NMR studies of the pbx1 TALE homeodomain protein free in solution and bound to DNA: proposal for a mechanism of HoxB1-Pbx1-DNA complex assembly.

The Hox homeodomain proteins are transcription factors involved in developmental regulation. Many of the vertebrate Hox proteins bind DNA cooperatively with the Pbx1 homeodomain protein. The crystal structure of a human HoxB1-Pbx1-DNA ternary complex revealed that interactions between the two proteins are mediated by the HoxB1 hexapeptide, which inserts into a hydrophobic pocket in Pbx1. It was also found that the Pbx1 DNA-binding domain is larger than the canonical three-helix homeodomain, containing an additional alpha-helix that is joined to the C terminus of the homeodomain by a turn of 310helix. These extra C-terminal residues had previously been shown to augment the cooperative interaction of Pbx1 with Hox partners, as well as enhancing the DNA binding of monomeric Pbx1. In order to characterize the role of the fourth Pbx1 helix in greater detail, we have examined the backbone structure of the enlarged Pbx1 DNA-binding domain in solution by(1)H,(15)N and(13)C multidimensional NMR spectroscopy. Our results show that the additional alpha-helix of Pbx1 is unfolded when the protein is free in solution and that its folding is triggered by binding of Pbx1 to DNA. In contrast, no change in conformation is observed upon mixing the HoxB1 protein with Pbx1 in the absence of DNA. This study suggests a model for the assembly of a stable HoxB1-Pbx1-DNA ternary complex.

Crystal structure of dimeric HIV-1 capsid protein.

X-ray diffraction analysis of a human immunodeficiency virus (HIV-1) capsid (CA) protein shows that each monomer within the dimer consists of seven alpha-helices, five of which are arranged in a coiled coil-like structure. Sequence assignments were made for two of the helices, and tentative connectivity of the remainder of the protein was confirmed by the recent solution structure of a monomeric N-terminal fragment. The C-terminal third of the protein is mostly disordered in the crystal. The longest helices in the coiled coil-like structure are separated by a long, highly antigenic peptide that includes the binding site of an antibody fragment complexed with CA in the crystal. The site of binding of the Fab, the position of the antigenic loop and the site of cleavage between the matrix protein and CA establish the side of the dimer that would be on the exterior of the retroviral core.

Structure of the amino-terminal core domain of the HIV-1 capsid protein.

The three-dimensional structure of the amino-terminal core domain (residues 1 through 151) of the human immunodeficiency virus-type 1 (HIV-1) capsid protein has been solved by multidimensional heteronuclear magnetic resonance spectroscopy. The structure is unlike those of previously characterized viral coat proteins and is composed of seven alpha helices, two beta hairpins, and an exposed partially ordered loop. The domain is shaped like an arrowhead, with the beta hairpins and loop exposed at the trailing edge and the carboxyl-terminal helix projecting from the tip. The proline residue Pro1 forms a salt bridge with a conserved, buried aspartate residue (Asp51), which suggests that the amino terminus of the protein rearranges upon proteolytic maturation. The binding site for cyclophilin A, a cellular rotamase that is packaged into the HIV-1 virion, is located on the exposed loop and encompasses the essential proline residue Pro90. In the free monomeric domain, Pro90 adopts kinetically trapped cis and trans conformations, raising the possibility that cyclophilin A catalyzes interconversion of the cis- and trans-Pro90 loop structures.

Comparison of NMR and molecular modeling results for a rigid and a flexible oligosaccharide.

