Proteolytic activation of proteose peptone component 3 by release of a C-terminal peptide with antibacterial properties.

The milk protein proteose peptone component 3 (PP3, also known as lactophorin) is a small phosphoglycoprotein, which is exclusively expressed in the lactating mammary gland. A 23-residue synthetic peptide (lactophoricin, Lpcin S), corresponding to the C-terminal amphipathic α-helix of PP3, has previously been shown to permeabilize membranes and display antibacterial activity. Lactophorin readily undergoes proteolytic cleavage in milk and during dairy processing, and it has been suggested that PP3-derived peptides are part of milk's endogenous defense system against bacteria. Here, we report that a 26-residue C-terminal peptide (Lpcin P) can be generated by trypsin proteolysis of PP3 and that structural and functional studies of Lpcin P indicate that the peptide has antibacterial properties. The Lpcin P showed α-helical structure in both anionic and organic solvents, and the amount of α-helical structure was increased in the presence of lipid vesicles. Oriented circular dichroism showed that Lpcin P oriented parallel to the membrane surface. However, the peptide permeabilized calcein-containing vesicles efficiently. Lpcin P displayed antibacterial activity against Streptococcus thermophilus, but not against Staphylococcus aureus and Escherichia coli. The PP3 full-length protein did not display the same properties, which could indicate that PP3 functions as a precursor protein that upon proteolysis, releases a bioactive antibacterial peptide.

Preparation and comparison of cytotoxic complexes formed between oleic acid and either bovine or human α-lactalbumin.

α-Lactalbumin is a ubiquitous calcium-binding milk protein with a well-characterized function in regulating the synthesis of lactose. An entirely different activity has been shown to occur when a complex is formed between calcium-free α-lactalbumin and oleic acid. This complex shows strong cytotoxic action against several cancer cells, and several mechanisms have been suggested to account for this cell-killing activity. Most studies have been performed using the human protein, but bovine α-lactalbumin shows similar activity. A new and simple 2-step method for purification of calcium-free α-lactalbumin has been developed, and the resulting highly purified preparation was used to generate a complex with oleic acid. Using 3 different cell lines and 2 types of cell viability assays, the bovine and human α-lactalbumin showed comparable cytotoxic activity. The effect was apparent after 15 min of incubation and was inhibited by the presence of fetal bovine serum or bovine serum albumin. The bovine protein might be a useful alternative to the human protein, but also raises the question whether cytotoxic activity could be generated in different kinds of food containing α-lactalbumin.

Considerable variation in the concentration of osteopontin in human milk, bovine milk, and infant formulas.

Osteopontin (OPN) is a multifunctional bioactive protein that is implicated in numerous biological processes such as bone remodeling, inhibition of ectopic calcification, and cellular adhesion and migration, as well as several immune functions. Osteopontin has cytokine-like properties and is a key factor in the initiation of T helper 1 immune responses. Osteopontin is present in most tissues and body fluids, with the highest concentrations being found in milk. In the present study, ELISA for human and bovine milk OPN were developed and OPN concentration in human breast milk, bovine milk, and infant formulas was measured and compared. The OPN concentration in human milk was measured to approximately 138 mg/L, which corresponds to 2.1% (wt/wt) of the total protein in human breast milk. This is considerably higher than the corresponding OPN concentrations in bovine milk (approximately 18 mg/L) and infant formulas (approximately 9 mg/L). Moreover, bovine milk OPN is shown to induce the expression of the T helper 1 cytokine IL-12 in cultured human lamina propria mononuclear cells isolated from intestinal biopsies. Finally, the OPN concentration in plasma samples from umbilical cords, 3-mo-old infants, and pregnant and nonpregnant adults was measured. The OPN level in plasma from 3-mo-old infants and umbilical cords was found to be 7 to 10 times higher than in adults. Thus, the high levels of OPN in milk and infant plasma suggest that OPN is important to infants and that ingested milk OPN is likely to induce cytokine production in neonate intestinal immune cells.

Characterization of human mucin (MUC15) and identification of ovine and caprine orthologs.

