Determination of the X-ray structure of the snake venom protein omwaprin by total chemical synthesis and racemic protein crystallography

Authors

  • James R. Banigan,

    1. Department of Biochemistry and Molecular Biology, The University of Chicago, Chicago. Illinois
    2. Department of Chemistry, The University of Chicago, Chicago. Illinois
    Search for more papers by this author
  • Kalyaneswar Mandal,

    1. Department of Biochemistry and Molecular Biology, The University of Chicago, Chicago. Illinois
    2. Department of Chemistry, The University of Chicago, Chicago. Illinois
    3. Institute for Biophysical Dynamics, The University of Chicago, Chicago. Illinois
    Search for more papers by this author
  • Michael R. Sawaya,

    1. Molecular Biology Institute, University of California, Los Angeles, California
    2. Department of Chemistry and Biochemistry, University of California Los Angeles, California
    Search for more papers by this author
  • Vilasak Thammavongsa,

    1. Department of Microbiology, The University of Chicago, Chicago. Illinois
    Search for more papers by this author
  • Antoni P. A. Hendrickx,

    1. Department of Microbiology, The University of Chicago, Chicago. Illinois
    2. Department of Medical Microbiology, University Medical Center Utrecht, Utrecht, The Netherlands
    Search for more papers by this author
  • Olaf Schneewind,

    1. Department of Microbiology, The University of Chicago, Chicago. Illinois
    Search for more papers by this author
  • Todd O. Yeates,

    1. Molecular Biology Institute, University of California, Los Angeles, California
    2. Department of Chemistry and Biochemistry, University of California Los Angeles, California
    3. University of California Los Angeles-United States Department of Energy Institute for Genomics and Proteomics
    Search for more papers by this author
  • Stephen B. H. Kent

    Corresponding author
    1. Department of Biochemistry and Molecular Biology, The University of Chicago, Chicago. Illinois
    2. Department of Chemistry, The University of Chicago, Chicago. Illinois
    3. Institute for Biophysical Dynamics, The University of Chicago, Chicago. Illinois
    • GCIS #W204, 929 East 57th Street, Chicago, IL 60637

    Search for more papers by this author

Abstract

The 50-residue snake venom protein L-omwaprin and its enantiomer D-omwaprin were prepared by total chemical synthesis. Radial diffusion assays were performed against Bacillus megaterium and Bacillus anthracis; both L- and D-omwaprin showed antibacterial activity against B. megaterium. The native protein enantiomer, made of L-amino acids, failed to crystallize readily. However, when a racemic mixture containing equal amounts of L- and D-omwaprin was used, diffraction quality crystals were obtained. The racemic protein sample crystallized in the centrosymmetric space group P21/c and its structure was determined at atomic resolution (1.33 Å) by a combination of Patterson and direct methods based on the strong scattering from the sulfur atoms in the eight cysteine residues per protein. Racemic crystallography once again proved to be a valuable method for obtaining crystals of recalcitrant proteins and for determining high-resolution X-ray structures by direct methods.

Introduction

Omwaprin is a small protein recently isolated and characterized from the venom of the inland taipan (Oxyuranus microlepidotus).1 It has been shown to have dose-dependent antimicrobial activity against a selected range of Gram-positive bacteria, such as Bacillus megaterium and Staph warneri, and is nontoxic in mice at doses up to 10 mg/kg, when injected intraperitoneally. A predicted structure of omwaprin has been suggested based on homology modeling1 but no nuclear magnetic resonance (NMR) or X-ray crystal structure has been reported for this protein.

We set out to determine the X-ray crystal structure of omwaprin. We have recently shown that protein crystals formed more readily from a racemic mixture (containing equal proportion of L- and D-protein enantiomers) resulting in centrosymmetric crystals.2, 3 It has been shown that these centrosymmetric crystals facilitate ab initio protein structure solution via direct methods for a combination of reasons. The crystals tend to be well ordered and thus give high resolution diffraction data, and the phases of the X-ray reflections are all quantized (related by 0, π radians).3, 4 Some of the potential advantages that have been observed for phasing and structure interpretation were anticipated by early studies on racemic macromolecules.5, 6

To obtain a racemic protein crystal, it is necessary to make the D-protein (the enantiomer of the L-protein); this can only be done by total chemical synthesis of the protein.7–9 With the use of modern synthetic methods based on native chemical ligation, synthesis of mirror image D-proteins is more practical than it was just a short time ago.10, 11 In this article, we discuss the total chemical synthesis of L- and D-omwaprin and the determination of the X-ray crystal structure of omwaprin (at 1.33 Å) by direct methods. We also report antimicrobial assays for both L- and D-omwaprin.

