Carbohydrate moieties in human secretory component.

Human secretory component has seven putative sites for N-linked glycosylation. From tryptic and Glu-C digests we have isolated peptides encompassing asparagines 65, 72, 117, 168, 403, 451 and 481. Analysis by on line HPLC-electrospray mass spectrometry indicated that these residues were fully glycosylated and that the major carbohydrate moieties were far less diversified in composition than expected. Fast atom bombardment mass spectrometry performed on oligosaccharides released by peptide-N-glycosidase F treatment of fractionated and unfractionated SC digests showed the following glycan compositions: Fuc(2)Hex(5)HexNAc(4), Fuc(3)Hex(5)HexNAc(4), NeuAcFucHex(5)HexNAc(4), NeuAcFuc(2)Hex(5)HexNAc(4), NeuAc(2)Hex(5)HexNAc4 and NeuAc(2)FucHex(5)HexNAc(4). Three of these oligosaccharides are the major carbohydrate moieties in human lactoferrin. A possible biological role of the secretory component glycans in the protection of mucosal surfaces is discussed.

Rapid structural characterisation of a murine monoclonal IgA alpha chain: heterogeneity in the oligosaccharide structures at a specific site in samples produced in different bioreactor systems.

A murine hybridoma cell line which secretes a monoclonal IgA antibody, directed against the LPS antigen of Vibrio cholerae, was grown in either a continuous stirred tank reactor, a fluidised bed reactor or a hollow fibre reactor. Different methods were used for the structural characterisation of the IgA alpha chains. A classical approach consisted of Edman sequencing and mass determination of peptides separated by reversed phase HPLC. Alternatively, peptides and glycopeptides from a tryptic digest of each alpha chain were identified directly by MALDI-TOF mass spectrometry. A detailed analysis of the oligosaccharide structures at an unique site on the alpha chain was made by labelling the oligosaccharides released by N-glycosidase F with 1-(p-methoxy)phenyl-3-methyl-5-pyrazolone. After separation by HPLC, mass measurements were made using matrix-assisted laser desorption time of flight mass spectrometry before and after digestion with specific exoglycosidases. The primary structure of the alpha chain of IgA was not affected by different cell culture conditions; in contrast, significant variations could be detected in the pattern of N-linked oligosaccharide structures, most prominently in the degree of sialylation. The efficiency of the analytical techniques in providing quality control of the identity, integrity and consistency of the glycoprotein is shown and discussed.

Thiol-disulfide redox buffers maintain a structure of immunoglobulin A that is essential for optimal in vitro binding to secretory component.

We have shown that human secretory component (SC) binds in vitro to different samples of human and murine dimeric immunoglobulin A (IgA). The binding ratio in the IgA/SC complex is 1:1. IgA which is stably bound to SC is separated from unreacted IgA by anion exchange chromatography. A part of IgA/SC complexes formed in vitro is unstable to this elution; the proportion varies between different samples of IgA; it increases following prolonged incubation of IgA at 37 degrees C. Incubation of IgA with glutathione/glutathione disulfide (GSH/GSSG) redox buffers increases the proportion able to form a stable complex with SC to approximately 90%. The presence of bound SC is not essential for this process but does allow it to occur at a lower GSH/GSSG concentration. The stable IgA/SC complex consists of a structure with a disulfide bond between IgA and SC apparently in equilibrium with a structure in which this bond is absent. The proportion bound covalently is similar for different samples of IgA and is insensitive to incubation with GSH/GSSG. It is significantly greater for secretory IgA (sIgA) and for IgA and SC incubated together with a starting mixture of cysteine/cystine. Monoclonal, antigen-specific IgA, all of which is optimally bound to SC in essentially the same way as in native sIgA, can be isolated in high yield. Our results support a mechanism for optimal binding of IgA to SC, that can occur both in vitro and in vivo, in which a thiol disulfide interchange occurs between a free IgA thiol and a sensitive SC disulfide following the initial non-covalent interaction.

Human free secretory component is composed of the first 585 amino acid residues of the polymeric immunoglobulin receptor.

The main objective of this work was to unequivocally determine the C-terminal sequence of human milk free secretory component (SC). It was found to end at arginine-585, i.e. 33 amino acids downstream from the major heterogeneous C-terminal residue previously identified for colostrum SC. In contrast, our data showed that the C-terminal end of SC was found to be homogeneous. Conflicting assignments, Asp/Gln, a missing Asn-211, Asp/Asn, Glu/Gln were corrected and found to agree with the cDNA sequence. An Ala/Val substitution at position 562 (domain VI) was identified. Its genetic significance is uncertain at present.

