Proteomic characterisation of leech microglia extracellular vesicles (EVs): comparison between differential ultracentrifugation and Optiprep™ density gradient isolation.

In Mammals, microglial cells are considered as the resident immune cells in central nervous system (CNS). Many studies demonstrated that, after injury, these cells are activated and recruited at the lesion site. Leech microglia present a similar pattern of microglial activation and migration upon experimental lesion of CNS. This activation is associated with the release of a large amount of extracellular vesicles (EVs). We collected EVs released by microglia primary culture and compared two different protocols of isolation: one with differential ultracentrifugation (UC) and one using an additional Optiprep™ Density Gradient (ODG) ultracentrifugation. Nanoparticles tracking analysis (NTA) and transmission electron microscopy (TEM) were used to assess vesicles size and morphology. The protein content of isolated EVs was assessed by mass spectrometry approaches. Results showed the presence of EV-specific proteins in both procedures. The extensive proteomic analysis of each single ODG fractions confirmed the efficiency of this protocol in limiting the presence of co-isolated proteins aggregates and other membranous particles during vesicles isolation. The present study permitted for the first time the characterisation of microglial EV protein content in an annelid model. Interestingly, an important amount of proteins found in leech vesicles was previously described in EV-specific databases. Finally, purified EVs were assessed for neurotrophic activity and promote neurites outgrowth on primary cultured neurons.

Compatibility in the Biomphalaria glabrata/Echinostoma caproni model: Potential involvement of proteins from hemocytes revealed by a proteomic approach.

As an approach to investigate the suspected involvement of cellular factors in Biomphalaria glabrata resistance/susceptibility to Echinostoma caproni, we compared protein patterns from hemocytes collected from susceptible and resistant snails. This proteomic approach revealed that twelve hemocytic proteins exhibited significant differences in their apparent abundance. The genes corresponding to five of them were characterized by a combination of mass spectrometry and molecular cloning. They encode an aldolase, an intermediate filament protein, a cytidine deaminase, the ribosomal protein P1 and the histone H4. Furthermore, we investigated their expression in parasite-exposed or -unexposed snails. These last experiments revealed changes in transcript levels corresponding to intermediate filament and histone H4 proteins post-infection.

Characterisation of proteins differentially present in the plasma of Biomphalaria glabrata susceptible or resistant to Echinostoma caproni.

Snail immune responses towards a trematode infection are known to rely on both plasmatic and cellular host factors. As an approach to further investigate the suspected involvement of plasmatic factors in Biomphalaria glabrata resistance/susceptibility to Echinostoma caproni, we compared protein patterns of plasma collected from susceptible and resistant snails. This proteomic approach revealed that 13 plasmatic proteins exhibited significant differences in their apparent representativity. The genes corresponding to five of them were characterised by a combination of mass spectrometry and molecular cloning. They encode two isoforms of a glycolytic enzyme, two isoforms of a calcium binding protein and an inhibitor of cysteine protease. Furthermore, we investigated gene expression in parasite-exposed or -unexposed snails as well as in various tissues by quantitative PCR. This study showed that: (i) differential representation of plasma proteins between the snail strains was correlated with a differential level of transcripts; (ii) expression of these genes after parasite exposure was differentially regulated in the two strains; and (iii) these genes were expressed predominantly in the albumen gland.

Association of ATP synthase alpha-chain with neurofibrillary degeneration in Alzheimer's disease.

Amyloid deposits and neurofibrillary tangles (NFT) are the two hallmarks that characterize Alzheimer's disease (AD). In order to find the molecular partners of these degenerating processes, we have developed antibodies against insoluble AD brain lesions. One clone, named AD46, detects only NFT. Biochemical and histochemistry analyses demonstrate that the labeled protein accumulating in the cytosol of Alzheimer degenerating neurons is the alpha-chain of the ATP synthase. The cytosolic accumulation of the alpha-chain of ATP synthase is observed even at early stages of neurofibrillary degenerating process. It is specifically observed in degenerating neurons, either alone or tightly associated with aggregates of tau proteins, suggesting that it is a new molecular event related to neurodegeneration. Overall, our results strongly suggest the implication of the alpha-chain of ATP synthase in neurofibrillary degeneration of AD that is illustrated by the cytosolic accumulation of this mitochondrial protein, which belongs to the mitochondrial respiratory system. This regulatory subunit of the respiratory complex V of mitochondria is thus a potential target for therapeutic and diagnostic strategies.

