| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Regular Article |
a Department of Zoology, College of Agriculture and Life Sciences, North Carolina State University, Raleigh, North Carolina 27695-7617
b Laboratory of Physiology, Department of Marine Biological Science, Faculty of Fisheries, Hokkaido University, Hakodate, Hokkaido 041-8611, Japan
c Department of Biochemistry and Molecular Biology, University of Florida, Gainesville, Florida 32610
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
20-kDa band plus some diffuse smaller bands that could be visualized by staining for phosphoprotein with Coomassie Brilliant Blue complexed with aluminum nitrate. Immunological and biochemical characteristics of YP1, YP2s, and YP3 identified them as white perch lipovitellin, ß'-components, and phosvitin, respectively. A novel receptor-binding assay for vitellogenin was developed based on digoxigenin (DIG)-labeled vitellogenin tracer binding to ovarian membrane proteins immobilized in 96-well plates. Lipovitellin from white perch and vitellogenin from perch and other teleosts effectively displaced specifically bound DIG-vitellogenin in the assay, but phosvitin and the ß'-component could not, demonstrating for the first time that the lipovitellin domain of teleost vitellogenin mediates its binding to the oocyte receptor. Lipovitellin was less effective than vitellogenin in this regard, suggesting that the remaining yolk protein domains of vitellogenin may interact with its lipovitellin domain to facilitate binding of vitellogenin to its receptor.
estradiol, female reproductive tract, gamete biology, oocyte development, ovary
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Although vitellogenin has been purified and biochemically characterized for many fish species, surprisingly little attention has been paid to identifying vitellogenin-related yolk proteins in fishes other than salmonids. Three yolk proteins derived from vitellogenin; namely, lipovitellin, phosvitin, and ß'-component, have been identified in salmonid oocytes [7, 8]. Salmonid lipovitellin and phosvitin resemble their homologues in chicken and Xenopus, but a relation of ß'-component, which contains neither lipid nor phosphorus, to YGP or other yolk components from higher vertebrates remains to be verified.
Oocytes sequester vitellogenin via a process of receptor-mediated endocytosis. Specific binding of vitellogenin to oocyte membrane preparations and solubilized membrane proteins has been demonstrated for insects, amphibians, and birds [9–14]. Stifani et al. [15] first demonstrated specific binding of vitellogenin to fish ovarian membrane preparations using a salmonid species, the coho salmon (Oncorhynchus kisutch), to confirm the existence of a receptor-mediated system for vitellogenin internalization in fishes. The salmon vitellogenin receptor was found to resemble the vitellogenin receptor of chicken and Xenopus with regard to its estimated mass, binding kinetics, ligand specificity, and localization to the ovary [16, 17]. Similar results were later obtained for more advanced teleosts such as tilapia (Oreochromis niloticus), white perch (Morone americana), and sea bass (Dicentrarchus labrax) [18–20]. More recently, cDNAs encoding the rainbow trout receptor were cloned and the deduced amino acid sequence revealed a type I membrane protein of size and sequence similar to and bearing the same characteristic suite of functional domains as the vitellogenin receptor of chickens and Xenopus [21, 22].
Previously, it was believed that the phosvitin domain of vitellogenin mediated its binding to the vitellogenin receptor on chicken and Xenopus oocytes [9, 10]. However, it was later verified that lipovitellin, but not phosvitin, from these species could displace vitellogenin from its receptor [13, 14]. To our knowledge, binding of YGPs to the vitellogenin receptor has not yet been explored. Aside from verification that yolk extracts can displace vitellogenin from its receptor [16, 19], comparable studies have not been performed using yolk proteins purified from any fish species.
Temperate basses (genus Morone), including the white perch and its relatives, are important species in fisheries and aquaculture and are becoming established models for basic research on the reproductive physiology of Perciformes [19, 23–25]. Their vitellogenin has been purified and characterized in detail, immunoassays for the protein have been developed and used to study oogenesis, and functional characteristics of their ovarian vitellogenin receptors have been described [19, 23, 26–29]. However, the yolk proteins derived from vitellogenin remained to be investigated in Morone species. The objectives of the present study were to 1) purify and biochemically characterize the yolk proteins derived from vitellogenin in white perch, 2) develop a nonisotopic assay of vitellogenin receptor-binding using digoxigenin (DIG)-labeled vitellogenin tracer and 96-well plates coated with ovarian membrane proteins, and 3) use the assay and purified yolk proteins to identify the specific yolk protein domain of vitellogenin responsible for recognition of the protein by its oocyte receptor.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
The experimental animals used in this study were mature white perch reared in ponds at the Pamlico Aquaculture Field Laboratory of North Carolina State University (NCSU) or juvenile and adult fish held in tanks at the NCSU Aquatic Research Laboratory [28, 30]. All experiments involving these animals were conducted in accordance with the 1996 Guide for Care and Use of Laboratory Animals published by the National Research Council. Blood plasma was obtained from male and immature fish injected with estradiol-17ß (E2) as described previously [28]. Ovaries were excised from vitellogenic females, immediately frozen in liquid nitrogen, and then kept in liquid nitrogen until use. Typically, 5 g of ovary were finely minced and then homogenized in a 55-ml Potter-Elvehjem type tissue grinder (Wheaton, Millville, NJ) with 10 ml of binding buffer (20 mM Tris-HCl, 2 mM CaCl2, and 150 mM NaCl pH 8.0) containing 1 mM PMSF and 4 IU L-1 aprotinin. After centrifugation for 15 min at 10 000 x g, the supernatant was filtered through Whatman No. 2 filter paper (Whatman International Ltd., Maidstone, U.K.) and collected as perch ovarian extract, the starting material for purification of yolk proteins. The resulting pellet was used for preparation of ovarian membranes.
