Peanut allergens are rapidly transferred in human breast milk and can prevent sensitization in mice.

BACKGROUND: Food allergens have been evidenced in breast milk under physiological conditions, but the kinetic and the role of this passage in food allergies are still unclear. We then aimed to analyze the passage of peanut allergens in human breast milk and their allergenicity/immunomodulatory properties. METHODS: Human breast milk was collected from two non-atopic peanut-tolerant mothers before and at different time points after ingestion of 30 g of commercial roasted peanut. Ara h 6, Ara h 6 immune complexes, and the IgE binding capacity of breast milk samples were measured using specific immunoassays. Their allergenic functionality was then assessed using cell-based assay. Finally, human breast milk obtained before or after peanut ingestion was administered intragastrically to BALB/c mice at different ages, and mice were further experimentally sensitized to peanut using cholera toxin. RESULTS: Ara h 6 is detected as soon as 10 min after peanut ingestion, with peak values observed within the first hour after ingestion. The transfer is long-lasting, small quantities of peanut allergens being detected over a 24-h period. IgG-Ara h 6 and IgA-Ara h 6 immune complexes are evidenced, following a different kinetic of excretion than free allergens. Peanut allergens transferred in milk are IgE reactive and can induce an allergic reaction in vitro. However, administration of human breast milk to young mice, notably before weaning, does not lead to sensitization, but instead to partial oral tolerance. CONCLUSION: The low quantities of immunologically active allergens transferred through breast milk may prevent instead of priming allergic sensitization to peanut.

Anandamide excites central terminals of dorsal root ganglion neurons via vanilloid receptor-1 activation.

Recently, the cannabinoid (CB) receptor agonist anandamide (AEA) has been shown to excite perivascular terminals of primary sensory neurons via activation of the vanilloid receptor-1 (VR-1). To determine whether AEA stimulates central terminals of these neurons, via VR-1 activation, we studied the release of calcitonin gene-related peptide (CGRP)- and substance P (SP)-like immunoreactivities (LI) from slices of rat dorsal spinal cord. Mobilization of Ca(2+) in rat dorsal root ganglion (DRG) neurons in culture was also studied. AEA (0.1-10 micrometer) increased the outflow of CGRP-LI and SP-LI from slices of the rat dorsal spinal cord in a Ca(2+)-dependent manner and increased [Ca(2+)](i) in capsaicin-sensitive cultured DRG neurons. Both effects of AEA were abolished by capsaicin pretreatment and by the VR-1 antagonist capsazepine but not affected by the CB receptor antagonists AM281 or AM630. Both neuropeptide release and Ca(2+) mobilization induced by electrical field stimulation (EFS) were inhibited by a low concentration of AEA (10 nm). Inhibition by AEA of EFS-induced responses was reversed by AM281 and AM630, but was not affected by capsazepine. Results indicate that stimulation of VR-1 with high concentrations of AEA excites central terminals of capsaicin-sensitive DRG neurons, thus causing neuropeptide release in the dorsal spinal cord. This novel activity opposes the CB receptor-mediated inhibitory action of low concentrations AEA. However, only if large amounts of endogenous AEA could be produced at the level of the dorsal spinal cord, they may not inhibit, but rather activate, nociceptive sensory neurons.

Immunological studies of human constitutive cyclooxygenase (COX-1) using enzyme immunometric assay.

Polyclonal antisera and six distinct monoclonal antibodies (mAbs) were raised against constitutive cyclooxygenase (COX-1) purified from ram seminal vesicles. Immunoblotting experiments revealed that the polyclonal antisera and 4 of the mAbs strongly recognized human COX in platelet extracts. Different two-site immunometric assays of ram COX-1 were established using different combinations of mAbs. The assays were performed in 96-well microtiter plates coated with one mAb, with another mAb (covalently labeled with acetylcholinesterase (AChE)) as tracer. One combination (solid phase CX-101 + CX-105-AChE) exhibited the best sensitivity, with significant detection of concentrations as low as 23 pg/ml (0.3 fmol/ml of sheep COX-1). Unfortunately, this assay poorly cross-reacted with human COX-1 from platelet extracts. Another combination (solid phase CX-111 + CX-110-AChE) exhibited good recognition of human COX-1 but poor cross-reactivity with ram COX-1. Finally, purified anti-COX-1 IgG coated and CX-110-AChE were chosen as the best compromise since both good sensitivity (limit of detection, 113 pg/ml of ram COX-1) and significant cross-reactivity between COX-1 from both species were observed. In parallel, polyclonal antibodies were raised in rabbits against a peptide of 12 amino acids corresponding to the aminoterminal part of human COX-1. These polyclonal antibodies were affinity-purified and used in development of another two-site immunometric assay of COX-1 with CX-110-AChE as tracer. These two assays were used to analyze the COX-1 content of human platelets and cultured human umbilical vein cells (HUVEC). The results obtained with each assay were compared in terms of sensitivity and specificity. The validity of both assays was checked by analyzing platelets and HUVEC extracts previously fractionated by molecular sieve chromatography.

