Characterization of bovine neutrophil gelatinase-associated lipocalin.

A protein of relative molecular mass of approximately 25,000 was purified from bovine colostrum by cation-exchange and size-exclusion chromatography. The N-terminus of the protein matched the sequence predicted by the National Center for Biotechnology Information for the bovine homolog of human neutrophil gelatinase-associated lipocalin, a glycoprotein of relative molecular mass 25,000 belonging to the family of lipocalins. The protein was further designated as bovine neutrophil gelatinase-associated lipocalin (bNGAL). Sodium dodecyl sulfate-PAGE of enzymically deglycosylated bNGAL indicated that the intact protein bears one N-linked glycan. Monosaccharide and mass spectrometric analyses of released N-linked carbohydrates revealed the presences of complex- and hybrid-type glycans, with galactose substituted with N-acetylgalactosamine. This substitution is typical for glycoproteins expressed in the bovine mammary gland. A specific ELISA revealed bNGAL concentrations in plasma and mature milk of about 0.05 and 1 microg/mL, respectively, whereas values as high as 51 microg/mL were measured in colostrum. Thus, we have isolated and characterized a novel bovine (milk) protein that is a new member of the lipocalin family.

Characterization of recombinant human lactoferrin secreted in milk of transgenic mice.

Human lactoferrin (hLF) is an iron-binding protein involved in host defense against infection and severe inflammation. Transgenic mice were produced harboring either hLF cDNA or genomic hLF sequences fused to regulatory elements of the bovine alphaS1 casein gene. Recombinant hLF expressed in the milk of transgenic mice (transgenic hLF) was compared with natural (human milk-derived) hLF. Immunological identity of the two forms was shown by double antibody immunoassays and the absence of an anti-hLF antibody response in transgenic mice on hyperimmunization with natural hLF. Mono S cation-exchange chromatography and N-terminal protein sequencing of transgenic and natural hLF revealed identical cationicity and N-terminal sequences. SDS-polyacrylamide gel electrophoresis and absorbance measurements of purified transgenic hLF showed this protein was 90% saturated with iron, whereas natural hLF is only 3% saturated. The pH-mediated release of iron from transgenic hLF was not different from that of iron-saturated natural hLF. Unsaturated transgenic hLF could be completely resaturated upon addition of iron. Slight differences in mobility between transgenic and natural hLF on SDS-polyacrylamide gel electrophoresis were abolished by enzymatic deglycosylation. Binding of transgenic and natural hLF to a range of ligands, including bacterial lipopolysaccharide, heparin, single-stranded DNA, Cibacron blue FG 3A, and lectins, was not different. Based on these observations, we anticipate that (unsaturated) rhLF and natural hLF will exert similar, if not identical, antibacterial and anti-inflammatory activity in vivo.

N-terminal stretch Arg2, Arg3, Arg4 and Arg5 of human lactoferrin is essential for binding to heparin, bacterial lipopolysaccharide, human lysozyme and DNA.

Human lactoferrin (hLF), a protein involved in host defence against infection and excessive inflammation, interacts with heparin, the lipid A moiety of bacterial lipopolysaccharide, human lysozyme (hLZ) and DNA. To determine which region of the molecule is important in these interactions, solid-phase ligand binding assays were performed with hLF from human milk (natural hLF) and N-terminally deleted hLF variants. Iron-saturated and natural hLF bound equally well to heparin, lipid A, hLZ and DNA. Natural hLF lacking the first two N-terminal amino acids (Gly1-Arg2) showed reactivities of one-half, two-thirds, one-third and one-third towards heparin, lipid A, hLZ and DNA respectively compared with N-terminally intact hLF. A lack of the first three residues (Gly1-Arg2-Arg3) decreased binding to the same ligands to one-eighth, one-quarter, one-twentieth and one-seventeenth respectively. No binding occurred with a mutant lacking the first five residues (Gly1-Arg2-Arg3-Arg4-Arg5). An anti-hLF monoclonal antibody (E11) that reacts to an N-lobe epitope including Arg5 completely blocked hLF-ligand interaction. These results show that the N-terminal stretch of four consecutive arginine residues, Arg2-Arg3-Arg4-Arg5, has a decisive role in the interaction of hLF with heparin, lipid A, hLZ and DNA. The role of limited N-terminal proteolysis of hLF in its anti-infective and anti-inflammatory properties is discussed.