Preparation of Ovarian Membranes
Ovarian membranes were prepared from ovaries of vitellogenic white perch according to our previous study with slight modification [19]. All procedures were performed at 4°C. After centrifugation of the ovary homogenate for 15 min at 10 000 x g, the pellet was resuspended in 20 ml of the same buffer and washed four times to completely remove yolk proteins. The resulting pellet was resuspended in 15 ml of the same buffer, homogenized with a Polytron PT10/35 tissue homogenizer (Brinkman, Westbury, NY) using two bursts at setting 5 for 30 sec and two bursts at setting 7 for 20 sec, and then centrifuged at 500 x g for 5 min. The resulting supernatant was filtered through 100-µm nylon mesh and ultracentrifuged at 100 000 x g for 1 h. The membrane pellet was resuspended in 1.5 ml of the same buffer by repetitive aspiration through a 22-gauge hypodermic needle and designated as the crude ovarian membrane preparation. The crude membrane preparation was stored at -80°C for up to 1 mo for later use. A crude preparation of erythrocyte membranes was prepared from the blood of juvenile striped bass (Morone saxatilis) as described previously [19].
Solubilization of the Membrane Preparation
The crude membrane preparation was mixed with an equal volume of binding buffer containing 2.4% n-octyl-ß-D-glucopyranoside and then stirred for 15 min on ice. The resulting extract was centrifuged for 15 min at 10 000 x g to remove insoluble materials. The supernatant was filtered through a 0.45-µm disposable syringe filter (Millex GV; Millipore, Bedford, MA), held on ice, and then used in the receptor binding assay within 24 h. Typically, approximately 900 µg of solubilized membrane protein could be obtained from 5 g of ovary.
Purification and Digoxigenin Labeling of White Perch Vitellogenin
White perch and striped bass vitellogenin were purified according to our previously reported methods [19, 27, 28]. Purified white perch vitellogenin was coupled to DIG using a DIG antibody labeling kit (Roche Diagnostics, Indianapolis, IN) at a molar ratio of 1:5 (vitellogenin:DIG). The labeling procedure otherwise followed the manufacturer's instructions for the product.
Purification of Vitellogenin from Several Teleosts
Vitellogenin from gag grouper (Mycteroperca microlepis); a hybrid sturgeon, bester (Huso huso x Acipencer luthenus); Sakhalin taimen (Hucho perryi); and cutthroat trout (Oncorhynchus clarki) was purified according to our previously reported methods [8, 31, 32].
Column Chromatography for Purification of Yolk Proteins
Ion exchange chromatography was performed on diethylaminoethyl (DEAE) cellulose (DE-52; Whatman International) at 4°C. The DEAE cellulose was equilibrated with starting buffer (20 mM Tris-HCl pH 8.0) and loaded into a 2.5 x 20 cm column. Samples were eluted by stepwise addition of NaCl solutions in increasing concentrations from 50 mM to 1.0 M at a flow rate of 100 ml/h. Eluted fractions were collected at a volume of 7 ml/tube.
Gel filtration was performed on a Superdex 200 column (Amersham Pharmacia Biotech, Uppsala, Sweden) fitted to a fast protein liquid chromatography system (Amersham) operated at 4°C. The column was eluted with 20 mM Tris-HCl buffer pH 8.0 containing 2% NaCl and 0.1% NaN3. Column flow rate was 1.0 ml/min and the eluted fractions were collected at a volume of 0.5 ml/tube.
For gel filtration on Sepharose 6B (Amersham), the column was eluted with the Tris-HCl buffer at a flow rate of 12.5 ml/h and the eluted fractions were collected at 3.15 ml/tube.
Molecular Weight Determination by Gel Filtration
The Superdex 200 column was calibrated with marker proteins (Sigma, St. Louis, MO) for determination of the molecular weight of purified yolk proteins according to our previous report [8].
Antisera
Antisera were raised in rabbits against vitellogenic white perch ovary extracts and purified yolk proteins by intradermal injection of each sample emulsified in an equal volume of Freund complete adjuvant. Injections were conducted four times at 7- to 10-day intervals. Antisera against white spotted-charr (Salvelinus leucomaenis) ß'-component (a-charr ß'), rainbow trout lipovitellin (a-trout Lv), and striped bass female-specific plasma proteins (a-FSPP) were prepared as described in previous reports [19, 28, 33, 34].
Electrophoresis and Immunological Procedures
Immunoelectrophoresis and double immunodiffusion were conducted in 1% agarose gels using 50 mM barbital buffer and 0.9% NaCl containing 0.1% NaN3, respectively, using routine procedures.
SDS-PAGE (5%–22.5% or 8%–25%) was carried out using a Tris-glycine buffer system [35] or Tricine buffer system [36]. The gels were stained with 0.1% Coomassie Brilliant Blue R 250 (CBB) in a solution of ethanol:acetic acid: distilled water (4:1:5 v/v; CBB stain) for protein or 0.05% CBB, 100 mM aluminum nitrate, and 1% Triton X-100 in a solution of isopropanol:acetic acid (5:2 v/v; CBB-Al stain) for phosphoprotein. The molecular mass of proteins banded in the gels was estimated using low molecular and high molecular weight markers (both from Amersham), and prestained protein markers (New England Biolabs, Beverly, MA).
Western blotting was conducted according to our previous report [37]. For detection of DIG-labeled vitellogenin blotted onto polyvinylidene fluoride (PVDF) membranes, antidigoxigenin conjugated to polymerized horseradish peroxidase (anti-DIG-poly-POD; Roche Diagnostics) was used as the primary antibody. The enzymatic color reaction was developed using a diaminobenzidine peroxidase substrate kit (Vector Laboratories, Burlingame, CA).
Phosphorus Quantification
During purification of phosvitin, alkaline-labile phosphorus in fractions eluted from the Superdex 200 gel filtration column was measured according to the method of Gamst and Try [38].
N-Terminal Amino Acid Sequencing and Amino Acid Composition
Purified proteins or peptides blotted onto PVDF membranes after SDS-PAGE were sequenced using an Applied Biosystems (Foster City, CA) model 494 HT Protein sequencer with an on-line phenylthiohydantoin amino acid analyzer [39]. Resulting sequences were compared to sequences for vitellogenin and yolk proteins obtained from primary publications or from databases at the National Center for Biotechnology Information using the basic local alignment search tool network service [40]. Amino acid composition analyses of white perch phosvitin were performed as described previously [41].