Isolation and characterization of a new peroxiredoxin from poplar sieve tubes that uses either glutaredoxin or thioredoxin as a proton donor.

A sequence coding for a peroxiredoxin (Prx) was isolated from a xylem/phloem cDNA library from Populus trichocarpa and subsequently inserted into an expression plasmid yielding the construction pET-Prx. The recombinant protein was produced in Escherichia coli cells and purified to homogeneity with a high yield. The poplar Prx is composed of 162 residues, a property that makes it the shortest plant Prx sequence isolated so far. It was shown that the protein is monomeric and possesses two conserved cysteines (Cys). The Prx degrades hydrogen peroxide and alkyl hydroperoxides in the presence of an exogenous proton donor that can be either thioredoxin or glutaredoxin (Grx). Based on this finding, we propose that the poplar protein represents a new type of Prx that differs from the so-called 2-Cys and 1-Cys Prx, a suggestion supported by the existence of natural fusion sequences constituted of a Prx motif coupled to a Grx motif. The protein was shown to be highly expressed in sieve tubes where thioredoxin h and Grx are also major proteins.

Therostasin, a novel clotting factor Xa inhibitor from the rhynchobdellid leech, Theromyzon tessulatum.

Therostasin is a potent naturally occurring tight-binding inhibitor of mammalian Factor Xa (K(i), 34 pm), isolated from the rhynchobdellid leech Theromyzon tessulatum. Therostasin is a cysteine-rich protein (8991 Da) consisting of 82 amino acid residues with 16 cysteine residues. Its amino acid sequence has been determined by a combination of techniques, including Edman degradation, enzymatic cleavage, and matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF MS) on the native and s-beta-pyridylethylated compound. Sequence analysis reveals that it shares no significant homology with other Factor Xa inhibitors except for the putative reactive site. Moreover, it contains a signature pattern for proteins of the endothelin family, potent vasoconstrictors isolated in mammal and snake venom. Therostasin cDNA (825 bp) codes for a polypeptide of 82 amino acid residues preceded by 19 residues, representing a signal peptide sequence. As for the other known inhibitors of Factor Xa, therostasin is expressed and stored in the cells of the leech salivary glands.

DNA-interacting proteins in the spermiogenesis of the mollusc Murex brandaris.

Sperm chromatin of Murex brandaris (a neogastropod mollusc) undergoes a series of structural transitions during spermiogenesis. The DNA-interacting proteins responsible for these changes as well as the mature protamines present in the ripe sperm nucleus have been characterized. The results reveal that spermiogenic nuclear proteins are protamine precursors that are subjected to a substantial number of small N-terminal deletions that gradually modify their overall charge. The composition of mature protamines is remarkably simple in turn, promoting an efficient and extremely tight packaging of DNA. The pattern of spermiogenic chromatin condensation in M. brandaris clearly departs from that corresponding to vertebrate chromatin.

NMR structures and activity of a novel alpha-like toxin from the scorpion Leiurus quinquestriatus hebraeus.

NMR structures of a new toxin from the scorpion Leiurus quinquestriatus hebraeus (Lqh III) have been investigated in conjunction with its pharmacological properties. This toxin is proposed to belong to a new group of scorpion toxins, the alpha-like toxins that target voltage-gated sodium channels with specific properties compared with the classical alpha-scorpion toxins. Electrophysiological analysis showed that Lqh III inhibits a sodium current inactivation in the cockroach axon, but induces in addition a resting depolarization due to a slowly decaying tail current atypical to other alpha-toxin action. Binding studies indicated that radiolabeled Lqh III binds with a high degree of affinity (Ki=2.2 nM) on cockroach sodium channels and that the alpha-toxin from L quinquestriatus hebraeus highly active on insects (LqhalphaIT) and alpha-like toxins compete at low concentration for its receptor binding site, suggesting that the alpha-like toxin receptor site is partially overlapping with the receptor site 3. Conversely, in rat brain, Lqh III competes for binding of the most potent anti-mammal alpha-toxin from Androctonus australis Hector venom (AaH II) only at very high concentration. The NMR structures were used for the scrutiny of the similarities and differences with representative scorpion alpha-toxins targeting the voltage-gated sodium channels of either mammals or insects. Three turn regions involved in the functional binding site of the anti-insect LqhalphaIT toxin reveal significant differences in the Lqh III structure. The electrostatic charge distribution in the Lqh III toxin is also surprisingly different when compared with the anti-mammal alpha-toxin AaH II. Similarities in the electrostatic charge distribution are, however, recognized between alpha-toxins highly active on insects and the alpha-like toxin Lqh III. This affords additional important elements to the definition of the new alpha-like group of scorpion toxins and the mammal versus insect scorpion toxin selectivities.

