Dedication: This manuscript is dedicated to the memory of the late Professor Gabor N. Than, whose inspiring leadership of his research team will be remembered forever.
Functional analyses of placental protein 13/galectin-13
Article first published online: 16 FEB 2004
European Journal of Biochemistry
Volume 271, Issue 6, pages 1065–1078, March 2004
How to Cite
Than, N. G., Pick, E., Bellyei, S., Szigeti, A., Burger, O., Berente, Z., Janaky, T., Boronkai, A., Kliman, H., Meiri, H., Bohn, H., Than, G. N. and Sumegi, B. (2004), Functional analyses of placental protein 13/galectin-13. European Journal of Biochemistry, 271: 1065–1078. doi: 10.1111/j.1432-1033.2004.04004.x
- Issue published online: 16 FEB 2004
- Article first published online: 16 FEB 2004
- (Received 8 December 2003, revised 14 January 2004, accepted 20 January 2004)
- brush border membrane;
- carbohydrate binding;
- placental protein
Placental protein 13 (PP13) was cloned from human term placenta. As sequence analyses, alignments and computational modelling showed its conserved structural and functional homology to members of the galectin family, the protein was designated galectin-13. Similar to human eosinophil Charcot–Leyden crystal protein/galectin-10 but not other galectins, its weak lysophospholipase activity was confirmed by 31P-NMR. In this study, recombinant PP13/galectin-13 was expressed and specific monoclonal antibody to PP13 was developed. Endogenous lysophospholipase activity of both the purified and also the recombinant protein was verified. Sugar binding assays revealed that N-acetyl-lactosamine, mannose and N-acetyl-glucosamine residues widely expressed in human placenta had the strongest binding affinity to both the purified and recombinant PP13/galectin-13, which also effectively agglutinated erythrocytes. The protein was found to be a homodimer of 16 kDa subunits linked together by disulphide bonds, a phenomenon differing from the noncovalent dimerization of previously known prototype galectins. Furthermore, reducing agents were shown to decrease its sugar binding activity and abolish its haemagglutination. Phosphorylation sites were computed on PP13/galectin-13, and phosphorylation of the purified protein was confirmed. Using affinity chromatography, PAGE, MALDI-TOF MS and post source decay, annexin II and beta/gamma actin were identified as proteins specifically bound to PP13/galectin-13 in placenta and fetal hepatic cells. Perinuclear staining of the syncytiotrophoblasts showed its expression in these cells, while strong labelling of the syncytiotrophoblasts' brush border membrane confirmed its galectin-like externalization to the cell surface. Knowing its colocalization and specific binding to annexin II, PP13/galectin-13 was assumed to be secreted to the outer cell surface by ectocytosis, in microvesicles containing actin and annexin II. With regard to our functional and immunomorphological results, PP13/galectin-13 may have special haemostatic and immunobiological functions at the lining of the common feto-maternal blood-spaces or developmental role in the placenta.
carbohydrate recognition domain
placental protein 13
PP13 purified from placenta
post source decay
Placental protein 13 (PP13) is a member of the group of the so-called ‘pregnancy-related proteins’ that might be highly expressed in placenta and some maternal/fetal tissues during pregnancy. The structural and functional characteristics of these proteins and their possible role in placental development and regulation pathways are receiving increased interest at present. PP13 was first isolated from human placenta and characterized by Bohn et al. in 1983. It was found to be comprised of two identical 16 kDa subunits held together by disulfide bonds, and to have the lowest carbohydrate content (0.6%) of any known placental proteins . Later, cloning of PP13 was performed in parallel by two research groups [3,4], and its sequence was deposited separately at the GenBank database (AF117383, AY055826). At that time, sequence analysis and alignment showed that PP13 shared the highest homology to human eosinophil Charcot–Leyden crystal (CLC) protein/galectin-10 , and similarly to CLC, PP13 purified from human placenta (PP13-B) showed weak lysophospholipase (LPLA) activity . However, conserved structural identity of PP13 to the members of the galectin family was also found . Subsequently, computational 3D modelling based on its primary structure and homology to prototype galectins  revealed a characteristic ‘jellyroll’ fold (deposited to Brookhaven Data Bank, Accession No. 1F87), a single conserved carbohydrate recognition domain (CRD) and predicted sugar binding capabilities of PP13, and it was therefore designated as galectin-13 .
As several galectins have recently proved to be very closely related to PP13/galectin-13 [8,9], and there were also some incongruities in its tissue expression in studies performed by polyclonal antibodies to PP13 and PP13 cDNA , more powerful, specific monoclonal antibodies to PP13 had to be developed. By the expression of recombinant PP13 protein (PP13-R), it became possible to perform more detailed functional studies on the protein. Because LPLA activity of CLC protein/galectin-10 has recently been assigned to its interaction with putative eosinophil LPLAs or their known inhibitors , elucidation of intrinsic LPLA, phospholipase (PLA) or sugar binding activities of PP13/galectin-13 had to be reconsidered. In this study, immunoaffinity purification and mass spectrometry (MS) studies indicated the binding of PP13/galectin-13 to proteins involved in phospholipid metabolism and cytoskeletal functions, but no intracellular LPLA was detectably bound to it. On the other hand, intrinsic LPLA activity for not only the purified PP13-B, but also the bacterially expressed PP13-R was confirmed. With sugar binding assays, the results of previous predictions on the sugar binding specificity of its CRD  were strongly underlined. In contrast to other known prototype galectins, PP13/galectin-13 was found to be a homodimer linked by disulphide bonds. Unlike most thiol-dependent galectins, reducing agents were shown to decrease its sugar binding activity and abolish its haemagglutination. In addition, putative phosphorylation sites were computed, and phosphorylation of the purified protein was empirically proved.
As not only the information on their structural and carbohydrate binding characteristics of galectins, but also their exact morphological localizations in cells and tissues are essential for the understanding of their interaction with glycoconjugates and diverse biological functions, to obtain better insight into the physiological role and involvement in placental development and functions of PP13/galectin-13, as well as its predicted role in different pregnancy complications , a detailed immunolocalizational study was also performed. Its in vitro characterization in a collaborative study between the leading groups of PP13/galectin-13 research adequately revealed the putative physiological functions of the protein, and gave a possible hypothesis for its importance in placental developmental processes and its conjunction with fetal haemopoetic tissues.