Differential measurement of constitutive (COX-1) and inducible (COX-2) cyclooxygenase expression in human umbilical vein endothelial cells using specific immunometric enzyme immunoassays.

We have produced and characterized monoclonal antibodies (mAbs) directed against a specific carboxyterminal sequence of human cyclooxygenase-2 (residues 580-598). A rabbit polyclonal antiserum was also raised against another sequence of 10 amino acids (residues 570-581) not present in human constitutive cyclooxygenase-1. Affinity-purified polyclonal antibodies, coated on microtiter plates, were used as capture antibodies in a two-site immunometric assay, with an mAb-acetylcholinesterase conjugate used as tracer. The detection limit was 500 fmol/ml of peptide C3-COX2 (residues 570-595). The assay was specific for the cyclooxygenase-2 (COX-2) isoform, since no immunoreactivity could be detected in platelet extracts known to be rich in cyclooxygenase-1 (COX-1). In contrast, extracts from cultured human umbilical vein endothelial cells challenged with 20 nM phorbol myristate acetate (PMA) showed an increase in COX-2 immunoreactivity related both to the increase in enzyme activity and the variations observed by Western blot analysis. Under these conditions, analysis of the same cell lysates with another immunometric assay specific for COX-1 revealed insignificant variation of this enzyme. The specificity of detection was further assessed by measuring the immunoreactivity of the fractions obtained after molecular sieve chromatography of control and stimulated cell extracts, and corroborated the marked enhancement of COX-2 by comparison with COX-1. Treatment of PMA-activated cells with H-7 or actinomycin D totally abolished the COX-2 signal and had little effect on COX-1. No significant variation in COX-2 immunoreactivity was observed using the inactive isomer 4 alpha-PMA, even at 100 nM. These assays constitute the first quantitative analysis of constitutive COX-1 and of inducible COX-2 in nucleated cells at the protein level.

Heme oxygenase-1 induction by nitric oxide in RAW 264.7 macrophages is upregulated by a cyclo-oxygenase-2 inhibitor.

Unstimulated RAW 264.7 macrophages express negligible heme oxygenase-1 (HO-1) protein but incubation with the nitric oxide (NO) donor spermine nonoate (SPNO) induced HO-1 and weakly cyclo-oxygenase-2 (COX-2) protein. This effect was potentiated by coincubation with the COX-2 selective inhibitor, SC58125. Cells incubated with SPNO showed a strong increase in HO-1 mRNA levels after 4 h with a significant potentiation in the presence of SC58125, which did not modify HO-1 mRNA stability. The induction of HO-1 by NO and its potentiation by anti-inflammatory agents may play a role in inflammatory and immune responses.

Acetylcholinesterases from Elapidae snake venoms: biochemical, immunological and enzymatic characterization.

