENCES
 
Peptide Design and Activity Assay—The original E. coli membrane anchor (6) is named peptide A1 (Table I). For comparison, we chose aurein 1.2 (9) because it is one of the shortest peptides in the GLFD family of antimicrobial peptides (8). In addition, the C-terminal residue, Phe13, of this peptide was found to be essential for anticancer activity (9). As a consequence, peptide A2 was designed by changing Asp13 of peptide A1 to Phe13 with Lys14-Lys15 deleted. Further, a hydrophobic residue, either Val or Ile, is found at position 5 of the antimicrobial peptides in the GLFD family. Thus, Lys5 in peptide A2 was further changed to an isoleucine residue. Meanwhile, we also changed Leu6 to Val6 to facilitate the proton assignments of methyls by homonuclear two-dimensional NMR spectroscopy. This new peptide is named peptide A3. Finally, the Lys7-Lys8 pair is conserved in all of the peptides starting with the sequence GLFD (8, 9). Therefore, Ser8 in peptide A3 was replaced by Lys8, leading to peptide A4. The overall sequence of peptide A4 is now comparable with aurein 1.2 (peptide A5, Table I). Like aurein 1.2, all designed peptides were amidated at the C terminus.

The antibacterial activity of these peptides was assayed using E. coli itself. As a functional domain of a signal transduction protein (6), one would assume that the E. coli membrane anchor (peptide A1) is not toxic to the bacterium. This is indeed the case based on the assay. The minimum inhibition concentration values for this group of peptides are given in Table I. The antibacterial activity of the peptides designed is in the following order: peptide A5 ≥ peptide A4 > peptide A2 > peptide A3. The lowest activity of peptide A3 to E. coli may be attributed to its net charge of 0, which makes it not as soluble in water as other peptides in this series.

Circular Dichroism—Fig. 1 shows the CD spectra of the peptides in water and in SDS. In the absence of detergents, the spectra of these peptides are characterized by a strong negative band at ~200 nm, indicating the disordered states. The CD spectra of peptide A5 changed slightly with concentration from 0.1 to 0.5 mM. At 0.5 mM, a shoulder was observed near ~222 nm (Fig. 1), suggesting a conformational change as a result of peptide aggregation. The spectrum for peptide A3 at 0.5 mM in water was not recorded because of limited solubility. In the presence of 40-fold SDS, the CD spectra for all peptides changed dramatically (Fig. 1). A combination of a positive band at ~195 nm and double minima at ~208 and 222 nm strongly suggests helical structures in all peptides (52). Because the 222 nm band is proportional to the degree of helix, we estimated the helicity for each peptide using the method of Jackson et al. (53). Thus, the helix percentages for peptides A2, A4, and A5 in SDS micelles are 53%, 70%, and ~100%, respectively.



View larger version (23K):
[in this window]
[in a new window]
  		FIG. 1.
CD spectra of the peptides in the absence and presence of lipid-mimetic micelles. From top to bottom: open diamonds, peptide A5 in water; filled circles, peptide A2/SDS 1:40; open squares, peptide A4/SDS 1:40; and filled triangles, peptide A5/SDS 1:40. Data shown here were recorded at a peptide concentration of 0.5 mM, pH 5.4, and 25 °C.

 
NMR Structural Analysis of Antimicrobial Peptides in Water—As observed previously with peptide A1 (6), peptides A2, A3, and A4 have no ordered backbone structure in water according to NOESY experiments. Peptide A5 is an exception. The one-dimensional NMR spectrum of peptide A5 is both pH- and concentration-dependent. At 0.8 mM and pH 3.3, the amide signals of the peptide showed a narrow dispersion from 8.51 to 8.03 ppm, indicating primarily a random coiled state. Increasing the pH to 5.4, however, improved the signal dispersion (8.6-7.86 ppm), suggesting that electrostatic interactions play a role in determining the peptide conformation in water (Table I). An increase of the peptide concentration to 2.5 mM at pH 5.4 further expanded the amide region to 8.75-7.73 ppm. Under these conditions, strong NOE cross-peaks between sequential amide protons of peptide A5, as well as weak short range NOE cross-peaks of (i, i + 2), (i, i + 3), and (i, i + 4) types were detected, suggesting a nascent helical structure in water (33).

Secondary Structures of Antibacterial Peptides in SDS or DOPG Micelles by NMR—To shed light on the molecular form in the bacterial membrane, we also elucidated the structures of these peptides in SDS micelles by NMR (10, 11, 34). Fig. 2 presents portions of the NOESY spectra regions for peptides A2 and A5. The cross-peaks of peptide A5 are better dispersed than those of peptide A2. In addition to SDS, short chain phosphatidylglycerols (PGs) were also shown to be useful for structural studies of membrane-associating peptides (6, 35, 36). PGs are the major anionic lipids in E. coli and arguably have higher biological relevance than SDS as a membrane-mimetic model. Because deuterated short chain PGs are not available commercially at present, it is advisable to utilize as few protonated PGs as possible in two-dimensional homonuclear NMR studies so that the interference of strong lipid signals with measurements can be diminished. As a result, we employed DOPG in this study (36). Fig. 3A shows the plots of the H{alpha} chemical shifts of peptide A4 in both SDS and DOPG. The two lines overlap nicely, indicating similar structures in both micelles. Peptide A1 has already been found to possess similar solution structures in SDS, DHPG (6), or DOPG (36). All of these structural data support that anionic SDS mimics short chain anionic PGs well in these particular cases. Consequently, we have performed detailed structural studies of antibacterial peptides A2-A5 in deuterated SDS micelles. Deuteration of detergents not only suppresses potential spin diffusion but also facilitates the collection of high quality HSQC spectra, especially between proton and carbon at a natural abundance (below).