PP13 antigen denoted here as PP13-B (Op. 234/266) and rabbit polyclonal antibody to PP13 (160 ZB) was prepared by H. Bohn (Behringwerke AG, Marburg/Lahn, Germany). NSO/1 myeloma cell line was kindly provided by C. Milstein (MRC, Cambridge, UK). We used anti-annexin II rabbit polyclonal IgG (Santa Cruz Biotechnology, Santa Cruz, CA, USA), fluorescein isothiocyanate (FITC) labelled anti-mouse IgG (Molecular Probes, Eugene, OR, USA) and FITC-labelled anti-rabbit IgG (BD Pharmingen, San Diego, CA, USA). We obtained WRL-68 human fetal hepatic cells (ATCC, Manassas, VA, USA); D2O (Isotec Inc., Miamisburg, OH, USA); pUC57-T vector (MBI Fermentas, St. Leon-Rot, Germany); pQE30 vector, M15 (pREP4) Escherichia coli and Ni-nitrilotriacetic acid column (Qiagen Inc., Valencia, CA, USA); Protein A column (Affiland, Ans-Liege, Belgium); bicinchoninic acid reagent (Pierce Biotechnology Inc., Rockford, IL, USA); ECL chemiluminescence system (Amersham Pharmacia Biotech, Buckinghamshire, UK); DRAQ5 dye (Biostatus Ltd, Shepshed, UK); Universal Kit (Immunotech, Marseille, France); Pro-Q Diamond phosphoprotein gel staining kit (Molecular Probes, Eugene, OR, USA); trypsin (Promega GmbH, Mannheim, Germany); ZipTipC18 pipette tips (Millipore, Bedford, MA, USA). N-acetyl-d-lactosamine, l-fucose, galactose, glucose, lactose, maltose, mannose, N-acetyl-d-galactosamine, N-acetyl-d-glucosamine; cyanogen-bromide activated sepharose 4B, l-fucose-agarose, glucose-agarose, lactose-agarose, maltose-agarose, mannose-agarose, N-acetyl-d-galactosamine-agarose, N-acetyl-d-glucosamine-agarose; 1,2-dioleoyl-sn-glycero-3-phosphocholine, 1,2-dioleoyl-sn-glycero-3-phospho-l-serin, l-phosphatidylinositol, l-phosphatidyl-ethanolamine; l-α-1-lysophosphatidylcholine, lysophosphatidylethanolamine, l-α-1-lysophosphatidylinositol, l-α-1-lysophosphatidyl-l-serin; isopropyl thio-β-d-galactoside (IPTG); antibiotic-antimycotic solution, bovine serum albumin (BSA), Dulbecco's modified Eagle's medium (DMEM), fetal bovine serum, N-(2-hydroxyethyl)piperazine-N-(2-ethanesulfonic acid) (Hepes), phenylmethylsulfonyl fluoride and horseradish peroxidase labeled anti-rabbit and anti-mouse IgGs were purchased from Sigma-Aldrich Co. (St. Louis, MO, USA).
PP13/galectin-13 cDNA and amino acid sequences were compared to various EST, genomic and protein databases by blast at NCBI (Bethesda, MD, USA) . Multiple sequence alignments were carried out with clustalw at EMBnet (Lausanne, Switzerland) . The PROSITE  and NetPhos  databases were searched for biologically significant patterns and putative phosphorylation sites. The carbohydrate binding moiety and cysteine residues potentially involved in intermolecular cross-linking were localized on the 3D model of PP13/galectin-13 (PDB 1F87) with rasmol.
Construction of bacterial PP13/galectin-13 expression plasmids
Full length PP13/galectin-13 cDNA was isolated by the standard RACE method [17,18] using 4 µg of total placental RNA and specific primers. The resulting PCR fragments were inserted into pUC57-T cloning vector. Insert-containing clones were selected and sequenced by automated DNA sequencing at the Biological Services of the Weizmann Institute (Rehovot, Israel). Subsequently, the whole open reading frame of the cDNA containing the consensus Kozak sequence at its 5′ end  was PCR amplified with (5′-CGATACGGATCCATGTCTTCTTTACCCGTGC-3′) and (5′-TAAGTCGAGCTCATTGCAGACACACACTGAGG-3′) primers. The resultant PCR product was cloned into the BamHI and Sac1 sites of the pQE30 expression vector.
Expression and purification of recombinant PP13/galectin-13
The PP13-R/pQE30 expression vector was transformed into M15 (pREP4) Escherichia coli host strain and the bacteria were induced with IPTG. The expressed protein was subsequently purified with Ni-nitrilotriacetic acid column in the presence of the His6-tag. The primary structure and purity of PP13-R was verified by sequence analysis  and by immunoblotting with both polyclonal and monoclonal antibodies to PP13. The specific antibody recognition of both PP13-R and PP13-B were investigated by sandwich ELISA performed with two different monoclonal anti-PP13 IgGs described below.
Preparation of monoclonal antibodies to PP13/galectin-13
Monoclonal antibodies to PP13 were produced at the Hybridoma Center of the Weizmann Institute. Female Balb/c mice (Jackson Laboratory, Bar Harbor, ME, USA) were immunized with 0.05 mg PP13-B. Hybridomas were prepared from mice spleen cells by hybridizing with NSO/1 myeloma cells as described previously . Cells were screened by direct ELISA using PP13-B as antigen. Anti-PP13 Ig producing clones were subsequently injected intraperitoneally into mice. Antibodies were isolated from the ascitic fluid, purified on Protein A column and checked for subclass and protein content by immunoblots and sandwich ELISA.
PP13/galectin-13 lysophospholipase and phospholipase activity detection by NMR
PP13-B purified from placenta and bacterially expressed PP13-R (20 µg each) were dissolved in 500 µL aqueous solutions (200 mm Hepes, 5.0 mm CaCl2 and 130 mm NaCl, pH 7.4) of 5.0 mg·mL−1 of the different lysophospholipids listed in Materials. Aliquots without PP13 proteins were used as controls. The solutions were prepared and stored at 37 °C in 5 mm (outside diameter) NMR tubes and their 31P-NMR spectra were recorded at various time intervals. During NMR measurements a 2 mm (outside diameter) insert tube filled with D2O was placed in the NMR tubes. To detect phospholipase activity of PP13-B and PP13-R, 7.2 mg·mL−1 of the phospholipids listed in Materials were used, and 25 µL Triton X-100 was added to the aliquots to enable dissolution of the substrate. 31P-NMR spectra were obtained on a Varian UNITYINOVA 400 WB spectrometer at 161.90 MHz, 37 °C. Proton decoupling provided 128 transients, using 30 °C flip angle pulses with 3.4 s delays and a 0.6 s acquisition time, in order for the peak integrals to represent the relative concentrations of the phosphorous containing species. The chemical shifts were referred to the deuterium resonance frequency of the D2O in the insert tube. The relative concentrations (in molar fractions) of the species observed during the whole course of the study were determined by deconvolution of the spectra, using the routine built into the NMR software (vnmr 6.1 b; Varian Inc., Palo Alto, CA, USA).