Receptor Binding Assay
Receptor binding assays were carried out in 96-well microtiter plates (Coster EIA plate 3590; Corning Life Sciences, Acton, MA).
Coating step One hundred microliters of solubilized ovarian membrane proteins diluted with binding buffer containing 1.2% n-octyl-ß-D-glucopyranoside (typically, 50 µg of membrane protein/ml) was coated in each well for 4 h at 25°C.
Blocking step After washing four times with 200 µl of binding buffer, 200 µl of 5% nonfat skim milk in binding buffer was added to each well and then incubated for 4 h or overnight at 4°C.
Binding step After again washing as described above, 50 µl of binding buffer (for total binding) or binding buffer containing different dilutions of competitive ligands was added to each well. After preincubation of the competitors for 30 min, 50 µl of DIG-vitellogenin (typically, 1 µg/ml) diluted in binding buffer containing 5% nonfat skim milk was added to each well and incubated for 16 h at 25°C. To estimate nonspecific binding, DIG-vitellogenin was incubated with a 250-fold molar excess of unlabeled, purified white perch vitellogenin at this step.
Anti-DIG-poly-POD step
After washing as described above with binding buffer containing 0.25% Triton X-100 (BB-T), 100 µl of a-DIG-poly-POD at a 1:500 dilution in binding buffer containing 1% bovine 
-globulin (from Cohn fraction II and III; Sigma Chemical Co.) was added to each well and incubated for 2 h at 25°C.
Color development step After washing as described above with BB-T, 100 µl of 3,3',5,5'-tetra-methylbenzidine (TMB) enzyme substrate (Kirrkegaard and Perry Laboratory, Gaithersburg, MD) was added to each well and incubated for 15 min at room temperature. The reaction was stopped by adding 100 µl of 1 M phosphoric acid, and absorbance was then read at 450 nm on a Bio-Rad (Hercules, CA) model 3550 microplate reader.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Three white perch yolk proteins were purified by precipitation in water or ammonium sulfate solution followed by gel filtration, ion exchange chromatography, or both. Ovary extracts from vitellogenic perch were dialyzed against distilled water (DW) and centrifuged, after which the precipitate was redissolved in Tris-HCl buffer and used for purification of yolk protein 1 (YP1), while the supernatant was used for purification of yolk protein 2 (YP2) and yolk protein 3 (YP3). Gel filtration of the DW-insoluble yolk fraction on Sepharose 6B yielded two distinct protein peaks. The second peak was pooled and subjected to gel filtration on Superdex 200, yielding a single symmetrical peak at the position corresponding to 310 kDa, which was collected as purified YP1. Purified YP1 generated one precipitin line in immunoelectrophoresis against a-trout lipovitellin or the antiserum raised against white perch ovary extracts (a-OE), but it showed no immunoreactivity to a-charr ßß (Fig. 1). After gradient SDS-PAGE with or without prior reduction of the sample, purified YP1 generated three bands corresponding to 100 kDa, 90 kDa, and 29 kDa (Fig. 2).
|
|
After precipitation of the DW-soluble yolk fraction in 67% saturated ammonium sulfate, the precipitate was collected by centrifugation at 10 000 x g for 20 min, dissolved in 20 mM Tris-HCl buffer (pH 8.0), and dialyzed against the Tris-HCl buffer overnight at 4°C. The resulting dialysate was subjected to anion exchange chromatography on DEAE cellulose using the Tris-HCl starting buffer. The pass-through fraction was concentrated by centrifugal ultrafiltration using a Centricon Plus 20 (Millipore) apparatus and subjected to gel filtration on Superdex 200, yielding several peaks. The second (27 kDa) and third (17 kDa) peaks were collected as purified YP2s (Fig. 3A). The 27-kDa peak (fraction 35) appeared as a single band of 29 kDa after SDS-PAGE under nonreducing conditions, whereas the 17-kDa peak (fractions 37 and 38) appeared as a single band of 16.5 kDa (Fig. 3B). The band of 29 kDa generated by the 27-kDa gel filtration peak resolved into a single band of 16.5 kDa when reduced with 2-mercaptoethanol before electrophoresis (data not shown). In Western blots, both of these bands clearly reacted to a-charr ß', but no other bands were detected by this antiserum (Fig. 3B). Based on the relative molecular masses estimated for these peaks after gel filtration and SDS-PAGE, the 27-kDa and 17-kDa gel filtration peaks were named YP2 dimer (YP2d) and YP2 monomer (YP2m), respectively. Immunoelectrophoresis of the purified YP2s using a-OE and a-charr ß' revealed a single precipitin line, but the YP2s did not generate any precipitin line when reacted against a-trout lipovitellin (Fig. 4).
|
|
After treatment of the DW-soluble fraction of ovary extracts with 67% ammonium sulfate and centrifugation, the supernatant was collected, dialyzed overnight against DW at 4°C, lyophilized, and used as starting material to purify YP3. The freeze-dried powder was dissolved in 200 mM ammonium bicarbonate and subjected to gel filtration on Superdex 200. The chromatography pattern yielded no apparent protein peak (OD 280), but one major and symmetrical phosphorus peak was observed when alkaline-labile phosphorus was measured in the chromatography fractions (data not shown). The peak phosphorus fraction was collected as purified YP3. After SDS-PAGE, YP3 could not be visualized by staining with CBB, but staining of the gel with CBB-Al revealed a major band with an apparent mass of 20 kDa and some diffuse bands located from the 15-kDa position to the dye front (Fig. 5). These latter bands included a faint band at the position corresponding to 15 kDa and two bands with apparent masses less than the smallest (14.4 kDa) marker protein (one migrated just below the 14.4-kDa marker and the other migrated on the dye front).