New toxins acting on sodium channels from the scorpion Leiurus quinquestriatus hebraeus suggest a clue to mammalian vs insect selectivity.

Two new toxins were purified from Leiurus quinquestriatus hebraeus (Lqh) scorpion venom, Lqh II and Lqh III. Lqh II sequence reveals only two substitutions, as compared to AaH II, the most active scorpion alpha-toxin on mammals from Androctounus australis Hector. Lqh III shares 80% sequence identity with the alpha-like toxin Bom III from Buthus occitanus mardochei. Using bioassays on mice and cockroach coupled with competitive binding studies with 125I-labeled scorpion alpha-toxins on rat brain and cockroach synaptosomes, the animal selectivity was examined. Lqh II has comparable activity to mammals as AaH II, but reveals significantly higher activity to insects attributed to its C-terminal substitution, and competes at low concentration for binding on both mammalian and cockroach sodium channels. Lqh II thus binds to receptor site 3 on sodium channels. Lqh III is active on both insects and mammals but competes for binding only on cockroach. The latter indicates that Lqh III binds to a distinct receptor site. Thus, Lqh II and Lqh III represent two different scorpion toxin groups, the alpha- and alpha-like toxins, respectively, according to the structural and pharmacological criteria. These new toxins may serve as a lead for clarification of the structural basis for insect vs mammal selectivity of scorpion toxins.

Disulphide bonds assignment in the inter-alpha-inhibitor heavy chains--structural and functional implications.

Human inter-alpha-inhibitor (IalphaI) is a plasma serine-proteinase inhibitor. It consists of three polypeptide chains covalently linked by a glycosaminoglycan: a light one named bikunin, carrying the antiproteinase activity and two heavy chains H1 and H2. The amino acid sequences of these heavy chains are highly similar; however when IalphaI is digested by neutrophil proteinases, their proteolytic susceptibility strongly differs [Balduyck, M., Piva, F., Mizon, C., Maes, P., Malki, N., Gressier, B., Michalski, C. & Mizon, J. (1993) Human leucocyte elastase (HLE) preferentially cleaves the heavy chain H2 of inter-alpha-trypsin inhibitor (ITI), Biol. Chem. Hoppe-Seyler 374, 895-901]. We mapped the disulphide topology of the IalphaI heavy chains in order to investigate whether or not disulphide bonds might be responsible for their differential susceptibility to proteolysis. Using amino acid sequencing and mass spectrometry analysis, we demonstrate that the H1 heavy chain contains one free thiol group and two disulphide bridges of which one links two largely spaced cysteine residues (Cys239 and Cys511). Thus H1 is clearly different from H2 which contains two disulphide bonds between closely located cysteine residues. However, using immunoprint analysis, we show that, when IalphaI is subjected to a limited digestion by Staphylococcus aureus V-8 proteinase, the two polypeptide chains are similarly susceptible to proteolysis. This enzyme preferentially cleaves the IalphaI heavy chains from their N-terminal extremity. These results are consistent with the circular dichroism (CD) analysis, suggesting that the conformation of the polypeptide backbone of H1 is not very different from that of H2, with calculated alpha-helicities of 24% and 28%, respectively. The CD measurements reveal that the aromatic amino acids of H1 and H2 are in a different asymmetrical environment. Inside the IalphaI molecule, the heavy chains are linked to the glycosaminoglycan chain via their C-terminal aspartic acid residue. Thus we suggest that the affinity of cationic neutrophil proteinases for the anionic glycosaminoglycan is responsible for the cleavage of the heavy chains (mainly H2) near their C-terminal end and the high susceptibility of IalphaI to these proteinases.

Isolation and characterization of a leech neuropeptide in rat brains: coupling to nitric oxide release in leech, rat and human tissues.