PP13/galectin-13 sugar binding assays
Binding of PP13-R to different sugars was studied essentially as described in , but protein binding was followed by the endogenous fluorescence of PP13/galectin-13 (excitation at 280 nm, emission at 360 nm). PP13-R (50 µg) was dissolved in 200 µL sodium phosphate buffer (50 mm, pH 7.3, containing 0.15 m NaCl, 20 mm EDTA) and added to 50 µL activated sugar-coupled agarose beads as listed in Materials. In parallel experiments, 1 mm dithiothreitol was also added to the mixture. The solutions were incubated in 0.5 mL microtubes at 37 °C for 1 h with vigorous shaking. Tubes were then centrifuged at 10 000 gfor 20 s to sediment agarose beads. For quantification of unbound PP13-R, fluorescence of the supernatants was determined in a protein concentration range of 2–100 µg·mL−1, measured by an LS50B PerkinElmer Luminescence Spectrometer (Shelton, CT, USA). For controls, uncoupled agarose beads (Sepharose 2B) were used. After removing the unbound PP13-R, specifically bound PP13-R was eluted with different sugars in different concentrations (1 mm−1 m) and fluorescence of the supernatants was measured by the same method. For positive controls, PP13-R (50 µg) was dissolved in buffer, for negative controls only buffer was used.
PP13/galectin-13 haemagglutination assay
Lectin activity of both PP13-B and PP13-R was determined by measurement of their capabilities to agglutinate human erythrocytes. Agglutination assays were performed in a 96 well microtiter plate with serial twofold dilutions (0.21–200 µg·mL−1) of the proteins in NaCl/Pi. Assays were also carried out by the addition of dithiothreitol, mannose or N-acetyl-lactosamine (1 mm each) to the mixtures. Samples (50 µL) were gently mixed with 2% suspension of erythrocytes (50 µL) and incubated at room temperature for 1 h. Agglutination activity was determined on the basis of the sedimentary state of the erythrocytes.
PP13/galectin-13 dimerization assay
For the detection of dimerization, PP13-R was diluted (0.16–0.6 mg·mL−1) in Laemmli solution prepared with or without 10% (v/v) 2-mercaptoethanol and subjected to 12% (w/v) SDS/PAGE, then visualized by Coomassie staining. Protein bands were identified by subsequent MALDI-TOF mass spectrometry.
Pro-Q Diamond phosphoprotein gel staining
PP13-B, PP13-R, ovalbumin (positive control) and BSA (negative control) (20 µg each) were pretreated in reducing conditions, run on 15% SDS/PAGE and stained with Pro-Q Diamond phosphoprotein gel stain according to the manufacturer's protocol. For detecting phosphoproteins, the gel was visualized and photographed in UV light. For detecting its total protein content, Coomassie staining was applied.
WRL-68 cells were grown on 100 mm dishes in standard DMEM containing 1% (v/v) antibiotic-antimycotic solution, supplemented with 10% (v/v) fetal calf serum, under 5% CO2 condition and 95% humidified air at 37 °C. Cells were harvested and low-speed centrifuged at 2000 g, then the pellet was dispersed by vortexing in lysis buffer (50 mm Tris pH 7.4, 1 mm phenylmethylsulfonyl fluoride) for 10 min at 4 °C. After further cell disruption in a Teflon/glass homogenizer, the homogenate was pelleted, and the supernatant was coupled to cyanogen-bromide activated Sepharose 4B by the instructions of the manufacturer.
One hundred milligram tissue blocks from a term human placenta obtained from Histopathology Ltd. (Pecs, Hungary) were homogenized in lysis buffer (50 mm Tris pH 7.4, 1 mm phenylmethylsulfonyl fluoride) for 10 min at 4 °C in a Teflon/glass homogenizer. After pelletting the homogenates, supernatants were either coupled to cyanogen-bromide activated, PP13-bound Sepharose 4B for immunoaffinity purification, or measured by bicinchoninic acid reagent and equalized for 1 mg·mL−1 protein content in 2× Laemmli solution for Western blotting. Other parts of the placenta were formalin fixed, paraffin embedded, cut for 4 µm sections, mounted on slides, dried at 37 °C overnight, dewaxed and rehydrated for immunohistochemistry and immunofluorescence confocal microscopy.
Affinity purification of PP13/galectin-13 bound proteins
Both PP13-B and PP13-R were coupled to cyanogen-bromide activated Sepharose 4B and incubated with protein extracts from human placenta or WRL-68 fetal hepatic cells at 24 °C for 1 h. For controls, samples were incubated with uncoupled Sepharose 4B. Gels were washed three times with 20 mm Tris/HCl buffer (pH 7.4, 150 mm NaCl) followed by four rinses with 20 mm Tris/HCl buffer (pH 7.4) to remove unbound proteins. Specifically bound proteins were removed by an equal volume of 2× Laemmli buffer, separated by 15% SDS/PAGE and visualized by Coomassie staining.
Protein identification by mass spectrometry
Bands of interest either in Coomassie stained PP13-B or PP13-R, as well as PP13-B or PP13-R bound and eluted protein extracts were excised from the gels, reduced, alkylated and gel digested with trypsin as described in . Proteins were identified by a combination of MALDI-TOF MS peptide mapping and MALDI-post source decay (PSD) MS sequencing. The digests were purified with ZipTipC18 pipette tips with a saturated aqueous solution of 2,5-dihydroxybenzoic acid matrix (ratio of 1 : 1). A Bruker Reflex IV MALDI-TOF mass spectrometer (Bruker-Daltonics, Bremen, Germany) was employed for peptide mass mapping in positive ion reflector mode with delayed extraction. The monoisotopic masses for all peptide ion signals in the acquired spectra were determined and used for database searching against a nonredundant database (NCBI, Bethesda, MD, USA) using ms fit program (UCSF, San Francisco, CA, USA) . Primary structure of tryptic peptide ions was confirmed by PSD MS sequencing.
