Characterization of folding pathways of the type-1 and type-2 periplasmic binding proteins MglB and ArgT

The family of periplasmic binding proteins (PBPs) is believed to have arisen from a common ancestor and to have differentiated into two types. At first approximation, both types of PBPs have the same fold pattern, reflecting their common origin. However, the connection between the main chains of a type 2 PBP is more complicated than a type 1 PBP's. We have been interested in the possibility that such structural changes affect the folding of PBPs. In this study, we have characterized the folding pathways of MglB (a type 1 PBP) and ArgT (a type 2 PBP) by using urea gradient gel electrophoresis, fast protein size-exclusion liquid chromatography and hydrophobic dye ANS binding assay. We found a distinct difference in folding between these two proteins. The folding of MglB followed a simple two-state transition model, whereas the folding of ArgT was more complicated.     

Regulation of type 1 protein phosphatase/inhibitor-2 complex by glycogen synthase kinase-3beta in intact cells

Inhibitor 2 (I-2) is a ubiquitous regulator of type 1 protein phosphatase (PP1). Previous in vitro studies suggested that its inhibitory activity towards PP1 is regulated by phosphorylation at Thr72 by glycogen synthase kinase-3beta (GSK-3beta), and at Ser86, Ser120, and Ser121 by casein kinase 2 (CK2). Here we report that GSK-3beta expressed in COS-7 cells phosphorylates wild-type I-2 but not an I-2 mutant carrying a T to A substitution at residue 72, showing that GSK-3beta phosphorylates I-2 at T72 in vivo as well. Co-immunoprecipitation study demonstrated that HA-GSK-3beta and I-2-FLAG co-exist in a same complex in the intact cells, but they do not bind directly. It is noteworthy that co-expression of Myc-PP1C significantly increased co-precipitation of HA-GSK-3beta with I-2-FLAG, showing a complex formation of HA-GSK-3beta/Myc-PP1C / I-2-FLAG in vivo. Further studies using a GSK-3beta kinase-dead mutant and LiCl, an inhibitor of GSK-3beta, showed that the enzyme activity of GSK-3beta is required for co-precipitation. IP-Western study using several I-2 mutants substituted at phosphorylation sites (T72, S86, S120, and S121) suggested that phosphorylation of I-2 by CK2 is also involved in enhancement of association between GSK-3beta and I-2 in vivo. This study is the first demonstration that GSK-3beta associates with PP1C/I-2 complex and phosphorylates I-2 at T72 in the intact cells.     

Characterization and preliminary crystallographic studies of EMS16, an antagonist of collagen receptor (GPIa/IIa) from the venom of Echis multisquamatus

EMS16 is a member of the snake venom-derived C-type lectin family of proteins (CLPs) found in the venom of Echis multisquamatus. It binds to glycoprotein Ia/IIa (integrin alpha2beta1), a major collagen receptor of platelets, acting as a potent antagonist of platelet aggregation and cell migration. Amino acid sequencing and cDNA cloning of EMS16 have revealed that it is composed of an A chain of 134 amino acid residues and a B chain of 128 residues. Crystals of EMS16 belong to space group P2(1)2(1)2(1), with unit-cell parameters a = 46.57, b = 59.93, and c = 115.74 A, and diffract to a resolution of 1.9 A. Phase determination is underway by means of molecular replacement with the structure of blood coagulation factor IX-binding protein (IX-bp) from habu snake venom (PDB code 1bj3) as the search model.     

Molecular cloning of a novel transmembrane protein MOLT expressed by mature oligodendrocytes

A novel oligodendrocyte (OL)-specific cDNA was isolated from brain capillary endothelial cells and characterized. The cDNA encodes a protein of 1099 amino acids that contains a signal peptide and a transmembrane domain. The protein was expressed in mature OLs in vivo and in vitro cell cultures and was thus designated as mature OL transmembrane protein (MOLT). RT-PCR analysis showed that MOLT mRNA was expressed in brain, lung, pancreas, and testis. A polyclonal antibody raised against a part of the mouse MOLT reacted specifically with multipolar OLs possessing radially oriented processes that penetrated into the gray matter. More cells were detected in the white matter, and these had longitudinally oriented processes. In a rat OL lineage culture system, oligodendrocyte precursor cells did not initially produce MOLT mRNA and protein, but when they begun to differentiate into mature OLs, they started expressing MOLT. Consequently, MOLT may function as OLs become mature and may serve as a cell-surface marker for OL differentiation.     

