Calcium is required for the substrate binding and for the chemical step of the interfacial catalytic turnover cycle of pancreatic phospholipase A2 (PLA2), but not for the binding of the enzyme to the interface. The role of calcium and other divalent cations (C) is analyzed for the effect on the substrate binding and kcat* for the chemical step. The cofactor role of 3d-cations(II) (C) for the hydrolysis of dimyristoylphosphatidylmethanol (DMPM) vesicles is characterized as an equilibrium dissociation constant for the interfacial binary (E*C) and ternary (E*CL) complexes of PLA2 and substrate mimics (L). Of the cations(II) that promote the binding of a mimic to the enzyme at the interface (E*), only a subgroup supports the chemical step. For example, Cd, Zn, and Cu form ternary E*CL complexes with kcat* of <1 s-1, compared to the rate of >100 s-1 with Ca, Fe, Mn, Co, and Ni. Oxygen exchange from H218O to the products of hydrolysis of DMPM incorporates one 18O in myristate. Incorporation of the first and second 18O occurs during the incubation of both the products of hydrolysis in H218O with PLA2 and Ca, but not with Zn. The cation-dependent changes in the UV difference spectrum, associated with the formation of E*C and E*CL, suggest that the changes are mainly due to catalytic His-48, and possibly Tyr-52 and Tyr-73, and are different with Ca as opposed to Zn. These results and simulations suggest considerable plasticity in the calcium binding and catalytic site environment. It is proposed that the higher ground state stability of the E*CS complex with the inhibitory cations increases the effective activation energy. For the chemical step, calcium coordinated with a nucleophilic water and the ester carbonyl oxygen facilitates the near-attack geometry in the E*CaS, and the His-48.Asp-99 pair acts as a proton acceptor. As a prelude to establishing the catalytic mechanism, factors controlling the energetically demanding transition state are also discussed.

The conformational properties in solution of the glycans on the alpha subunit of recombinant human chorionic gonadotropin are described, using high-resolution multinuclear NMR studies on uniformly 13C, 15N-enriched recombinant glycoprotein expressed in CHO cells. The glycan important for full biological activity of hCG, namely, that at Asn 52, appears to extend into solution both in the isolated alpha subunit and in complex with the beta subunit. The disposition of this glycan with respect to the protein backbone suggests that glycosylation maintains full biological activity of hCG either by interacting with a lectin-like region of the hCG receptor or by reducing the affinity of the hormone for the hCG receptor and preventing its down-regulation.

Abstract:We have analyzed the known three-dimensional structures of trimeric porins from bacterial outer membranes. The distribution of surface-exposed residues in a direction perpendicular to the membrane is similar to that in helical membrane proteins, with aliphatic residues concentrated in the central 20 8, of the bilayer. Outside these residues is a layer of aromatic residues, followed by polar and charged residues. Residues in the trimer interface are more conserved than residues not in the interface. By comparing the interface and noninterface residues, an interface preference scale has been derived that may be used as a basis for predicting interface surfaces in monomer models.Keywords: P-barrel; membrane protein; porin; three-dimensional structure As far as is known from high-resolution structural data, integral membrane proteins belong either to the helix-bundle or the /?-barrel class (von Heijne, 1997). In both cases, hydrophobic amino acids provide the interface to the surrounding lipid bilayer, but, beyond this, the two architectures are very different. Thus, while transmembrane a-helices are more or less uniformly hydrophobic (Bowie, 1997;Wallin et al., 1997), transmembrane P-strands tend to be amphiphilic, with polar and charged residues projecting into a central pore. Moreover, most transmembrane a-helices are believed to be individually stable in the bilayer (Hunt et al., 1997;Popot & Engelman, 1997) whereas individual transmembrane @strands certainly are not. Their respective modes of membrane insertion are thus very different (von Heijne, 1997).All P-barrel outer membrane proteins (porins), for which highresolution structures are available form oligomers, and, thus, bury part of their outer surface in a protein-protein interface (Fig. 1). In contrast to water-soluble proteins, little is known about the factors that drive the association between protein subunits in a lipid environment. To provide a basis for understanding general aspects of membrane protein structure, we have analyzed the known highReprint requests to: Arne Elofsson, Department of Biochemistry, Stockholm University, s-106 91 Stockholm, Sweden; e-mail: arne@biokemi.su.se. resolution porin structures in terms of which kinds of residues are exposed to the lipid environment, buried within monomers, exposed to the central pore and buried between subunits. In particular, we show that subunit interfaces can be roughly identified on the basis of amino acid composition and residue conservation, suggesting that it may be possible to extend current methods for predicting transmembrane P-strands and overall monomer fold (Schirmer & Cowan, 1993;Gromiha et al., 1997) to also include the oligomeric structure. Table 1 presents some basic statistics for the proteins used in this study. The porins of known structure can be divided into three groups according to the number of P-strands forming the pore. Porins from Rhodopseudomonas blastica (Ipm) and from Rhodobacter capsulatus (2por and 3por, two different crystal forms of the same protein) both...