We analyzed 45 batches of venom from 20 different species belonging to 11 genera from the 3 main families of venomous snakes (Elapidae, Viperidae and Crotalidae). We found high acetylcholinesterase (AChE) activity in all venoms from Elapidae, except in those from the Dendroaspis genus. AChE was particularly abundant in Bungarus venoms which contain up to 8 mg of enzyme per gram of dried venom. We could not detect acetylcholinesterase activity in any batch of venom from Viperidae or Crotalidae. Titration of active sites with an organophosphorous agent (MPT) revealed that the AChE of all venoms have similar turnovers (6000 to 8000 s(-1)) which are clearly higher than those of Torpedo and mammalian enzymes but lower than that of Electrophorus. AChEs from the venom of elapid snakes of the Bungarus, Naja, Ophiophagus and Haemacatus genera were purified by affinity chromatography. SDS-PAGE analysis and sucrose gradient centrifugation demonstrated that AChE is exclusively present as a nonamphiphilic monomer. These enzymes are true AChEs, hydrolyzing acetylthiocholine faster than propionylthiocholine and butyrylthiocholine and exhibiting excess substrate inhibition. Twenty-seven different monoclonal antibodies directed against AChE from Bungarus fasciatus venom were raised in mice. Half of them recognized exclusively the Bungarus enzyme while the others cross-reacted with AChEs from other venoms. Polyspecific mAbs were used to demonstrate that venoms from Dendroaspis, which contain the AChE inhibitor fasciculin but lack AChE activity, were also devoid of immunoreactive AChE protein. AChE inhibitors acting at the active site (edrophonium, tacrine) and at the peripheral site (propidium, fasciculin), as well as bis-quaternary ligands (BW284C51, decamethonium), were tested against the venom AChEs from 11 different species. All enzymes had a very similar pattern of reactivity with regard to the different inhibitors, with the exception of fasciculin. AChEs from Naja and Haemacatus venoms were relatively insensitive to fasciculin inhibition (IC50 >> 10(-6) M), while Bungarus (IC50 approximately 10(-8) M) and especially Ophiophagus (IC50 < 10(-10) M) AChEs were inhibited very efficiently. Ophiophagus and Bungarus AChEs were also efficiently inhibited by a monoclonal antibody (Elec-410) previously described as a specific ligand for the Electrophorus electricus peripheral site. Taken together, these results show that the venoms of most Elapidae snakes contain large amounts of a highly active non-amphiphilic monomeric AChE. All snake venom AChEs show strong immunological similarities and possess very similar enzymatic properties. However, they present quite different sensitivity to peripheral site inhibitors, fasciculin and the monoclonal antibody Elec-410.

Pharmacological in vitro evaluation of new substance P-cyclodextrin derivatives designed to drug targeting towards NK1-receptor bearing cells.

Some biological properties of new bifunctional conjugates designed for drug targeting were evaluated through in vitro experiments. Eight peptidylcyclodextrin compounds were used, which correspond to modified beta- or gamma-cyclodextrin (CD) grafted on neuropeptide substance P (SP) or a shorter derivative (SP(4-11)). Using anti-SP and anti-CD antibodies as molecular probes, we showed that the main structural features of the two moieties of these adducts were preserved. Binding experiments, using CHO cells expressing the human SP-specific NK1 receptor, demonstrated the functionality of all peptidylcyclodextrin derivatives, which exhibited IC50 values in a 10(-9)-10(-7) M range. All compounds were able to induce a pharmacological response, triggering phosphatidylinositol turnover with EC50 values in the same range as the natural ligand. Moreover, autoradiography analysis of rat spinal corn sections proved that [125I]SP binding was dose-dependently displaced by one selected compound (a gamma-CD-SP), showing a similar affinity of this adduct for the rat neurokinin 1 receptor. Our observations demonstrate that these peptidylcyclodextrins efficiently target NK1 receptor-expressing cells.

Homologous and heterologous phosphorylation of the vasopressin V1a receptor.

The vasopressin V1a receptor undergoes homologous and heterologous desensitizations which can be mimicked by activation of protein kinase C. This suggests that phosphorylation of the V1a receptor may be involved in the desensitization mechanisms. Such a phosphorylation was presently investigated in HEK 293 cells stably transfected with rat vasopressin V1a receptor. Metabolic labelling and immunoprecipitation of epitope-tagged V1a receptor evidenced a 52-kDa band and a 92-kDa band. Glycosidase treatments and immunoblotting experiments suggest that the 52-kDa band corresponds to an immature unprocessed receptor protein, whereas the 92-kDa band would correspond to a highly glycosylated form of the mature V1a receptor. Exposure of the cells to vasopressin induced a selective 32P phosphate incorporation in the 92-kDa form of the receptor. This homologous ligand-induced phosphorylation was dose dependent with maximal phosphate incorporation corresponding to four times the basal level. Stimulation of the endogenous phospholipase C-coupled m3 muscarinic receptor by carbachol-induced heterologous phosphorylation of the V1a receptor whose amplitude was half that of the homologous phosphorylation. This heterologous phosphorylation was associated with a reduced vasopressin-dependent increase in intracellular calcium.