The osmoregulator peptide (leech osmoregulatory factor, LORF; IPEPYVWD) was first found in the leech central nervous system (CNS). Given the fact that certain peptides can be found in mammals and invertebrates, e.g., opioid, we examined rat brains to determine if LORF was present. This peptide was found and isolated by successive reversed-phase HPLC purification steps and characterized by electrospray mass spectrometry measurement. It was sequenced by Edman degradation and quantified in different tissues by ELISA. Our results demonstrate the presence of LORF in the hypothalamus, thalamus, and striatum (6 pmol/mg of protein extract) and in other brain areas at lower levels. This octapeptide is also present in the rat duodenum and liver (10 to 14 pmol/mg) and at lower levels in heart, lung, pancreas and caudal spinal cord (< 5 pmol/mg). The testes, adrenals and kidneys have the lowest levels of all the tissues examined (ca. 0.5 pmol/mg of protein). Furthermore, we also demonstrate that LORF is coupled to nitric oxide (NO) release in leech CNS, rat hypothalamus and human saphenous vein in a manner which is inhibited by a nitric oxide synthase inhibitor as well as an antibody directed toward LORF. The study demonstrates that LORF, and its function in relation to NO release, has been conserved over more than 400 million years of evolution.

Glycosylation pattern of human inter-alpha-inhibitor heavy chains.

Human inter-alpha-inhibitor (IalphaI) is a plasma serine-proteinase inhibitor. It consists of three polypeptide chains covalently linked by a glycosaminoglycan chain: a light chain named bikunin carrying the anti-proteinase activity and two heavy chains, H1 and H2, which exhibit specific properties, e.g. they interact with hyaluronan thus stabilizing the extracellular matrix. In this study, using matrix-assisted laser desorption ionization-time-of-flight MS and amino acid sequencing of tryptic peptides, we provide a detailed analysis of the glycosylation pattern of both heavy chains. H1 carries two complex-type N-glycans of predominantly biantennary structure linked to asparagine residues at positions 256 and 559 respectively. In contrast, the oligosaccharides attached to H2 are a complex-type N-glycan in the N-terminal region of the protein (Asn64) and three to four type-1 core-structure O-glycans mono- or di-sialylated, clustered in the C-terminal region. We propose that these O-glycans might function as a recognition signal for the H2 heavy chain. The biological implications of this hypothesis, notably for the biosynthetic pathway of IalphaI, are discussed.

The DNA-binding protein II from Zymomonas mobilis. Complete amino acid sequence and interaction with DNA.

The primary structure of the DNA-binding protein II from Zymomonas mobilis has been determined from data provided by automated Edman degradation of the intact protein and of peptides derived from cleavage at aspartic acid and arginine residues. When compared with the homologous protein isolated from other bacteria, the DNA-binding protein II from Z mobilis shows many substitutions. Several non-conservative substitutions at positions usually highly conserved in this type of protein probably account for the weaker DNA-binding activity of this protein compared to that of the E coli protein.

Nuclear basic proteins in spermiogenesis.

In animal species, spermiogenesis, the late stage of spermatogenesis, is characterized by a dramatic remodelling of chromatin which involves morphological changes and various modifications in the nature of the nuclear basic proteins. According to the evolution of species, three situations can be observed: a) persistence of somatic histones or appearance of sperm-specific histones; b) direct replacement of histones by generally smaller and more basic proteins called protamines; and c) occurrence of a double nuclear basic protein transition: histones are not directly replaced by protamines but by intermediate basic proteins which are themselves replaced by one or several protamines. However, in some species, two kinds of intermediate basic proteins can be distinguished in spermatid nuclei: transition proteins and protamine precursors. Whereas transition proteins are not structurally related either to histones or to protamines, protamine precursors are further processed at the end of spermiogenesis to give rise to the mature protamine. The molecular characteristics of the protamines as well as number of protamine types present in the spermatozoon vary from species to species. In some cases, protamine-encoding genes, although present, are not expressed to a significant level. The diversity and the precise function of intermediate basic proteins remain open to discussion. Some of them are the precursors of protamines but the mechanism, sequential or not, as well as the enzyme(s) involved in the proteolytic processing, remain to be discovered.

The amino acid sequence of the ram spermatidal protein 3--a transition protein TP3 or TP4?