View larger version (28K):
[in this window]
[in a new window]
  		FIG. 2.
The fingerprint regions of the NOESY spectra of peptide A2 (A) and peptide A5 (B) in deuterated SDS at a peptide/SDS ratio of 1:40, pH 5.4, and 25 °C. The HN-H{alpha} NOE cross-peak for each residue is labeled with the single-letter amino acid code. Interresidue NOE cross-peaks are labeled with two numbers separated by a slash. Each number corresponds to the residue number with the first one for H{alpha} and the second one for the amide proton of the residue.

 



View larger version (26K):
[in this window]
[in a new window]
  		FIG. 3.
Chemical shift plots. A, the H{alpha} chemical shift of peptide A4 as a function of residue number in SDS (solid ellipses) and DOPG (open ellipses) at 25 °C. The NMR data in SDS were recorded at a peptide/detergent molar ratio of 1:40 at pH 5.4, and the data in DOPG were collected at a peptide/lipid ratio of 1:5 at pH 5.9. A slightly higher pH was used in the case of DOPG to increase the solubility of the complex. The chemical shifts were extracted from two-dimensional NMR spectra. B, plots of the secondary amide proton shifts of the peptides in SDS versus the residue number. Solid diamonds, peptide A1; solid triangles, peptide A2; open triangles, peptide A3; open squares, peptide A4; and solid circles, peptide A5. Secondary shifts are calculated as the differences between the measured shifts and the random shifts tabulated by Wüthrich (12). All chemical shifts are obtained from two-dimensional NMR spectra in SDS. The chemical shifts for peptide A1 were obtained from Ref. 6.

 
A combination of medium (i, i + 1), (i, i + 3), weak (i, i + 2), and (i, i + 4) types of NOE cross-peak indicates helical structures (12) for peptides A2 and A5, almost covering the entire sequence (Fig. 2). Similar results were observed for peptides A3 and A4 as well. The helical structures for these peptides are also supported by secondary shifts as calculated previously (6). The helicity of each peptide in SDS was quantified based on secondary shifts of H{alpha} chemical shifts, and the results are listed in Table II. Apparently, the helicity increases steadily from peptide A1 (64%) through peptide A5 (95%). It is evident that the overall trend in the helix content of these peptides in SDS is identical to that found by CD (above). To compare the effect of different environments on the helicity, we also calculated the helix content of peptide A5 under other conditions. In water, peptide A5 has a helix percentage of 84% at 2.5 mM and pH 5.4, which is actually higher than that (78%) found in trifluoroethanol at pH 2 reported previously (9). Thus, peptide A5 is most helical (95%) in micelles, probably because of more favorable hydrophobic peptide-detergent interactions (6).

Fig. 3B presents the plots for the secondary shift of amide protons for this series of peptides. Interestingly, the amide secondary shift of residue 10 is getting more positive from peptide A1 to peptide A5. Because the periodicity of the amide plot is an indication of amphipathic structure (37, 38), we also calculated the amphipathicity for each peptide in a manner similar to helicity. We define the amphipathicity as the ratio between an arbitrary reference value of -0.205 and the average amide secondary shift of a peptide. The reference value -0.205 is the average amide shift for peptide A5, which has the highest helicity in this series of peptides bound to micelles. The trend for amphipathicity is the same as the helicity (Table II) with a correlation coefficient of 0.92 between them.

Three-dimensional Structure Refinement Using Chemical Shift-derived Backbone Angles—The total number of distance restraints for each peptide (Table II) varies slightly as a result of the differences in primary sequence and spectral dispersion. In addition, we have also obtained dihedral angle restraints based on a set of heteronuclear chemical shifts using the NMR program TALOS (13). The 1H{alpha}, 15N, 13C{alpha}, and 13C{beta} chemical shifts were obtained from two-dimensional HSQC spectra between 15N and 1H (Fig. 4A) as well as between 13C and 1H (Fig. 4B), both at a natural abundance. Carbonyl carbons have no directly attached protons, and their chemical shifts are not available in these spectra. However, a comparative analysis of the chemical shifts of histidine-containing protein, HPr, revealed that 92% of the TALOS predictions are the same with or without the carbonyl shifts.



View larger version (51K):
[in this window]
[in a new window]
  		FIG. 4.
Portions of natural abundance two-dimensional HSQC spectra of peptide A4 in deuterated SDS at pH 5.4 and 25 °C. Shown in A are the correlations between the backbone amide 1H and 15N and in B are cross-peaks between 1H and 13C. Peaks are labeled. The unlabeled cross-peak at the cross of 4 and 63 ppm belongs to the H{beta}/C{beta} of Ser11. C{beta} chemical shifts for other residues were obtained from other regions of the same spectrum.