SDS/PAGE and Western blotting
Ten nanograms each of PP13-B and PP13-R (or 50 ng each in the case of monoclonal antibodies) and 10 µg of human placental protein extract was subjected to 15% (w/v) SDS/PAGE followed by immunoblotting with polyclonal or monoclonal antibodies to PP13 and horseradish peroxidase labeled secondary IgGs as described in . Protein bands were revealed by ECL chemiluminescence system.
Formalin fixed, paraffin-embedded tissue sections were incubated either with monoclonal or polyclonal antibodies to PP13. Immunostaining was carried out according to the streptavidin/biotin/peroxidase technique using Universal Kit . Control sections were incubated only with secondary IgGs. Visual evaluation of hematoxylin counterstained slides was performed with an Olympus BX50 light microscope with incorporated photography system (Hamburg, Germany).
Immunofluorescence confocal microscopy
Paraffin embedded tissue sections were deparaffinated and treated with either monoclonal or polyclonal antibodies to PP13 followed by FITC-labelled secondary anti-mouse or anti-rabbit IgGs and 20 µm DRAQ5 nucleus labelling dye in NaCl/Pi containing 0.1/0.1% (v/v) saponin and BSA. To visualize the localization of annexin II, anti-annexin II primary and FITC-labelled secondary IgGs were used. Control sections were incubated with only secondary IgGs, and antigen depletion was carried out on distinct slides. Fluorescence was scanned with a Bio-Rad MRC-1024ES laser confocal attachment (Herefordshire, UK) mounted on a Nikon Eclipse TE-300 inverted microscope (Kingstone, UK).
Values in the figures and text were expressed as mean ±SEM of n observations. Statistical analysis was performed by analysis of variance followed by Student's t-test and chi-square test. P < 0.05 was considered to be statistically significant.
PP13/galectin-13 is a member of a new subfamily among prototype galectins
From the GenBank search of related EST sequences, it could be assumed that PP13/galectin-13 mRNA was expressed only in placenta, fetal liver and spleen . PP13/galectin-13 gene mapped to chromosome 19 (19q13.1) in the close vicinity of genes of four known (galectin-10 , galectin-7 , galectin-4  and placental protein 13-like protein ) and three putative (‘similar to placental protein 13’ at locus XP_086001/AC005515-I, ‘unnamed protein’ at locus BAC85631/AC005515-II and ‘Charcot–Leyden Crystal 2 protein’ at locus AAP97241) galectins at 19q13.1–13.2 with similar exon structures, indicating their common genetic origin. PP13/galectin-13 was found to have a close relationship with the predominantly placental expressed ‘similar to placental protein 13’ (69% identity, 80% similarity) and placental protein 13-like protein (68% identity, 79% similarity) as well as CLC protein (56% identity, 69% similarity). The putative ‘Charcot–Leyden Crystal 2 protein’ and ‘unnamed protein’ also had a considerably high relationship to PP13/galectin-13. Putative serine and tyrosine kinase phosphorylation sites localized on the outer surface of PP13/galectin-13 were predicted at positions 44–52 (Ser48), 37–45 (Tyr41) and 76–84 (Tyr80) by computations (Fig. 1A). With rasmol, four cysteine residues were revealed on the surface of PP13/galectin-13 (Fig. 1B). By clustalw alignments, Cys136 and Cys138 on beta-sheet F1 were found to be missing from all homologues. Cys19 and Cys92 on beta-sheets F2 and F3 were missing from distant homologues, but some of the newly described closest homologues contained them (Fig. 1A).
PP13/galectin-13 possesses weak endogenous LPLA activity
For both PP13-B and PP13-R, the highest degree of transformation was found for l-α-lysophosphatidylcholine (1-acyl-glycero-3-phosphorylcholine, LPC); other lysophospholipids showed at most 5% (molar) transformation during the same period (data not shown). In the course of LPC transformation, four species could be distinguished and quantified in the 31P-NMR spectra (Fig. 2A), and their relative concentrations showed similar time-dependence (Fig. 2B); however, the reaction rates varied in the three solutions (PP13-B, PP13-R and control) (see below). The peak at 0.72 p.p.m. could be assigned to the starting material, which was involved in an isomerization equilibrium with 2-acyl-glycero-3-phosphorylcholine (iLPC) (δ = 0.56 p.p.m) , independent from the presence of PP13 proteins. In a slower reaction, LPC was transformed into two other species, one appearing at 1.00 p.p.m., and the other at 0.82 p.p.m. The former signal could be assigned to glycero-3-phosphorylcholine (GPC) based on its chemical shift [3,30–32]. The relative concentrations of the three major species (δ = 1.00, 0.82 and 0.72, respectively), expressed in molar fractions, are shown in Fig. 2B. The relative concentration of iLPC fluctuated between 0.10 and the limit of quantitation over the whole course of the reactions, roughly following the change of LPC (data not shown). The kinetics of the transformation of LPC could not be exactly described by classical models. However, the reaction appeared to move toward equilibrium, as judged by the time-dependence of the relative concentrations of the major species. The species appearing at 0.82 p.p.m. might well be an intermediate in the transformation, as its molar fraction increased in the first period and decreased after reaching a maximum value. Attempts are underway to identify this presumed intermediate. Determination of the enzymatic activities was difficult because the concentrations of both the intermediates and the products remained under the limit of quantitation for several tens of hours. However, a rough estimate could be made by the first spectra showing GPC in quantifiable concentrations: PP13-B showed 4.8 mol% transformation in 306 h, PP13-R gave 4.5 mol% in 210 h whereas control samples showed 1.1 mol% in 272 h. In terms of specific activity, these data read as 0.69, 0.94 and 0.18 nmol·min−1·mg−1, respectively, whereas approximately 1300 µmol·min−1·mg−1 was found for human brain LPLA  and 2.5 µmol·min−1·mg−1 for an LPLA isolated from human amnionic membrane . Phospholipase (PLA) activity of PP13-B and PP13-R was tested analogously, using phospholipids as substrates. No change could be observed in 31P-NMR spectra for any phospholipids, thus neither PP13-B nor PP13-R appeared to possess detectable PLA activity under these circumstances.
