1H-, 13C- and 31P-NMR studies of dioctanoylphosphatidylcholine and dioctanoylthiophosphatidylcholine

Coupling constants and chemical shifts were measured for dioctanoylphosphatidylcholine and its thio analogue in a CDCl3/CD3OD solvent mixture. Replacing the bridging oxygen atom of the CH-CH2-O-P portion of the phosphatidylcholine molecule with a sulfur atom affects chemical shifts and coupling constants in the glycerol backbone portion of the molecule as well as in the choline head group region. Preferred conformations about selected bonds in the phospholipids were determined from the vicinal 1H-1H, 31P-1H and 31P-13C coupling constants. A reduction of the 31P T2* (effective spin-spin relaxation time) for the thio analogue, as well as changes in the relative chemical shifts of 13C nuclei in the acyl chains, suggest a somewhat greater degree of aggregation for the thio analogue. The quadrupolar coupling constant 1J(14N-13C) for the choline methyls of either analogue, however, indicates that aggregation of these phospholipids in the CDCl3/CD3OD solvent mixture is not significant. Differences in conformation between dioctanoylphosphatidylcholine and its thio analogue may be responsible for their differences in chemical and physical properties.     

Measurement of methyl 13C-1H cross-correlation in uniformly 13C-, 15N-, labeled proteins

An understanding of side chain motions in protein is of great interest since side chains often play an important role in protein folding and intermolecular interactions. A novel method for measuring the dynamics of methyl groups in uniformly ¹³C-, ¹⁵N-labeled proteins has been developed by our group. The method relies on the difference in peak intensities of ¹³C quartet components of methyl groups, in a spectrum recording the free evolution of ¹³C under proton coupling in a constant-time period. Cross-correlated relaxation rates between ¹³C-¹H dipoles can be easily measured from the intensities of the multiplet components. The degree of the methyl restrictions (S ²) can be estimated from the cross-correlated relaxation rate. The method is demonstrated on a sample of human fatty acid binding protein in the absence of fatty acid. We obtained relaxation data for 33 out of 46 residues having methyl groups in apo-IFABP. It has been found that the magnitude of the CSA tensor of spin ¹³C in a methyl group could be estimated from the intensities of the ¹³C multiplet components.     

Structural comparison between oxidized and reduced Escherichia coli thioredoxin. Proton NMR and CD studies

Escherichia coli thioredoxin (Mr 11,700) usually functions as a hydrogen carrier protein that undergoes reversible oxidation/reduction reactions of its active-site disulfide linkage. By use of a number of assigned and identified resonances in one- and two-dimensional 1H NMR spectra, the two forms of the protein have been compared. Only groups that are relatively close to the active-site Cys-32, Cys-35 linkage such as Trp-28, Trp-31, Phe-27, Ala-29, and Val-25 undergo substantial changes in their 1H NMR chemical shift upon reduction. Various residues that are further removed from the active site, like Tyr-49, Tyr-70, His-6, Phe-12, Phe-81, and Phe-102, appear to be little affected (less than 0.02 ppm) by the reduction, suggesting that the rest of the protein structure is not much affected. Thus, the structural changes that occur upon reduction appear to be localized to the disulfide-containing turn and the central strand of the twisted beta-sheet that directly leads to this turn. Notwithstanding the apparent similarity in the secondary and tertiary structures of the oxidized and reduced forms of the protein, the thermal stability of the protein decreases by 10 degrees C upon the reduction of the single disulfide. This was found by both 1H NMR and near- and far-ultraviolet circular dichroism studies. Oxidized thioredoxin was also more resistant to alkaline denaturation. Furthermore, the exchange rate of the relatively stable slow-exchanging backbone amide protons that are part of the core of the twisted five-stranded beta-sheet of thioredoxin increases substantially after reduction.(ABSTRACT TRUNCATED AT 250 WORDS)     

Assignment of the 15N NMR spectra of reduced and oxidized Escherichia coli thioredoxin

As a necessary first step in the use of heteronuclear correlated spectra to obtain high resolution solution structures of the protein, assignment of the 15N NMR spectra of reduced and oxidized Escherichia coli thioredoxin (Mr 12,000) uniformly labeled with 15N has been performed. The 15N chemical shifts of backbone amide nitrogen atoms have been determined for both oxidation states of thioredoxin using 15N-1H correlated and two-dimensional heteronuclear single-quantum coherence (HSQC) TOCSY and NOESY spectra. The backbone assignments are complete, except for the proline imide nitrogen resonances and include Gly33, whose amide proton resonance is difficult to observe in homonuclear 1H spectra. The differences in the 15N chemical shift between oxidized and reduced thioredoxin, which occur mainly in the vicinity of the two active site cysteines, including residues distant in the amino acid sequence which form a hydrophobic surface close to the active site, are consistent with the differences observed for proton chemical shifts in earlier work on thioredoxin.     