 
According to TALOS (13), residues Leu2-Ser11 of peptide A2 were predicted as "good," and their backbone dihedral angles are located in the helical region. Similarly, residues Leu2-Ser11 of peptide A3, Leu2-Asp12 of peptide A4, and Phe3-Glu11 of peptide A5 were predicted to have helical conformations. Each good prediction (at least 9 of 10) (13) allows the generation of two angle restraints ({varphi}, {varphi}) (Table II). Fig. 5 shows the Ramachandran plots for peptide A2 (A and B) and peptide A4 (C and D) with and without the backbone angle restraints. In both cases, the plots are scattered without the TALOS-derived angles (Fig. 5, A and C). In the presence of the angle restraints, however, nearly all dots are clustered in the most favored region of the plots (Fig. 5, B and D). A quantitative analysis by the program Procheck (39) was performed on the 50 low energy structures. For peptides A2 and A4 without the use of the chemical shift-derived angles, the residues in the most favored region of the Ramachandran plot are 64 and 84%, respectively. However, the percentages increased to 99 and 100% with the backbone angle restraints. In both cases, the structure quality improved significantly because of the addition of the angle restraints. In the absence of the angle restraints, the slightly lower percentage in the most favored region for peptide A2 (64%) than peptide A4 (84%) may be attributed to the differences in their NOE-derived distance restraints. In the case of peptide A5, where a better set of distance restraints was obtained, the additional backbone angles for residues Phe3-Glu11 showed little effect on the Ramachandran plot (Table II).



View larger version (46K):
[in this window]
[in a new window]
  		FIG. 5.
Ramachandran plots (46) for the ensembles of 100 peptide structures determined with and without TALOS-derived angle restraints. Plots are shown for peptide A2 (A) and peptide A4 (C) in the presence of NOE-derived distance restraints only. B and D are plots for peptides A2 and A4, respectively, in the presence of both distance and TALOS-derived (12) dihedral angle restraints. All structures represent SDS-bound forms. The figures were generated using MOLMOL (47).

 
In addition, both the calculation convergence and success rate have increased by including backbone angle restraints. For example, without the additional angle restraints, the ensemble of 100 structures of peptide A2 has a wide distribution in helix length, ranging from 3 to 11 with the majority of the peptides having helix lengths at 7 (29%) and 9 (30%). In the presence of the angle restraints, the predominant structures (62%) have a helix length of 10. Hence, the helix length for peptide A2 converged significantly upon the addition of angle restraints. Similar effects were observed with other peptides.

Displayed in Fig. 6 are the ensembles of the backbone structures for these peptides. The coordinate precision, as measured by the backbone r.m.s.d. for the final accepted 50 structures, is 0.40 Å for peptide A2, 0.27 Å for peptide A3, 0.19 Å for peptide A4, and 0.21 Å for peptide A5. A trend of the r.m.s.d. drop from peptide A2 to peptide A5 may reflect the quality of distance restraints (e.g. Fig. 2) because a similar number of angle restraints was utilized (Table II). In general, the backbone r.m.s.d. for the ensemble of structures reduced slightly by including the angle restraints. All of these results indicate that high quality structures for peptides A2-A5 have been determined by using all NMR restraints in Table II. Hence, chemical shift-based angle restraints provide a practical approach for structural refinement of antimicrobial peptides that are chemically synthesized without isotope enrichment.



View larger version (28K):
[in this window]
[in a new window]
  		FIG. 6.
The backbone views of the three-dimensional structures of the peptides in SDS micelles. In each case, the backbone atoms for residues 2-12 were superimposed using MOLMOL (47). Provided (from top to bottom) are the ensembles of the final accepted structures for peptide A2 (A), peptide A3 (B), peptide A4 (C), and peptide A5 (D).

 
Structure Description—All peptides have an amphipathic helical structure with a clear segregation of the hydrophobic and hydrophilic sectors (Fig. 7). On one hand, the hydrophobic surfaces are very similar as measured by hydrophobic transfer free energy (Table II) based on the hydrophobicity scale of Karplus (40). On the other hand, the hydrophilic surfaces of these peptides differ. This is mainly reflected in the number of potential salt bridges with an (i, i + 3) or (i, i + 4) relationship in the sequence (41) (Table I). For peptides A2 (Fig. 7A) and A3 (Fig. 7B), there are two possible salt bridges, namely between the N terminus of Gly1 and Asp4, and between Asp4 and Lys7. Nevertheless, there are four potential salt bridges on the hydrophilic surfaces of peptide A4 (Fig. 7C) and five on peptide A5 (Fig. 7D). The one less potential salt bridge in peptide A4 results from the alteration of the Glu11-Ser12 pair in peptide A5 by the Ser11-Asp12 pair in peptide A4. On average, however, the distances between these pairs of charged residues are generally greater than 6 Å, indicating that salt bridges (3.5 Å, Ref. 42) do not exist between these residues of the peptides bound to SDS micelles.