PP13/galectin-13 has strong sugar binding capabilities
Nonmodified agarose beads (Sephadex 2B) did not bind PP13-R at all, while all types of sugar-coupled agarose beads bound more than 95% of PP13-R after 1 h incubation. Different sugars (1 mm−1 m) eluted the protein from various sugar-coupled agarose in different manners, with the following elution capacity: N-acetyl-lactosamine >mannose > N-acetyl-galactosamine > maltose > glucose >galactose > fucose > lactose (Fig. 3A). In 1 m concentration, N-acetyl-lactosamine had significantly the highest efficacy (95–100%) to elute PP13-R from all kinds of beads, while mannose was less effective, having an elution capacity between 15 and 30%. On average, N-acetyl-galactosamine was the third most effective to specifically compete with PP13-R binding (12–19%). The elution capacities for other sugars were determined to be below 8% in the following order: maltose (0–8%), glucose (0–4%), galactose (0–7%), fucose (0–4%), lactose (0–2%). These latter sugars had higher elution efficacy only in some special combinations: maltose/fucose-agarose (21%) and maltose-agarose (42%); glucose/maltose-agarose (23%); lactose/lactose-agarose (7%) (Fig. 3A). In the presence of 1 mm dithiothreitol during the 1 h binding period, approximately half of PP13-R was bound to different sugar-coupled agaroses (e.g. lactose-agarose: 60%, glucose- and mannose-agarose: 55% each), and the elution of specifically bound PP13-R with different sugars was four-times more effective compared to nonreducing conditions. 100 mm mannose eluted 31–100% of PP13-R from glucose-, mannose- or lactose-agarose, while without the presence of dithiothreitol the elution was only 8–16% (Fig. 3B). The order of the elution capacity of the different sugars for PP13-R from the various sugar-coupled agaroses was the same in reducing and nonreducing conditions, but in the presence of dithiothreitol, sugar elution of specifically bound PP13-R from sugar-coupled agaroses was significantly higher. Mannose (100 mm) eluted all bound PP13-R from lactose-agarose, 50% from mannose-agarose and 43% from glucose-agarose.
PP13/galectin-13 possesses lectin activity
Lectin activity of PP13-B and PP13-R was confirmed by measurements of their agglutination capabilities of human erythrocytes. In nonreducing conditions, very small amounts of both PP13-B and PP13-R induced haemagglutination, and strong agglutination was detected at and above 50 µg·mL−1 applied protein concentrations (Fig. 4), which was very similar to the phenomenon seen in cases of other galectins . The pattern and effectiveness of both PP13-B and PP13-R were identical in agglutination of erythrocytes. However, no haemagglutination occurred in reducing conditions with the addition of 1 mm dithiothreitol to the mixture. Different sugars also had an inhibitory effect on haemagglutination capabilities of PP13-R. At and above concentrations of ≈ 1 mmN-acetyl-lactosamine and mannose, previously found to be the best ligands of PP13-R, abolished its haemagglutination activity (Fig. 4).
PP13/galectin-13 dimerizes via disulphide bonds
Galectins were known to be dimerized by noncovalent interactions [6,9]. From earlier data , as well as in our experiments, PP13/galectin-13 was found to be composed of two identical subunits held together by disulphide bonds. In nonreducing conditions, dimerization occurred at and above 0.21 mg·mL−1 PP13-R concentrations (Fig. 5A). When PP13-R was dissolved in Laemmli solution containing 10% (v/v) 2-mercaptoethanol, no dimerization of PP13-R was found at all, even at higher protein concentrations (Fig. 5B).
Placental expressed PP13/galectin-13 is phosphorylated
Pro-Q Diamond phosphoprotein gel stain specific for phosphorylated protein side chains was used to detect previously predicted putative phosphorylation of PP13/galectin-13. Both placental purified and bacterially expressed PP13 was examined along with ovalbumin (positive control) and BSA (negative control). A strong signal of phosphorylated groups in the lane of ovalbumin and a weak signal in the lane of PP13-B purified from placenta could be specifically detected. No signal in the lanes of albumin and bacterially expressed PP13-R was found (Fig. 6A). An equal amount of protein content for each lane was verified by Coomassie staining (Fig. 6B).
PP13/galectin-13 binds annexin II and beta/gamma actin
By Coomassie staining after SDS/PAGE, in cases of PP13-B and PP13-R, major bands at 16 or 18 kDa were detected. No additional bands in lower or higher molecular mass regions could be identified, indicating high purity of both protein preparations. Bands were cut from the gels, then MALDI-TOF MS peptide mapping with MALDI-PSD MS sequencing was performed, recognizing both PP13-B and PP13-R as PP13/galectin-13. Next, human term placental tissue and fetal hepatic cell (WRL-68) extracts were bound either to PP13-B or PP13-R coupled to Sepharose 4B, or to Sepharose 4B alone. Again, using Coomassie staining, the same major protein bands at 16 kDa (in the case of PP13-B), at 18 kDa (in the case of PP13-R), or at 38 and 41 kDa (in cases of both PP13-B and PP13-R) could be detected either in placental or in fetal hepatic protein extracts bound to either PP13-B (data not shown) or to PP13-R (Fig. 7, lanes 1–2), while Sepharose 4B did not specifically bind any proteins at all (Fig. 7, lanes 3–4). By MALDI-TOF MS peptide mapping and MALDI-PSD MS sequencing, all protein bands yielded good quality peptide maps, and most of the input masses matched the candidate protein sequences. The eluted 16 or 18 kDa proteins were identified as PP13-B or PP13-R subunits dimerized with PP13-B or PP13-R subunits coupled to Sepharose 4B. MALDI-TOF MS data of the 38 kDa protein in both cases permitted the identification of human annexin II (Accession No. NM_004039) (Table 1A), while the mass map of the 41 kDa protein matched beta/gamma actin (Table 1B) in both cases (Accession No. NM_001101 and NM_001614). PSD data obtained for precursors also confirmed the identity of these proteins.