The glycoprotein MUC15 (mucin 15) was initially isolated from the bovine milk fat globule membrane. The present work demonstrates the existence of immunologically similar proteins ( approximately 130 kDa) in ovine, caprine, porcine, and buffalo milk samples. Purification and N-terminal amino acid sequencing confirmed the presence of ovine and caprine MUC15 orthologs in milk fat globule membranes. Expression of MUC15 in human milk was demonstrated by immunostaining ( approximately 150 kDa) as well as by mass spectrometry. Screening of a human multiple tissue expression array showed abundant MUC15 gene expression in placenta, salivary gland, thyroid gland, trachea, esophagus, kidney, testis, and the leukemia K-562 cell line. Furthermore, moderate expression was seen in the pancreas, adult and fetal lung, fetal kidney, lymph node, adult and fetal thymus, and parietal lobe. Structural motifs for interactions (epidermal growth factor receptor and Src homology 2 domains) are identified in the intracellular region. Implication of the mucin in signal transduction and the potential physiological function of MUC15 are discussed.

Characterization of carbohydrate structures of bovine MUC15 and distribution of the mucin in bovine milk.

The present work reports the characterization of carbohydrate structures and the distribution of the newly identified mucin MUC15, a highly glycosylated protein associated with the bovine milk fat globule membrane (MFGM). Distribution of MUC15 was investigated in various fractions of bovine milk by densitometric scanning of Western blots. In raw milk, MUC15 was shown to constitute 0.08% (wt) of the protein and approximately 1.5% (wt) of the MFGM-associated proteins. Surprisingly, this study showed that in addition to the fat-containing fractions, such as MFGM and buttermilk, MUC15 was present in nonfat-containing fractions as well, such as skim milk and whey. Compositional and structural studies of the carbohydrates of bovine milk MUC15 showed that the glycans are composed of fucose, galactose, mannose, N-acetylgalactosamine, N-acetylglycosamine, and sialic acid. The carbohydrate was shown to constitute 65% of the total molecular weight, and the molar ratios of the individual sugars to protein of the O-linked glycans were determined. The glycan structures of MUC15 were further studied by enzymatic deglycosylation experiments using different endo- and exoglycosidases as well as a panel of lectins. The N-linked glycans were shown to contain mainly hybrid-type N-glycans. In addition, the N-glycans were shown to be sialylated and contain terminal poly-lactosamine structures. The O-linked glycans were found to constitute some unsubstituted Core-1 structures and a substantial number of sialylated Core-1 O-linked glycans. By comparing the results of peanut agglutinin lectin binding, enzymatic deglycosylation, and monosaccharide composition analysis, we concluded that bovine MUC15 also contains more complex O-glycans containing high amounts N-acetylglucosamine residues. Furthermore, a small subset of the O-linked glycans is decorated with lactosamine on their terminal ends.

Inhibitory activities of bovine macromolecular whey proteins on rotavirus infections in vitro and in vivo.

Rotavirus is a major cause of infantile viral gastroenteritis and can lead to severe and sometimes lethal dehydration. Previous studies have shown that breast-fed children are better protected against symptomatic infections, and that the milk fat globule protein lactadherin might be at least partly responsible for this effect. In vitro studies have shown that human lactadherin, in contrast to the bovine ortholog, could inhibit rotavirus infectivity, and that bovine MUC1 and a commercially available bovine macromolecular whey protein (MMWP) fraction proved to be effective. The present work describes the versatility of MMWP against the infection of 2 human intestinal cell lines (Caco-2 and FHs 74 Int) by 4 different rotavirus strains (Wa, RRV, YM, RF). Isolation of a protein fraction (CM3Q3) from MMWP that effectively inhibits rotavirus infectivity in vitro is documented. Purification was achieved by monitoring the rotaviral inhibitory activity in fractions obtained from 2 consecutive steps of ion-exchange chromatography. The major component of CM3Q3 was shown to be bovine IgG, and the attenuating capacity of this fraction is most properly linked to this component. The capacity of MMWP, MUC1, lactadherin, and the CM3Q3 fraction to inhibit the infectivity of the murine EMcN rotavirus strain was analyzed in adult BALB/c mice by using 2 different amounts of virus (10 and 100 times more than 50% the viral shedding doses). Only CM3Q3 was able to significantly affect the shedding of rotavirus in the stools of experimentally infected mice when the high viral dose was given. Detection of rotavirus-specific serum antibodies showed that the high dose infected all groups of mice. Experiments with the low dose of virus implied that all the tested milk proteins could affect the viral shedding in stools; in addition, use of MUC1, MMWP, and CM3Q3 prevented the appearance of serum viral antibodies. The advantages of using bovine immunoglobulins to induce passive immunity against rotavirus have been substantially investigated, although studies have mainly focused on the use of derivatives from immunized cows, especially colostrum. This report associates considerable activity against rotavirus infectivity with an ordinary whey product, suggesting that there might be alternatives to colostral-derived products.