View larger version (43K):
[in this window]
[in a new window]
  		FIG. 7.
Ribbon diagrams of the peptide structures. The structures were elucidated using all of the NMR restraints as listed in Table II. The ribbon (red and yellow) width does not reflect the real size but clearly shows the helical feature for each peptide. Side chains (blue) are labeled with the single-letter amino acid code.

 
To substantiate this further, we also collected NMR data for peptide A5 at pH 2.2. A plot of the chemical shift deviations from pH 5.4 to 2.2 against residue number, using the combined proton and nitrogen shifts in the natural abundance HSQC spectra (43), found that the effect of pH is local. The most perturbed residues are Phe3, Asp4, and Ile5 (150-300 Hz). In addition, the protonations of Asp4 and Glu11 appear to accelerate further the acid-catalyzed local amide exchange. This is evidenced by the even broader cross-peaks for Leu2, Lys8, and Glu11 (25-33 Hz) compared with other cross-peaks (17-22 Hz) at pH 2.2, which are, in turn, broader than those at pH 5.4. However, the helicity of peptide A5 did not change from pH 5.4 (95%) to pH 2.2 (95%), as calculated based on the H{alpha} shifts (see Table II legend). Therefore, we conclude that salt bridges are not significant, if there are any, in stabilizing the helical structure of peptide A5 bound to micelles. It is the hydrophobic interactions that dominate (44).

Comparison with the Bacterial Membrane Anchor—Because the bacterial membrane anchor (peptide A1) is not toxic to E. coli, it would be useful to compare its structure determined previously (6) with those of antibacterial peptides elucidated here (Fig. 7). Peptide A1 has a short helical region from residues Phe3 to Val10 with a narrow hydrophobic surface (green, Fig. 8E). The addition of Phe13 made the hydrophobic surfaces of peptides A2-A5 longer and broader (Fig. 8, A-D). The total solvent-accessible surfaces for the three-dimensional structures of peptides A2-A5 are comparable, ~1,300 Å2 (Table II). Note that the hydrophobic surface of the peptide is protected by detergents in the NMR samples as supported by NOE cross-peaks between the peptide and lipid (below). They are "exposed" in the three-dimensional structures displayed here (Figs. 7, 8, 9) as a result of the absence of lipids in structural calculations. The hydrophobic portion of solvent-accessible surfaces, however, can be employed as an estimation of the size of the hydrophobic surface (Table II in parentheses). Consistent with Fig. 8, peptides A2-A5 have a slightly larger exposed hydrophobic surface (~760 ± 20 Å2) than peptide A1 (529 Å2). A larger hydrophobic surface would allow the peptide to penetrate deeper into the bacterial membrane.



View larger version (37K):
[in this window]
[in a new window]
  		FIG. 8.
Space-filling models of the antibacterial peptides (viewed from the N terminus of the helix). Green, hydrophobic residues; yellow, cationic residues; mixed colors of red (oxygen), blue (nitrogen), gray (carbon), and white (hydrogen), hydrophilic residues. The structure most resembling the average is shown. The figure was generated using RASMOL (www.umass.edu/microbio/rasmol).

 



View larger version (59K):
[in this window]
[in a new window]
  		FIG. 9.
Potential surfaces of peptide A2 (A), peptide A3 (B), peptide A4 (C), and peptide A5 (D). Hydrophobic grooves bordered by positive charges, as shown in D, have a high membrane-perturbation potential. Such a potential (D) is reduced in peptides A2, A3, and A4 as a result of the presence of an acidic residue in the vicinity of the positive charge (C), the lack of a hydrophobic groove (A), or the absence of the positive charge completely (B). Blue, basic residues; red, acidic residues; and white, hydrophobic and neutral residues. The figure was made using MOLMOL (47).

 
Also of note is that the sum of solvent-accessible surfaces for Phe3 and Phe13 accounts for up to ~50% of the exposed hydrophobic portion in peptides A2-A5, suggesting their importance in membrane binding. Indeed, aromatic residues are good membrane anchors (44). To provide direct evidence for peptide-lipid interactions, we also collected intermolecular NOESY spectra (6, 45). In the presence of protonated lipids such as DOPG, the aromatic rings of both Phe3 and Phe13 of peptide A4 showed NOE cross-peaks with the lipid acyl chains but not with the glycerol backbone protons. Similar results were observed with peptide A5. This is different from our previous NMR data for peptide A1, which showed intermolecular NOE cross-peaks with both the acyl chains and the glycerol backbone of short chain PGs, including DHPG, DOPG, and didecanoyl phosphatidylglycerol (6, 36). Thus, these peptide-lipid cross-peaks not only substantiate the lipid binding of both the membrane anchor and antibacterial peptides, but also suggest that peptide A4 has inserted slightly deeper into the lipid environment than peptide A1.