|Measured mass (MH+)||Calculated mass (MH+)||Delta (p.p.m.)||Modifications||Fragment||Missed cleavages||Database sequence|
|1035.6177||1035.5297||85||–||213–220||0||(K) WISIMTER (S)|
|1086.5769||1086.4856||84||–||29–37||0||(K) AYTNFDAER (D)|
|1086.5769||1086.6821||−97||–||287–295||1||(K) VLIRIMVSR (S)|
|1094.5963||1094.5271||63||pyroGlu||69–77||0||(R) QDIAFAYQR (R)|
|1111.6201||1111.5536||60||–||69–77||0||(R) QDIAFAYQR (R)|
|1244.6868||1244.6235||51||–||136–145||0||(R) TNQELQEINR (V)|
|1439.8798||1439.7238||108||2Met-ox||291–302||1||(R) IMVSRSEVDMLK (I)|
|1460.7615||1460.6732||60||–||234–245||0||(K) SYSPYDMLESIR (K)|
|1542.9514||1542.8491||66||–||50–63||0||(K) GVDEVTIVNILTNR (S)|
|1588.8890||1588.7681||76||–||234–246||1||(K) SYSPYDMLESIRK (E)|
|1778.0156||1777.8642||85||–||120–135||0||(K) GLGTDEDSLIEIICSR (T)|
|1909.0682||1908.8827||97||–||180–196||0||(R) AEDGSVIDYELIDQDAR (D)|
|2065.2063||2064.9838||108||–||179–196||1||(R) RAEDGSVIDYELIDQDAR (D)|
|1198.7517||1198.5228||191||–||44–54||0||(K) DSYVGDEAQSK (R)|
|1198.7517||1198.7061||38||–||22–32||0||(R) AVFPSIVGRPR (H)|
|1203.6632||1203.5614||85||2Met-ox||33–43||0||(R) HQGVMVGMGQK (D)|
|1499.7963||1499.6767||80||pyroGlu||353–365||0||(K) QEYDESGPSIVHR (K)|
|1515.8512||1515.7497||67||–||78–88||0||(K) IWHHTFYNELR (V)|
|1628.0516||1627.7716||172||pyroGlu||353–366||1||(K) QEYDESGPSIVHRK (C)|
|1791.0558||1790.8925||91||–||232–247||0||(K) SYELPDGQVITIGNER (F)|
|1954.2281||1954.0650||83||–||89–106||0||(R) VAPEEHPVLLTEAPLNPK (A)|
|2215.3066||2215.0705||107||–||285–305||0||(K) DLYANTVLSGGTTMYPGIADR (M)|
|2807.5914||2807.3119||100||–||207–231||1||(K) EKLCYVALDFEQEMATAASSSSLEK (S)|
Polyclonal and monoclonal antibodies to PP13 have specific recognition to PP13/galectin-13
To investigate and compare the specificity of polyclonal and newly developed monoclonal antibodies to PP13, Western blot testing was performed utilizing PP13-B, PP13-R proteins and human placental tissue extracts. As previously shown, polyclonal antibody to PP13 bound specifically to PP13-B extracted from human term placenta and also reacted with the same size protein in some fetal tissues such as liver and spleen . Here it was observed that polyclonal antibody to PP13 could recognize PP13-R in a similar pattern as purified PP13-B and placental expressed PP13/galectin-13, with no other proteins recognized (Fig. 8A). From the newly developed monoclonal antibodies to PP13, clone 215 developed against a PP13/galectin-13 specific epitope had the strongest reaction with PP13-B and PP13-R, and also recognized the placental expressed PP13 with no cross-reaction to other proteins of the placenta (Fig. 8B).
PP13/galectin-13 is localized predominantly on the brush border membrane of placental syncytiotrophoblasts
In human term placental tissue, special localization of PP13/galectin-13 was found by different immunological techniques. Monoclonal antibody to PP13 gave a significantly weaker staining on immunohistochemical sections, while it had stronger staining with confocal imaging than polyclonal antibody to PP13. With both antibodies, labelling mainly on the brush border membrane of the syncytiotrophoblasts could be seen by immunohistochemistry, with a parallel weak staining of the cells (Fig. 9A,B). By the more sensitive immunofluorescence confocal imaging, a similar, but more intense PP13 staining of the brush border membrane was detected, also with a discrete perinuclear labelling of the syncytiotrophoblasts by both monoclonal and polyclonal antibodies (Fig. 9C,D). Parallel annexin II staining of the syncytiotrophoblasts as well as intense staining on the brush border membrane could be seen (Fig. 9E).
Although PP13 was first isolated and cloned from human term placenta [2,3], its expression in human fetal liver and spleen tissues has also been detected . As PP13 showed conserved sequential, structural and computed functional homology to members of the growing β-galactoside-binding galectin family , it was designated as galectin-13 . In this study it was verified that PP13/galectin-13 mRNA and related ESTs were predominantly expressed in placenta, but also in fetal liver and spleen tissues. The PP13/galectin-13 gene mapped to the close vicinity of genes of four known and three putative galectins [8,9,27–29] with similar exon structures and surrounding untranslated regions in a tight cluster on chromosome 19. The encoded proteins also proved to share 80% of the highly conserved galectin residues, which suggested a gene multiplication event in this galectin subfamily. In contrast to the evolutionarily ancient galectins expressed in many tissues, this subfamily comprising PP13/galectin-13 appeared to have already developed in nonprimates but expanded in primates, as members are predominantly expressed in specific tissues, with many of them abundant only in placenta. Not only this fact but also their specific transcriptional regulation underlined the differential placental expression of these genes, as numerous placenta-specific transcriptional factor binding sites were found in the promoter regions . An analogous gene duplication event on chromosome 11 occurred in the case of eosinophil major basic proteins, of which human major basic protein-2 is present only in eosinophils, while human major basic protein-1 is abundant in placenta, and both are involved in immune functions [9,36]. Similarly, genes of mannose-specific C-type lectins, DC-SIGN and DC-SIGNR, and their homologue CD23 (FcERII) were described to be evolutionary duplicated on chromosome 19p13.3. Their concomitant expression was shown in placenta and dendritic cells with specific immunobiological functions [37,38]. As several other galectins are also involved in inflammation and immune defences , our findings suggest that the newly evolved and differentially expressed PP13/galectin-13 with its homologues might have special immune functions at the fetomaternal interface. In the near future this phenomenon must be analyzed in the light of previous clinical data on PP13 serum levels in different disorders of pregnancy .
Because of the highly conserved homology with several other galectins, it was likely that PP13/galectin-13 exhibited sugar binding activity. Indeed, in our previous report based on homology modelling , the possible functional and structural characteristics of PP13/galectin-13 were predicted, including a CRD which resembled the β-galactoside-binding site of galectins. In this study, binding experiments showed that PP13/galectin-13 was effectively bound to different sugar containing agarose gels, and that various sugars could compete this effect with different affinities to the PP13/galectin-13 binding site. As in the case of most galectins when similar sugar concentrations were applied, N-acetyl-lactosamine had the highest affinity to its CRD. Similarly to CLC protein/galectin-10 but not other previously analyzed galectins, PP13/galectin-13 also had high affinity to mannose, which could be understood in terms of the similarities in their CRDs [7,9].
