Characterization of oligosaccharides from an antigenic mannan of Saccharomyces cerevisiae.

Mannans of the yeast Saccharomyces cerevisiae have been implicated as containing the allergens to which bakers and brewers are sensitive and also the antigen recognized by patients with Crohn's disease. A fraction of S. cerevisiae mannan, Sc500, having high affinity for antibodies in Crohn's patients has been characterized by NMR spectroscopy followed by fragmentation using alkaline elimination, partial acid hydrolysis and acetolysis. The released oligosaccharides were separated by gel filtration on a Biogel P4 column and analyzed by fluorescence labeling, HPLC and methylation analysis. The relationship between structure and antigen activity was measured by competitive ELISA. The antigenic activity of the original high molecular weight mannan could be ascribed to terminal Manalpha1-->3Manalpha1-->2 sequences which are rarely found in human glycoproteins but were over-represented in Sc500 compared to other yeast mannans.

Analysis of phosphorylated metabolites in crayfish extracts by two-dimensional 1H-31P NMR heteronuclear total correlation spectroscopy (heteroTOCSY).

In vivo and extract analyses by one-dimensional 31P NMR have been a key tool in investigating energy-related metabolism. Although many phosphorylated metabolites have been observed, many of them have yet to be identified. This reflects the difficulty in identifying them using 31P NMR alone. Two-dimensional 1H-31P correlation experiments have been shown to be useful for assigning phosphorylated metabolites. To obtain better sensitivity and structure information, 1H-detected 31P-1H heteronuclear total correlation spectroscopy (heteroTOCSY) was implemented and a complete chemical shift assignment for a number of phosphorylated standards was made. The time courses of 1D heteroTOCSY signal intensity versus spin-locking time were established for these standards to aid the optimization of the 2D experiment. This method was applied to crayfish extracts for the assignment of glucose 6-phosphate, alpha-glycerophosphate, ribose 5-phosphate, fructose 1,6-bisphosphate, phosphocholine, phosphoethanolamine, glucose 1-phosphate, glycerophosphoethanolamine, glycerophosphocholine, ATP, ADP, and AMP. An alkyl phosphate, a hexose 1-phosphate, and a UDP-hexose were also observed. These assignments allowed the identification of many changes in the 31P NMR spectra of crayfish extracts elicited by treatment with the organophosphate pesticide chlorpyrifos. The assignment of an in vivo 31P spectrum of a live crayfish was also made based on the extract assignment. This approach should be a powerful tool for examining stress-associated changes in the metabolism of phosphorylated compounds.

Estimation and relevance of depth correction in paediatric renal studies.

Measurement of absolute renal function by gamma camera techniques requires knowledge of kidney depth to correct for soft tissue attenuation, there is debate about the need to take depth into account when only relative renal function is estimated. The aim of this study was to derive a formula for renal depth in children and to assess the importance of depth correction when relative renal function is assessed with dimercaptosuccinic acid (DMSA) on the gamma camera. In this study, kidney depths were derived from measurements on abdominal computerised tomography (CT) images in 57 children in the supine position with two normally located kidneys. Using best-subset regression analysis, one formula for both left and right kidney depth (KD, cm) was developed based on the easily measured parameters of height (H, cm) and body weight (W, kg). The inclusion of extra variables was found to significantly improve the model compared with a model using weight alone (P<0.005). A second group of 19 children who underwent technetium-99m DMSA scans, had differential function estimated from both anterior and posterior views and the geometric mean method. The mean difference in differential renal function calculated by the geometric mean method versus the posterior image was only 1.2%. In conclusion, we present a new formula for the estimation of paediatric kidney depth for the absolute quantitation of kidney uptake. Further, for normally located kidneys it appears unnecessary to use the geometric mean method or to correct for individual renal depth when calculating differential function.

Structural characterisation of two hexasaccharides and an octasaccharide from chondroitin sulphate C containing the unusual sequence (4-sulpho)-N-acetylgalactosamine-beta1-4-(2-sulpho)-glucuronic acid-beta1-3-(6-sulpho)-N-acetylgalactosamine.

