| dc.description.abstract | Lectins form a diverse group of specific carbohydrate-binding proteins of non-immune origin and are characterized by their ability to agglutinate cells. In addition to their usefulness in several applications, these proteins have attracted wide attention due to their unique ability to specifically bind to cell-surface carbohydrates. They, however, constitute a comparatively underexplored family of proteins.
The introductory chapter outlines the origin, biological role, and uses of lectins and briefly discusses the crystallographic studies already carried out on them when the X-ray analysis of peanut lectin was taken up in this laboratory.
Peanut lectin exhibits specificity for the T-antigen (D-Gal 1-3 D-GalNAc). At physiological pH, it exists as a tetramer of molecular weight 110,000 with four apparently identical polypeptide chains. The crystallization and the preliminary X-ray studies of four different forms of the protein, carried out in this laboratory, and the initial low-resolution self-rotation function calculations, in which the author participated, are outlined in the second chapter. These rotation function studies and the further investigations described in this thesis were carried out on an orthorhombic form, space group P222, with a = 129.3 Å, b = 126.9 Å, and c = 76.9 Å. The self-rotation function and chemical cross-linking experiments, carried out in parallel, suggested the molecule to be a dimer of a dimer with 222 symmetry.
The details of high-resolution intensity data collection (2.4 Å) using oscillation photography followed by computer-controlled microdensitometry and the subsequent processing of the data are described in Chapter 3. Intensities of 40,899 unique reflections were derived from a total of 153,631 reflections recorded, with a merging R value of 0.114 on intensities for reflections with I > 2(I). Several self-rotation functions were computed employing data from different resolution shells and more than one radius of integration. Significant features in the functions were in the neighborhood of angles appropriate for three crystallographic twofold or 2 screw axes. It appeared that the molecular diads were nearly, but not exactly, parallel to the crystallographic ones.
Following reports that legume lectins exhibit circularly permuted sequence homology among themselves, the available partial amino acid sequence of peanut lectin was compared with the sequences of Concanavalin A (ConA), pea lectin, and favin, the three legume lectins whose three-dimensional structures are available. The detailed comparison, discussed in Chapter 5, led to the conclusion that peanut lectin was likely to have a three-dimensional structure similar to those of ConA, pea lectin, and favin, although differences were anticipated, especially in the carbohydrate-binding region.
Having established the homology of peanut lectin with the other three legume lectins, attempts to determine the orientation and position of the peanut lectin molecule in the unit cell were initiated with ConA and pea lectin molecules as search models (Chapter 6). The structure of favin is very close to that of pea lectin, and therefore the use of additional models involving favin coordinates was unnecessary. Cross-rotation functions involving ConA indicated two possible solutions to the rotation problem. This twofold ambiguity persisted in calculations using ConA and pea lectin dimers, although one of the solutions appeared more likely than the other.
Preliminary R-factor and correlation coefficient searches with the ConA model at several orientations in the neighborhood of the two orientations indicated by cross-rotation functions failed to yield a solution to the translation problem. It was felt that one of the reasons for the failure could be errors in the search models. Plausible tetrameric, dimeric, and monomeric search models were constructed after careful examination of the two sets of ConA coordinates available in the Protein Data Bank and the pea lectin coordinates kindly supplied by F. Sudda. In these search models, only the regions of the molecule corresponding to ConA and pea lectin were used. Cross-rotation functions using these models confirmed the two possible solutions to the rotation problem identified earlier. The regions of the crystal asymmetric unit to be searched for the solution of the translation problem were demarcated by packing analysis using a novel program developed by the author.
Subsequent exhaustive R-factor and correlation coefficient searches using two plausible tetrameric models again failed to yield the position of the molecule in the cell. The possibility of a solution with the molecule located on crystallographic twofold axes was also carefully explored, but again without any positive result.
While the rotation and translation searches outlined above were being pursued by the author, attempts were also underway in the laboratory for preparing isomorphous heavy-atom derivatives. Data from six derivatives, up to resolutions varying between 3.6 to 3.2 Å, and fresh native data up to a resolution of 3 Å, all collected on a Nicolet-Siemens area detector system, were made available to him at a stage when he had almost reached the end of the tether as far as the translational search was concerned.
The determination and refinement of heavy-atom parameters in the derivatives and the low-resolution structure of peanut lectin determined using them are discussed in the final chapter. The occupancies of heavy-atom sites in all the derivatives were rather low. Also, the heavy-atom constellations in the three platinum derivatives, two involving KPtCl and one involving KPtCl, were related to one another and had no symmetry among the sites. The other three derivatives, involving samarium nitrate, iodophenyl galactose, and gold chloride, had four heavy-atom sites, each related to one another by very approximate 222 symmetry.
The results of the phase angle calculations indicated phasing to be good at low resolution but poor at high resolution; therefore, an electron density map was calculated at 5.5 Å resolution. The tetrameric molecule could be clearly distinguished from the solvent regions in the map. The center of the density corresponding to the molecule occurred at a position found to be allowed in the earlier packing analysis. The density had non-crystallographic 222 symmetry, and the orientations of the three twofold axes were in excellent conformity with the more probable of the two solutions obtained earlier from the cross-rotation function studies.
The low-resolution map unambiguously establishes the gross structure of peanut lectin, which is entirely consistent with different rotation functions and packing analysis, and sets the stage for a future high-resolution analysis in terms of the polypeptide chain conformation.
The thesis has two annexures. One of them describes the program used in R-factor search. The second is concerned with a fast algorithm developed by the author and used in the work on peanut lectin for macromolecular packing calculations. The use of this algorithm results in considerable saving in computation time. This is achieved by reducing the three-dimensional search to a set of three two-dimensional searches. | |