A glutathione reductase mutant of yeast accumulates high levels of oxidized glutathione and requires thioredoxin for growth.

A glutathione reductase null mutant of Saccharomyces cerevisiae was isolated in a synthetic lethal genetic screen for mutations which confer a requirement for thioredoxin. Yeast mutants that lack glutathione reductase (glr1 delta) accumulate high levels of oxidized glutathione and have a twofold increase in total glutathione. The disulfide form of glutathione increases 200-fold and represents 63% of the total glutathione in a glr1 delta mutant compared with only 6% in wild type. High levels of oxidized glutathione are also observed in a trx1 delta, trx2 delta double mutant (22% of total), in a glr1 delta, trx1 delta double mutant (71% of total), and in a glr1 delta, trx2 delta double mutant (69% of total). Despite the exceptionally high ratio of oxidized/reduced glutathione, the glr1 delta mutant grows with a normal cell cycle. However, either one of the two thioredoxins is essential for growth. Cells lacking both thioredoxins and glutathione reductase are not viable under aerobic conditions and grow poorly anaerobically. In addition, the glr1 delta mutant shows increased sensitivity to the thiol oxidant diamide. The sensitivity to diamide was suppressed by deletion of the TRX2 gene. The genetic analysis of thioredoxin and glutathione reductase in yeast runs counter to previous studies in Escherichia coli and for the first time links thioredoxin with the redox state of glutathione in vivo.     

1H-, 13C-, 31P-NMR studies and conformational analysis of NADP+, NADPH coenzymes and of dimers from electrochemical reduction of NADP+

All H,H, H,P and several C,P coupling constants, including those between C-4′ and the vicinal phosphorus atom, have been determined for NADP⁺, NADPH coenzymes and for a 4,4-dimer obtained from one-electron electrochemical reduction of NADP⁺. From these data the preferred conformation of the ribose, that of the 1,4-dihydronicotinamide rings, and the conformation about bonds C(4′)C(5′) and C(5′)O(5′) were deduced. The preferred form of the 1,4- and 1,6-dihydropyridine rings and the conformation about the ring-ring junction were also obtained for all the other 4,4- and 4,6-dimers formed in the same reduction. All the dimers show a puckered structure, i.e., a boat form for the 1,4- and a twist-boat for the 1,6-dihydronicotinamide ring; both protons at the ring-ring junctions are equatorial and have preferred gauche orientation. On the contrary, the reduced coenzyme NADPH displays a planar or highly flexible conformation, rapidly flipping between two limiting boat structures. The conformation of the ribose rings, already suggested for the NADP coenzymes to be an equilibrium mixture of C(2′)-endo (S-type) and C(3′)-endo (N-type) puckering modes, has been reexamined by using the Altona procedure and the relative proportion of the two modes has been obtained. The S and N families of conformers have almost equal population for the adenine-ribose, whereas for the nicotinamide-ribose rings the S-type reaches the 90%. The rotation about the ester bond C(5′)O(5′) and about C(4′)C(5′), defined by torsion angles β and γ respectively, displays a constant high preference for the trans conformer β^t (75–80%), whereas the rotamers γ are spread out in a range of different populations. The values are distributed between the gauche γ⁺ (48–69%) and the rans γ^t forms (28–73%). The γ⁺ conformer reaches a 90% values in the case of NADP⁺ and NMN⁺. The conformations of the mononucleotides 5′-AMP, NMN⁺ and NMNH were also calculated from the experimental coupling constant values of the literature.     

Structure of oxidized bacteriophage T4 glutaredoxin (thioredoxin). Refinement of native and mutant proteins.