Structure-Activity Relationship—As shown in Tables I and II, the following parameters do not correlate exactly with the order of the peptide activity: net charge, the number of positively charged residues, helicity, amphipathicity, hydrophobic transfer free energy, and the size of the hydrophobic surface. Then, what kind of structural feature is correlated with the peptide activity? On one hand, a narrow hydrophobic surface, such as what was observed with the bacterial membrane anchor (peptide A1) (Fig. 8E), has limited penetration power (6) and is not toxic to the bacteria. On the other hand, a simple deep penetration is probably not sufficient to kill the bacteria either (Fig. 9B). The poor activity of peptides A2 and A3 indicates that both positive charges and hydrophobic residues are critical. By assuming that the bacterial killing is a result of membrane perturbation by cationic peptides, we introduce the concept of membrane perturbation potential (MPP) based on the three-dimensional structures (Fig. 9). MPP can be described by a combination of structural elements, which are hydrophobic grooves, positively and negatively charged residues in the interface. Both hydrophobic and basic residues are positive factors toward negatively charged membranes, whereas interfacial acidic residues, if found in the vicinity of a positively charged residue, may counteract the effect. Based on this idea, peptide A5 is the most active peptide in this group because the hydrophobic grooves of the structure are bordered by positive charges with negative charges in the middle of the hydrophilic face (Fig. 9D). Such a structure is perfect for the docking of anionic lipids, leading to lipid reorganization in the bacterial membrane. Peptide A4 possesses an amino acid sequence very similar to that of peptide A5 (Table I), with the exception of the switch in position of an acidic residue from 11 in peptide A5 to 12 in peptide A4. It appears that Asp12 in peptide A4 is deviated from the hydrophilic face and closer to the interfacial region near Lys8 (Fig. 9C). Because there is no salt bridge between Lys8 and Asp12 (above), the offset effect of acidic Asp12 to basic Lys8 appears to be insignificant, consistent with the marginal activity difference between peptides A4 and A5 (Table I). Although peptide A2 has the same number of basic residues as peptide A4, the hydrophobic defect (Fig. 9A) underneath Lys5 hinders its ability to attract anionic lipids. The worst case is peptide A3, where Lys8 is absent, and the hydrophobic grooves in the vicinity of Ile5 are bordered by the negative charge of Asp12 (Fig. 9B). Such a structure is not attractive to anionic lipids at all and may further explain the lowest activity of peptide A3. The opposite side of the surfaces of those structures in Fig. 9, which we did not discuss, is more or less similar. The common feature is the Lys7 and Phe3 pair, which appears to be an excellent membrane-targeting motif found initially in the bacterial membrane anchor (6).

Interactions with Phospholipid Bilayers Consisting of a Mixture of the E. coli Lipids—To test the above MPP hypothesis further, we have also investigated the interactions of peptides A2, A3, A4, and A5 with lipid vesicles, consisting of dioleoyl phosphatidylethanolamine (75%) and dioleoyl phosphatidylglycerol (25%), mimicking E. coli membranes. The solid state 31P NMR spectra of the lipids in the presence and absence of these peptides at 25 °C are presented in Fig. 10. In the absence of the peptides, the 31P nuclei in lipid vesicles are axially symmetric and have a chemical shift anisotropy of ~40 ppm, typical of the liquid-crystalline lamellar phase (Fig. 10A) (51, 54-56). When peptide A5 was added until a peptide/lipid ratio of 1:16, a power-pattern spectrum similar to Fig. 10D was obtained. This spectrum displayed an opposite chemical shift anisotropy, which might result from the transition of a lipid bilayer to a hexagonal phase (56). Similar spectra were obtained for the peptide-vesicle complexes at a peptide A4/lipid ratio of 1:16 and at a peptide A2/lipid ratio of 1:8 (Fig. 10D). At a peptide A5/lipid ratio of 1:8, however, the chemical shift anisotropy (~40 ppm) turned positive again. In addition, an extra peak showed up at ~0 ppm (pointed by an arrow in Fig. 10B). This peak is attributed to the formation of isotropic nonlamellar structures (probably inverted cubic phases) (54-56). At the peptide A4/lipid ratio of 1:8, the spectrum (Fig. 10C) is almost identical to Fig. 10B, and the isotropic peak in Fig. 10C is only slightly smaller. To confirm this further, we also increased the temperature to 35 °C. The isotropic peaks indeed increased to a similar extent for both peptides A4 and A5 (1:8 ratio). Therefore, peptide A5 and peptide A4 have more or less similar lysis effects on the lipid vesicles. Because the behavior of peptide A2 at a ratio of 1:8 (Fig. 10D) is similar to those of peptides A4 and A5 at a lower ratio of 1:16, it can be concluded that peptide A2 is less active than either peptide A4 or peptide A5. A similar study on peptide A3 was not performed because of its limited solubility. Thus, these 31P NMR spectra of the E. coli lipids qualitatively suggest that the vesicle perturbation capability of these peptides is peptide A5 ≥ peptide A4 > peptide A2 > peptide A3, consistent with the order of antibacterial activity of these peptides. These parallel results from bacteria assay to lipid vesicle perturbation provide additional support that these peptides target bacterial membranes. They also support that MPP is a useful concept in understanding the antibacterial activity of antimicrobial peptides.