|
N-Terminal Amino Acid Sequences for Purified Vitellogenin and Yolk Proteins
The N-terminal amino acid sequence (20 residues) of native white perch YP1 was identical to that of white perch vitellogenin. The YP1 sequence was compared with corresponding sequences for striped bass vitellogenin (sVg) and barfin flounder vitellogenin A (bVg A) and vitellogenin B (bVg B), as well as N-terminal amino acid sequences deduced from cDNAs encoding mummichog (Fundulus heteroclitus) vitellogenin I (fVg I) and vitellogenin II (fVg II), rainbow trout vitellogenin (tVg), and haddock (Melanogrammus aeglefinus) vitellogenin A (hVg A) and vitellogenin B (hVg B). The various sequences were aligned to yield the best fit with the sequence for white perch YP1 used as the template (Fig. 6A). Considering the alignment positions between amino acids 1 and 15 of white perch YP1, the N-terminus of white perch YP1 (or vitellogenin) was 80% identical to the aligned sequences of sVg or fVgI and similar (67% or 60% identical) to corresponding sequences for the remaining proteins. For the sequence NFAPEFAAGKTY, there was 100% identity between YP1 (residues 4–15), sVg (residues 4–15), and fVg I (positions 17–28).
|
Peptides from purified YP1 appearing in SDS-PAGE at the 100-kDa, 90-kDa, and 29-kDa positions (see Fig. 2) were separately sequenced at their N-termini. The 100-kDa and 90-kDa peptides had N-terminal sequences identical to that of the native form of YP1, whereas the corresponding sequence of the 29-kDa peptide, YETKFTKNHI, was nearly completely different (9 of 10 residues). The 29-kDa YP1 peptide sequence was aligned to yield the best fit with corresponding sequences for mummichog, trout, and haddock vitellogenin, as well as barfin flounder vitellogenic lipovitellin light chains A (bvLvL A) and B (bvLvL B). The 29-kDa YP1 sequence was 70% identical to bvLvL A (residues 1–10) and 50% identical to bvLvL B (residues 1–10), sharing an identical five amino acid sequence, TKNHI, with the latter peptide (Fig. 6B). It was 60% identical to fVg II (positions 1176–1185) and 50% identical to tVg (positions 1143–1152), hVg A (positions 1157–1166), and hVg B (positions 1133–1142), but only 30% identical to fVg I (positions 979–988).
The N-terminal amino acid sequence of purified white perch YP2 was compared with sequences for barfin flounder ß'-components (bß' A and bß' B) and masu salmon ß'-component (masu ß'), as well as sequences for fVg I and II, hVg A and B, and tVg deduced from cDNA.
The YP2 sequence (17 residues) was used as the template for comparison by best fit (Fig. 6C). The YP2 sequence was highly similar (82%–71% identical) to the sequences for hVgB (positions 1371–1387), bß' A (residues 3–17), masu ß' (residues 2–17), and tVg (positions 1385–1401), and similar (65%–59% identical) to corresponding sequences for the remaining proteins, except hVg A (positions 1393–1409), with which it shared limited homology (41% identical).
Amino Acid Composition of YP3
Analysis of the amino acid composition of perch YP3 (Fig. 7A) indicated that it is very similar in composition to phosvitin from Sakhalin taimen (Fig. 7B). Both proteins had an extremely high serine content (perch/taimen, 57.3%/56.1%).
|
Antigenic Relation Among Vitellogenin, YP1, and YP2s
In double immunodiffusion (Fig. 8), purified YP1 and YP2s both reacted with the antiserum against striped bass vitellogenin (a-FSPP), giving rise to one precipitin line. Precipitin lines of YP2s appeared to fuse with one another, but crossed the YP1 precipitin line. Yolk protein 3 did not react with a-FSPP in double immunodiffusion (data not shown).
|
Plasma from a control male and an E2-injected juvenile perch were subjected to SDS-PAGE and Western blotting using the antisera raised against purified YP1 and YP2 (Fig. 9). Two major bands corresponding to the 180-kDa and 110-kDa positions on the gel were specifically observed when plasma from E2-injected perch was subjected to SDS-PAGE. We consider these bands to represent the primary subunit and a degradation product of white perch vitellogenin, respectively [27]. In Western blotting, a-FSPP showed immunoreactivity to the 180-kDa band and another band at the position corresponding to 74 kDa, as well as some additional minor bands considered to be degradation products of vitellogenin. A similar pattern of immunostaining was observed when the antiserum raised against white perch YP2 (a-YP2) was used for Western blotting. On the other hand, the antiserum to white perch YP1 (a-YP1) reacted to the bands at the 180-kDa and 110-kDa positions but did not immunostain the band corresponding to 74 kDa.
|
Digoxigenin-Labeled Vitellogenin
Unlabeled white perch vitellogenin and vitellogenin labeled with DIG were subjected to SDS-PAGE and then blotted onto a PVDF membrane (data not shown). There were no apparent differences in the electrophoregrams between labeled and unlabeled vitellogenin when they were stained with CBB. Western blotting using a-DIG-poly-POD gave rise to several bands for the DIG-labeled vitellogenin, which were identical to the vitellogenin banding patterns found on CBB-stained gels as reported previously [27]. Nonlabeled vitellogenin could not be visualized after Western blotting using a-DIG-poly-POD.
Binding of DIG-Vitellogenin to Immobilized Ovarian Membrane Preparations
When immobilized ovarian membrane protein preparations were incubated in the presence of a constant amount of DIG-vitellogenin (100 ng/well), the quantity of DIG-vitellogenin bound (total binding) increased linearly with the quantity of membrane protein introduced into the well up to a concentration of 50 µg protein/ml (R2 = 0.98). At a higher membrane protein concentration (100 µg protein/ml), DIG-vitellogenin binding appeared to be approaching a plateau (Fig. 10A).
|
Nonspecific binding of DIG-vitellogenin was relatively constant, being about 12% of total binding at 100 µg/ml membrane concentration. However, there was some nonspecific binding of DIG-vitellogenin even in the absence of membrane protein. We refer to this minor component of binding, which could not be displaced by excess unlabeled vitellogenin, as plate nonspecific binding.
Typical dilution curves for DIG-vitellogenin in the binding assay are shown in Figure 10B. The specific vitellogenin-binding sites (5 µg of membrane protein/well) neared saturation in the presence of 500 ng/ml of DIG-vitellogenin (50 ng/well).