The structure of wild-type bacteriophage T4 glutaredoxin (earlier called thioredoxin) in its oxidized form has been refined in a monoclinic crystal form at 2.0 A resolution to a crystallographic R-factor of 0.209. A mutant T4 glutaredoxin gives orthorhombic crystals of better quality. The structure of this mutant has been solved by molecular replacement methods and refined at 1.45 A to an R-value of 0.175. In this mutant glutaredoxin, the active site residues Val15 and Tyr16 have been substituted by Gly and Pro, respectively, to mimic that of Escherichia coli thioredoxin. The main-chain conformation of the wild-type protein is similar in the two independently determined molecules in the asymmetric unit of the monoclinic crystals. On the other hand, side-chain conformations differ considerably between the two molecules due to heterologous packing interactions in the crystals. The structure of the mutant protein is very similar to the wild-type protein, except at mutated positions and at parts involved in crystal contacts. The active site disulfide bridge between Cys14 and Cys17 is located at the first turn of helix alpha 1. The torsion angles of these residues are similar to those of Escherichia coli thioredoxin. The torsion angle around the S-S bond is smaller than that normally observed for disulfides: 58 degrees, 67 degrees and 67 degrees for wild-type glutaredoxin molecule A and B and mutant glutaredoxin, respectively. Each sulfur atom of the disulfide cysteines in T4 glutaredoxin forms a hydrogen bond to one main-chain nitrogen atom. The active site is shielded from solvent on one side by the beta-carbon atoms of the cysteine residues plus side-chains of residues 7, 9, 21 and 33. From the opposite side, there is a cleft where the sulfur atom of Cys14 is accessible and can be attacked by a nucleophilic thiolate ion in the initial step of the reduction reaction.     

1H-, 13C-, 31P-NMR studies and conformational analysis of NADP+, NADPH coenzymes and of dimers from electrochemical reduction of NADP+.

All H,H, H,P and several C,P coupling constants, including those between C-4' and the vicinal phosphorus atom, have been determined for NADP+, NADPH coenzymes and for a 4,4-dimer obtained from one-electron electrochemical reduction of NADP+. From these data the preferred conformation of the ribose, that of the 1,4-dihydronicotinamide rings, and the conformation about bonds C(4')-C(5') and C(5')-O(5') were deduced. The preferred form of the 1,4- and 1,6-dihydropyridine rings and the conformation about the ring-ring junction were also obtained for all the other 4,4- and 4,6-dimers formed in the same reduction. All the dimers show a puckered structure, i.e., a boat form for the 1,4- and a twist-boat for the 1,6-dihydronicotinamide ring; both protons at the ring-ring junctions are equatorial and have preferred gauche orientation. On the contrary, the reduced coenzyme NADPH displays a planar or highly flexible conformation, rapidly flipping between two limiting boat structures. The conformation of the ribose rings, already suggested for the NADP coenzymes to be an equilibrium mixture of C(2')-endo (S-type) and C(3')-endo (N-type) puckering modes, has been reexamined by using the Altona procedure and the relative proportion of the two modes has been obtained. The S and N families of conformers have almost equal population for the adenine-ribose, whereas for the nicotinamide-ribose rings the S-type reaches the 90%. The rotation about the ester bond C(5')-O(5') and about C(4')-C(5'), defined by torsion angles beta and gamma respectively, displays a constant high preference for the trans conformer beta t (75-80%), whereas the rotamers gamma are spread out in a range of different populations. The values are distributed between the gauche gamma + (48-69%) and the trans gamma t forms (28-73%). The gamma + conformer reaches a 90% value in the case of NADP+ and NMN+. The conformations of the mononucleotides 5'-AMP, NMN+ and NMNH were also calculated from the experimental coupling constant values of the literature.     

Production in two-liter beverage bottles of proteins for NMR structure determination labeled with either 15N- or 13C-15N