View larger version (22K):
[in this window]
[in a new window]
  		FIG. 10.
31P NMR spectra of lipid bilayers with and without antimicrobial peptides. The vesicles were made of a mixture of dioleoyl phosphatidylglycerol (25%) and dioleoyl phosphatidylethanolamine (75%), mimicking the E. coli membrane (5). A, E. coli lipids; B, peptide A5/lipid molar ratio of 1:8; C, peptide A4/lipid molar ratio of 1:8; and D, peptide A2/lipid molar ratio of 1:8.

 

   CONCLUSIONS
TOP
 ABSTRACT
 INTRODUCTION
 EXPERIMENTAL PROCEDURES
 RESULTS AND DISCUSSION
 CONCLUSIONS
REFERENCES
 
In this study, we have taken a new avenue in understanding the relationship between structure and activity of antimicrobial peptides by comparison with the E. coli membrane anchor of similar sequence (Table I). This nontoxic membrane anchor could be converted to a toxic peptide with very few changes. The membrane anchor has a shorter helix and narrower hydrophobic surface, which primarily interacts with the glycerol backbone of short chain PGs, as supported by intermolecular NOE data between the peptide and lipid (6). In contrast, antibacterial peptides possess a broader hydrophobic surface (Fig. 8), allowing them to penetrate deeper into the membrane. Peptide A3 is a good case to support that deeper penetration is necessary but not sufficient to kill E. coli. However, the antibacterial activity does not correlate directly with the size of hydrophobic surface. It does not correlate with net charge, hydrophobic transfer free energy, helicity, or amphipathicity (Tables I and II). The determination of high quality three-dimensional structures for these peptides in micelles (Figs. 4, 5, 6), including the use of chemical shift-derived backbone angles, revealed the hidden surfaces for these peptides, which are remarkably different (Fig. 9). This finding allows us to arrive at the MPP concept, the potential of a peptide in attracting anionic lipids from the bacterial membrane. It is proposed that a structural pattern, consisting of hydrophobic grooves bordered by basic side chains, possesses a high MPP (Fig. 9D). The concept of MPP allows us to correlate nicely the three-dimensional structures of this group of peptides with their antibacterial activity as well as their disruption ability to lipid vesicles (Fig. 10). Because MPP can be applied to any structure, irrespective of the specific structural scaffold underneath, this concept may be of general use for us to understand the activity of antimicrobial peptides of a variety of three-dimensional structures.


   FOOTNOTES
 
The atomic coordinates and NMR restraints (codes 1VM2, 1VM3, 1VM4, and 1VM5) have been deposited in the Protein Data Bank, Research Collaboratory for Structural Bioinformatics, Rutgers University, New Brunswick, NJ (http://www.rcsb.org/).

* This work was supported by the startup fund from the Eppley Institute, University of Nebraska Medical Center (to G. W.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact. Back

{ddagger} To whom correspondence should be addressed: Eppley Institute, Rm. 3018, University of Nebraska Medical Center, 986805 Nebraska Medical Center, Omaha, NE 68198-6805. Tel.: 402-559-4176; Fax: 402-559-4651; E-mail: gwang{at}unmc.edu.

1 The abbreviations used are: IIAGlc, glucose-specific enzyme IIA involved in phosphotransfer from phosphoenolpyruvate to glucose; DHPG, dihexanoyl phosphatidylglycerol; DOPG, dioctanoyl phosphatidylglycerol; DQF-COSY, double quantum-filtered correlated spectroscopy; DSS, 2,2-dimethyl-2-silapentane-5-sulfonate sodium salt; HSQC, heteronuclear single quantum correlation; MPP, membrane perturbation potential; NOE, nuclear Overhauser effect; NOESY, nuclear Overhauser effect spectroscopy; PG, phosphatidylglycerol; ROESY, rotating frame Overhauser effect spectroscopy; r.m.s.d., root mean square deviation; TALOS, torsion angle likelihood obtained from shift and sequence similarity; TOCSY, total correlated spectroscopy. Back


   ACKNOWLEDGMENTS
 
We thank Frank Delaglio and Dan Garrett (National Institutes of Health) for NMR software and Paul Keifer for assistance with 31P NMR as well as maintenance of the NMR Core Facility at the University of Nebraska Medical Center. We are indebted to Mike Owen and Sandor Lovas (Creighton University) for help with the collection of the CD data. The CD machine is part of the Structural Proteomics Facility at Creighton University, which is supported by National Institutes of Health Grant 1P20RR16469 from the Institutional Development Award Network of Biomedical Research Excellence Program of the National Center for Research Resource. We thank Alan Peterkofsky (National Institutes of Health) for the generous gift of the wild type E. coli used for antibacterial assays. We appreciate Ken Cowan and Angie Rizzino for allowing us to use their instruments. Finally, we thank Kristi Berger for text editing of the manuscript.