When a constant amount of ovarian membrane protein (5 µg/well) was incubated with a constant amount of DIG-vitellogenin (1 µg/ml) and increasing concentrations (from 0.048 to 500 µg/ml) of unlabeled vitellogenin, bound DIG-vitellogenin was displaced by unlabeled vitellogenin in a dose-dependent manner, covering a range from 90% to 7% of total (maximum) DIG-vitellogenin binding (Fig. 10C).
To verify some tissue specificity of specific DIG-vitellogenin binding, erythrocyte membrane protein preparations were substituted in the assay (data not shown). No specific binding of DIG-vitellogenin to the erythrocyte membrane protein preparations was observed, and total binding of DIG-vitellogenin in incubations containing erythrocyte membrane proteins was less than 5% of total binding measured when oocyte membrane protein preparations were used.
In experiments involving selective omission of all possible combinations of membrane protein, DIG-vitellogenin, a-DIG-poly-POD, and TMB substrate for color development, it was discovered that the low level of plate nonspecific binding was entirely attributable to DIG-vitellogenin binding directly to the plate and not due to nonspecific binding of a-DIG-poly-POD (data not shown). Plate nonspecific binding was not altered by first coating wells with membrane protein. Accordingly, to accurately measure specific displacement of DIG-vitellogenin from the immobilized membrane proteins by competitive ligands (Fig. 11), we used the OD 450 value for wells incubated with DIG-vitellogenin but not coated with membrane protein as "zero" or baseline values for the experiment. The data shown previously in Figure 10 are uncorrected for the low levels of plate nonspecific binding.
|
Ligand specificity of the binding site for vitellogenin on ovarian membranes was evaluated by incubation of membrane protein preparations (coated at 5 µg/well) with constant amounts of DIG-vitellogenin (1 µg/ml) alone (total binding) or in the presence of several-fold molar excesses (relative to DIG-vitellogenin) of competitive ligands. The competitive ligands included unlabeled white perch vitellogenin (400 kDa), striped bass vitellogenin (400 kDa), gag grouper vitellogenin (439 kDa), bester vitellogenin (400 kDa), Sakhalin taimen vitellogenin (540 kDa), and cutthroat trout vitellogenin (540 kDa). Ligands used as negative controls for vitellogenin receptor-binding included BSA (66 kDa), normal goat serum (NGS), keyhole limpet hemocyanin (KLH), and bovine -globulin (BGG, 150 kDa). To eliminate from consideration the possibility that vitellogenin-binding in the assay resulted from specific, nonreceptor binding of lipoproteins to the plate or ovarian membrane protein preparation, human lipoproteins were also used as additional control ligands. These included high-density lipoprotein (HDL; 175–500 kDa), low-density lipoprotein (LDL; 3500 kDa) and very low-density lipoprotein (VLDL; 6000–27 000 kDa). To calculate molar ratios of DIG-vitellogenin to competitive ligands in this experiment, the "mass" of NGS was arbitrarily set at 400 kDa (equal to white perch vitellogenin), that of KLH was set at 400 kDa as a tentative average for its various isoforms, and the average molecular weight of HDL (337.5 kDa) or VLDL (16 500 kDa) was used.
The control proteins (BSA, NGS, KLH, and BGG) had no ability to displace DIG-vitellogenin bound to the membrane preparations even when more than a 1000-fold molar (or protein concentration for NGS) excess of the proteins was added to the incubations (Fig. 11). In addition, none of the human lipoprotein controls showed any ability to displace DIG-vitellogenin, even at the maximum molar excesses relative to DIG-vitellogenin (HDL, 1180; LDL, 114; VLDL, 24.2) that were used (data not shown). Unlabeled vitellogenin from striped bass and grouper effectively displaced DIG-vitellogenin bound to the ovarian membrane protein preparations, and they were indistinguishable from white perch vitellogenin in their potency for this effect. Suramin (10 mM) also completely eliminated specific binding of the DIG-vitellogenin tracer (data not shown). Vitellogenins from salmonids (Sakhalin taimen or cutthroat trout) could decrease DIG-vitellogenin binding by 80% when present at a 500-fold molar excess, whereas sturgeon (bester) vitellogenin was less potent in this regard, decreasing specific binding by only 50% when present in a 1000-fold molar excess.
Purified yolk proteins and ovary extracts were also introduced into the binding assay as competitors with DIG-vitellogenin for specific binding to the ovarian membrane protein preparations (Fig. 12). Our objective was to identify the yolk protein components of vitellogenin that mediate binding of the protein to its oocyte receptor. A 10-fold to 1250-fold molar excess of lipovitellin (310 kDa), phosvitin (20 kDa used as the tentative molecular mass for native phosvitin, see Discussion), and the two forms of ß'-components (17-kDa monomer and 27 kDa dimer) were used in this experiment. For ß'-components, only the results obtained using the 27-kDa dimer are shown (Fig. 12) because results obtained using the monomer were identical. The lipovitellin, phosvitin, and ß'-component preparations were added to wells coated with 5 µg of oocyte membrane protein preparation and then incubated with 500 ng/ml (50 ng/well) of DIG-vitellogenin. White perch phosvitin and ß'-components did not show any ability to displace DIG-vitellogenin from the immobilized membrane protein preparations. Among the purified white perch yolk proteins, only lipovitellin exhibited any ability to compete with DIG-vitellogenin for specific binding sites in the assay. The potency of lipovitellin for displacing DIG-vitellogenin from its binding sites was similar but slightly less than that of unfractionated aqueous extracts of ovaries from vitellogenic white perch. White perch lipovitellin and ovarian extract proteins were able to reduce DIG-vitellogenin binding in the assay by from 45%–50% (250-fold molar excess) to 75%–80% (1250-fold molar excess), respectively, of total binding. It should be recalled that 400 kDa was used arbitrarily as the average molecular weight of extracted ovarian proteins to calculate the molar ratios for this experiment, and this is clearly an overestimate based on results stated above. Therefore, any apparent difference in potency between lipovitellin and ovarian extract proteins is likely to be an experimental artifact.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Another focus of our research was to identify the yolk protein domain that bears the moiety required for binding of vitellogenin to its receptor on growing oocytes. Only two of three major yolk proteins (lipovitellin and phosvitin) from two classes of vertebrates have been examined with respect to vitellogenin receptor-binding. It was verified that the lipovitellin domain of vitellogenins from chickens and Xenopus can displace vitellogenins from their receptors [13, 14]. In both cases, lipovitellin showed some ability to displace labeled vitellogenin from specific binding sites in ovarian membrane or membrane protein preparations, but phosvitin did not. However, to our knowledge, the yolk plasma glycoproteins have not been investigated with regard to vitellogenin receptor-binding, nor have ß'-components or other yolk proteins from fishes and other vertebrates. Collectively, this information was insufficient to focus our attention on lipovitellin as the sole yolk protein domain on vitellogenin responsible for binding of the protein to its receptor.
