Search

The glomerular basement membrane (GBM) is damaged in diabetes through complex mechanisms that are not fully understood. Prominent among them is nonenzymatic protein glycation leading to the formation of so-called advanced glycation end products (AGEs). We examined the effects of in vitro glycation of intact collagen type IV in bovine lens capsule (LBM) and kidney glomerular (GBM) basement membranes on their susceptibility to matrix metalloproteinases, using stromelysin 1 (MMP-3) and gelatinase B (MMP-9). Sites of cleavage of unmodified LBM collagen were located in the triple helical region. In vitro glycation by glucose severely inhibited the release of soluble collagen cleavage peptides by MMP-3 and MMP-9. The distribution of AGEs within the three domains of collagen IV (7S, triple helical, and noncollagenous NC1) were compared for LBM glycation using AGE fluorescence, pentosidine quantitation, and immunoreactivity towards anti-AGE antibodies that recognize the AGE carboxymethyllysine (CML). Marked asymmetry was observed, with the flexible triple helical domain having the most pentosidine and fluorescent AGEs but the least CML. The in vivo relevance of these findings is supported by preliminary studies of AGE distribution in renal basement membrane (RBM) collagen IV domains from human kidneys of two insulin-dependent diabetics and one normal subject. Pentosidine and fluorescent AGE distributions of diabetic RBM were similar to LBM, but the CML AGE in diabetic kidney was less in the triple helical domain than in NC1. Our results support the hypothesis that nonenzymatic glycation of collagen IV contributes to the thickening of basement membranes, a hallmark of diabetic nephropathy.

Nonenzymatic glycation (Maillard reaction) of long-lived proteins is a major contributor to the pathology of diabetes and possibly aging and Alzheimer's disease. We report here kinetic studies of the glycation of the model protein ribonuclease A by glucose and ribose leading to the formation of antigenic advanced glycation end products ("AGEs"), detectable by AGE-specific polyclonal antibodies, and pentosidine, an acid-stable fluorescent AGE. As anticipated, the kinetics of glycation by ribose were considerably faster than by glucose, and the rate of AGE formation initially increased with increasing sugar concentrations. However, ribose above 0.15 M appeared to paradoxically slow the kinetics of AGE formation, suggesting ribose inhibits the conversion of "early" Amadori rearrangement products to "late" AGEs and thus favors the accumulation of reactive Amadori intermediates. The facile isolation of such protein intermediates was achieved by an "interrupted glycation" protocol which free and reversibly bound (Schiff base) ribose was removed following a short (24h) initial incubation of 0.5 M ribose at 37 degrees C. The kinetics of buildup of the Amadori intermediates and the kinetics of their post-Amadori conversion to antigenic AGEs were independently studied. A rapid and reversible inhibition of the post-Amadori kinetics by free ribose was verified by direct re-addition of ribose to the isolated, sugar-free intermediate. The pH dependence of the kinetics of antigenic AGE formation from such intermediates was measured and exhibited an unusual bell-shaped profile over the pH range of 5.0-9.5 with a maximum near pH 8.0. Aminoguanidine, a pharmacological AGE inhibitor, was found to moderately or weakly inhibit antigenic AGE formation in such post- Amadori steps. The isolation of the glycated ribonuclease intermediate thus simplifies kinetic and mechanistic studies of AGE formation, permits AGE studies in the absence of complications arising from free or Schiff base bound sugar, and provides a novel methodology for evaluating the mechanism and efficacy of therapeutic agents that may inhibit AGE formation.

The primary structures of two proteins that comprise PA28, an activator of the 20S proteasome, have been determined by cDNA cloning and sequencing. These protein subunits, termed PA28 alpha and PA28 beta, are about 50% identical to one another and are highly conserved between rat and human. PA28 alpha and PA28 beta are homologous to a previously described protein, Ki antigen, whose function is unknown. PA28 alpha, but neither PA28 beta nor Ki antigen, contains a 'KEKE motif', which has been postulated to promote the binding of proteins having this structural feature. PA28 alpha and PA28 beta were coordinately regulated by gamma-interferon, which greatly induced mRNA levels of both proteins in cultured cells. The mRNA level of the Ki antigen also increased in response to gamma-interferon treatment, but the magnitude of the increase was less than that for the PA28s, and the effect was transient. These results demonstrate the existence of a new protein family, at least two of whose members are involved in proteasome activation. They also provide the basis for future structure/function studies of PA28 subunits and the determination of their relative physiological roles in the regulation of proteasome activity.