The use of 2-L polyethylene terephthalate beverage bottles as a bacterial culture vessel has been recently introduced as an enabling technology for high-throughput structural biology [Sanville Millard, C. et al., 2003. Protein Express. Purif. 29, 311-320]. In the article following this one [Stols et al., this issue, pp. 95-102], this approach was elaborated for selenomethionine labeling used for multiwavelength anomalous dispersion phasing in the X-ray crystallographic determinations of protein structure. Herein, we report an effective and reproducible schedule for uniform 15N- and 13C-labeling of recombinant proteins in 2-L beverage bottles for structural determination by NMR spectroscopy. As an example, three target proteins selected from Arabidopsis thaliana were expressed in Escherichia coli Rosetta (DE3)/pLysS from a T7-based expression vector, purified, and characterized by electrospray ionization mass spectrometry and NMR analysis by 1H-15N heteronuclear single quantum correlation spectroscopy. The results show that expressions in the unlabeled medium provide a suitable control for estimation of the level of production of the labeled protein. Mass spectral characterizations show that the purified proteins contained a level of isotopic incorporation equivalent to the isotopically labeled materials initially present in the growth medium, while NMR analysis of the [U-15N]-labeled proteins provided a convenient method to assess the solution state properties of the target protein prior to production of a more costly double-labeled sample.     

Purification of thioredoxin, thioredoxin reductase, and glutathione reductase by affinity chromatography.

A scheme is described for the large scale purification of thioredoxin, thioredoxin reductase, and glutathione reductase. The scheme is based on an initial separation of thioredoxin from the two reductases by affinity chromatography on agarose-bound N6-(6-aminohexyl)-adenosine 2',5'-bisphosphate (agarose-2',5'-ADP). The two reductases were then separated by hydrophobic chromatography and purified separately to homogeneity. Thioredoxin was purified to homogeneity by immunoadsorption to agarose containing immobilized goat anti-thioredoxin. Overall yields for thioredoxin, thioredoxin reductase, and glutathione reductase exceeded 80% in each case. Both reductases exhibit an absorption band at approximately 320 nm which appears due to a residual amount of tightly bound NADP. Presence of this absorption band has no apparent effect on the specific activity of either enzyme.     

Mapping of trxB, a mutation responsible for reduced thioredoxin reductase activity.

The mutation (trxB) responsible for reduced thioredoxin reductase activity has been mapped on the Escherichia coli K-12 chromosome clockwise from aroA between 20 and 21 min. The gene order in this region of the E. coli chromosome was found to probably be serC-aroA-trxB. The location of gshA, the structural gene for gamma-glutamylcystein synthetase, relative to srl and recA also was determined.     

Determination of solid-state NMR structures of proteins by means of three-dimensional 15N-13C-13C dipolar correlation spectroscopy and chemical shift analysis

In this paper, a three-dimensional (3D) NMR-based approach for the determination of the fold of moderately sized proteins by solid-state magic-angle spinning (MAS) NMR is presented and applied to the alpha-spectrin SH3 domain. This methodology includes the measurement of multiple (13)C-(13)C distance restraints on biosynthetically site-directed (13)C-enriched samples, obtained by growing bacteria on [2-(13)C]glycerol and [1,3-(13)C]glycerol. 3D (15)N-(13)C-(13)C dipolar correlation experiments were applied to resolve overlap of signals, in particular in the region where backbone carbon-carbon correlations of the C(alpha)-C(alpha), CO-CO, C(alpha)-CO, and CO-C(alpha) type appear. Additional restraints for confining the structure were obtained from phi and psi backbone torsion angles of 29 residues derived from C(alpha), C(beta), CO, NH, and H(alpha) chemical shifts. Using both distance and angular restraints, a refined structure was calculated with a backbone root-mean-square deviation of 0.7 A with respect to the average structure.     

Metabolic state of Zymomonas mobilis in glucose-, fructose-, and xylose-fed continuous cultures as analysed by 13C- and 31P-NMR spectroscopy

The reasons for the well-known significantly different behaviour of the anaerobic, gram-negative, ethanologenic bacterium Zymomonas mobilis during growth on fructose (i.e. decreased growth and ethanol yields, increased by-product formation) as compared to that on its second natural substrate, glucose, have remained unexplained. A xylose-fermenting recombinant strain of Z. mobilis that was recently constructed in our laboratory also unexpectedly displayed an increased formation of by-products and a strongly reduced growth rate as compared to the parent strain. Therefore, a comprehensive study employing recently developed NMR-based methods for the in vivo analysis of intracellular phosphorylated pool sizes and metabolic fluxes was undertaken to enable a global characterization of the intracellular metabolic state of Z. mobilis during growth on ¹³C-labelled glucose, fructose and xylose in defined continuous cultures. The ¹³C-NMR flux analysis indicated that ribose 5-phosphate is synthesized via the nonoxidative pentose phosphate pathway in Z. mobilis, and it identified a metabolic bottleneck in the recombinant xylose-fermenting Z. mobilis strain at the level of heterologous xylulokinase. The ³¹P-NMR analyses revealed a global alteration of the levels of intracellular phosphorylated metabolites during growth on fructose as compared to that on glucose. The results suggest that this is primarily caused by an elevated concentration of intracellular fructose 6-phosphate. Received: 7 January 1999 / Accepted: 22 March 1999     