In the present study, white perch YP1 and YP2s were highly purified immunologically and chromatographically. YP1 was identified as lipovitellin based on its exhibition of several common properties of a teleost lipovitellin. Specifically, it was the major egg yolk protein, a large molecule of 310 kDa in native form with subunit structure similar to that of lipovitellin from other species, and immunoreactive to a-trout lipovitellin [1–3, 8, 33]. White perch lipovitellin was endowed with an N-terminal amino acid sequence identical to perch vitellogenin (Fig. 6). It has long been known that the lipovitellin heavy-chain (Lv I) includes the N-terminus of vertebrate vitellogenin [1–3]. The perch YP1 (and vitellogenin) sequence was clearly homologous with corresponding sequences for vitellogenin from a phylogenetically diverse array of teleosts [44–49]. There was considerable identity among these sequences, especially between residues 4 and 12 (NFAPEFAAG) of the YP1 template. These results confirm our identification of white perch YP1 as lipovitellin and corroborate our previous report that the N-terminal sequence of teleost vitellogenins (and lipovitellins) is highly conserved [44].
In salmonid fishes, as in other vertebrates, lipovitellin is composed of a heavy chain and a light chain [8]. In the present study, the perch lipovitellin heavy chain seemed to be heterogeneous, represented by two apparent subunits (100 kDa and 90 kDa), whereas the light chain appeared to be a 29-kDa monomer (Fig. 2). It is likely that the two heavy chain bands apparent after SDS-PAGE of white perch lipovitellin result from partial degradation of a single major lipovitellin, because the N-termini of the 100-kDa and 90-kDa lipovitellin peptides are identical. The N-terminal amino acid sequence of the perch vitellogenin purified in this and our prior study [44] indicates that it belongs to the vitellogenin family including vitellogenin A or vitellogenin I of barfin flounder and mummichog, respectively, rather than to the group including barfin vitellogenin B and mummichog vitellogenin II (Fig. 6).
The N-terminal sequence of the white perch lipovitellin light chain (29 kDa) was aligned to sequences for the light chains of barfin flounder lipovitellin and corresponding sequences from the various teleost vitellogenins. Although only 10 residues were available for comparison, relatively high homology was observed among the perch and barfin flounder lipovitellin light chains. This comparison also yielded good fit between the perch lipovitellin light chain and fVg II, tVg, hVg A, and hVg B at amino acid positions 1176–1185, 1385–1401, 1157–1166, and 1133–1142, respectively. These positions are located just after the polyserine (i.e., phosvitin) domains of each vitellogenin, in a region that corresponds to the N-terminus of lipovitellin II (Lv light chain) in chicken and Xenopus vitellogenin [50].
In native form, purified white perch YP2 apparently existed either as a monomer (17 kDa in gel filtration, 16.5 kDa in SDS-PAGE) or as a dimer (27 kDa in gel filtration and 29 kDa in SDS-PAGE) composed of 16.5-kDa monomers. The YP2 dimer appeared to be converted into the YP2 monomer when reduced prior to SDS-PAGE, and the dimer and monomer displayed the same antigenicity in double immunodiffusion. Furthermore, a-charr ß' reacted to both forms of white perch YP2.
When the N-terminal amino acid sequence of white perch YP2 was aligned with corresponding sequences for barfin flounder ß'-components A and B, masu salmon ß'-component, and vitellogenins from diverse teleosts, a relatively high degree of identity was observed among the perch YP2 and the ß'-components (75%–71%). Homology of YP2 with the aligned vitellogenin sequences also was high (82%–59%), with the possible exception of the sequence for haddock vitellogenin A (41%). Based on its apparent mass, subunit structure, antigenicity, and sequence at the amino terminus, YP2 was definitively identified as white perch ß'-component. Collectively with the prior reports on vitellogenin-derived, ß'-component yolk proteins in salmonids [8] and flounder [51], our discovery of ß'-component in an advanced perciform fish leads us to speculate that this class of yolk protein may be generally, or at least widely, present among teleost species.
The coding sequence of the vertebrate vitellogenin gene is arranged in linear fashion with respect to yolk protein domains as follows: NH2-lipovitellin I (lipovitellin heavy chain)-phosvitin-lipovitellin II (lipovitellin light chain)-C-terminal coding sequence-COOH [52]. Sequence alignments of barfin flounder and white perch ß'-components with Fundulus vitellogenin I place the ß'-component sequences just downstream from the lipovitellin II (lipovitellin light chain) domain of the vertebrate vitellogenin gene, in the C-terminal coding sequence of vitellogenin cDNAs ([45, 52–54], this study). This vitellogenin cDNA region had not been fully accounted for by any known yolk protein until Yamamura et al. [6] purified a novel yolk protein from chicken eggs, yolk plasma glycoprotein 40, and identified it as the cysteine-rich, 40 kDa, C-terminus of the major chicken vitellogenin, vitellogenin II. A second yolk plasma glycoprotein (YGP 42) derived from vitellogenin I was isolated from chicken yolk plasma in this same study.