N-acetyl-galactosamine also had a certain affinity to PP13/galectin-13 CRD, in contrast to other sugar derivates, which only slightly displaced the protein from sugar-coupled agaroses. Interestingly, homology modelling data had also indicated that N-acetyl-lactosamine would bind the most effectively to the PP13/galectin-13 binding site, and in the case of other sugars, there were only minor discrepancies between the previously suggested and experimentally observed binding affinities . Strong lectin activity of PP13-B and PP13-R was also proven by their haemagglutination activity and by haemagglutination inhibition assays, where excess sugar molecules competed with red blood cell sugar residues and proved to be more likely to bind to PP13 proteins leaving red blood cells to sediment. These experimental data and the specific and predominant localization of PP13/galectin-13 on the brush border membrane of syncytiotrophoblasts were in close agreement with a systematic study on the structure and distribution of specific glycans in human placenta, which showed that residues containing N-acetyl-lactosamine, mannose and N-acetyl-glucosamine were widely expressed on villous surfaces . This may provide an explanation of the binding specificity of PP13/galectin-13, and suggests a similar binding pattern of its newly described, mainly placental-expressed homologues.
In vitro, PP13/galectin-13 dimerization occurred at and above 0.21 mg·mL−1 concentrations in nonreducing conditions, while in the presence of dithiothreitol, no dimerization was detected at all. Furthermore, in reducing conditions, approximately half of PP13-R was bound to sugar-coupled agaroses, and the protein's haemagglutination activity was also abolished, all of which could be explained by the loss of dimerization. Cystein residues in galectins (formerly known as ‘S-type’ or ‘thiol-dependent’ lectins) were considered to be important, because some but not all galectins, such as galectin-1 and galectin-2, might lose their sugar binding activity under nonreducing conditions [40,41]. Compared to the structure of CLC protein/galectin-10, which is known to be a monomer , four additional cysteine residues were found in PP13/galectin-13, which might localize on a putative dimerization surface. Interestingly, Cys136 and Cys138 on beta-sheet F1 in PP13/galectin-13 were found to be missing from all homologues, whereas Cys19 and Cys92 on beta-sheets F2-F3 were missing from homologues including CLC/galectin-10, only the newly described closest relatives contained them. This data may vindicate the surprising experimental findings on its dimerization via disulphide bonds, a phenomenon yet not described for other galectins. Superposing the 3D model of PP13/galectin-13 monomer on well known models of galectin dimers [43–45] with rasmol revealed, that if PP13/galectin-13 dimerizes such as galectin-1 and galectin-2 in a two fold rotation perpendicular to the beta-sheets, then only beta-sheets S1 and F1 containing Cys136 might participate in the dimerization. If PP13/galectin-13 dimerizes more like to galectin-7, then beta-sheets F1-F5 might comprise the putative dimerization interface containing Cys19, Cys92 and Cys136, as well. As our data is still not enough to describe an exact dimerization interface, X-ray crystallographic experiments have been started to reveal the proper dimeric structure and carbohydrate binding of the protein.
Our data also showed that dithiothreitol equally decreased the binding strengths between PP13-R and different sugars, but no change occurred in the order of their binding capabilities. As the CRD of PP13/galectin-13 was situated just opposite the cysteine-rich region, it is understandable that the monomeric form could not cross-link red blood cells, but the mechanism of how the reducing agent could decrease the sugar binding potential of the protein remains to be elucidated. Our sequence alignment also suggested that the new subfamily members might also dimerize via disulphide bonds, which should be critically analyzed later. As these proteins were primarily expressed in the low blood flow organ placenta, special alterations in the oxygenation could easily affect their biological functions through their dimerization status and sugar binding affinities.
It is important to mention that computations localized putative serine and tyrosine kinase phosphorylation sites on the outer surface of PP13/galectin-13 at positions 44–52 (Ser48), 37–45 (Tyr41) and 76–84 (Tyr80), in close vicinity to its CRD, similarly to placental protein 13-like protein . Experimental data showed that in vivo placental expressed and purified PP13-B was phosphorylated, while the in vitro bacterial expressed PP13-R was not. Knowing that galectin-3 was reported to be phosphorylated at Ser6, and phosphorylation modulated its carbohydrate affinity and biological functions as an ‘on/off’ switch [46,47], our preliminary data raised the possibility of phosphorylation having an influence on PP13/galectin-13 functional properties. In our conditions, carbohydrate binding affinities were similar both in cases of PP13-B and PP13-R; the only difference was detected in their lysopholypase activity. Further detailed experiments must be performed to establish the importance and the exact mechanism of this phenomenon.
Formerly it was found that PP13/galectin-13 had a weak lysophospholipase activity , which had previously been observed in the case of CLC protein/galectin-10 . However, it was shown later that this enzymatic activity might not be derived from CLC protein/galectin-10, but from another protein associated with it . Although PP13-B purified from human term placenta did not contain a level of impurity comparable to that which the CLC protein/galectin-10 preparation probably had, we overexpressed the cloned and His-tagged PP13-R in E. coli, and purified it on a Ni-NTA column. Next, the enzymatic activity of PP13-B and PP13-R were compared. 31P-NMR analysis showed that both PP13-B and PP13-R had a very weak endogenous LPLA activity, even PP13-R showed higher catalytic activity than the purified one, indicating that PP13/galectin-13 itself possesses LPLA activity. PLA activity of neither the purified nor the recombinant PP13/galectin-13 could be confirmed under these circumstances. The exact LPLA catalytic centre of PP13/galectin-13 must to be further analyzed.
Although the possibility that the weak LPLA activity of PP13-B coming from an associated and copurified protein could be excluded, it still remained an interesting question as to which intracellular proteins were interacting PP13/galectin-13. By immobilizing PP13-B and PP13-R, the proteins extracted from term placental tissue and fetal hepatic cells and specifically bound to PP13 proteins were determined. By means of MALDI-TOF MS peptide mapping and sequencing, the 38 kDa and 41 kDa proteins, which bound both to PP13-B and PP13-R, were identified as human annexin II and beta/gamma actin, respectively. The 16 or 18 kDa proteins were identified as the eluted PP13-B or PP13-R subunits.