Results

Chemical synthesis of L- and D-omwaprin

Omwaprin is a protein of 50 amino acid residues with eight cysteines, which form four disulfide bonds in the folded state.1 Its small size and large number of cysteine residues make omwaprin ideal for total synthesis using native chemical ligation, which involves the reaction of a peptide thioester with a cysteine–peptide to form a product covalently linked by a native peptide bond.10 The target amino acid sequence and the synthetic strategy for the preparation of omwaprin are shown in Scheme 1.

Scheme 1.

Omwaprin synthetic design. A: Amino acid sequence and disulfide bond connectivity of omwaprin.1 B: Convergent synthetic strategy: native chemical ligation of two unprotected peptide segments.10

The peptide thioester and the Cys-peptide segments were each prepared from either L- or D-amino acids by stepwise solid phase peptide synthesis using Boc chemistry “in situ neutralization” protocols.12 After cleavage from the resin with concomitant deprotection, the purified peptides were ligated in 6M guanidinium hydrochloride in the presence of 100 mM 4-mercaptophenylacetic acid (MPAA), 100 mM mercaptoethanesulfonic acid Na salt (MESNA), and 30 mM triscarboxyethylphosphine HCl salt (TCEP·HCl) at pH 6.9. The MESNA was used to help combat thiolactone formation involving an internal cysteine in the peptide thioester segment. The ligated product was subsequently purified using preparative high pressure liquid chromatography (HPLC). Analytical data for the ligation reactions are shown in Supporting Information Figure S-3.

Folding was optimized using the 50 residue L-polypeptide. The lyophilized, purified product was folded first using an aqueous buffer containing 1M guanidinium hydrochloride and a redox buffer made up of 8 mM cysteine and 1 mM cystine hydrochloride at a pH of 7.9, with a peptide concentration of 1 mg/mL. This method resulted in some product formation within 1 day, but large amounts of peptide–cysteine adducts were formed (data not shown). Next, folding was attempted under similar conditions with a redox buffer made up of 3 mM reduced glutathione and 0.3 mM oxidized glutathione. This method of folding produced similar results to the cysteine/cystine folding, with a major product being the formation of adducts between glutathione and the starting polypeptide chain. Finally, air oxidation was used in 0.1M ammonium bicarbonate buffer pH 8.1 at a concentration of 0.5 mg/mL peptide. The folding was essentially completed within 3 days, as shown in Figure 1. The major component from the air oxidation folding reaction was then purified by preparative reverse phase HPLC and used for further study. D-Omwaprin was made in essentially similar fashion (see Supporting Information Fig. S-5).

Figure 1.

Folding of L-omwaprin. A: Air oxidation of L-linear peptide to form L-omwaprin. B: Day 5 of air oxidation with the extracted ion chromatogram (EIC) for the 1401.5 [M + 4H]+4 ion, corresponding to the mass of the folded protein with four disulfide bonds. Analytical HPLC used a Vydac C18, 300 Å, 5 μm, 4.6 × 150 mm column with a 2% B per min gradient over 20 min starting at 1% Solvent B, 99% Solvent A and ending at 41% Solvent B, 59% Solvent A. Solvent A is water + 0.1% TFA; Solvent B is acetonitrile + 0.08% TFA.

Characterization

L-Omwaprin had an observed mass of 5601.4 ± 1.0 Da, and D-omwaprin had an observed mass of 5601.3 ± 0.2 Da (calc. 5602.5 Da (average isotopes)). The protein enantiomers had identical elution times as determined by LC-MS analysis [Fig. 2(A)]. The circular dichroism (CD) spectra of D- and L-omwaprin were equal in magnitude but opposite in sign, as shown in Figure 2(B).

Figure 2.

Characterization of synthetic omwaprins: A: Analytical LC-MS of D-omwaprin, L-omwaprin, and cochromatography of {D-omwaprin + L-omwaprin}, using a self-packed C18, 2.1 × 50 mm, 300 Å, 3 μm Microsorb column, and a 2% B per min gradient over 20 min (1% B, 99% A to 41%B, 59%A). Solvent A is water + 0.1% TFA, solvent B is acetonitrile + 0.08% TFA; B: CD spectra of D-omwaprin (upper curve) and L-omwaprin (lower curve); each protein was separately dissolved at 50 μM in MilliQ water; a 0.1 cm path length cuvette was used.