Aggregation analysis of the microtubule binding domain in tau protein by spectroscopic methods

The microtubule-associated protein tau is a highly soluble protein that shows hardly any tendency to assemble under physiological conditions. In the brains of Alzheimer's disease (AD) patients, however, tau dissociates from the axonal microtubule and abnormally aggregates to form paired helical filaments (PHFs). One of the priorities in Alzheimer research is to clarify the mechanism of PHF formation. In recent years, several factors regulating tau assembly have come to light, yet some important questions remain to be answered. In this work, the His-tagged gene constructs of the four-repeat microtubule binding domain (4RMBD) in tau protein and its three mutants, 4RMBD S305N, N279K, and P301L, were expressed in E. coli and purified. Gel filtration chromatography and dynamic light scattering measurement yielded a Stokes radius of 3.1 nm, indicating that the His-tagged 4RMBD normally exists in buffer solution in a dimer state, which is formed by non-covalent intermolecular interactions. This non-covalent dimer can further polymerize to form filaments in the presence of polyanions such as heparin. The kinetics of the in vitro aggregation was monitored by thioflavine S dye fluorescence and CD measurements. The aggregation of 4RMBD was suggested to be a nucleation-dependent process, where the non-covalent dimer acts as an effective structural unit. The aggregation rate was strongly affected by the point mutation. Among the 4RMBD mutants, the rate of S305N was exceptionally fast, whereas N279K was the slowest, even slower than the wild-type. The aggregations were optimal in a weakly reducing environment for all the mutants and the wild type. However, the aggregations were affected differently by buffer pH, depending on the 4RMBD mutation.     

Distinct sites regulating grayanotoxin binding and unbinding to D4S6 of Na(v)1.4 sodium channel as revealed by improved estimation of toxin sensitivity

Grayanotoxin (GTX) exerts selective effects on voltage-dependent sodium channels by eliminating fast sodium inactivation and causing a hyperpolarizing shift in voltage dependence of channel activation. In this study, we adopted a newly developed protocol that provides independent estimates of the binding and unbinding rate constants of GTX (k(on) and k(off)) to GTX sites on the sodium channel protein, important in the molecular analysis of channel modification. Novel GTX sites were determined in D2S6 (Asn-784) and D3S6 (Ser-1276) by means of site-directed mutagenesis; the results suggested that the GTX receptor consists of the S6 transmembrane segments of four homologous domains facing the ion-conducting pore. We systematically introduced at two sites in D4S6 (Na(v)1.4-Phe-1579 and Na(v)1.4-Tyr-1586) amino acid substituents with residues containing hydrophobic, aromatic, charged, or polar groups. Generally, substitutions at Phe-1579 increased both k(on) and k(off), resulting in no prominent change in dissociation constant (K(d)). It seems that the smaller the molecular size of the residue at Na(v)1.4-Phe-1579, the larger the rates of k(on) and k(off), indicating that this site acts as a gate regulating access of toxin molecules to a receptor site. Substitutions at Tyr-1586 selectively increased k(off) but had virtually no effect on k(on), thus causing a drastic increase in K(d). At position Tyr-1586, a hydrophobic or aromatic amino acid side chain was required to maintain normal sensitivity to GTX. These results suggest that the residue at position Tyr-1586 has a more critical role in mediating GTX binding than the one at position Phe-1579. Here, we propose that the affinity of GTX to Na(v)1.4 sodium channels might be regulated by two residues (Phe and Tyr) at positions Phe-1579 and Tyr-1586, which, respectively, control access and binding of GTX to its receptor.     

Domains in tropoelastin that mediate elastin deposition in vitro and in vivo

Elastic fiber assembly is a complicated process involving multiple different proteins and enzyme activities. However, the specific protein-protein interactions that facilitate elastin polymerization have not been defined. To identify domains in the tropoelastin molecule important for the assembly process, we utilized an in vitro assembly model to map sequences within tropoelastin that facilitate its association with fibrillin-containing microfibrils in the extracellular matrix. Our results show that an essential assembly domain is located in the C-terminal region of the molecule, encoded by exons 29-36. Fine mapping studies using an exon deletion strategy and synthetic peptides identified the hydrophobic sequence in exon 30 as a major functional element in this region and suggested that the assembly process is driven by the propensity of this sequence to form beta-sheet structure. Tropoelastin molecules lacking the C-terminal assembly domain expressed as transgenes in mice did not assemble nor did they interfere with assembly of full-length normal mouse elastin. In addition to providing important information about elastin assembly in general, the results of this study suggest how removal or alteration of the C terminus through stop or frameshift mutations might contribute to the elastin-related diseases supravalvular aortic stenosis and cutis laxa.     

Crystal structure of rice alpha-galactosidase complexed with D-galactose

alpha-Galactosidases catalyze the hydrolysis of alpha-1,6-linked galactosyl residues from galacto-oligosaccharides and polymeric galacto-(gluco)mannans. The crystal structure of rice alpha-galactosidase has been determined at 1.5A resolution using the multiple isomorphous replacement method. The structure consisted of a catalytic domain and a C-terminal domain and was essentially the same as that of alpha-N-acetylgalactosaminidase, which is the same member of glycosyl hydrolase family 27. The catalytic domain had a (beta/alpha)8-barrel structure, and the C-terminal domain was made up of eight beta-strands containing a Greek key motif. The structure was solved as a complex with d-galactose, providing a mode of substrate binding in detail. The d-galactose molecule was found bound in the active site pocket on the C-terminal side of the central beta-barrel of the catalytic domain. The d-galactose molecule consisted of a mixture of two anomers present in a ratio equal to their natural abundance. Structural comparisons of rice alpha-galactosidase with chicken alpha-N-acetylgalactosaminidase provided further understanding of the substrate recognition mechanism in these enzymes.