The structures of two hexasaccharides and an octasaccharide, isolated from shark cartilage chondroitin sulphate C after partial chondroitin ABC lyase digestion, have been elucidated by disaccharide analysis, liquid secondary-ion mass spectrometry, and 1H/13C NMR spectroscopy. In the case of the octasaccharide, the order of the internal disaccharide units was deduced from a novel one-dimensional TOCSY/ROESY excitation sculpting experiment [Gradwell M. J., Kogelberg, H. & Frenkiel, T. A. (1997) J. Magn. Reson. 124, 267-270] which incorporates two double-pulsed field-gradient spin-echo sequences. The oligosaccharides each contain the unusual motif GalNAc(4SO3-)beta1-4GlcA(2SO3-)beta1-3GalNAc(6SO3++ +-). From the adjoining sequences in the oligosaccharides, it is apparent that this motif occurs interspersed with typical chondroitin sulphate-C disaccharides, GlcAbeta1-3GalNAc(6SO3-). Such differentially sulphated domains along the glycosaminoglycan chains are potential sites for specific recognition processes.

Validation of the use of intermolecular NOE constraints for obtaining docked structures of protein-ligand complexes.

The use of intermolecular NOEs for docking a small ligand molecule into its target protein has been investigated with the aim of determining the effectiveness and methodology of this type of NOE docking calculation. A high-resolution X-ray structure of a protein-ligand complex has been used to simulate loose distance constraints of varying degrees of quality, typical of those estimated from experimental NOE intensities. These simulated data were used to examine the effect of the number, distribution and representation of the experimental constraints on the precision and accuracy of the calculated structures. A standard simulated annealing protocol was used, as well as a more novel method based on rigid-body dynamics. The results showed some analogies with those from similar studies on complete protein NMR structure determinations, but it was found that more constraints per torsion angle are required to define docked structures of similar quality. The effectiveness of different NOE-constraint averaging methods was explored and the benefits of using 'R-6 averaging' rather than 'centre averaging' with small sets of NOE constraints were shown. The starting protein structure used in docking calculations was obtained from previous X-ray or NMR structure studies on a related complex. The effects on the calculated conformations of introducing structural differences into the binding site of the initial protein structure were also considered.

Role of the N-terminal region of the skeletal muscle myosin light chain kinase target sequence in its interaction with calmodulin.

The binding of calmodulin (CaM) to four synthetic peptide analogues of the skeletal muscle myosin light chain kinase (sk-MLCK) target sequence has been studied using 1H-NMR. The 18-residue peptide WFF is anchored to CaM via the interaction of the Trp 4 side chain with the C-domain and the Phe 17 side chain with the N-domain of the protein. A peptide corresponding to the first 10 residues (WF10) does not provide the second anchoring residue and is not long enough to span both domains of CaM. 1H-NMR spectroscopy indicates that the WF10 peptide interacts specifically with the C-domain of CaM, and the chemical shifts of the bound Trp side chain are very similar in the CaM:WF10 and CaM:WFF complexes. Binding of the C-domain of CaM to the strongly basic region around Trp 4 of this MLCK sequence may be an important step in target recognition. Comparison of 1H-NMR spectra of CaM bound to WFF, a Trp 4-->Phe analogue (FFF), or a Trp 4-->Phe/Phe 17-->Trp analogue (FFW) suggests that all three peptides bind to CaM in the same orientation, i.e., with the peptide side chain in position 4 interacting with the C-domain and the side chain in position 17 interacting with the N-domain. This indicates that a Trp residue in position 4 is not an absolute requirement for binding this target sequence and that interchanging the Trp 4 and Phe 17 residues does not reverse the orientation of the bound peptide, in confirmation of the deduction from previous indirect studies using circular dichroism (Findlay WA, Martin SR, Beckingham K, Bayley PM, 1995, Biochemistry 34:2087-2094). Molecular modeling/energy minimization studies indicate that only minor local changes in the protein structure are required to accommodate binding of the bulkier Trp 17 side chain of the FFW peptide to the N-domain of CaM.