Crystallization

Crystallization of L-omwaprin was attempted using a screen of 96 conditions from the Hampton Index (HR2-144). Purified L-omwaprin was dissolved in water to a concentration of 35 mg/mL. Each condition used 1 mL of well solution with hanging drops consisting of 1 μL well solution and 1 μL protein solution. The plates were kept in an incubator at 19°C and monitored over the course of days to weeks. This screen yielded no crystals even after 4 months. However, using the same 96 conditions, a racemic mixture of {D-omwaprin + L-omwaprin} at 25 mg/mL (12.5 mg/mL of L-omwaprin + 12.5 mg/mL of D-omwaprin) produced crystals from eight conditions in as little as 3 days. We optimized four of these conditions by varying precipitant concentrations. The crystal that led to the best diffraction results [Fig. 3(A)] was grown from 61% Tacsimate®,* pH 7.0, with a drop size of 4 μL: 2 μL from the well solution + 2 μL of 25 mg/mL racemic protein solution.

Figure 3.

Racemic crystallography of {D-omwaprin + L-omwaprin}. A: Racemic omwaprin crystal in a loop; B: Unit cell of racemic omwaprin in P21/c in cartoon representation; L-omwaprin is red, D-omwaprin is cyan. C: Cartoon representation of L-omwaprin (PDB code 3NGG). D: Initial electron density map calculated from observed structure factors and phases from 404 atoms obtained by direct methods, contoured at 1.2σ. The sticks correspond to the final model and indicate the quality of the initial electron density map. E: 2FoFc electron density map at a σ level of 2 for Asn-36, Tyr-37, and Gly-38.

X-ray crystal structure determination

Diffraction data were collected at an energy of 12.6 keV using the NE-CAT beamline 24-ID at the Advanced Photon Source at Argonne National Laboratory. Diffraction was observed to a resolution of 1.15 Å, and a complete data set was collected to a resolution of 1.33 Å (Supporting Information Table S-1). Data reduction statistics revealed that the protein racemate had crystallized in the P21/c space group with unit cell a = 44.09 Å, b = 46.56 Å, c = 39.65 Å, and β = 92.25°. A consideration of the volume of the asymmetric unit13 suggested that there were likely two protein molecules per asymmetric unit and eight molecules per unit cell. The solvent content was calculated as 32%, which falls toward the lower end of the normal range for protein crystals.

To solve the structure of omwaprin, we attempted molecular replacement using the predicted structure reported by Nair et al.,1 as a search model, but this proved to be unsuccessful. We also attempted to use the NMR structure of nawaprin (PDB ID: 1UDK) and the X-ray crystal structure of elafin (PDB ID: 1FLE) as search models because of the homologous amino acid sequences of these two proteins, but those models also proved ineffective for solving the structure by molecular replacement methods. Finally, the structure was successfully solved by direct methods using ShelxD. Trial structures were seeded from a list of Patterson peaks based on the expectation that these peaks resulted from the significantly stronger scattering of the sulfur atoms contributed by 16 cysteines in the asymmetric unit. Subsequent cycles of automated atom picking led to an essentially complete structure of the protein with statistics: Rwork = 0.174 and Rfree = 0.209. A cartoon representation of L-omwaprin is shown in Figure 3(C).

Searches of the RCSB Protein Data Bank using the programs MATRAS14 and DALI15 reported that omwaprin is closest in structure to the 50 amino acid residue C-terminal WAP domain of the human protein secretory leukocyte protease inhibitor (SLPI) (PDB ID 2Z7F).16 This domain of SLPI has eight Cys residues that align well with the Cys residues of omwaprin, suggesting a similar fold.1 We found that the tertiary structures of omwaprin and the WAP domain of SLPI are similar, with an RMS deviation of 1.19 Å over 36 aligned alpha carbons. Shared features include a pair of central antiparallel strands, a single turn of helix, and four disulfide bonds in equivalent locations. The largest deviations between the two proteins involve a loop preceding the first helix (between residues 10 and 20 of omwaprin) which in SLPI contains four additional residues. Despite the similarity in tertiary structure to SLPI, omwaprin itself is reported to have no inhibitory activity toward proteases.1