The secalosides, novel tumor cell growth inhibitory glycosides from a pollen extract.

The pollen of rye (Secale cereale) was shown to contain a biologically highly active family of glycosides called the secalosides. Secalosides A and B (1), both of molecular formula C46H51-NO24, were found to be epimeric esters of (2-oxo-3-indolyl)acetic acid (4). They are made up, in addition to this heterocyclic aglycon I (4), of three hexose building blocks and a carbocyclic aglycon II, which is an indan-derived dicarboxylic acid (5). In aqueous solution, secalosides A and B interchanged by epimerization at the chiral center of 4. A further epimeric pair, secalosides C and D (2), contain one additional glucose building block. Secalosides A and B, the racemic aglycon I (4), and 2-oxo-1,2,3, 4-tetrahydroquinoline-4-carboxylic acid (3), which results from 4 by hydrolytic rearrangement, exhibited significant antitumor activity against S180 sarcoma in vivo. IC50 values obtained were about 5 micrograms/mouse for the secalosides and 1 microgram/mouse for 3 and 4.

cDNA cloning of a human 100 kDa de-ubiquitinating enzyme: the 100 kDa human de-ubiquitinase belongs to the ubiquitin C-terminal hydrolase family 2 (UCH2).

The full length cDNA encoding a 100 kDa human de-ubiquitinating enzyme, referred to as de-ubiquitinase was obtained using one clone selected from a randomly sequenced human brain cDNA library and specific primers. The sequence of 18 peptides generated from the de-ubiquitinase isolated from out-dated human erythrocytes matched perfectly with the predicted amino acid sequence, which would encode a protein containing 858 amino acids (calculated M(r) = 95,743 Da). Homology search disclosed that the protein is a member of a large family of ubiquitin C-terminal hydrolases (UCH2), that was defined on the basis of the presence of two specific patterns, 'the Cys- and His-domains', which are likely to be involved in the de-ubiquitinating activity [7]. An additional conserved region, 'the aspartic acid domain', was also identified, the functional role of which is unknown.

Identification of a prostate inhibitory substance in a pollen extract.

Recently, much attention has focused on the treatment of BPH with the pollen extract, Cernilton. The present investigation was designed to identify the active component in this agent which might be responsible for the symptomatic relief of BPH as previously reported. Sequential purification of the active component present in the pollen extract was carried out by a combination of dialysis, gel filtration, and reverse phase chromatography. To monitor the biological activity of each of the purified fractions, a biological assay employing the human prostate cancer cell line DU145 was undertaken. While we have identified a number of constituent components in the pollen extract, only one fraction designated V-7 (FV-7) maintained a strong inhibitory effect on the growth of DU145 cells. The inhibition was time- and dose-dependent, and the concentrations of FV-7 required to reduce the cell numbers by 50% (IC50) after 2 days of exposure was 5 micrograms/ml. FV-7 was also inhibitory towards the primary culture of prostate stroma and epithelial cells, with the stroma/fibroblast showing greater sensitivity towards the HPLC-purified component. However, it should be noted that this inhibitory activity measured in the primary culture cells was only achieved at higher concentrations of FV-7. Preliminary characterization of the active ingredient identified FV-7 as DIBOA which is a cyclic hydroxamic acid. FV-7 and DIBOA induce similar inhibitory effects on the growth of DU145 cells.

Isolation and characterization of a cyclic hydroxamic acid from a pollen extract, which inhibits cancerous cell growth in vitro.

One fraction, designated FV-7, in the water soluble ingredient of the pollen extract Cernilton was found to be inhibitory to the growth of a prostate cancer cell line. Characterization of FV-7 by high-resolution mass spectrometry and nuclear magnetic resonance identified the fraction as hydroxamic acid, 2,4-dihydroxy-2H-1,4-benzoxazin-3(4H)-one (DIBOA). To confirm this further, we synthesized an authentic sample of DIBOA and found subsequently that the synthetic DIBOA was structurally indistinguishable from FV-7. Furthermore, in a separate experiment we compared the in vitro effects of FV-7 and DIBOA on the growth of a prostate cancer cell line and found that in both cases the effect was inhibitory and that the inhibition curves obtained for both compounds were virtually identical.

A human de-ubiquitinating enzyme with both isopeptidase and peptidase activities in vitro.