Global fold determination from a small number of distance restraints.

We have designed a distance geometry-based method for obtaining the tertiary fold of a protein from a limited number of structure-specific distance restraints and the secondary structure assignment. Interresidue distances were predicted from patterns of conserved hydrophobic amino acids deduced from multiple alignments. A simple model chain representing the protein was then folded by projecting its distance matrix into Euclidean spaces with gradually decreasing dimensionality until a final three-dimensional embedding was achieved. Tangled conformations produced by the projection steps were eliminated using a novel filtering algorithm. Information on various aspects of protein structure such as accessibility and chirality was incorporated into the conformation refinement, increasing the robustness of the algorithm. The method successfully identified the correct folds of three small proteins from a small number of restraints, indicating that it could serve as a useful computational tool in protein structure determination from NMR data.

Solution structure of bound trimethoprim in its complex with Lactobacillus casei dihydrofolate reductase.

Two- and three-dimensional (2D and 3D) NMR techniques have been used to assign the signals from nearly all of the protons in Lactobacillus casei dihydrofolate reductase (DHFR) (M(r) 18,300) in its 1:1 complex with the antibacterial drug trimethoprim. A sample of uniformly 15N-labeled protein was examined using 3D 15N/1H experiments [nuclear Overhauser, heteronuclear multiple quantum coherence (NOESY-HMQC) and total correlation, heteronuclear multiple quantum coherence (TOCSY-HMQC) experiments]. Twenty-two intermolecular NOEs between trimethoprim and protein protons and four intramolecular NOEs in the ligand have been detected. Some were obtained by using heteronuclear editing and 2D HMQC-NOESY experiments on complexes formed with 15N-and 13C-labeled trimethoprim molecules ([1,3-15N2,2-amino-15N]-and [7-13C,4'-methoxy-13C]trimethoprim) bound to unlabeled protein. The ligand-protein NOEs were used as distance constraints in conjunction with minimum energy and simulated annealing calculations (carried out with X-PLOR) to dock the trimethoprim ligand into dihydrofolate reductase, using as a starting structure the crystal coordinates from a related complex with a similar overall protein structure. The restrained minimum energy calculations and the simulated annealing calculations gave 83 calculated structures with distance violations of < 0.1 A. In all of these, the two aromatic rings of trimethoprim occupied essentially the same region of conformational space in the binding site (RMSD = 0.63 A). The protein residues nearest to the bound trimethoprim were found to be very similar in all of the structures and agreed well with corresponding contact residues observed in the X-ray crystal studies on trimethoprim complexes formed with Escherichia coli and chicken liver DHFRs.

Solution structure of a trefoil-motif-containing cell growth factor, porcine spasmolytic protein.

The porcine spasmolytic protein (pSP) is a 106-residue cell growth factor that typifies a family of eukaryotic proteins that contain at least one copy of an approximately 40-amino acid protein domain known as the trefoil motif. In fact, pSP contains two highly homologous trefoil domains. We have determined the complete three-dimensional solution structure of pSP by using a combination of two- and three-dimensional 1H NMR spectroscopy and distance geometry calculations. pSP is a relatively elongated molecule, consisting of two compact globular domains joined via a small interface. The protein's two trefoil domains adopt the same tertiary structure and contain a core C-terminal two-stranded antiparallel beta-sheet, preceded by a 6-residue helix that packs against the N-terminal beta-strand. The remainder of the protein backbone is taken up by two short loops that lie on either side of the beta-hairpin and are linked by an extended region that wraps around the C-terminal beta-strand. The topology of the protein backbone observed for the trefoil domains in pSP represents an unusual polypeptide fold. A striking feature of both trefoil domains is a surface patch formed from five conserved residues that have no obvious structural role. The two patches are located at the far ends of the protein molecule, and we propose that these residues form at least part of the receptor binding site, or sites, on pSP.