Omwaprin inhibits bacterial growth

L-Omwaprin has previously been shown to inhibit the growth of the soil bacterium B. megaterium.1 Consistent with these reports, we found that chemically synthesized L-omwaprin is also able to inhibit growth of B. megaterium; moreover, D-omwaprin also exhibited inhibitory activity toward this bacterium. Both the L- and D-forms of omwaprin failed to inhibit growth of Bacillus anthracis, thus showing biological specificity toward different microorganisms. To examine whether inhibition of bacterial growth was due to omwaprin-dependent cell wall destabilization,1 we treated B. megaterium with 5 μg/μL of either L- or D-omwaprin and analyzed the cells under scanning electron microscopy (SEM). Cells that were treated with buffer alone showed healthy, rod-shaped bacilli with numerous pili appendages. However, when the cells were incubated with L-omwaprin, smoothening of the cells was observed along with the disappearance of pili. An even more dramatic difference in cellular structure was seen in the presence of D-omwaprin compared with buffer alone. D-Omwaprin–treated cells seem to have lost all cellular integrity and resembled lysed cells releasing their intracellular contents. Morphological differences in cells treated with either L- or D-omwaprin may indicate differences in the molecular mechanism of bacterial inhibition or may reflect differing stabilities of the L- and D-protein molecules under assay conditions. The inhibition assay images and SEM images are shown in Supporting Information Figures S-6 and S-7, respectively.

Discussion

The protein enantiomers L- and D-omwaprin were prepared by total chemical synthesis using modern ligation methods.11 High yield total chemical synthesis of the 50 residue polypeptide was achieved using native chemical ligation of two unprotected synthetic peptides. There was significant difficulty in folding/disulfide bond formation. Folding in either cysteine/cystine buffer or glutathione (reduced/oxidized) buffer resulted in a major product being the formation of adduct in which two cysteine or two glutathione molecules were covalently attached to the polypeptide chain, while only two disulfide bonds were formed. Folding via air oxidation, as shown in Figure 1(A), was moderately efficient giving product with four disulfide bonds over the course of a few days; two isomeric products of the same (correct) mass were formed, one dominant and the other minor [Fig. 1(B)]. The major peak of the extracted ion chromatogram corresponded to the major peak in the UV chromatogram. This major product was isolated for structure determination.

Attempted crystallization of L-omwaprin failed in a standard initial screen of 96 conditions. However, when a racemic protein mixture consisting of L-omwaprin + D-omwaprin in equal amounts was used, protein crystals formed within a few days under 8 of the same 96 conditions. This is further evidence of facile crystal formation from racemic protein mixtures,2, 3 supporting an earlier prediction of such an advantage.17 It was also predicted that for racemic protein crystals, the space group, P1, would be the one most frequently observed17; crystallographic experiments by us and by others have so far confirmed this prediction.2, 4, 8, 18–20 The present study adds another example of a racemic protein crystal that is not in P1 but in another space group, P21/c.3 The number of proteins that have been crystallized in racemic form is still small, so an accurate statistical analysis of space groups is not possible yet. Structures of additional racemic proteins should begin to reveal additional trends regarding their crystallization behavior.

Racemic crystals of omwaprin were sufficiently well ordered to provide a complete data set to a resolution of 1.33 Å. This high resolution, combined with the presence of an unusually large number of sulfur atoms, enabled a structure solution by direct methods. In this case, the 16 sulfur atoms in the asymmetric unit provided a starting point for ab initio phasing based on their high X-ray scattering magnitudes, without relying on their potential anomalous scattering behavior. The successful location of the sulfur atoms in ShelxD enabled the location of the remaining protein atoms by an iterative dual space (i.e., real and reciprocal space) procedure. The method was powerful enough to allow the placement of 732 nonhydrogen protein atoms in the asymmetric unit. The contribution of the sulfur atoms to the ab initio structure determination of omwaprin parallels an earlier experiment in which four iron atoms contributed to the ab initio structure determination of Cyt c3.21 Because the omwaprin structure determination relied on the stronger scattering of multiple sulfur atoms, we examined retrospectively whether phases calculated from the sulfur atoms alone would have allowed an interpretation of electron density maps. On the contrary, the uninterpretable nature of those electron density maps (not shown) confirmed that the dual space procedure used was required. The potential importance of the reciprocal space component of the procedure in the present case is further suggested by the relatively low solvent content of the crystal, a feature that might be expected to weaken the power of the real space component.