   REFERENCES
TOP
 ABSTRACT
 INTRODUCTION
 EXPERIMENTAL PROCEDURES
 RESULTS AND DISCUSSION
 CONCLUSIONS
 REFERENCES
 

  1. Boman, H. G. (1995) Annu. Rev. Immunol. 13, 61-92[CrossRef][Medline] [Order article via Infotrieve]
  2. Hancock, R. E. W. (2001) Lancet Infect. Dis. 1, 156-164[CrossRef][Medline] [Order article via Infotrieve]
  3. Ganz, T., and Lehrer, R. I. (1999) Mol. Med. Today 5, 292-297[CrossRef][Medline] [Order article via Infotrieve]
  4. Zasloff, M. (2002) Nature 415, 389-395[CrossRef][Medline] [Order article via Infotrieve]
  5. Wang, G., Peterkofsky, A., and Clore, G. M. (2000) J. Biol. Chem. 275, 39811-39814[Abstract/Free Full Text]
  6. Wang, G., Keifer, P. A., and Peterkofsky, A. (2003) Protein Sci. 12, 1087-1096[Abstract/Free Full Text]
  7. Szeto, T. H., Rowland, S. L., Rothfield, L. I., and King, G. F. (2002) Proc. Natl. Acad. Sci. U. S. A. 99, 15693-15698[Abstract/Free Full Text]
  8. Wang, Z., and Wang, G. (2004) Nucleic Acids Res. 32, D590-D592[Abstract/Free Full Text]
  9. Rozek, T., Wegener, K. L., Bowie, J. H., Olver, I. N., Caver, J. A., Wallace, J. C., and Tyler, M. J. (2000) Eur. J. Biochem. 267, 5330-5341[Medline] [Order article via Infotrieve]
  10. Gesell, J., Zaslof, M., and Opella, S. J. (1997) J. Biomol. NMR 9, 127-135[CrossRef][Medline] [Order article via Infotrieve]
  11. Rozek, A., Friedrich, C. L., and Hancock, R. E. W. (2000) Biochemistry 39, 15765-15774[CrossRef][Medline] [Order article via Infotrieve]
  12. Wüthrich, K. (1986) NMR of Proteins and Nucleic Acids, John Wiley and Sons, New York
  13. Cornilescu, G., Delaglio, F., and Bax, A. (1999) J. Biomol. NMR 13, 289-302[CrossRef][Medline] [Order article via Infotrieve]
  14. Yu, K., Park, K., Kang, S.-W., Shin, S. Y., Hahm, K.-S., and Kim, Y. (2002) J. Peptide Res. 60, 1-9[CrossRef][Medline] [Order article via Infotrieve]
  15. Marion, D., Ikura, M., Tschudin, R., and Bax, A. (1989) J. Magn. Reson. 85, 393-399
  16. Jeener, J., Meier, B. H., Bachmann, P., and Ernst, R. R. (1979) J. Chem. Phys. 71, 4546-4553[CrossRef]
  17. Braunschweiler, L., and Ernst, R. R. (1983) J. Magn. Reson. 53, 521-528
  18. Bax, A., and Davis, D. G. (1985) J. Magn. Reson. 65, 355-360
  19. Griesinger, C., Otting, G., Wüthrich, K., and Ernst, R. R. (1988) J. Am. Chem. Soc. 110, 7870-7872[CrossRef]
  20. Bax, A., and Grzesiek, S. (1996) Encyclopedia of Nuclear Magnetic Resonance (Grant, D. M., and Harris, R. K., eds) pp. 4157-4167
  21. Rance, M., Sørensen, O. W., Bodenhausen, G., Wagner, G., Enrst, R. R., and Wüthrich, K. (1983)) Biochem. Biophys. Res. Commun. 117, 479-485[CrossRef][Medline] [Order article via Infotrieve]
  22. Kay, L. E., Keifer, P. A., and Saarinen, T. (1992) J. Am. Chem. Soc. 114, 10663-10665[CrossRef]
  23. Delaglio, F., Grzesiek, S., Vuister, G. W., Zhu, G., Pfeifer, J., and Bax, A. (1995) J. Biomol. NMR 6, 277-293[Medline] [Order article via Infotrieve]
  24. Zhu, G., and Bax, A. (1990) J. Magn. Reson. 90, 405-410
  25. Hartel, A. J., Lankhorst, P. P., and Altona, C. (1982) Eur. J. Biochem. 129, 343-357[Medline] [Order article via Infotrieve]
  26. Markley, J. L., Bax, A., Arata, Y., Hilbers, C. W., Kaptein, R., Sykes, B. D., Wright, P. E., and Wüthrich, K. (1998) J. Biomol. NMR 12, 1-23[CrossRef][Medline] [Order article via Infotrieve]
  27. Garrett, D. S., Powers, R., Gronenborn, A. M., and Clore, G. M. (1991) J. Magn. Reson. 95, 214-220
  28. Nilges, M., Gronenborn, A. M., Brünger, A. T., and Clore, G. M. (1988) Protein Eng. 2, 27-38[Abstract/Free Full Text]
  29. Brünger, A. T. (1992) X-PLOR: A system for X-ray Crystallography and NMR, Yale University Press, New Haven, CT
  30. Schwieters, C. D., Kuszewski, J., Tjandra, N., and Clore, G. M. (2002) J. Magn. Reson. 160, 65-73