As in other mammals, several nuclear basic proteins replace histones during the differentiation of germinal cells into spermatozoa in the ram. These proteins called transition proteins (TP) are later replaced by protamines. The amino acid sequence of the ram spermatidal protein 3 has been established by Edman degradation of the protein and of its fragments generated from digestion with endoproteinase Lys-C and pepsin and from the coding sequence of the gene and of the cDNA. The ram protein 3 is a basic protein of 109 residues (calculated Mr 13,200) with arginine and lysine residues uniformly distributed along the polypeptide chain. Of the 13 serine and threonine residues, 9 are located in structural motifs where they could be phosphorylated and, thus, modulate the binding of the protein to DNA. The tyrosine residues at position 33 and position 93, located in a basic environment, and the tryptophan residue at position 29 could be involved in the interactions of the protein with DNA through the stacking of their aromatic ring between the nucleotide bases. The ram protein 3 differs completely from the two well-defined transition protein families TP1 and TP2, which are also synthesised transiently during mammal spermiogenesis. In contrast with the rat TP3, ram protein 3 does not correspond to a precursor of a protamine. However, it shares structural similarities with both transition proteins TP3 and TP4 of the boar. The ram protein 3 and the boar transition proteins TP3 and TP4 probably belong to the same transition-protein group and would play similar functions in the chromatin remodelling during spermiogenesis. As protamine and transition-protein TP1 and TP2 genes from mammals, the coding sequence of the gene of ram protein 3 is interrupted by one intron but its organisation is different.

Microheterogeneity of the oligosaccharides carried by the recombinant bovine lactoferrin expressed in Mamestra brassicae cells.

The development of therapeutic glycoprotein production using the baculovirus expression system depends on the ability of insect cell lines to reproduce site specific mammalian-like N-glycans. A combination of 1H-NMR and mass spectrometry techniques (MALD-MS, ES-MS, and CID-MS-MS) allowed us to elucidate the N-linked oligosaccharides microheterogeneity on three different N-glycosylation sites, Asn233, Asn476, and Asn545, of a baculovirus-expressed recombinant bovine lactoferrin produced in Mamestra brassicae. Two families of N-glycan structures have been found: first, oligomannosidic glycans (Man[9-5]GlcNAc2) and secondly, short truncated partially fucosylated glycans (Man(3-2)[Fuc(0-1)]GlcNAc2). These results indicate that Mamestra brassicae cell line is not able to synthesize complex N-glycans, even if an alpha1,6-linked fucose residue is frequently present on the asparagine-bound N-acetylglucosamine residue of short truncated structures. Nevertheless, we have shown that Mamestra brassicae ensures the same N-glycosylation pattern as found on natural bovine lactoferrin showing the same distribution between complex and high-mannose type glycans on the different glycosylation sites. Sites which are naturally occupied by high-mannose glycans (Asn233 and Asn545) are substituted essentially by the same type of N-glycans in the recombinant counterpart, and the site Asn476,which carries sialylated complex type chains in the natural glycoprotein, is substituted by short, truncated, partially fucosylated chains in Mamestra brassicae-expressed bovine lactoferrin. These various results lead us to the conclusion that bovine lactoferrin is an interesting model to determine the potential of glycosylation of the baculovirus/insect cell expression systems.

Purification of a new dimeric protein from Cliona vastifica sponge, which specifically blocks a non-L-type calcium channel in mouse duodenal myocytes.

Marine sponges are synthesizing a wide variety of peptidic and organic molecules with biological activities. Multiple-step purification of Cliona vastifica extract led to a new dimeric peptide (mapacalcine; M(r) = 19,064) that is composed of two homologous chains, each containing nine cysteins. This protein has been found to selectively block a new calcium conductance characterized in mouse duodenal myocytes with an IC50 value of approximately 0.2 microM. The mapacalcine-sensitive current was a non-L-type calcium current activated from a holding potential of -80 mV that persisted during stimulation of the cell at high frequencies (0.1-0.2 Hz) within 5-10 min. Time constants of inactivation were similar for both L-type and non-L-type calcium currents. The non-L-type calcium current of duodenal myocytes was not blocked by the pharmacological agents specific for N-, L-, P-, or Q-type calcium channels. Mapacalcine was unable to block T-type calcium current in portal vein myocytes as well as voltage-dependent potassium currents and calcium-activated chloride currents in duodenal and portal vein cells. Mapacalcine did not affect caffeine-induced calcium responses, indicating that it did not interfere with intracellular calcium stores. Competition experiments on mouse intestinal membranes showed that mapacalcine did not interact with dihydropyridines receptors. These data suggest that mapacalcine may be a specific inhibitor of a new type of calcium current, first identified in duodenal myocytes.