Some enzymatic and physicochemical properties of a human ubiquitin-specific isopeptidase are reported. The enzyme was purified to homogeneity from red blood cells and its specificity towards polymeric ubiquitin substrates suggests a de-ubiquitinating activity capable of cleaving 'head-to-tail' polyUb chains as well as isoamide 'branched' Ub dimers. KM values show a 10 fold preference for the cleavage of branched Ub dimers over head-to-tail Ub dimers. The enzymatic activity can be strongly inhibited by various peptides containing either of the cleavage site sequences found in Ub polymers, but not by unrelated peptides. The enzyme is monomeric under reducing conditions and exhibits a globular shape with an average diameter of 9 nm, an S20,w value of 5.2 S and a molar mass of 110 kDa +/- 10%. Because the enzyme cleaves both peptide-linked and isopeptide-linked Ub moieties from substrates, we propose to name it de-ubiquitinase rather than isopeptidase.

The nude mutation results in impaired primary antibody repertoire.

We have studied the effect of the nude mutation and/or T lymphocytes on the development of V gene germ-line repertoire in neonatal athymic (nu/nu) and euthymic (+/nu) littermates. A total of 2.35 x 10(6) and 1.47 x 10(6) B lymphocyte clones from nu/nu and +/nu neonates, respectively, were examined for the expression of select VH (J558, J606, S107, 36-60, 7183 and Q52) and Vx (1, 2, 8 and 9) gene families as well as VH (J558, S107) + Vx (1, 9) associations. Data showed that the nude mutation, whether homozygous or heterozygous, significantly affects VH and Vx gene expression as well as VH and Vx pairings and, thus, provide evidence for a defective development of B cell repertoire in both athymic nude (nu/nu) and euthymic (+/nu) mice. In addition, an analysis of 3.34 x 10(6) B lymphocyte clones from adult C57BL/6 mice showed non-stochastic association between VHJ558 + Vx1 gene families that suggests restrictions on clonal population in order to maintain homeostasis in the immune system. Studies outlined here, therefore, describe an hitherto unknown defect in the development of B lymphocyte repertoire as a result of the nude mutation which is independent of thymic dysgenesis.

Disulfide bond assignment in human J chain and its covalent pairing with immunoglobulin M.

The assignment of disulfide bonds in human J chain and its covalent pairing with immunoglobulin M was determined under conditions which minimize disulfide bond interchange. We show that in J chain the three intradisulfide bridges are formed between Cys 12 and 100, Cys 71 and 91, and Cys 108 and 133. Previous reports [reviewed by Koshland, M. E. (1985) Annu. Rev. Immunol. 3, 425-453] have proposed that cysteines 12, 14, or 68 were linked to the penultimate cysteine 575 of two mu chain tails. In this work, we demonstrate that cysteines 14 and 68 are disulfide-bridged to mu chains. A revised, albeit putative, model of J chain folding is presented which takes into account the correct disulfide pairing and the predictive secondary structure assignment.

Structural and electron-microscopic studies of jacalin from jackfruit (Artocarpus integrifolia) show that this lectin is a 65 kDa tetramer.

The 133-amino-acid sequences of the alpha-subunit of jacalin (a lectin from Artocarpus integrifolia) and of the slightly larger alpha'-subunit were determined. The alpha'- and alpha-subunits, in the approximate ratio of 1:3, were found to be virtually identical in their primary structures, except for one valine for isoleucine substitution at position 113. Although both alpha'- and alpha-chains were glycosylated, the extent of glycosylation in the alpha'-chain was much greater than that in the alpha-subunit. In the alpha'-polypeptide, all molecules contained an N-linked oligosaccharide at position 74 and some contained sugar at position 43. The alpha- and alpha'-subunits were found to be strongly non-covalently associated with three distinct beta-subunits containing 20 amino acids each. Electron-microscopic visualization of native jacalin disclosed a structure composed of four alpha-type subunits with a clear-cut 4-fold symmetry. Analytical-ultracentrifugation studies of jacalin revealed an average molecular mass of 65 kDa, a value compatible with a tetrameric structure of the alpha(alpha')-subunits. The recalculated number of sugar-binding sites per jacalin molecule, given a molecular mass of 65 kDa, would yield 0.8 sites per alpha(alpha')-promoter, i.e. about twice the value previously determined [Appukutan & Basu (1985) FEBS Lett. 180, 331-334; Ahmed & Chatterjee (1989) J. Biol. Chem. 264, 9365-9372].

Contribution of the VH11 gene family to mitogen-responsive B cell repertoire in C57BL/6 mice.