Crystal structures of proteins larger than omwaprin have rarely been solved by direct methods in the absence of atoms heavier than sulfur. Examples include a scorpion toxin (518 nonhydrogen protein atoms including eight sulfur atoms; space group P212121; 0.96 Å resolution),22 hen egg white lysozyme (1001 nonhydrogen protein atoms including 10 sulfur atoms; space group P1; 0.85 Å resolution),23 and feglymycin (863 nonhydrogen protein atoms with no sulfur atoms; space group P65; 1.1 Å resolution).24 Note, the resolution limits of these data sets were 1.1 Å or better. The ability to solve the structure of omwaprin at lower resolution (1.33 Å) was likely aided by the centrosymmetric nature of the crystal formed by the racemic protein; the phases for X-ray reflections are restricted (e.g., to 0° or 180°) in centrosymmetric space groups.

The present structure was solved using only native diffraction data at a single wavelength without making use of anomalous scattering information. Using anomalous scattering information from native sulfur atoms would make it possible to solve even more challenging structures. Although anomalous scattering differences (i.e., between Friedel mates) are absent in X-ray diffraction of crystals in centrosymmetric space groups like P21/c, dispersive differences could in principle be used to obtain accurate phase information in cases where the resolution of the diffraction data might not be high enough to proceed by ordinary direct methods.25, 26 For successful structure determination using only dispersive differences, it might be expected that the smallest ratio of anomalous scattering atoms per protein amino acid residue necessary would be about the same as the ratio required to successfully use single wavelength anomalous dispersion data: that is, approximately one selenomethionine per 50 residues, with data extending to 1.3 Å resolution.27 Higher ratios would be required for lower resolution data. The extrapolation of this rule to dispersive data is reasonable as the dispersive signal measured between data sets collected at inflection and high remote wavelengths is comparable in strength to the anomalous signal. For example, the difference in f′ between the inflection and high energy remote of selenium (f12660eVf12760eV = 4.1 electrons) is only moderately smaller than twice the value of f″ at the peak wavelength (f12662eV =3.8 electrons); note that the maximum difference between structure factor magnitudes of Friedel pairs is two times f″. Furthermore, taking into consideration that the 1 per 50 rule is derived from observations made in noncentrosymmetric protein crystals, it might be expected that even fewer anomalous scatterers would be required for phasing in a centrosymmetric space group as the values of the phases are related by 0 or π radians.

In the present case, the asymmetric unit has two molecules of omwaprin containing a total of 16 sulfur atoms in 100 amino acid residues. The asymmetric unit can be chosen to contain one L molecule and one D molecule or two of either enantiomer. As the crystal has a glide plane, any L molecule can be related by a crystallographic operation to a D molecule, so one could choose an asymmetric unit having 1L + 1D, 2L, or 2D [It is interesting to see how these racemic crystals bring up issues we (and others) have not previously thought about]. We chose the asymmetric unit containing two L molecules. These were related by a local noncrystallographic twofold rotation symmetry axis nearly aligned with the b-axis. The unit cell contains eight molecules—4 L- and 4 D-omwaprin—as shown in Figure 3(B). Within the unit cell, there are three interfaces where L- and D-protein molecules interact. For reasons of symmetry, at interfaces that include an inversion center, each interacting element is obligately doubled. Thus, at one interface, there are two direct hydrogen bonds between backbone atoms from Val-49 in the L-enantiomer and Val-49 in the D-enantiomer. At another interface, there are two direct hydrogen bonds between the residues, Asp-26 in one enantiomer and Ser-27 in the other enantiomer. At a third interface, there are several water-mediated hydrogen bonds between the enantiomers and one direct hydrogen bond between a lysine and a backbone carbonyl. These interactions are shown in Supporting Information Figure S-1.

The experimentally determined structure of omwaprin is significantly different from the predicted structure of omwaprin as published by Nair et al. (Fig. 4).1 Although the two structures share similar shapes, they differ significantly at the atomic level. Superimposing the two structures resulted in an average root mean square deviation (RMSD) over Cαs of 5.46 Å. The deviations for each pair of Cαs can be found in Supporting Information Graph 1 and Table S-2. Intriguingly, the snake venom–derived omwaprin molecule has a fold that is closely similar to that of the C-terminal WAP domain of the human protein SLPI. Given the divergent origins of these two proteins and the lack of common protease inhibitory activity, the significance of this similarity at a molecular level is not obvious.

Figure 4.

Superposition of predicted1 and experimental structures of omwaprin in walleyed stereo. L-Omwaprin cartoon structure (green) is superimposed on the predicted omwaprin structure (magenta); disulfides are highlighted in yellow.