Heterogeneity in utilization of N-glycosylation sites Asn624 and Asn138 in human lactoferrin: a study with glycosylation-site mutants.

Human lactoferrin (hLF) is a glycoprotein involved in the host defence against infection and excessive inflammation. Our objective was to determine to what extent each of the three sequons for N-linked glycosylation in hLF is actually used. Human kidney-derived 293(S) cell lines expressing recombinant hLF (rhLF) or glycosylation-site mutants were produced. The mutations involved replacement of asparagine residues with glutamine at one or more sequons for N-glycosylation (Asn138, Asn479 and Asn624). Comparative SDS/PAGE analyses of rhLF, mutated rhLF and human-milk-derived (natural) hLF led us to propose that glycosylation of hLF occurs at two sites (at Asn138 and Asn479) in approx. 85% of all hLF molecules. Glycosylation at a single site (Asn479) or at all three sites occurs in approx, 5% and 9% of hLF respectively. The extent of glycosylation at Asn624 was increased to approx. 29% and 40% of Asn479 and Asn138/479 mutant molecules respectively, which indicates that glycosylation at Asn624 in natural hLF might be limited by glycosylation at Asn479. The presence in supernatant of unglycosylated hLF (approx. 60% of the total) after mutations of Asn138 and Asn479 suggests that glycosylation of hLF is not an absolute requirement for its secretion. The pronounced degradation of unglycosylated hLF in supernatant after mutation at all three glycosylation sites (Asn138/479/624 mutant) but not after mutation at both Asn138 and Asn479 suggests that an altered conformation rather than the lack of glycosylation has rendered the Asn138/479/624 mutant susceptible to intra- and/or extra-cellular degradation.

Glycosylated and unglycosylated human lactoferrins both bind iron and show identical affinities towards human lysozyme and bacterial lipopolysaccharide, but differ in their susceptibilities towards tryptic proteolysis.

We studied the role of N-glycosylation of human lactoferrin (hLF) with respect to properties that are relevant to its antibacterial and anti-inflammatory activities. A human kidney-derived 293(S) cell line that constitutively expresses recombinant hLF (rhLF) was produced. The reactivity towards various antibodies of rhLF that had been expressed in the absence or presence of tunicamycin (which blocks N-linked glycosylation) did not differ from that of natural (human milk-derived) hLF. Cation-exchange chromatography and N-terminal protein sequencing showed identical cationic properties and an intact N-terminal sequence for rhLF and natural hLF. SDS/PAGE of rhLF expressed in the presence of tunicamycin revealed a protein with the same M(r) as that of enzymically deglycosylated natural hLF. Both glycosylated and unglycosylated rhLF appeared to be completely saturated with iron. The affinity of natural hLF, glycosylated and non-glycosylated rhLF for both human lysozyme (Kd 4.5 x 10(-8) M) and bacterial lipopolysaccharide did not differ. SDS/PAGE of hLF species subjected to trypsin indicated that unglycosylated rhLF was much more susceptible to degradation. Furthermore, this analysis suggests that N-glycosylation heterogeneity in natural hLF and rhLF resides in the C-lobe. Thus our results provide no argument for differential antibacterial and/or anti-inflammatory activity of natural and (glycosylated) rhLF and suggest that a major function of glycosylation in hLF is to protect it against proteolysis.

Induction of T-cell proliferation by recombinant mouse and chimeric mouse/human anti-CD3 monoclonal antibodies.

The isotype of anti-CD3 mAb has a dramatic effect on anti-CD3 induced T-cell activation, as was previously reported for switch variants (IgG2b to IgA) of a high-avidity IgG1 anti-CD3 mAb (CLB-T3/4.1). In order to study and compare the isotype dependency of T-cell activation with anti-CD3 mAb of various mouse and human subclasses, we now prepared recombinant anti-CD3 mAb. The variable region of the anti-CD3 Ig heavy chain was cloned, joined with genes for the heavy chain constant region and expressed in a cell line only secreting autologous mouse chi light chains. Thus we obtained cell lines that produced mouse (m) IgM, mIgG3 and chimaeric mouse/human (h) IgM, hlgG1, hlgG2, hlgG3, hlgG4, hlgE and hlgA2 anti-CD3. The matched set of mouse and mouse/human chimaeric anti-CD3 isotypes switch variants was then used to study activation of T cells in an accessory cell-dependent system. hlgG1, hlgG4, hlgE, mlgG2a and mlgE induced T-cell proliferation in PBMC of all donors tested, whereas PBMC from a subset of donors were unresponsive to stimulation with hlgG2, hlgG3, hlgA2, mlgG1 and mlgG2b anti-CD3 mAb. hlgM, mlgM and mlgA were only able to induce T-cell mitogenesis in combination with PMA. Our panel of anti-CD3 mAb variants may prove a powerful tool to study mouse and human isotype-dependent effector functions and their influence on T-cell activation requirements in detail.