The contribution of VH11 gene family to the development of the primary B cell repertoire has been studied by analyzing 1.8 x 10(4) mitogen induced B lymphocyte colonies. The data demonstrate that VH11 family is predominantly expressed among neonatal splenic as well as adult peritoneal B cell colonies, both rich in Ly-1+ B cells. VH11 gene family expression among B splenocytes decreases during ontogeny and VH11 family pairs stochastically with different V kappa families among mitogen-activated neonatal B cell colonies, which are representative of an antigen unselected B cell repertoire. Thus, an increased VH11 expression among peritoneal and neonatal B cells points towards its biased expression among Ly-1+ B lymphocytes. The restricted V gene rearrangements and VH11-V kappa 9 pairing observed among anti-bromelain-treated mouse red blood cells autoantibodies are likely to be an outcome of both intrinsic gene recombination processes per se as well as selection by an autoantigen and/or local selective environmental factors.

cDNA clones encoding immunoglobulin lambda chains from rabbit expressing the phenotype c7.

A cDNA library derived from spleen cells of an unimmunized rabbit expressing the c7 phenotype of Ig lambda chains (c7+, c21-) was screened with V lambda or C lambda probes of a lambda light chain bearing c21 epitopes. The nucleotide sequences of three hybridizing clones were found to be identical within the V lambda, J lambda and C lambda regions. The V lambda region was 97% similar to that of the functional germ-line gene V lambda 2, and the C lambda region was identical to that of gene C lambda 6, recently identified. Gene C lambda 6 exhibited four codon differences when compared with gene C lambda 5, the latter encoding c21 epitopes. The data presented here and in the accompanying report (Jaton, J.-C. et al., Eur. J. Immunol. 1990, 20:2713) support the view that gene C lambda 6 encodes the C region of c7 lambda chains and that c7 and c21 markers designate two distinct isotypic forms of lambda chains. On the basis of comparative Southern blotting analyses and restriction maps of cloned genomic regions containing V lambda and C lambda genes, a scheme is proposed to account for the c7- and c21- phenotypes.

Chemical and immunochemical identification of the second major rabbit immunoglobulin lambda chain polypeptide bearing c7 epitopes.

The purification of lambda chain-containing IgG fraction from pooled sera of Basilea rabbits, which were bred and selected for the expression of a high level of lambda chains positive for c7 but negative for c21, was carried out. On the basis of specific binding to anti-c7 antiserum, the c7 lambda chain fraction in serum IgG was shown to account for up to 70% of the total immunoglobulin light chains, the remaining 30%, bearing the expected k2 bas isotype. Peptide mapping of the mixed light chains (lambda + k2 bas) followed by microsequencing of the constant region fragments indicated that the C lambda region originated from the cDNA sequence derived from a c7+,c21- Basilea rabbit (i.e. identical to gene C lambda 6), as described in the preceding report (Hayzer, D. J. et al., Eur. J. Immunol. 1990. 20: 2707). In addition, a synthetic peptide encompassing residues 139-159 of the constant region derived from the predicted sequence of gene C lambda 6 was shown to partially inhibit the c7-anti-c7 binding reaction in a sensitive enzyme-linked immunosorbent assay. Taken together, the chemical and immunochemical data clearly demonstrate that gene C lambda 6 indeed encodes c7 epitopes.

Organization of the murine immunoglobulin VH complex: placement of two new VH families (VH10 and VH11) and analysis of VH family clustering and interdigitation.

During B cell development, there is an ordered expression of heavy chain variable region (VH) genes during ontogeny such that JH proximal VH genes are rearranged and expressed before the more JH distal VH genes. Thus, the relative chromosomal position of VH genes is biologically significant. We have previously employed deletion mapping to order the nine described murine VH gene families as follows: 3609-J558-(J606/VGAM3-8/S107)-3660-(X24/Q52/7183 ). (Families within parentheses were not mapped relative to each other.) In this report we continue this analysis by mapping two recently described heavy chain variable region gene families (VH10 and VH11). VH10 is located at the JH proximal end of the major cluster of J558 VH gene segments. VH11 (a very small family) is intermingled with the 3660 family. Although in general VH genes are thought to be clustered, we and others have reported some interspersion between families. To further address this issue, we have analyzed 80 recombinant phage clones containing J558 VH gene segments for the presence of other VH family genes. Our data indicate that the J558 and 3609 VH families are extensively intermingled as has recently been described for the most JH proximal Q52 and 7183 families.

The amino acid sequence of rabbit J chain in secretory immunoglobulin A.