Primary structure of a myohemerythrin-like cadmium-binding protein, isolated from a terrestrial annelid oligochaete.

Two isoforms of a cadmium-binding protein (Cd-BP 14a and Cd-BP 14b) were isolated from the terrestrial oligochaete annelid, Allolobophora caliginosa. The complete amino acid sequence of the major isoform Cd-BP 14a (molecular mass: 13441 Da; 119 residues) and the amino-terminal sequence (57 residues) of Cd-BP 14b were determined. The sequence of Cd-BP 14a is highly similar to that of myohemerythrins present in marine invertebrates. Furthermore, as myohemerythrins, Cd-BP 14a and Cd-BP 14b bind two atoms of iron and their ultraviolet/visible spectra are typical of non-heme iron-binding proteins. Three substitutions were found in the amino-terminal half of the proteins at positions 19, 21 and 41. The substitutions at positions 19 and 21 are conservative, whereas that at position 41 consists of the replacement of an aspartate residue in isoform a by a lysine residue in isoform b. To our knowledge, it is the first report of a protein belonging to the hemerythrin family in a terrestrial invertebrate.

Characterization of a new family of toxin-like peptides from the venom of the scorpion Leiurus quinquestriatus hebraeus. 1H-NMR structure of leiuropeptide II.

To extend our knowledge about the structural features of short scorpion toxins, the ion-exchange fractions obtained from Leiurus quinquestriatus hebraeus venom were investigated by plasma desorption mass spectrometry in order to select low molecular mass polypeptides. Three toxin-like peptides with molecular mass close to 3 kDa, named leiuropeptides I, II and III, were purified and found devoid of any significant toxicity against mammals and insects. Their amino acid sequences revealed a cysteine pattern analogous to that of short-chain scorpion toxins. The solution structure of leiuropeptide II was determined by 2D 1H-NMR spectroscopy and indicated the presence of a helix accommodating a proline, connected to a two-standard beta-sheet by three disulfide bonds. The overall fold of leiuropeptide II is found to be similar to that of leiurotoxin I, a 31-residue toxin present in the same scorpion venom which acts on K+ channels. In order to rationalize the absence of toxicity, the electrostatic potential of leiuropeptide II was compared to that of leiurotoxin I. The peptide is characterized by a large negative zone around Glu4, Asp5 and Asp8 residues, beginning in the neighbourhood of the beta-turn and extending along the helix. In the same area, leiurotoxin I exhibits a positive surface, around Arg6 and Arg13 basic residues, which are essential for its receptor affinity.

Human pre-alpha-inhibitor: isolation from a by-product of industrial scale plasma fractionation and structural analysis of its H3 heavy chain.

Pre-alpha-inhibitor (P alpha I) is a serine proteinase inhibitor from human plasma. It comprises bikunin (BK) responsible for antiprotease activity, covalently linked to a heavy chain H3. Here we describe its isolation from a side fraction of an industrial preparation of plasma clotting factors. By using a highly specific polyclonal antiserum prepared from rabbit immunized with a H3P polypeptide obtained in a bacterial expression system, we were able to identify the fractions containing P alpha I. Then, taking advantage of the differential affinity of the members of the inter-alpha-inhibitor family (I alpha I) for heparin-Sepharose and blue-Sepharose, we isolated P alpha I. Its specific antitryptic activity was 580 IU/g, higher than that of I alpha I: 420 IU/g. Its M(r), determined by sodium dodecyl sulfate polyacrylamide gel electrophoresis, with or without prior reduction, was 130,000. Its peptide chains were identified by N-terminal sequencing. The H3 heavy chain was isolated from P alpha I by alkaline dissociation and anion-exchange chromatography. Its electrophoretic mobility was compared to that of the HI and H2 heavy chains of I alpha I. In reducing conditions, it was quite similar to that of H2 (M(r) 85,000) but clearly different from that of H1 (M[r] 78,000). Thus, the so-determined apparent M(r) of H3 was overestimated since its molecular mass determined by MALDI-TOF was 74,100. This result agrees with the proposed structure for H3. Indeed, by carbohydrate analysis and PNGase F digestion, we demonstrate that the two potential N-glycosylation sites present in the core-protein (theoretical mass: 69,454) are really occupied by two N-glycans, probably of biantennary type.