Three-bond heteronuclear coupling constants (3JCH) are extremely useful in describing flexible models for oligosaccharides. We show that antiphase methods for measuring 3JCH in oligosaccharides have limited reliability but that the coupling constants can be reliably measured in natural abundance by quantitative J-correlation methods. Interpretation of 3JCH data for a pentasaccharide (lacto-N-fuco-pentaose 2) from human milk are consistent with a rigid model for the Lewis(a) trisaccharide epitope but for an antigenic tetrasaccharide fragment from the cell wall polysaccharide of viridans streptococci, 3JCH data imply a considerably more flexible model. Nuclear Overhauser effect (NOE) data are reported for a heptasaccharide repeating unit isolated from the cell wall polysaccharide of Streptococcus gordonii 38. The results for a tetrasaccharide fragment are similar to data reported for the same fragment in the cell wall polysaccharide from S.mitis J22. This result implies a similar conformation for the tetrasaccharide fragment in the polysaccharide and in the heptasaccharide and also implies that anisotropy of motion is not significant in the interpretation of the nuclear Overhauser effects in the polysaccharide. Interpretation of the NOE results for the tetrasaccharide fragment, like the 3JCH data, implies a flexible model with three conformations in fast exchange. The results of the two experimental techniques are combined with molecular modeling results including molecular dynamics simulation to provide a clear delineation between flexible and rigid oligosaccharide epitopes. The blood group Lewis(a) trisaccharide antigenic determinant is highly restricted in its motions by steric interactions while the antigenic tetrasaccharide fragment of the S.gordonii 38 heptasaccharide is considerably more mobile. We propose that some branched oligosaccharides are relatively rigid and some are flexible depending on subtle details of the linkages.

Measurement of long-range2 13C-1H coupling constants of 95% uniformly 13C-labeled polysaccharide from Streptococcus mitis J22.

The coaggregation of Streptococcus mitis strain J22 in the early stages of dental plaque formation has been shown to result from interaction of cell wall polysaccharides with lectins on the surface of other oral bacterial species. This bacterium was grown in a medium containing 13C as the sole carbon source. We have isolated the lectin receptor polysaccharide from this strain with full enrichment in 13C and have determined a number of two-bond and three-bond 13C-1H coupling constants from measurements of the offsets in two-dimensional homonuclear nmr spectra [exclusive correlated spectroscopy (E-COSY) method]. A scheme for reliable extraction of these coupling constants from homonuclear Hartmann-Hahn and nuclear Overhauser effect spectroscopy spectra is tested in model compounds. We interpret the three-bond coupling across the glycosidic linkage in terms of dihedral angles in order to provide conformational information to supplement molecular modeling and nuclear Overhauser effect data. We show that the E-COSY method works well even for coupling constants smaller than the nmr line width and that a number of the 3JCH across the glycosidic linkage are in the range of 1-2 Hz, which is much smaller than many previously reported values.

Structure of ten free N-glycans in ripening tomato fruit. Arabinose is a constituent of a plant N-glycan.

The concentration-dependent stimulatory and inhibitory effect of N-glycans on tomato (Lycopersicon esculentum Mill.) fruit ripening was recently reported (B. Priem and K.C. Gross [1992] Plant Physiol 98: 399-401). We report here the structure of 10 free N-glycans in mature green tomatoes. N-Glycans were purified from fruit pericarp by ethanolic extraction, desalting, concanavalin A-Sepharose chromatography, and amine-bonded silica high performance liquid chromatography. N-Glycan structures were determined using 500 MHz 1H-nuclear magnetic resonance spectroscopy, fast atom bombardment mass spectrometry, and glycosyl linkage methylation analysis by gas chromatography-mass spectrometry. A novel arabinosyl-containing N-glycan, Man alpha 1-->6(Ara alpha 1-->2)Man beta 1-->4GlcNAc beta 1-->4(Fuc alpha 1-->3)GlcNAc, was purified from a retarded concanavalin A fraction. The location of the arabinosyl residue was the same as the xylosyl residue in complex N-glycans. GlcNAc[5']Man3(Xyl)GlcNAc(Fuc)GlcNAc and GlcNAc[5']Man2GlcNAc(Fuc)GlcNAc were also purified from the weakly retained fraction. The oligomannosyl N-glycans Man5GlcNAc, Man6GlcNAc, Man7GlcNAc, and Man8GlcNAc were purified from a strongly retained concanavalin A fraction. The finding of free Man5GlcNAc in situ was important physiologically because previously we had described it as a promoter of tomato ripening when added exogenously. Mature green pericarp tissue contained more than 1 microgram of total free N-glycan/g fresh weight. Changes in N-glycan composition were determined during ripening by comparing glycosyl and glycosyl-linkage composition of oligosaccharidic extracts from fruit at different developmental stages. N-Glycans were present in pericarp tissue at all stages of development. However, the amount increased during ripening, as did the relative amount of xylosyl-containing N-glycans.