We contrast the adsorption of human serum albumin (HSA) onto two solid substrates previously primed with the same polyelectrolyte of net opposite charge to form one of two alternative structures: randomly adsorbed polymer and the "brush" configuration. These structures were formed either by the adsorption of quaternized poly-4-vinylpyridine (QPVP) or by end-grafting QPVP chains of the same chemical makeup and the same molecular weight to surfaces onto which QPVP segments did not adsorb. The adsorption of HSA was quantified by using Fourier transform infrared spectroscopy in attenuated total reflection (FTIR-ATR). The two substrates showed striking differences with regard to HSA adsorption. First, the brush substrate induced lesser perturbations in the secondary structure of the adsorbed HSA, reflecting easier conformational adjustment for longer free segments of polyelectrolyte upon binding with the protein. Second, the penetration of HSA into the brush substrate was kinetically retarded relative to the randomly adsorbed polymer, probably due to both pore size restriction and electrostatic sticking between charged groups of HSA and QPVP molecules. Third, release of HSA from the adsorbed layer, as the ionic strength was increased from a low level up to the high level of 1 M NaCl, was largely inhibited for the brush substrate, but occurred easily and rapidly for the substrate with statistically adsorbed QPVP chains. Finally, even after addition of a strong polymeric adsorption competitor (sodium polystyrene sulfonate), HSA remained trapped within a brush substrate though it desorbed slowly from the preadsorbed QPVP layer. This method to produce irreversible trapping of the protein within a brush substrate without major conformational change may find application in biosensor design.