Heat-induced enthalpy changes in different forms of bovine lactoferrin in water were examined by differential scanning calorimetry. Two thermal transitions with varying enthalpies were observed, depending on the iron-binding status of the protein. Iron-saturated lactoferrin was more resistant to heat-induced changes than was the apolactoferrin. Native lactoferrin had two transitional peaks, and pasteurization affected only the low temperature transition. Iron-saturated lactoferrin revealed a single transitional peak that was resistant to pasteurization. However, both protein forms were completely denatured by UHT. The effect of pasteurization and UHT on the protein interaction capacity with bacteria was examined in a 125I-labeled lactoferrin binding-inhibition assay. The ability of native and iron-saturated lactoferrins to bind various bacterial species was unaffected by pasteurization. However, UHT treatment decreased this interaction capacity. Native lactoferrins, both unheated and pasteurized, showed similar antibacterial properties and moderately inhibited Escherichia coli. However, this inhibitory capacity was lost after UHT treatment. Finally, iron-saturated lactoferrin did not inhibit bacterial growth; neither pasteurization nor UHT could change this property. Thus, UHT seems to affect structural as well as certain biological properties of both native and iron-saturated bovine lactoferrins, and pasteurization seems to be a treatment of choice for products containing this protein.

The effect of lactoferrin (Lf) on bacterial growth was tested by measuring conductance changes in the cultivation media by using a Malthus-AT system and was compared with the magnitude of "2'I-labeled Lf binding in 15 clinical isolates of Escherichia coli. The binding property was inversely related to the change in bacterial metabolic rate (r = 0.91) and was directly related to the degree of bacteriostasis (r = 0.79). The magnitude of Lf-bacterium interaction showed no correlation with the MIC of Lf. In certain strains, Lf at supraoptimal levels reduced the bacteriostatic effect. Thus, the Lf concentration in the growth media was critical for the antibacterial effect. The cell envelopes of Salmonella typhimurium 395MS with smooth lipopolysaccharide (LPS) and its five isogenic rough mutants revealed 38-kDa porin proteins as peroxidaselabeled-Lf-reactive components in sodium dodecyl sulfate-polyacrylamide gel electrophoresis and Western blot (ligand blot) analysis. However, in the whole cell binding assay, parent strain 395MS demonstrated a very low interaction with 1251I-Lf. On the other hand, Lf interaction gradually increased in correspondence with the decrease in LPS polysaccharide moiety in the isogenic rough mutants. Conductance measurement studies revealed that the low-level-Lf-binding (low-Lf-binding) strain 395MS with smooth LPS was relatively insusceptible to Lf, while the high-Lf-binding mutant Rd was more susceptible to Lf. These data suggested a correlation between Lf binding to porins and the Lf-mediated antimicrobial effect. The polysaccharide moiety of LPS shielded porins from the Lf interaction and concomitantly decreased the antibacterial effect.

Summary. An enzyme-linked ligand binding assay (ELBA) was devised to measure the interaction between bacteria and human (H) or bovine (B) lactoferrin (Lf) linked to horseradish peroxidase. Reagents were calibrated for optimum colour development with ophenylenediamine as chromophore and organisms that were either positive or negative in a radioisotope-labelled ligand binding assay (RLBA) with 12%Lf. Good correlation of Lf binding (r = 0.89) was found between ELBA and RLBA with 169 randomly selected strains of Escherichia coli. A semi-quantitative scoring system for ELBA, corresponding to a similar system for RLBA, was established and shown to be valid for 517 strains from seven species of bacterial pathogens. ELBA was used to measure bacterial Lf binding-saturation and displacement kinetics and shown to be comparable with RLBA. ELBA may be a suitable method for examining the binding of Lf to bacteria without the need for radioactive isotopes.

The interaction of lactoferrin (Lf) with Aeromonas hydrophila (n = 28) was tested in a 125I-labeled protein-binding assay. The mean per cent binding values for human Lf (HLf) and bovine Lf (BLf) were 13.4 +/- 2.0 (SEM), and 17.5 +/- 2.7 (SEM), respectively. The Lf binding was characterized in type strain A. hydrophila subsp. hydrophila CCUG 14551. The HLf and BLf binding reached a complete saturation within 2 h. Unlabeled HLf and BLf displaced 125I-HLf binding in a dose-dependent manner, and more effectively by the heterologous (1 microgram for 50% inhibition) than the homologous (10 micrograms for 50% inhibition) ligand. Apo- and holo-forms of HLf and BLf both inhibited more than 80%, while mucin caused approx. 50% inhibition of the HLf binding. Various other proteins (including transferrin) or carbohydrates did not block the binding. Two HLf-binding proteins with an estimated molecular masses of 40 kDa and 30 kDa were identified in a boiled-cell-envelope preparation, while the unboiled cell envelope demonstrated a short-ladder pattern at the top of the separating gel and a second band at approx. 60 kDa position. These data establish a specific interaction of Lf and the Lf-binding proteins seem to be porins in A. hydrophila.

The ability of Shigelia flexneri to interact with lactoferrin (Lf) was examined with a '251-labeled proteinbinding assay. The percent binding of human lactoferrin (HLf) and bovine lactoferrin (BLf) to 45 S. flexneri strains was 19 ± 3 and 21 ± 3 (mean ± standard error of the mean), respectively. '25I-labeled HLf and BLf binding to strain M9OT reached an equilibrium within 2 h. Unlabeled HLf and BLf displaced the 251-HLf-bacteria interaction in a dose-dependent manner. The Lf-bacterium complex was uncoupled by KSCN or urea, but not by NaCl. The interaction was specific, and -4,800 HLf binding sites (affinity constant [Kal, 690 nM) or -5,700 BLf binding sites (Ka, 104 nM) per cell were estimated in strain M9OT by a Scatchard plot analysis. The native cell envelope (CE) and outer membrane (OM) did not reveal Lf-binding components in sodium dodecyl sulfate-polyacrylamide gel electrophoresis. However, after being boiled, the CE and OM preparations showed three distinct horseradish peroxidase-Lf reactive bands of about 39, 22, and 16 kDa. The 39-kDa component was also reactive to a monoclonal antibody specific for porin (Pol) proteins of members of the family Enterobacteriaceae. The Lf-binding protein pattern was similar with BLf or HLf, for Crb+ and Crbstrains. The protein-Lf complex was dissociable by KSCN or urea and was stable after treatment with NaCl.Variation (loss) in the 0 chain of lipopolysaccharide (LPS) markedly enhanced the Lf-binding capacity in the isogenic rough strain SFL1070-15 compared with its smooth parent strain, SFL1070. These data establish that Lf binds to specific components in the bacterial OM; the heat-modifiable, anti-Pol-reactive, and LPS-associated properties suggested that the Lf-binding proteins are porins in S. flexneri.