Lack of IL-10 synthesis by murine alveolar macrophages upon lipopolysaccharide exposure. Comparison with peritoneal macrophages.

The central role of alveolar macrophages in the establishment of lipopolysaccharide (LPS)-induced lung inflammation is well demonstrated. They produce and release numerous proinflammatory molecules, among which is tumor necrosis factor alpha (TNF-alpha), a cytokine responsible in part for the neutrophilic alveolitis. Interleukin-10 (IL-10) produced by LPS-activated mononuclear phagocytes is a major anti-inflammatory cytokine that down-regulates TNF-alpha synthesis. We studied the ability of murine alveolar macrophages to produce IL-10 in vivo and in vitro, in response to LPS. Unexpectedly, the IL-10 protein was not detected in the whole lung and airspaces after LPS intranasal instillation. In addition, no IL-10 protein was found in supernatants of isolated and LPS-stimulated alveolar macrophages. The lack of IL-10 synthesis was confirmed by the absence of specific RNA transcripts. By contrast and as expected, autologous peritoneal macrophages produced IL-10 upon LPS challenge. Drugs that usually modify the TNF-alpha/IL-10 balance in favor of IL-10 were used without success. Thus, maneuvers allowing an increase in intracellular cAMP concentrations did not reverse this unexpected phenotype. Moreover, direct activation of protein kinase C with PMA was unable to trigger IL-10 formation by alveolar, by contrast to peritoneal, macrophages. The current findings describe a specific phenotype for murine alveolar macrophages during LPS-induced inflammation.

Importance of hydropathic complementarity for the binding of the neuropeptide substance P to a monoclonal antibody: equilibrium and kinetic studies.

In a previous study (Boquet et al., 1995, Molec. Immunol. 32, 303-308) we observed remarkable inversions of hydropathic profiles between complementarity determining regions (CDRs) of an anti-substance P monoclonal antibody (SP31) and the corresponding 5 amino acid C-terminal peptide epitope. Here we demonstrate the importance this hydropathic complementarity by measuring the immunoreactivity of SP-related peptides which have been modified in their C-terminal parts so that hydropathic profile has been conserved (by substituting amino acids in the epitope) or modified (by mixing the sequence of amino acids in the epitope). Experiments performed in equilibrium conditions using a competitive enzyme immunoassay showed that most of the peptides conserving the hydropathic profile of SP epitope were recognized by mAb SP31 (even if marked variations in affinity were observed), while those for which the hydropathic profile was modified exhibited very low or undetectable affinity. The kinetic parameters (ka and kd) of peptide-antibody interactions were determined using Surface Plasmon Resonance technology (BIACORE 2000). These measurements showed that all peptides recognized by mAb SP31 had similar association rate constants (close to 2 x 10[5] M[-1] s[-1]), differences in binding affinities essentially resulting from differences in dissociation rate constants (ranging from 1.61 x 10[-4] to 1.15 x 10[-1] s[-1]). From these results, it was concluded that hydropathic complementarity between the epitope and the paratope could be a necessary but not sufficient condition for establishing high-affinity binding. We hypothesize that hydropathic interactions may play a critical role during the first contacts between antibody CDRs and the peptide, possibly by favouring reorganization of water molecules at the antibody-peptide interface.

A monoclonal antibody directed against the neurokinin-1 receptor contains a peptide sequence with similar hydropathy and functional properties to substance P, the natural ligand for the receptor.

Monoclonal antibody (mAb) PS12, obtained using the complementary peptide methodology, mimics the neuropeptide substance P (SP) in recognizing the SP-binding domain of the neurokinin-1 receptor (NK1R) and eliciting production of polyclonal antibodies cross-reacting with SP with a high affinity (Déry et al., 1997. J. Neuroimmunol. 76, 1-9). The aim of the present study was to investigate which structural features of mAb PS12 might account for this molecular mimicry. Cloning and sequencing of variable regions of both light (VL) and heavy (VH) chains of this 'SP-like' antibody did not indicate any primary sequence homology between SP and any antibody region. Instead, they revealed a striking similarity between the hydropathic profile of SP and that of an 11-amino-acid region in the light chain encompassing the second complementarity determining region (CDR2). When applied to CHO cells expressing the human NK1R, a synthetic extended 17-amino-acid peptide (denoted CDR2L) corresponding to this VL region inhibited the high-affinity binding of radiolabeled SP and antagonized the SP-induced inositol phosphate production. Moreover, a re-examination of the sequences of several antibodies that previously served in the design of CDR-derived bioactive peptides indicated that these antibodies also carried the hydropathic image of the respective ligands that they mimic. In agreement with previous observations on artificial synthetic peptides, our data thus suggest that the molecular mimicry between natural proteins (i.e. antibody and hormone, for example) could be understood on a structural level directly related, at least in part, to hydropathic homology. These results could then guide the search for bioactive paratope-derived peptides of potential pharmacological interest. We also observed inverse hydropathy between multiple CDRs of mAb PS12 (including CDR3H and CDR3L) and the peptide epitope, confirming the importance of hydropathic complementarity in antigen-antibody interactions.