The observed biological activity of L-omwaprin was what was to be expected based on the original work on omwaprin by Kini and coworkers. The observed biological activity of D-omwaprin suggests an achiral mechanism of action. As stated by Kini and coworkers1 in the original work, SEM images show that L-omwaprin works by membrane disruption. Based on the SEM images of D-omwaprin (Supporting Information Fig. S-7), it works in a similar fashion, although the exact mechanism of action for each of the proteins remains unknown.

Conclusions

We have reported here the total chemical synthesis of omwaprin and the determination of its X-ray structure by racemic protein crystallography. High yield chemical synthesis of omwaprin was achieved using native chemical ligation of two unprotected peptide segments, followed by air oxidation to give the biologically active protein molecule containing four disulfide bonds. L-Omwaprin was difficult to crystallize, but crystals formed readily from a racemic mixture of {D-omwaprin + L-omwaprin}. The highly ordered crystals gave a complete X-ray diffraction data set to 1.33 Å; this enabled direct methods to be used to solve the structure of omwaprin. These results provide one further example of racemic crystallography as a viable alternative for obtaining crystals of proteins which are recalcitrant to crystallization.

Materials and Methods

Reagents

2-(1H-Benzotriazol-1-yl)-1,1,3,3-tetramethyluronium hexafluorophosphate and Nα-Boc protected amino acids (Peptide Institute, Osaka) were obtained from Peptides International. Side chain protecting groups used were Arg(Tos), Asn(Xan), Asp(OcHex), Cys(4-CH3Bzl), Glu(OcHex), Lys(2Cl-Z), Ser(Bzl), and Tyr(2Br-Z). Aminomethyl-resin was prepared from Biobeads S-X1 (BioRad, CA).28 Boc-L-Ala-OCH2-phenylacetic acid and Boc-Gly-OCH2-phenylacetic acid were purchased from NeoMPS, Strasbourg France. N,N-Diisopropylethylamine was obtained from Applied Biosystems. N,N-Dimethylformamide and dichloromethane were purchased from Burdick and Jackson. Diethyl ether and guanidine hydrochloride were purchased from Fisher. HPLC-grade acetonitrile was purchased from OmniSolv. Trifluoroacetic acid (TFA) was obtained from Halocarbon Products (NJ). Anhydrous hydrogen fluoride (HF) was purchased from Matheson. All other reagents were analytical reagent grade and were purchased from Sigma–Aldrich.

Peptide synthesis

The L- and D-amino acid Cys28–Gly50 peptide segments were synthesized on Boc-Gly-OCH2-Pam-resin, using manual in situ neutralization Boc chemistry protocols for stepwise solid phase peptide synthesis (SPPS),12 at 0.25 and 0.4 mmol scale for the L-peptide and 0.2 mmol scale for the D-peptide. The L- and D-amino acid Lys1–Ser27 peptide thioester segments were synthesized on trityl-SCH2CH2CO-Ala-OCH2-Pam-resin at 0.25 and 0.4 mmol scale for the L-peptide thioester and at 0.2 mmol scale for the D-peptide thioester. After removal of the Nα-Boc group, peptides were cleaved from the resin and simultaneously deprotected by treatment with anhydrous HF + 5% v/v p-cresol at 0°C for 1 h. Crude peptides were precipitated and triturated with chilled diethyl ether, then dissolved in water:acetonitrile (1:1; v/v) containing 0.1% TFA.

HPLC and LC-MS analysis

LC-MS data for the purified synthetic peptides is shown in the Supporting Information Fig. S-2. HPLC and LC-MS used linear gradients of decreasing percentage of Solvent A (MilliQ water with 0.1% TFA) and increasing percentage of Solvent B (acetonitrile with 0.08% TFA). Observed masses and yields of the synthetic peptide segments were: L-(1-27)-αCOSCH2CH2Ala-COOH 90.4 mg, 11% (based on 0.25 mmol starting aminoacyl-resin) and 187.6 mg, 14% (based on 0.4 mmol starting aminoacyl-resin); mass obs. 3224.0 ± 1.1 Da, calc. 3223.8 Da (av isotopes). L-(28-50)-COOH 60.3 mg, 9.4% (based on 0.25 mmol starting aminoacyl-resin) and 185.2 mg, 18% (based on 0.4 mmol starting aminoacyl-resin); mass obs. 2563.8 ± 0.2 Da, calc. 2563.9 Da (av isotopes). D-(1-27)-αCOSCH2CH2Ala-COOH 92.2 mg, 14.2% (based on 0.2 mmol starting aminoacyl-resin); mass obs. 3225.1 ± 0.7 Da, calc. 3223.8 Da (av isotopes). D-(28-50)-COOH 51.3 mg, 10% (based on 0.2 mmol starting aminoacyl-resin); mass obs. 2564.8 ± 0.2 Da, calc. 2563.9 Da (av isotopes).