Composition and structure of lipopolysaccharide-human plasma low density lipoprotein complex. Analytical ultracentrifugation, 31P-NMR, ESR and fluorescence spectroscopy studies

Complexes of Salmonella typhimurium lipopolysaccharide toxin (LPS) with low density lipoproteins (LDL) prepared in vitro have been analyzed. LPS-LDL complexes were found to comprise approx. 0.24 mg LPS/mg LDL protein. The major protein of complexes was apolipoprotein apoB-100 (greater than or equal to 90-95%). Incorporation of LPS molecules into LDL was accompanied by small changes in lipid composition, i.e. the phosphatidylcholine content was diminished by approx. 11% and the free fatty acid concentration was raised 2-fold. Analytical ultracentrifugation showed that insertion of LPS into LDL results in the increase of a portion of particles with higher density (lower flotation coefficient) compared to initial LDL. As was evidenced by ESR, in LPS-LDL complexes, the phospholipid hydrocarbon chains are more ordered than in LDL. 31P-NMR spectra indicated that in LPS-LDL complexes the mobility of phospholipid polar headgroups is restricted in comparison with LDL. Application of the shift reagent (Pr3+) revealed that phospholipid molecules form a monolayer structure on the surface of complexes. Upon binding of LPS to LDL, a maximum of the apoB intrinsic fluorescence was slightly red-shifted (1-2 nm) which may testify that the localization of apoB remains nearly unchanged. For LPS-LDL complexes, the accessibility of apoB fluorophores to quenchers (I-, Cs+, acrylamide) did not dramatically differ from that of LDL. It is concluded that rather large amounts of LPS (about 9-10 molecules) can accommodate in one LDL particle without severely perturbing its original composition and structure. Moreover, in the LPS-LDL complexes, oligosaccharide chains of LPS screen notably neither phospholipid polar headgroups nor, what is very important, apoB. LPS-LDL complexes are suggested to be able in vivo to bind to cellular apoB/E receptors, possible LPS receptors and scavenger-receptors of macrophages (monocytes).     

Selenodiglutathione is a highly efficient oxidant of reduced thioredoxin and a substrate for mammalian thioredoxin reductase.

Selenium compounds like selenite (SeO3(2-) may form a covalent adduct with glutathione (GSH) in the form of selenodiglutathione (GS-Se-SG), which is assumed to be important in the metabolism of selenium. We have isolated GS-Se-SG and studied its reactions with NADPH and thioredoxin reductase from calf thymus or with thioredoxin reductase and thioredoxin from Escherichia coli. Incubation of 0.1 microM calf thymus thioredoxin reductase or 0.1 microM thioredoxin reductase and 1 microM thioredoxin from E. coli with 5, 10, or 20 microM GS-Se-SG resulted in a fast initial reaction, followed by a large and continued oxidation of NADPH. However, anaerobic incubation of 0.1 microM calf thymus thioredoxin reductase and 20 microM GS-Se-SG resulted only in oxidation of a stoichiometric amount of NADPH; admission of oxygen started continuous NADPH oxidation. Contrary to the mammalian enzyme, GS-Se-SG was not a substrate for thioredoxin reductase from E. coli. The rate of the oxygen-dependent reaction between calf thymus thioredoxin reductase and GS-Se-SG was increased 2-fold in the presence of 4 mM GSH, indicating that HSe- was the reactive intermediate. Glutathione reductase from rat liver reduced GS-Se-SG with a very slow continued oxidation of NADPH, and the presence of the enzyme did not affect the oxygen-dependent nonstoichiometric oxidation of NADPH by GS-Se-SG and thioredoxin reductase. Fluorescence spectroscopy showed GS-Se-SG to be a very efficient oxidant of reduced thioredoxin from E. coli and kinetically superior to insulin disulfides. Thioredoxin-dependent reduction of CDP to dCDP by ribonucleotide reductase was effectively inhibited by GS-Se-SG.     