The yolk plasma glycoprotein domain of the vitellogenin gene specifies a distinct motif of repeated cysteine residues, 14 out of 15 of which are completely conserved between vitellogenins from rainbow trout, Xenopus, and chicken [6]. Based on its apparent mass in SDS-PAGE (16.5 kDa), assuming that it is a simple polypeptide, the perch ß'-component monomer likely consists of approximately 152 amino acids. Considering the sequence alignments noted above, the domain of the vitellogenin gene encoding perch ß'-component should be homologous to that encoding the amino terminus of YGP 40 beginning at residue 10, localized between amino acid residue 1576 and 1726 of the chicken vitellogenin II amino acid sequence deduced from cDNA [6], and including 5 of the 14 conserved cysteine residues (data not shown). This putative ß'-component domain of the vitellogenin gene accounts for only about half of the C-terminus of vitellogenin represented by YGP 40 and, excluding the cysteine residues, is poorly conserved among vertebrates, being only 37% and 35% identical, or 23% and 19% similar (discounting conservative amino acid substitutions), between chicken (vitellogenin II) and Xenopus (vitellogenin A2) or trout, respectively, using the chicken vitellogenin II sequence as the template. As Matsubara et al. [45] pointed out, ß'-components and additional cysteine-rich yolk polypeptides derived from vitellogenin could be generally present in fishes and other oviparous vertebrates as well.
YP3 also was highly purified and had the following characteristics in common with vertebrate phosvitins described in previous studies: a high content of phosphorus correlated with more than 50% serine residues in its amino acid composition, nonantigenicity, low absorbancy at 280 nm, and low stainability with CBB but not CBB-Al [8, 51, 55–59]. Based on these characteristics, YP3 was identified as white perch phosvitin.
In the present study, purified white perch phosvitin subjected to SDS-PAGE generated a major band corresponding to 20 kDa and some minor or diffuse bands ranging from the 15-kDa position to the dye front. We suspect that the minor, diffuse phosvitin bands might resolve into a single 20-kDa band after dephosphorylation, and that their appearance may be due to the different degrees of phosphorylation of phosvitin, as reported for salmonids [8].
The a-FSPP used in the present study was confirmed to be an antiserum specifically immunoreactive to vitellogenin from fishes of the genus Morone [19, 27]. The result of double immunodiffusion using a-FSPP verified that white perch lipovitellin and ß'-components were yolk proteins derived from vitellogenin. Conversely, the results from Western blotting using a-YP1 (a-Lv) and a-YP2 (a-ß') proved that vitellogenin was the parent molecule of the perch lipovitellin and ß'-components. As is typical of vertebrate phosvitins [8], white perch phosvitin displayed no antigenicity to the various antisera. However, phosvitin was purified as the dominant source of water-soluble phosphoprotein phosphorus from perch ovaries and, on this basis [8], is highly likely to be derived from the major phosphoprotein in blood plasma of vitellogenic females; namely, vitellogenin.
The antisera raised against lipovitellin and ß'-component (a-Lv and a-ß') showed different spectra of immunoreactivity to product bands generated by vitellogenin in plasma from vitellogenic white perch subjected to SDS-PAGE. Neither antiserum reacted with any protein in plasma from male fish. However, when plasma from vitellogenic (E2-injected) fish was used, both antisera reacted to the main 180-kDa band, which we previously identified as the primary polypeptide subunit of perch vitellogenin [27]. However, the bands migrating at 110 kDa and 74 kDa, which we consider to be degradation products of vitellogenin, were selectively stained by a-Lv and a-ß', respectively. Similar differences in specificity of various antisera raised against vitellogenin and yolk proteins have been reported for Sakhalin taimen [9].
Tao et al. [19] previously demonstrated the presence of a specific receptor for vitellogenin on ovarian membranes of white perch, consistent with several other recent reports on teleosts [15, 17, 20, 60, 61]. As in chickens and Xenopus, there appears to be a system for receptor-mediated endocytosis of vitellogenin by fish oocytes. In order to understand the physiological mechanism of oogenesis, considerable attention has been paid to assessing changes of the affinity and maximum binding capacity of the receptor (vitellogenin receptor) preparations during ovarian growth and maturation [17, 18, 60]. These investigations relied heavily on techniques involving use of radioisotopes to quantify vitellogenin binding to receptors in ovarian membrane and membrane protein preparations. Although the vitellogenin receptor-binding assay used in the present study was not designed to generate data for calculating binding affinity or capacity, the characteristics of DIG-vitellogenin binding were grossly similar to those observed using a more quantitative assay of the white perch vitellogenin receptor. When we examined specific binding of 125I-vitellogenin to ovarian membranes from white perch at various maturational stages (early vitellogenic-preovulatory), the median dissociation constant (Kd) for the specific binding was 336 nM, a value similar to those reported for other fish vitellogenin receptors [19]. Given our estimate of the molecular weight of perch vitellogenin (400 kDa), this Kd predicts that ovarian vitellogenin receptors would be 50% occupied at prevailing vitellogenin concentrations of approximately 134 µg/ml. This value is fairly similar to the concentration of unlabeled vitellogenin required to reduce total binding of DIG-vitellogenin by 50% in the vitellogenin receptor-binding assay used in the present study (Fig. 10).
Our ultimate goal is to identify the receptor-binding site on the vitellogenin molecule, as has been accomplished for the related lipoproteins, apolipoprotein B-100, and apolipoprotein E [62, 63]. An objective of the present study was to definitively verify which yolk protein domain of vitellogenin mediates receptor binding. For these purposes, it is useful to have a simple and safe, solid-phase binding assay using nonisotopic ligand for large-scale screening of functional properties (e.g., receptor-binding) of numerous peptides or subunits derived from vitellogenin. In the present study, a receptor (vitellogenin receptor)-binding assay was developed and validated for white perch vitellogenin using DIG-labeled tracer and a 96-well microtiter plate format. Nonisotopic binding assays using 96-well plates have been developed for, as examples, receptors for benzodiazepine drugs [64] or nicotinic acetylcholine [65], a binding protein (zona pellucida ZP3 alpha glycoprotein) on sperm membranes [66], and an inhibitor of serine proteinase [67].
