The determinants of glomerular capillary wall (GCW) permeability to proteins have been subject of controversial discussion. To study this question we have developed a modified isolated perfused rat kidney model in which tubular transport processes are completely blocked by perfusion fixation with glutaraldehyde. This model allows to directly titrate the charge density of the GCW using albumin solutions buffered over a wide pH-range, a manipulation that cannot be performed in the intact kidney. Analyzing the results of these experiments helped to determine a fixed charge density of the GCW of 43 mEq/L. In the present work, we used the isolated perfused fixed rat kidney model to study the influence of this fixed charge on the transglomerular passage of proteins. To do this, the fixed kidney was perfused with albumin solutions containing different isoforms of horseradish peroxidase. The lowest sieving coefficient was obtained with the acidic isoform (0.035+/-0.008, n = 7), while the isoforms at pI 6.85 and 8.45 showed higher sieving coefficients (0.059+/-0.008, n = 7 and 0.090+/-0.008, n = 4, respectively). The highest sieving coefficient (0.59+/-0.031, n = 6) was observed in perfusion experiments of the fixed kidney with cationic HRP (pI > or = 9.30). However, when comparing the sieving coefficients, the highly cationic isoform was excluded because it has a lower molecular weight than the other isoforms. The sieving coefficients of the other isoforms were significantly different (p < 0.05. ANOVA, Scheffé test). In conclusion, the presence of a discrete (even if lower than previously thought) "fixed" charge on the GCW of 43 mEq/L restricts the transglomerular passage of isoforms of horseradish peroxidase by a factor 2-3. These results imply that the influence of charge selectivity has been overstated in the literature.