Functional polymorphisms of Fc receptors in human monocyte-mediated cytotoxicity towards erythrocytes induced by murine isotype switch variants.

Human monocytes can be triggered to antibody-dependent cell-mediated cytotoxicity (ADCC) by murine antibodies. In this study, a series of H chain isotype switch variant antibodies against glycophorin A on human RBC was used to study the influence of isotype on the induction of ADCC. Furthermore, it was studied whether the functional heterogeneity in responsiveness to IgG1 and IgG2b anti-CD3 antibodies, as found among different donors in T cell proliferation induction experiments, was reflected in ADCC. Whereas IgG2a induced ADCC to the same extent in monocytes from all donors, IgG1 showed a heterogeneous pattern, which corresponded to the heterogeneity in T cell proliferation studies. IgG1 anti-CD3 nonresponder monocytes could, however, be induced to ADCC by IgG1 antiglycophorin, although they needed a much higher antibody density on the target cell than did responder monocytes. IgG2b antiglycophorin at a high density induced ADCC in monocytes from all donors irrespective of responsiveness to IgG2b anti-CD3, whereas IgE and IgA antiglycophorin were barely effective in monocytes from all donors. By specific blocking with mAb, the FcR that were involved in ADCC directed by the various isotypes were characterized. ADCC by IgG2a was predominantly mediated by FcRI and could be specifically enhanced by culturing the monocytes with rIFN-gamma. ADCC by IgG1 was predominantly mediated through FcRII in both anti-CD3 responder and nonresponder monocytes. FcRII was also involved in ADCC by IgG2b, although other receptors seemed to contribute significantly to ADCC. When FcRII or FcRI were blocked, IgG1 and IgG2a could also functionally interact with FcRI and FcRII, respectively, provided that the target cells were sensitized to a high degree. These findings indicate that FcRI and both forms of FcRII can mediate cytotoxicity and that the specificity of human FcR for murine isotypes is relative.

Murine monoclonal isotype switch variants. Detection with rat monoclonal antibodies in ELISA and isolation by sequential sublining.

Isotype switch variants, which arise in monoclonal antibody-producing cell lines, can be detected and selected on the basis of sensitive isotype-specific assays. In this study we used a series of enzyme-linked immunosorbent assays specific for murine IgG1, IgG2b, IgG2a, IgE or IgA, which permitted the detection of low frequency switch variants of hybridoma cell lines, irrespective of the specificity of the secreted antibody. In these assays two rat monoclonal antibodies were combined: one specific for the particular heavy-chain isotype, the other for the light-chain isotype, which was identical in all variants. The value of rat monoclonal antibodies for the detection of isotype switch variants is illustrated by the isolation of a series of variant antibodies specific for the CD3 complex present on human T lymphocytes.

Monoclonal anti-peroxidase isotype switch variants. Applications in studies of protein A binding and characterization of rat monoclonal antibodies.

A series of heavy chain isotype switch variants was derived from a hybridoma cell line secreting monoclonal antibodies specific for horseradish peroxidase. By the combined use of sensitive isotype-specific ELISAs and sequential sublining IgG2b, IgG2a, IgE and Iga anti-peroxidase-producing variants were successively isolated out of IgG1-secreting parental cells. The anti-peroxidase isotype variant antibodies are particularly appropriate for use in studies of the influence of heavy chain isotype in the effector functions of immunoglobulins. The use of variant antibodies with specificity for an enzyme favors their application in immunoassays because an enzyme-conjugated second antibody is not needed. Here we describe two applications of the anti-peroxidase switch variants. First, the variants are compared with respect to their affinity for Staphylococcus protein A. While IgG1 anti-peroxidase showed weak binding, both IgG2 variants strongly bound to protein A, whereas IgE and IgA variants had no affinity for protein A. Next, the switch variants were used to determine the isotype specificity of rat monoclonal antibodies generated to murine IgE.