It is well known that cells differ widely in their capacity to produce and secrete galectins, and galectin secretion is also responsive to developmental events. During the complex mechanisms of human placentation, correlating with the differentiation pathways of the trophoblasts, changes in the distribution patterns of galectin-1 and galectin-3, homologues of PP13/galectin-13, were already seen . Studies also revealed that actin filaments might play an important role in translocation of lectins during differentiation processes , and could also be involved in focal concentration of cytosolic galectin at specific cytoskeletal regions, for example at evaginating plasma membrane domains during ectocytosis . The exact mechanisms of how PP13/galectin-13 is transported to the outer surface of the syntitiotrophoblasts' plasma membrane have not yet been studied, but it is assumed to be released through an alternative, nonclassical route, similarly to other galectins . Fibroblasts were previously shown to secrete galectins by ectocytosis in microvesicles, which also contained actin (derived from disassembling microfilaments) and annexin II in high amounts . As PP13/galectin-13 was bound to these molecules very specifically and colocalized with annexin II on the brush border membrane, PP13/galectin-13 potentially utilizes this ectocytotic pathway for externalization, in the same way as the secretion of galectin-3 . In vivo, microvesicles were found to be labile and disrupted spontaneously, releasing their lectin cargo very rapidly. Although PLAs were considered to be candidates for catalyzing the hydrolysis of the sn-2 fatty acyl bond of phospholipids to liberate free fatty acids and lysophospholipids, and to release soluble extracellular galectins, to date no particular PLA has been found to carry out this exact function . PP13/galectin-13 possesses a relatively weak but definitive LPLA activity, which is probably insufficient for catabolizing significant amounts of phospholipids, but which may contribute to its penetration through the vesicle membrane.
Annexin II, a member of a Ca2+ and phospholipid binding protein family, is present as a heterotetramer on the apical extracellular surface of the syncytiotrophoblasts , colocalized with PP13/galectin-13. It is thought to play a role in the differentiation of the placenta and in the functions of the mature microvilli . It was recognized as a profibrinolytic coreceptor for tissue plasminogen activator and plasminogen on endothelial cells, and was shown to stimulate the tissue plasminogen activator-dependent conversion of plasminogen to plasmin [55,56]. In addition, annexin II was also found to promote plasmin inactivation, regulate ion channels, PLA2 and prothrombin activation . Therefore, its interaction with PP13/galectin-13 may play an important role in placental haemostatic processes.
In conclusion, PP13/galectin-13 was localized mostly in the brush border membrane of the syncytiotrophoblasts' lining at the common feto-maternal blood spaces of the placenta. As a ‘prototype’ galectin, it has a single sugar binding domain, which emerges into the extracellular space. Sugar binding assays proved that PP13/galectin-13 had high affinity to sugar residues widely expressed in placenta, and bound to surface antigens of red blood cells. The protein dimerized via disulphide bonds, which unique feature might affect its biological activity upon the oxygenation changes in the low blood flow organ placenta. Phosphorylation of the protein could be another regulator for its special biological functions, the mechanism of which must also be elucidated. With regard to the fact that red blood cells do not show a remarkable rate of aggregation to syncytiotrophoblasts in the case of normal placental function, and plasma levels of PP13/galectin-13 did not correlate with the intensity of the placental synthesis of the protein, a transitory stay of PP13/galectin-13 in the cell membrane followed by the release from the brush border membrane of the syncytiotrophoblasts' is involved. It is not yet known whether the release is catalyzed by an autonomous PLA, or whether PP13/galectin-13 has its own ability to split the chemical bonds. As protein separation techniques did not identify any PLA bound to PP13/galectin-13, and measurements proved a weak LPLA activity of the protein itself, a hypothesis could be set up that the low self-LPLA activity of the protein is enough to satisfy the demands on securing a slow, constant PP13/galectin-13 release, which is capable of preventing erythrocyte adhesion in the area with reduced blood flow. A considerable amount of detected PP13/galectin-13 also deserved attention in fetal liver and spleen, which are also known to be elemental haemopoietic organs with reduced blood flow. In contrast, as with other galectins involved in immunobiological functions, our data suggested that PP13/galectin-13 and its newly described, differentially expressed close homologues might have special immune functions at the feto-maternal interface. Involvement of PP13/galectin-13 and its homologues in developmental processes of the placenta is also hypothesized.
The authors are grateful to Dr Ilana Maor at Diagnostic Technologies Ltd. for her assistance in the collaboration. Special thanks to Dr Gyorgy Szekeres, Rita Keszthelyi and Zsuzsa Halas (Histopathology Ltd, Hungary) for their technical assistance in immunostaining, Erzsebet Osz (University of Pecs, Hungary) for her assistance in NMR measurements, Prof Nathan Sharon (The Weizmann Institute of Science, Israel) for his helpful discussions in methodology and Steve Starkey for critical reading of the manuscript. Zoltan Berente is grateful to the Hungarian Academy of Sciences for the Bolyai Janos Scholarship. Experiments were carried out at the facilities of the University of Pecs except PP13-R expression and the preparation of monoclonal antibody to PP13 (Diagnostic Technologies Ltd, Israel), immunhistology with monoclonal antibodies (Yale University, USA) and protein identification with MS (University of Szeged, Hungary). This work was supported by Hungarian Grants ETT T-09 163/01; FKFP 0166/2001; OMFB-BIO 00201/2002; and OTKA T/020622, T/023076, T/029824, M/36996. It was also funded in part by the Office of the Chief Scientist, Israel Ministry of Industry and Trade Grant #31851.
- 11993) Advances in Pregnancy-Related Protein Research, 1st edn. pp. 1–333. CRC Press, Boston., & (
- 41999) Placental Protein P3 (PP13), Patent: IL (WO/99/38970)-A., , & (
- 111993) Method for the detection of pregnancy disorders. United States Patent 5,198,366.(
- 171990) PCR Protocol: a Guide to Methods and Applications In (Innis, M.A., Gelfand, D.H., Sninsky, J.J. & White, T.J., eds), p28. Academic Press, San Diego.(
- 251989) Molecular Cloning: a Laboratory Manual, 3rd edn. Cold Spring Harbor Laboratory, Cold Spring Harbor, NY., & (
- 261994) Immunohistochemistry: antigen detection in tissue. In Advanced Laboratory Methods in Histology and Pathology (Mikel, U.V., ed.), pp. 1–40. Armed Forces Institute of Pathology and American Registry of Pathology, Washington DC.& (
- 281992) Localization of the human eosinophil Charcot–Leyden crystal protein (lysophospholipase) gene (CLC) to chromosome 19 and the human ribonuclease 2 (eosinophil-derived neurotoxin) and ribonuclease 3 (eosinophil cationic protein) genes (RNS2 and RNS3) to chromosome 14. Genomics 13, 240–242., , , , , & (