Transcobalamin from cow milk: isolation and physico-chemical properties.

The concentration of endogenous cobalamin (Cbl) in cow milk was 3.3 nM while the Cbl-binding capacity was 0.05 nM. Both endogenous and newly added Cbl showed similar quantitative distribution between a 280 kDa protein complex (45%) and a 43 kDa Cbl-binder (55%). Long time incubation, as well as urea treatment, was accompanied by a slow release of the 43 kDa Cbl-binder from the 280 kDa fraction. No other Cbl-binding proteins appeared after these procedures. The 43 kDa binder from cow milk, depleted of the ligand by urea treatment, reacted with Cbl even in the presence of a B12-analogue cobinamide (Cbi) at the ratio Cbl:Cbi = 1:40. The stokes radius of the binder changed from 2.7 nm for the Cbl-free protein to 2.5 nm for the Cbl-saturated form and the Cbl-saturated binder was able to displace human transcobalamin (TC) from the TC-receptor. The interaction between the protein and Cbl was significantly suppressed at pH 2.0. The N-terminal sequence of the purified 43 kDa Cbl-binder revealed homology with TC from human and rabbit plasma. In conclusion we have shown that TC is the main Cbl-binding protein in cow milk. This is surprising, since previous studies on human and rat milk have shown another Cbl-binder, apo-haptocorrin, to be the dominating Cbl-binding protein.

Structural characterization of bovine CD36 from the milk fat globule membrane.

Bovine CD36 from milk fat globule membranes was characterized and a full-length CD36 cDNA of 2772 nucleotides was isolated from a bovine mammary gland cDNA library. The deduced protein sequence contains 472 amino acid residues with 82-84% identity to the amino acid sequences of CD36 from other species. Peptides corresponding to 43% of the protein were sequenced. All eight potential N-glycosylation sites were glycosylated and the carbohydrate compositions of the individual sites were determined.

Characterization of a bovine mammary gland PP3 cDNA reveals homology with mouse and rat adhesion molecule GlyCAM-1.

A full length PP3 (Proteose-Peptone component 3) cDNA of 679 bp was isolated from a bovine mammary gland cDNA library. The cDNA encodes a signal peptide of 18 amino acids followed by the mature PP3 sequence of 135 amino acids. This polypeptide showed homology with mouse and rat GlyCAM-1 (Glycosylation dependent Cell Adhesion Molecule 1) a protein which has been shown to act as a ligand for lymphocytes. The similarity was most profound between the signal peptides and three short regions of the mature polypeptides. Additionally structural conservation was predicted by computer analysis in the shape of a C-terminal amphipathic helix. PP3 was found to be expressed in mammary gland but not in peripheral lymph nodes, Peyer's pathes, lung, spleen, heart, and muscle.

Molecular biology of Haemophilus influenzae IgA1 proteases.

IgA1 proteases of two distinct specificities were demonstrated among 95 isolates of Haemophilus influenzae and nine isolates of H. aegyptius. The two enzymes cleaved two different peptide bonds in the hinge region of the alpha chain of IgA1: a prolyl-seryl bond located at position 231-232 (type A cleavage) and a prolyl-threonyl peptide bond between residues 235 and 236 (type B cleavage). Each strain of H. influenzae produced either one or both of these types of enzymes, whereas all H. aegyptius strains produced type A enzyme only. The application of enzyme-neutralizing antibodies to the study of IgA1 proteases produced by the 104 strains of H. influenzae and H. aegyptius revealed at least 15 different types of protease activities based on inhibition patterns in nine selected antibody preparations. The types of IgA1 proteases closely correlated with the serotype of encapsulated strains of H. influenzae. The study suggests that H. influenzae strains produce at least two serologically different IgA1 proteases with distinct or identical enzymatic activities.