  31. Clore, G. M., and Gronenborn, A. M. (1989) CRC Crit. Rev. Biochem. Mol. Biol. 24, 479-564
  32. Clore, G. M., and Gronenborn, A. M. (1998) Trends Biotechnol. 16, 22-34[Medline] [Order article via Infotrieve]
  33. Dyson, H. J., Rance, M., Houghten, R. A., Wright, P. E., and Lerner, R. A. (1988) J. Mol. Biol. 201, 201-217[CrossRef][Medline] [Order article via Infotrieve]
  34. Henry, G. D., and Sykes, B. D. (1994) Methods Enzymol. 239, 515-535[Medline] [Order article via Infotrieve]
  35. Keifer, P. A., Peterkofsky, A., and Wang, G. (2004) Anal. Biochem. 331, 33-39[Medline] [Order article via Infotrieve]
  36. Wang, G., Keifer, P. A., and Peterkofsky, A. (2004) Spectroscopy 18, 257-264
  37. Bruix, M., Parello, M., Herranz, J., Rico, M., and Nieto, J. L. (1990) Biochem. Biophys. Res. Commun. 167, 1009-1014[CrossRef][Medline] [Order article via Infotrieve]
  38. Szilagyi, L. (1995) Prog. NMR Spectrosc. 27, 325-443
  39. Laskowski, R. A., MacArthur, M. W., Moss, D. S., and Thornton, J. M. (1993) J. Appl. Crystallogr. 26, 283-291[CrossRef]
  40. Karplus, P. A. (1997) Protein Sci. 6, 1302-1307[Abstract]
  41. Scholtz, J. M., Qian, H., Robbins, V. H., and Baldwin, R. L. (1993) Biochemistry 32, 9668-9676[CrossRef][Medline] [Order article via Infotrieve]
  42. Baker, E. N., and Hubbard, R. E. (1984) Prog. Biophys. Mol. Biol. 44, 97-179[CrossRef][Medline] [Order article via Infotrieve]
  43. Wang, G., Sondej, M., Garrett, D. S., Peterkofsky, A., and Clore, G. M. (2000) J. Biol. Chem. 275, 16401-16403[Abstract/Free Full Text]
  44. Wang, G., Pierens, G. K., Treleaven, W. D., Sparrow, J. T., and Cushley, R. J. (1996) Biochemistry 35, 10358-10366[CrossRef][Medline] [Order article via Infotrieve]
  45. Wang, G., Treleaven, W. D., and Cushley, R. J. (1996) Biochim. Biophys. Acta 1301, 174-184[Medline] [Order article via Infotrieve]
  46. Ramachandran, G. N., Sasisekharan, V., and Ramakrishnan, C. J. (1963) J. Mol. Biol. 7, 95-99[Medline] [Order article via Infotrieve]
  47. Koradi, R., Billeter, M., and Wüthrich, K. (1996) J. Mol. Graphics 14, 51-55[CrossRef][Medline] [Order article via Infotrieve]
  48. Waddell, W. J. (1956) J. Lab. Clin. Med. 48, 311-314[Medline] [Order article via Infotrieve]
  49. Rizo, J., Blanco, F. J., Kobe, B., Bruch, M. D., and Gierasch, L. M. (1993) Biochemistry 32, 4881-4894[CrossRef][Medline] [Order article via Infotrieve]
  50. Wishart, D. S., Sykes, B. D., and Richards, F. M. (1991) J. Mol. Biol. 222, 311-333[CrossRef][Medline] [Order article via Infotrieve]
  51. Banerjee, U., Zidovetzki, R., Birge, R. R., and Chan, S. I. (1985) Biochemistry 24, 7621-7627[CrossRef][Medline] [Order article via Infotrieve]
  52. Holzwarth, G. M., and Doty, P. (1965) J. Am. Chem. Soc. 87, 218-228
  53. Jackson, R. L., Morrisett, J. D., Pownall, H. J., and Gotto, A. M., Jr. (1973) J. Biol. Chem. 248, 5218-5224[Abstract/Free Full Text]
  54. Prenner, E. J., Lewis, R. N. A. H., Neuman, K. C., Gruner, S. M., Kondejewski, L. H., Hodges, R. S., and McElhaney, R. N. (1997) Biochemistry 36, 7906-7916[CrossRef][Medline] [Order article via Infotrieve]
  55. Bechinger, B., Kinder, R., Helmle, M., Vogt, T. C. B., Harzer, U., and Schinzel, S. (1999) Biopolymers 51, 174-190[CrossRef][Medline] [Order article via Infotrieve]
  56. Seelig, J. (1978) Biochim. Biophys. Acta 515, 105-140[Medline] [Order article via Infotrieve]



This article has been cited by other articles:


Home page
 Biophys. JHome page
Y.-L. Pan, J. T.-J. Cheng, J. Hale, J. Pan, R. E. W. Hancock, and S. K. Straus
Characterization of the Structure and Membrane Interaction of the Antimicrobial Peptides Aurein 2.2 and 2.3 from Australian Southern Bell Frogs
Biophys. J., April 15, 2007; 92(8): 2854 - 2864.
[Abstract] [Full Text] [PDF]

This Article
Right arrow Abstract Freely available