Molecular basis of IgE cross-reactivity between human beta-casein and bovine beta-casein, a major allergen of milk.

Twenty patients allergic to cow's milk proteins and with high levels of specific IgE directed against bovine whole casein were selected to evaluate reactivity of their IgE antibodies with human beta-casein. Highly purified human and bovine beta-caseins were prepared by selective precipitations and FPLC separation. Their identity and purity were assessed by HPLC, analysis of amino acid composition, sequencing of the five N-terminal amino acid residues and immunochemical tests. Direct and indirect ELISAs were performed using human and bovine beta-casein coated into microtiter plates and monoclonal anti-human IgE antibody AChE labelled for revelation. Seven sera contained specific IgE directed against human beta-casein. Inhibition studies using native human and bovine beta-caseins as well as bovine beta-casein-derived peptides demonstrated that, depending on the sera, one or several common epitopes located in different parts of the molecule were shared by the two homologous proteins.

Expression of cyclooxygenase-1 and -2 in ovine endometrium during the estrous cycle and early pregnancy.

In this study we investigated the expression of the two cyclooxygenases, cox-1 and -2, in sheep uterine tissues during the estrous cycle and early pregnancy. We identified the cox-2 isoform in the ovine uterus by Western blot and demonstrated that the two cyclooxygenases exhibited different patterns of expression. Cox-1 was expressed at steady state levels in the endometrium during the estrous cycle and comparable stages of pregnancy. In contrast, cox-2 was highly and transiently expressed from days 12-15 of the estrous cycle and declined thereafter to undetectable levels. Endometrium from early pregnant ewes showed a similar pattern of cox-2 expression, although there was a slower decrease beyond day 15. Immunohistochemical studies demonstrated that cox-1 was localized in both epithelial and stromal cells, whereas cox-2 was localized solely in the luminal epithelium and to a lesser extent in the superficial glands. Treatment of ovariectomized ewes with steroids indicated that expression of cox-1 remained at constant levels whatever the treatment. In contrast, endometrial cox-2 was highly induced by a 10-day progesterone treatment. Estradiol slightly increased cox-2 expression but only after progesterone priming. Collectively these results suggest that the developing ability of the uterus to synthesize PGs is due to the induction of cox-2.

Characterization of human NOV in biological fluids: an enzyme immunoassay for the quantification of human NOV in sera from patients with diseases of the adrenal gland and of the nervous system.

Immunochromatography has shown that human NOV (NOVH), a member of the CCN (CTGF/CYR61/NOV) family, forms a physiological complex with fibulin-1 in blood. We developed an enzyme immunoassay specific for NOVH and showed for the first time that the concentration of NOVH differs in each of these biological fluids. The normal concentration of NOVH circulating in the blood is 350-400 ng/ml, but this concentration varies with age. By using sera from patients with adrenal gland diseases we found that in vivo ACTH or glucocorticoids are not responsible for the high concentration of NOVH in this endocrine gland. However, the NOVH concentration was significantly modified in malignant adrenocortical tumors, but not in benign adrenocortical tumors. The concentration of NOVH was significantly decreased in patients suffering from astrocytomas or multiple sclerosis, two diseases of the nervous system. Thus, NOVH is a potentially useful marker for the diagnosis of these diseases.

Distribution of a negative regulator of haematopoietic stem cell proliferation (AcSDKP) and thymosin beta 4 in mouse tissues.