Binding of DIG-labeled vitellogenin to ovarian membrane proteins immobilized on 96-well plates exhibited several characteristics expected of a vitellogenin receptor. The quantity of specifically bound DIG-vitellogenin increased in direct proportion to the concentration of membrane proteins used to coat the plates, and specific binding depended on the concentration of DIG-vitellogenin in the incubation. The DIG-vitellogenin binding sites also appeared to be saturable. Tissue specificity of the binding site for vitellogenin was examined by comparing binding of DIG-vitellogenin to oocyte membrane proteins with its binding to erythrocyte membrane proteins, used as a negative control [19]. No specific binding of DIG-vitellogenin could be detected when erythrocyte membrane proteins were substituted into the assay. The specific binding sites for vitellogenin also exhibited the ligand specificity expected of a vitellogenin receptor. As in our previous study [19], specific binding of the vitellogenin tracer was completely displaced by unlabeled perch vitellogenin or suramin, a sulphated, polycyclic hydrocarbon known to block binding of vitellogenin and other lipoproteins to their receptors [19, 68, 69]. In contrast, the various proteins used as negative controls (BSA, NGS, KLH, and BGG) were completely unable to displace labeled vitellogenin from its receptor, even if more than a 1000-fold molar excess of these competitive ligands was added to the incubations. Specific binding of vitellogenin in the assay could not have resulted from specific, nonreceptor interactions of lipoproteins with the plate or ovarian membrane protein preparations because a variety of human lipoproteins (HDL, LDL, and VLDL) showed no ability to compete with DIG-vitellogenin for specific binding sites in the assay.
Unlabeled vitellogenins purified from highly advanced teleosts close and distantly related to white perch (striped bass and gag grouper) were as effective as white perch vitellogenin in displacing the DIG-vitellogenin specifically bound to its receptor. Unlabeled vitellogenin from more primitive teleosts, the salmonids (Sakhalin taimen and cutthroat trout), displaced specifically bound DIG-vitellogenin less potently, about as effectively as the crude preparation of perch egg yolk proteins. Although vitellogenin from a preteleost, the hybrid sturgeon (bester), had the lowest apparent affinity for the perch vitellogenin receptor, it could effectively displace up to 50% of specific DIG-vitellogenin binding when present at a 1000-fold molar excess in the assay. These results indicate, as has been reported earlier [14, 19, 70], that the receptor-binding domain of vitellogenin must be well-conserved among oviparous animals. Furthermore, they suggest that the degree of conservation of receptor binding affinity may, to some extent, follow phylogenetic relationships among fishes.
Ligand blotting experiments using chicken 125I-lipovitellin and membrane proteins prepared from chicken oocytes or coho salmon ovaries indicated that binding of vitellogenin to its receptor is likely mediated by the lipovitellin domain, but not the phosvitin domain, of vitellogenin [15]. However, there have been no reports of such studies using a homologous system to evaluate binding of purified yolk proteins to the vitellogenin receptor in fishes. Furthermore, no attention has been paid to potential vitellogenin receptor-binding by yolk proteins derived from the C-terminus of vitellogenin, including YGPs or ß'-components, the latter of which may be widely distributed in fishes ([8, 51], this study). In the present study, purified yolk proteins, including lipovitellin, phosvitin, and ß'-components were tested in the vitellogenin receptor-binding assay as competitive ligands. Lipovitellin and ovary extracts were able to reduce DIG-vitellogenin binding to the ovarian membrane proteins by 75%–80% when present at a 1250-fold molar excess relative to DIG-vitellogenin. However, ß'-components and phosvitin exhibited no capability to displace bound DIG-vitellogenin over the same range of excess concentration. In contrast, only a 10-fold molar excess of unlabeled vitellogenin was able to reduce DIG-vitellogenin binding by 80%, suggesting that specific conversion of vitellogenin to lipovitellin in vitellogenic oocytes might modify critical elements of the structural environment surrounding the receptor-binding site on lipovitellin. In fishes, these modifications could conceivably include removal of portions of phosvitin, ß'-components, or cysteine-rich yolk polypeptides derived from the C-terminus of vitellogenin that cooperate with the receptor-binding domain on lipovitellin to increase binding affinity. Further investigations are needed to explain the difference between lipovitellin and vitellogenin in their ability to compete with DIG-vitellogenin for receptor-binding.
In summary, four vitellogenin-derived egg yolk proteins, YP1, YP2 monomer, YP2 dimer, and YP3, were isolated from ovaries of vitellogenic white perch, and identified as lipovitellin, ß'-components, and phosvitin, respectively. Collectively, with reports on ß'-components from salmonids and flounder, discovery of this yolk protein in an advanced perciform indicates that ß'-component-like yolk proteins may be widely distributed among fishes. The ß'-component domain of white perch vitellogenin was mapped to the amino terminus of chicken YGP 40, covering about half of the YGP sequence representing the C-terminus of vitellogenin, suggesting that additional cysteine-rich yolk polypeptides derived from vitellogenin may remain to be discovered in fishes. A novel, nonisotopic, receptor binding assay for vitellogenin, based on DIG-vitellogenin tracer and ovarian membrane proteins immobilized in 96-well plates, was developed and validated for comparing the vitellogenin receptor-binding potency of various ligands. Using this assay it was proven, for the first time in a completely homologous system, that the lipovitellin domain of a teleost vitellogenin is primarily responsible for mediating receptor-binding. However, lipovitellin was much less potent than vitellogenin in this regard, indicating that the other YP domains of vitellogenin may make some contribution to receptor binding. These findings, coupled with development of the simple, high through-put, colorimetric assay for vitellogenin binding, set the stage for identification of the specific moiety on lipovitellin and vitellogenin that is recognized by the vitellogenin receptor.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
1 This work was supported by grant NA86RG0036 (project R/MBT-3) from the National Sea Grant College Program to the North Carolina Sea Grant College Program, by grant 9805-ARG-0444 from the North Carolina Biotechnology Center to C.V.S., and by a research fellowship to N.H. from the Japan Society for the Promotion of Science. 
2 Correspondence: Craig V. Sullivan, David Clark Laboratories, Room 115, Department of Zoology, College of Agriculture and Life Sciences, North Carolina State University, Raleigh, NC 27695-7617. FAX: 919 515 2698; craig_sullivan{at}ncsu.edu
Accepted: March 19, 2002.
Received: January 29, 2002.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
,20ß,21-trihydroxy-4-pregnen-3-one (20ß-S) on ovarian membranes of stripe