31P-NMR and 13C-NMR studies of mannose metabolism in Plesiomonas shigelloides. Toxic effect of mannose on growth.

The metabolism of mannose was examined in resting cells in vivo using 13C-NMR and 31P-NMR spectroscopy, in cell-free extracts in vitro using 31P-NMR spectroscopy, and by enzyme assays. Plesiomonas shigelloides was shown to transport mannose by a phosphoenolpyruvate-dependent phosphotransferase system producing mannose 6-phosphate. However, a toxic effect was observed when P. shigelloides was grown in the presence of mannose. Investigation of mannose metabolism using in vivo 13C NMR showed mannose 6-phosphate accumulation without further metabolism. In contrast, glucose was quickly metabolized under the same conditions to lactate, ethanol, acetate and succinate. Extracts of P. shigelloides exhibited no mannose-6-phosphate isomerase activity whereas the key enzyme of the Embden-Meyerhof pathway (6-phosphofructokinase) was found. This result explains the mannose 6-phosphate accumulation observed in cells grown on mannose. The levels of phosphoenolpyruvate and Pi were estimated by in vivo 31P-NMR spectroscopy. The intracellular concentrations of phosphoenolpyruvate and Pi were relatively constant in both starved cells and mannose-metabolizing cells. In glucose-metabolizing cells, the phosphoenolpyruvate concentration was lower, and about 80% of the Pi was used during the first 10 min. It thus appears that the toxic effect of mannose on growth is not due to energy depletion but probably to a toxic effect of mannose 6-phosphate.     

Structure of the lipopolysaccharide/human plasma low density lipoprotein complex. 31P-NMR and fluorescence spectroscopy studies

Complexes of Salmonella typhimurium lipopolysaccharide toxin (LPS) with low density lipoproteins (LDL) containing various amounts of LPS were prepared in vitro. The 31P-NMR spectra showed that in the LDL-LPS complexes as well as in native LDL all phosphate groups of phospholipids are accessible to the paramagnetic shift reagent, Pr3+. Besides, the low frequency mobility of phospholipid phosphates in the complex is diminished. It was supposed that the phospholipid molecules in the LDL/LPS complex as in native LDL form a monolayer structure on the surface of LDL. The intrinsic fluorescence spectra of tryptophan residues of the apoprotein (apo B-100) revealed that the incorporation of LPS molecules into LDL particles is accompanied by minor changes in the conformation and orientation of the apo B molecule. As a result of these changes, certain fragments become exposed to a more hydrophilic environment and become more accessible to fluorescence quenchers. The use of charged (I-, Cs+) and uncharged (acrylamide) quenchers permitted to identify in the apo B molecule different tryptophan residues, some of which are localized in the vicinity of negatively charged groups, whereas others are neighbouring positively charged groups. It is suggested that the LPS molecules incorporated into LDL particle do not screen the apo B molecule to such an extent that it would hinder the LDL/LPS complex binding to apo B/E cellular receptors.     

Three-dimensional 15N-1H-1H and 15N-13C-1H nuclear-magnetic resonance studies of HPr a central component of the phosphoenolpyruvate-dependent phosphotransferase system from Escherichia coli. Assignment of backbone resonances.

We have performed three-dimensional NMR studies on a central component of the phosphoenolpyruvate-dependent phosphotransferase system of Escherichia coli, denoted as HPr. The protein was uniformly enriched with 15N and 13C to overcome spectral overlap. Complete assignments were obtained for the backbone 1H, 15N and 13C resonances, using three-dimensional heteronuclear 1H NOE 1H-15N multiple-quantum coherence spectroscopy (3D-NOESY-HMQC) and three-dimensional heteronuclear total correlation 1H-15N multiple-quantum coherence spectroscopy (3D-TOCSY-HMQC) experiments on 15N-enriched HPr and an additional three-dimensional triple-resonance 1HN-15N-13C alpha correlation spectroscopy (HNCA) experiment on 13C, 15N-enriched HPr. Many of the sequential backbone 1H assignments, as derived from two-dimensional NMR studies [Klevit, R.E., Drobny, G.P. & Waygood, E.B. (1986) Biochemistry 25, 7760-7769], were corrected. Almost all discrepancies are in the helical regions, leaving the published antiparallel beta-sheet topology almost completely intact.