Posttranslational modifications of bovine osteopontin: identification of twenty-eight phosphorylation and three O-glycosylation sites.

Osteopontin (OPN) is a multiphosphorylated glycoprotein found in bone and other normal and malignant tissues, as well as in the physiological fluids urine and milk. The present study demonstrates that bovine milk osteopontin is phosphorylated at 27 serine residues and 1 threonine residue. Phosphoamino acids were identified by a combination of amino acid analysis, sequence analysis of S-ethylcysteine-derivatized phosphopeptides, and mass spectrometric analysis. Twenty-five phosphoserines and one phosphothreonine were located in Ser/Thr-X-Glu/Ser(P)/Asp motifs, and two phosphoserines were found in the sequence Ser-X-X-Glu/Ser(P). These sequence motifs are identical with the recognition sequences of mammary gland casein kinase and casein kinase II, respectively. Examination of the phosphorylation pattern revealed that the phosphorylations were clustered in groups of approximately three spanned by unphosphorylated regions of 11-32 amino acids. This pattern is probably of importance in the multiple functions of OPN involving interaction with Ca2+ and inorganic calcium salts. Furthermore, three O-glycosylated threonines (Thr 115, Thr 124, and Thr 129) have been identified in a threonine- and proline-rich region of the protein. Three putative N-glycosylation sites (Asn 63, Asn 85, and Asn 193) are present in bovine osteopontin, but sequence and mass spectrometric analysis showed that none of these asparagines were glycosylated in bovine mammary gland osteopontin. Alignment analysis showed that the majority of the phosphorylation sites in bovine osteopontin as well as all three O-glycosylation sites were conserved in other mammalian sequences. This conservation of serines, even in otherwise less well-conserved regions of the protein, indicates that the phosphorylation of osteopontin at specific sites is essential for the function of the protein.

Cloning of a cDNA encoding murine tetranectin.

A full-length cDNA encoding murine tetranectin (TN) was isolated and cloned from a murine lung lambda ZAPII cDNA library. The complete nucleotide sequence was determined revealing an open reading frame encoding 202 amino acids (aa) including a signal peptide of 21 aa. An overall aa identity of 79% exists between the deduced aa sequences of human and murine TN, revealing a high evolutionary conservation of the protein. The highest expression of mouse TN was found in lung and skeletal muscle.

Cloning of the murine tetranectin gene and 5'-flanking region.

The gene encoding murine tetranectin (Tna) and its 5'-flanking region was isolated and cloned from a EMBL3 SP6/T7 genomic library. The compiled nucleotide sequence was determined by sequencing, revealing a conserved Tna structure in man and mouse. Mapping of the transcription start point (tsp) suggests that murine Tna has more than one of these. In addition, no consensus TATA-box was found uptream for the putative tsp(s) in the 5'-flanking region of the gene, indicating that the murine Tna promoter belongs to the TATA-less class of genes. The cloned murine Tna was mapped to region F1-F3 on mouse chromosome 9 by fluorescence in situ hybridization (FISH).

The bovine PP3 gene is homologous to the murine GlyCAM 1 gene.

Proteose peptone component 3 (PP3) is a protein synthesized in the bovine mammary gland. A genomic 4.5-kb clone has been sequenced comprising a 2.9-kb PP3 transcriptional unit plus 1 kb of 5' and 0.6 kb of 3' flanking sequence. Bovine retroposons of the short interspersed nuclear element sequence class (SINES) and one microsatellite were localized in the PP3 gene. PP3 is homologous with the murine glycosylation-dependent cell-adhesion molecule 1 (GlyCAM 1) and this homology also extends to the exon/intron organization of the genes.

The amino-terminal domain of thrombomodulin and pancreatic stone protein are homologous with lectins.