The primary structure of rabbit J chain, which occurs covalently bound to secretory IgA, was determined. J chain was isolated in its S-carboxymethylated form, in one step, by SDS/PAGE followed by electro-elution; 5 nmol of protein (approx. 75 micrograms), in all, was necessary for the determination of the complete sequence by the 'shot-gun' microsquencing technique; with the use of several site-specific endoproteinases, the various digests of S-carboxymethylated J chain were separated by micro-bore reverse-phase h.p.l.c. and the partial N-terminal sequences of all peptides were analysed. From the sequence alignment, gaps were filled by further extensive sequencing of the relevant overlapping fragments isolated from selected digests. Rabbit J chain comprises 136 amino acid residues, out of which eight are conserved cysteine residues, and is more closely similar to the human sequence (73.5% identify) than to the mouse sequence (68% identity). There is one unique glycosylation site at asparagine-48.

Variable region sequences of pathogenic anti-mouse red blood cell autoantibodies from autoimmune NZB mice.

New Zealand Black (NZB) mice spontaneously develop a severe autoimmune hemolytic anemia due to the production of anti-mouse red blood cell (MRBC) autoantibodies. The contribution of variable region genes and somatic mutations in the pathogenicity of anti-MRBC autoantibodies was investigated by mRNA sequencing of eight NZB anti-MRBC monoclonal autoantibodies, among which five are capable of inducing anemia in BALB/c mice. Here we report that at least three VH gene families (J558, J606 and 3609) and five Vchi subgroups (V chi 8, 9, 19, 21 and 28), in combination with several D, JH and Jchi gene segments, encode anti-MRBC autoantibodies. Thus, the NZB anti-MRBC autoantibodies, whether pathogenic or not, are encoded by a large number of immunoglobulin gene elements and by members of known VH and Vchi gene families with preferential usage of VH gene families most distal to the D regions. The presence of several mutations in the JH gene segments of both IgM and IgG anti-MRBC autoantibodies, whether pathogenic or not, strongly suggests that their VH regions may be highly mutated and that the mechanism of somatic diversification might be important in the generation of anti-MRBC autoantibodies. Our results support the idea that anti-MRBC autoimmune responses are likely to be generated by an antigen-driven mechanism.

Rheumatoid factor autoantibody-binding site: a molecular analysis using monoclonal antibodies with dual anti-TNP and anti-IgG activities.

Two out of five murine IgG3 anti-trinitrophenyl (TNP) monoclonal antibodies (mAb) obtained either by immunization with TNP-keyhole limpet hemocyanin (KLH) (CB1, CB5, CB6 and 4H10) or with dinitrophenyl-lipopolysaccharide (9A6), exhibited anti-IgG rheumatoid factor (RF) activity (CB6 and 4H10). The anti-IgG activity of these two anti-TNP RF was specifically inhibited by murine IgG as well as by the hapten TNP. In order to identify the structural basis for the anti-IgG activity, the nucleotide sequences encoding the VH and VL regions were determined. By comparing the V regions of the non-RF and RF anti-TNP mAb, it was found that one anti-TNP RF antibody, CB6, shares virtually identical VL and VH regions with two anti-TNP antibodies, CB1 and CB5, but markedly differs from these in the D region. Furthermore, the light chain framework region 2 (FR2) and FR3 of non-RF mAb, CB1, CB5 and 9A6, have amino acid sequences almost identical to those claimed for anti-IgG1 RF activity (Shlomchik et al., J. Exp. Med. 1986. 164: 407). Our findings suggest, at least in the case of CB6 mAb, the involvement of CDR, but not light chain FR residues, in IgG binding.

Inactivation of rabbit immunoglobulin lambda chain variable region genes by the insertion of short interspersed elements of the C family.

Two rabbit germ-line genes encoding immunoglobulin lambda light chain V regions were cloned from a rabbit genomic liver DNA library and characterized. One, V lambda 1, is separated by at least 8 kb from any other V lambda gene. The second, V lamdba 4, forms part of a three-gene cluster with two functional V lambda genes recently reported. Both V lambda 1 and V lambda 4 have structural features rendering them pseudogenes. The coding regions have frame-shift mutations which would yield defective protein products; both genes are also interrupted by the insertions of short, interspersed repetitive elements of the C family. In the V lambda 1 gene, the 369-bp insert is located upstream of the gene between the putative TATA box and the leader exon, whereas in gene V lambda 4, the 360-bp insert interrupts the FR2 at codon 48c. In addition, the sequence of the complement-determining region 3 of gene V lambda 1 is very similar to the mouse DSP2.6 sequence.