Chemical synthesis of L-omwaprin

L-Omwaprin was prepared by native chemical ligation of two unprotected peptide segments, purification of the product 50 residue polypeptide, followed by folding/disulfide bond formation. L-(1-27)-αCOSCH2CH2 Ala-COOH (30 mg, 9.3 μmol) and L-(28-50)-COOH (15.6 mg, 6.1 μmol) were dissolved in ligation buffer (3.185 mL) containing 6M guanidine hydrochloride, 200 mM Na2HPO4, 30 mM TCEP hydrochloride, 100 mM MPAA, and 100 mM MESNA. The pH was adjusted to 6.6. The reaction mixture was purged with helium for 10 min and sealed. After completion of the reaction (48 h, LC-MS monitoring), the ligation solution was acidified with dilute HCl to a pH of 2 and purified by prep-HPLC using a C18 semi-prep column and isocratic gradient of 0.25% Solvent B (acetonitrile + 0.08% TFA) per minute. Fractions were analyzed by LC and LC-MS and those fractions containing purified peptide were combined and lyophilized. Additional ligations were performed following the same procedure using limiting C-terminal peptide amounts of: 18.7 mg; 23.4 mg; 35.1 mg; and 97.9 mg.

Purified linear (1–50) polypeptide was then dissolved in 0.1M NH4HCO3 at a concentration of 0.5 mg/mL. The pH was adjusted using NaOH(aq) to 8.13 and the plastic container was allowed to sit. The folding was essentially complete after 5 days at room temperature, as determined by analytical LC-MS. The reaction mixture was acidified to pH 2 with dilute HCl and the folded protein product was purified by reverse phase HPLC. The overall yield of the purified L-omwaprin was 24.4 mg (7.2% based on limiting peptide). Mass obs. 5601.4 ± 1.0 Da, calc. 5602.5 Da (av isotopes).

Chemical synthesis of D-omwaprin

For the synthesis of D-omwaprin, the same procedure was followed as used for the chemical synthesis of L-omwaprin. The ligation had a limiting C-terminal at 51.3 mg (20 μmol). The overall yield after chemical synthesis of the purified D-omwaprin was 11.3 mg (10.1%). Mass obs. 5601.3 ± 0.2 Da, calc. 5602.5 Da (av isotopes).

Circular dichroism

CD spectra were recorded on an Aviv CD spectrometer, model 202. L-Omwaprin (50 μM) and D-omwaprin (50 μM) in water were separately transferred to CD cuvettes of path length 0.1 cm. CD spectra were measured at room temperature (25°C) over the range 190–260 nm with 1 nm steps with averaging of five scans. Molar elipticities were calculated by equation image, where [θ]obs is the observed elipticity, Mw is the molecular weight of omwaprin, c is the concentration used, l is the path length, and Nr is the number of residues.

Crystallization

Crystallization of racemic omwaprin was performed by mixing equal amounts (by weight) of purified lyophilized L- and D-omwaprin in water at the following concentrations: 25 mg/mL (2 mg of L + 2 mg of D in 160 μL) and 35 mg/mL (3 mg of L and 3 mg of D in 170 μL). Each solution was centrifuged to remove any particulate matter present and then used directly for crystallizations. Crystallization screening was conducted at 19°C using the commercially available Hampton Index (HR2-144). Crystallization screens were performed by the hanging drop vapor diffusion method using 2 μL drops consisting of 1 μL of well solution and 1 μL of protein solution. Wells contained 1 mL of the desired condition. X-ray diffraction data were collected from crystals grown at 25 mg/mL protein concentration in (1) 56–64% Tacsimate, pH 7.0; (2) 1.0M succinic acid, pH 7.0, 0.1M HEPES, pH 7.0, 1% w/v polyethylene glycol monomethyl ether 2000; (3) 0.1M HEPES, pH 7.5, 1.9M ammonium sulfate; and (4) 0.1M HEPES, pH 7.5, 0.6–1.3M sodium citrate tribasic dehydrate. Diffraction data obtained from a crystal grown in 61% Tacsimate and 4 μL drop size (2 μL protein solution and 2 μL well solution) was used for structure solution.