Amino acid sequence alignment of the amino-terminal of thrombomodulin and pancreatic stone protein with the hepatic asialoglycoprotein receptor shows that these proteins are homologous. From the known disulfide bridge pattern of other proteins belonging to the same family two disulfide bonds can be predicted. The homology raises the question whether the amino-terminal part of thrombomodulin and the pancreatic protein binds carbohydrate or perhaps like tetranectin have a specific affinity for other proteins.

The gene structure of tetranectin, a plasminogen binding protein.

The gene for human tetranectin was isolated from a genomic library with a mixture of degenerate oligonucleotide probes. The gene is about 12 kbp and contains two intervening sequences. The gene encodes a protein of 202 amino acid residues, with a signal peptide of 21 amino acid residues, followed by the tetranectin sequence of 181 amino acid residues. Northern blot analysis revealed that tetranectin mRNA was present in all eight tissues tested with the highest concentration in lung. Southern blot analysis showed hybridization to two genes. Further investigations are needed to determine whether the genes are allelic or non-allelic.

Procathepsin D cannot autoactivate to cathepsin D at acid pH.

The amino acid sequence of the propart of bovine procathepsin D was determined at the protein level. Incubation of the isolated procathepsin D at pH 3.5-5.0 for 30-120 min leads to a 2 kDa reduction in its molecular mass, as seen by SDS-PAGE. The activation product is pseudocathepsin D and is the result of a proteolytic cleavage between LeuP26 and IleP27 in the propart. Incubation at pH 5.0 for 20 h of either procathepsin D or pseudocathepsin D results in both cases in approximately equal amounts of pseudocathepsin D and a further processed intermediate, nine amino acids shorter than pseudocathepsin D. No reaction products corresponding to cathepsin D with a mature amino terminus were observed, showing that autoproteolysis alone cannot generate the mature form found in the lysosomes.

Amino acid sequence of bovine protein Z: a vitamin K-dependent serine protease homolog.

The amino acid sequence of protein Z has been determined from sequence analysis performed on fragments obtained by chemical and enzymatic degradations. The polypeptide consists of a single chain containing 396 amino acid residues (Mr 43 677). Comparison with the vitamin K-dependent plasma proteins reveals an extensive homology. The N-terminal part, containing 13 gamma-carboxyglutamic acid and one beta-hydroxyaspartic acid residue, is extensively homologous to and of similar length to the light chain of factor X. The remainder of protein Z is homologous to the serine proteases and of similar size to the heavy chain of factor Xa, but of the active site residues only aspartic acid-102 is present. Histidine-57 and serine-195 are replaced in protein Z by threonine and alanine, respectively. The physiological function of protein Z is still uncertain.

Binding sites in fibronectin for an enterotoxigenic strain of E. coli B342289c.

The binding of fibronectin and fibronectin fragments to the enterotoxigenic strain E. coli B34289c was studied. E. coli cells bound to two distinct sites of fibronectin, one being the N-terminal domain, which also contains the binding sites for staphylococci and streptococci, and the other located within the central heparin binding region. In addition, the N-terminal and the heparin binding domain mediated the attachment of bacteria in a solid phase binding assay. E. coli cells expressed two classes of receptors, the first, a 17 kDa protein, recognized by the N-terminal fragment and the second, having a mol. mass of 55 kDa, which interacts with the internal heparin binding domain. Bacterial receptors, which bind the N-terminal end of fibronectin, may be structurally related.

Determination of the disulphide bridge arrangement of bovine histidine-rich glycoprotein.

Histidine-rich glycoprotein (HRG) was purified from bovine plasma and the disulphide bridge arrangement established. Disulphide-bridged peptides were obtained from peptic and tryptic degradation of native bovine HRG. Twelve half-cystine residues were found in bovine HRG (compared to sixteen cysteines in human HRG), all involved in the formation of six disulphide bridges connecting Cys-1 to Cys-12, Cys-2 to Cys-3, Cys-4 to Cys-5, Cys-6 to Cys-11, Cys-7 to Cys-8, and Cys-9 to Cys-10. Additional sequence analysis of 14C-carboxymethylated chymotryptic and Staphylococcus aureus V8 protease generated peptides and CNBr-fragments of bovine HRG yielded a partial amino acid sequence of bovine HRG constituting 78% of the sequence when compared to the human cDNA sequence.