A competitive enzyme immunoassay using acetylcholinesterase as tracer for thymosin beta 4, has been developed. Using this assay and a previously described EIA for AcSDKP, a negative regulator of pluripotent haematopoietic stem cell proliferation, the levels of these two peptides were determined in mouse tissue extracts. The combination of EIAs with different HPLC procedures validated these methods and clearly demonstrated the ubiquity of these peptides in mouse tissues. Similar results are reported for rabbit thymus which suggest different hypotheses for AcSDKP biosynthesis.

Post-translational phosphorylation affects the IgE binding capacity of caseins.

IgE response specific to those molecular regions of casein that contain a major phosphorylation site was analyzed using native and modified caseins and derived peptides. This study included (i) the naturally occurring common variants A1 and A from beta- and alphas2-caseins, respectively, which were purified in the native form and then dephosphorylated, (ii) a purified rare variant D of alphas2-casein which lacks one major phosphorylation site, and (iii) the native and dephosphorylated tryptic fragment f(1-25) from beta-casein. Direct and indirect ELISA using sera from patients allergic to milk showed that the IgE response to caseins is affected by modifying or eliminating the major phosphorylation site.

Role of peptide substrate structure in the selective processing of peptide prohormones at basic amino acid pairs by endoproteases.

Three putative processing enzymes, each with defined action in a prohormone system, a 'pro-ocytocin-neurophysin convertase' from bovine neurohypophysis secretory granules, a 'Leu-enkephalin Arg6 generating enzyme' from human CSF and the endoprotease from the 'S-28 convertase' complex of rat brain cortex, were tested for their ability to hydrolyze peptides deriving from pro-ocytocin, pro-enkephalin B and pro-somatostatin, respectively at pairs of basic amino acids. The observations suggest that structural parameters specified by the peptide region around the dibasic moieties govern recognition by the enzyme and define which peptide bond is hydrolyzed.

Identification in the NK1 tachykinin receptor of a domain involved in recognition of neurokinin A and septide but not of substance P.

The three mammalian tachykinins, substance P (SP), neurokinin A (NKA) and neurokinin B (NKB), exert their physiological effects through specific receptors, NK1, NK2 and NK3, respectively. However, homologous binding studies have recently demonstrated that, contrary to the generally accepted belief, NKA could bind NK1 receptor with high affinity (Hastrup and Schwartz, 1996). Using COS-7 cells expressing the human NK1 receptor, we show that two simultaneous point mutations (E193L and V195R) in a restricted five amino acid sequence (the (193-197) region), selected because of its hydropathic complementarity with the common C-terminal extremity of tachykinins, abolish both the high-affinity binding and highly potent biological activity of NKA, without affecting those of SP. In addition, the same mutations also suppressed the high functional activity of septide, a synthetic SP atypical agonist ([pGlu6-Pro9] SP 6-11). These results suggest that the (193-197) region, located at the end of the second extracellular loop of the receptor, could be part of a common high-affinity binding domain for both NKA and septide, distinct from the SP binding site.

Role of the carboxyl-terminal region, di-leucine motif and cysteine residues in signalling and internalization of vasopressin V1a receptor.

The structural requirements for internalization and signalling of the vasopressin V1a receptor were investigated in stably transfected HEK-293 cells. Removal of the 51 C-terminal amino acids did not affect vasopressin binding, calcium signalling, heterologous desensitization or internalization of the receptor. Deletion of 14 additional amino acids reduced vasopressin-dependent calcium increase and impaired receptor internalization. Substitution of cysteines 371-372 did not affect intracellular signalling, but decreased endocytosis by 26%. Substitution of the 361-362 leucine by alanine residues reduced by 56% V1a receptor sequestration without affecting calcium signalling. These results indicate that di-cysteine and mostly di-leucine motifs present in the C-terminal region of the V1a receptor are involved in its internalization.

Acetylcholinesterase from Bungarus venom: a monomeric species.

The venom of Bungarus fasciatus, an Elapidae snake, contains a high level of AChE activity. Partial peptide sequences show that it is closely homologous to other AChEs. Bungarus venom AChE is a non-amphiphilic monomeric species, a molecular form of AChE which has not been previously found in significant levels in other tissues. The composition of carbohydrates suggests the presence of N-glycans of the 'complex' and 'hybrid' types. Ion exchange chromatography, isoelectric focusing and electrophoresis in non-denaturing and denaturing conditions reveal a complex microheterogeneity of this enzyme, which is partly related to its glycosylation.