Data collection

For data collection, selected crystals were briefly transferred to the cryoprotectant [reservoir solution plus 20% (v/v) glycerol] and flash frozen in liquid nitrogen. The X-ray diffraction data were collected at 100 K using 0.9795-Å wavelength synchrotron radiation at the Argonne National Laboratory (Advanced Photon Source, beamline 24-ID C & E). Crystal diffraction images were integrated, scaled, and merged with HKL2000.29 Data were reduced in Laue symmetry P2/m, and a subsequent examination of systematic absences indicated that the automatically chosen unit cell corresponded to the nonstandard space group setting P21/a. The final data set was therefore reindexed to correspond to the standard space group setting P21/c to facilitate subsequent structure determination. Cell dimensions were (a) 44.1 Å, (b) 46.6 Å, (c) 39.7 Å, (α) 90°, (β) 92.8°, and (γ) 90°. The calculated total solvent content of the unit cell was 32%.

X-ray structure determination

The structure of racemic omwaprin in the P21/c space group was solved by direct methods using SHELXD.30, 31 The Patterson seeding algorithm was used with instructions to find 16 sulfur atoms participating in disulfide bonds. All data to 1.33 Å resolution were used. Three thousand trial structures were generated. The 1828th trial was the best with a correlation coefficient of 45.88. Coordinates for 404 of 732 protein atoms (nonhydrogen) were obtained by SHELXD. The electron density map was of excellent quality, allowing unambiguous tracing of the two enantiomeric omwaprin molecules in the asymmetric unit with the program COOT32 [Fig. 3(D)].

The omwaprin model was refined with REFMAC5.33 After each refinement step, the model was visually inspected in COOT using both 2FoFc and FoFc difference maps. All hydrogen atoms connected to carbon atoms and backbone nitrogen atoms were included at their geometrically calculated positions and refined using a riding model. All models were validated with the following structure validation tools: PROCHECK,34 ERRAT,35 and VERIFY3D.36 PROCHECK reported that 90% of the residues are in the most favored region of the Ramachandran plot, 10% of the residues are in additionally allowed regions. No residues were found in the generously allowed or in the disallowed regions of the Ramachandran plot. ERRAT reported that 85% of the residues were within the 95% confidence limit for rejection. The coordinates of the final model and the merged structure factors have been deposited in the RCSB Protein Data Bank with the code 3NGG.

Inhibition of bacterial growth

Radial diffusion assays were performed using midexponential-phase bacteria from overnight cultures (D620 = 0.4), which were centrifuged at 900g for 10 min, washed, and resuspended in 10 mM sodium phosphate buffer (pH 7.4) at 4°C. The resuspended bacteria (∼2 × 106 cfu/mL) were added to autoclaved warm (40–50°C) half-strength Luria-Bertani (LB) medium with 10 mg/mL low-melting point agarose prepared in 10 mM sodium phosphate buffer, pH 7.4. After mixing, the agarose-bacteria gel was poured into sterile 90 × 15 mm2 Petri dishes to form a uniform layer of 2-mm depth. After the agarose had solidified (∼20 min), 10 μL of either buffer alone or omwaprin (0–5 μg/μL) or ampicillin (0.1 μg/μL) was spotted onto the agar. The plates were then incubated for 5 h at 30° and images taken.

Scanning electron microscopy

Bacterial cells were fixed for 30 min in 2% glutaraldehyde at room temperature and postfixed for 30 min in glutaraldehyde onto poly-L-lysine–coated glass coverslips. Samples were washed twice with phosphate buffered saline (PBS) and subsequently serially dehydrated by successive incubations in 25 and 50% ethanol/PBS, 75 and 90% ethanol/H2O, 2× 100% ethanol, followed by 50% ethanol/hexamethyldisilazane (HDMS) and 100% HDMS. After overnight evaporation of HDMS, samples were mounted and sputter coated with 80% Pt/20% Pd to 8 nm using a Cressington 208HR Sputter Coater at 20 mA prior to examination with a Fei Nova NanoSEM 200 scanning electron microscope (FEI Co., Hillsboro, OR).

Acknowledgements

The authors thank Prof. Kini and coworkers, National University of Singapore, for providing the coordinates of the predicted omwaprin structure. They also thank Dr. Kay Perry, Dr. Kanagalaghatta Rajashankar, and Dr. Vladimir Torbeev for useful discussions.

  • *

    Tacsimate is a solution of 1.8305M malonic acid, 0.25M ammonium citrate tribasic, 0.12M succinic acid, 0.3M DL-malic acid, 0.4M sodium acetate trihydrate, 0.5M sodium formate, and 0.16M ammonium tartrate dibasic.

Ancillary