| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Department of Animal Resource Production,3 the United Graduate School of Agricultural Science,
Departmentof Biology, Faculty of Education,4
Department of Biological Diversity and Resources,5 Faculty of Agriculture, Gifu University, Gifu 501-1193, Japan
Hokkaido National Fisheries Research Institute,6 Kushiro, Hokkaido 085-0802, Japan
Department of Zoology,7 College of Agriculture and Life Sciences, North Carolina State University, Raleigh, North Carolina 27695-7617
Graduate School of Fisheries Sciences,8 Hokkaido University, Hakodate, Hokkaido 041-8611, Japan
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
ß'-component, estradiol, gamete biology, gametogenesis, lipovitellin, oocyte development, ovary, phosvitin, teleost, vitellogenin, yolk protein
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
In higher oviparous vertebrates, multiple forms of Vg have been discovered. Three types of Vg protein and their corresponding cDNAs have been found in the chicken [7, 14]. In Xenopus laevis, three distinct Vg proteins [15] and four Vg genes [16–18] were characterized. Molecular cloning studies have revealed the primary structures of dual Vgs in several teleost species. LaFleur et al. [19, 20] isolated complete cDNAs encoding two distinct Vgs (VgI and VgII) in the mummichog, Fundulus heteroclitus. Full-length cDNAs encoding two separate Vgs were later obtained from haddock, Melanogrammus aeglefinus (VgA and VgB) [21] and medaka, Oryzias latipes (Vg1 and VgII; GenBank accession numbers AB064320 and AB074891). Dual Vg genes also were cloned as cDNA from barfin flounder, Verasper moseri (GenBank accession numbers AB181833 and AB181834) following biochemical and functional characterization of the two Vg proteins (VgA and VgB) in this species [22]. All of these dual teleostean Vgs can be classified into two groups based on their functional properties and similarities of their N-terminal and internal amino acid sequences or corresponding peptide sequences deduced from cDNA [23]. The VgA group includes mummichog (Fun) VgI, haddock (Had) VgA, medaka (Med) Vg1, and barfin flounder (Bar) VgA; while the Vg B group includes Fun VgII, Had VgB, Med VgII, and Bar VgB. The coding sequences of Vgs in both groups is arranged in a linear fashion with respect to yolk protein domains as follows: NH2-LvH-Pv-LvL-ß'c-C-terminal coding region-COOH [2, 13, 23]. We refer to Vgs with this structure as complete Vgs to distinguish them from novel forms of Vg lacking a Pv domain.
Two forms of Vg protein, one complete and one incomplete, have been discovered in tilapia species, Oreochromis aureus [24] and O. mossambicus [25, 26]. When compared with the other tilapia Vg, one Vg had an unusually low native mass (
300 kDa) and a low phosphorous content [24, 25]. An incomplete form of Vg with similar characteristics also has been described for medaka, O. latipes [27]. Recently, Wang et al. [28] identified cDNAs transcribed from multiple Vg genes by probing an expressed sequence tag liver library of adult female zebrafish, Danio rerio. One of the zebrafish cDNAs (vg3) encoded an unusual Vg that lacked a polyserine domain and had low sequence similarity to other piscine Vgs. With regard to these characteristics, this phosvitinless (Pvl) Vg appeared to be of an ancient type similar to insect Vgs. A PvlVg protein (Vg-320) [29] and its cDNA (GenBank accession number AB088473) also were discovered in estrogen-treated Japanese common goby, Acanthogobius flavimanus. Very recently, Hiramatsu et al. [23] purified and identified three different forms of Vg circulating in estrogen-treated white perch, Morone americana. The two complete perch Vgs were classified into the Vg groups A and B, while the third was classified as a C-type Vg, along with the PvlVgs described above. To our knowledge, there has been no species for which all three types of Vg protein (A, B, and C) and their corresponding full-length cDNAs have been identified and characterized.
Further research is needed to better understand the evolution of multiple piscine Vgs and their physiological significance in reproduction. Current information on specialized physiological functions of different forms of Vg is limited to a few oviparous teleosts, including marine species (barfin flounder and haddock) [21, 22] and brackish water fishes (white perch) [23] whose oocytes undergo remarkable hydration during final maturation. In these species, it appears that differential processing of yolk proteins derived from each type of Vg generates free amino acids (FAAs) that drive oocyte hydration during maturation and may provide a pool of diffusible nutrients to support early embryonic nutrition (see reviews in [3, 30]). In barfin flounder, most of the yolk proteins derived from the A-type Vg are cleaved completely into FAAs during oocyte maturation, whereas the Lv derived from the B-type Vg remains largely intact and may be selectively used by late stage embryos and larvae [22, 31]. Aside from an apparently reduced capacity to transport phosphorus or calcium, specific physiological functions and product yolk proteins of C-type teleost Vgs remain to be definitively identified.
Comparatively little is known about ovoviviparous or viviparous teleosts with regard to the occurrence of multiple Vgs or the potential roles of yolk proteins derived from different forms of Vg during gestation. Prior investigations of these species have involved identification and characterization of only a single immunoreactive form of Vg (viviparous blenny, Zoarces viviparus [32]; ovoviviparous white-edged rockfish, Sebastes taczanowskii [33]; viviparous eelpout, Z. elongates [34]; ovoviviparous mosquitofish, Gambusia affinis [35]). The ovoviviparous mosquitofish is an established model for reproductive physiology and is considered to be a valuable bioindicator species for monitoring estrogenic contamination of aquatic environments [35]. Although Vg has become accepted as a biomarker of fish exposure to endogenous or exogenous estrogens in such research (reviews in [36–41]), with the exception of our preliminary report on the subject [42], the potential multiplicity of Vg in mosquitofish has not been previously explored.
The objectives of the present study were to identify and characterize multiple forms of Vg in mosquitofish and to discover the fate of each Vg during oogenesis. In pursuit of these objectives, we undertook a complete description and classification of the three different forms of transcribed and translated Vg products in mosquitofish and evaluated molecular alterations of the three types of Vg (A, B, and C) during their processing into product yolk proteins.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Mosquitofish were collected from an irrigation canal at Nagashima, Mie, Japan, and were subsequently kept in a pond at Gifu University, Gifu, Japan under natural ambient conditions of photoperiod and water temperature. Adult fish were transferred to indoor, 50-L glass aquaria and held under a 16L:8D photoperiod at 25°C. After 2 wk of acclimation, the fish initiated a natural reproductive cycle. More than 100 females were given an i.p. injection of 5 µl of estradiol-17ß (E2) solution (0.8 mg E2/ml propylene glycol) to stimulate hepatic Vg synthesis. Seven days later, blood and tissue samples were taken from anesthetized (ethyl-4-aminobenzoate) males, naturally vitellogenic females, and E2-treated females. Blood samples were taken from the severed caudal vein using heparinized microhematocrit tubes (TERUMO, Tokyo, Japan) and then centrifuged at 5000 x g for 5 min to separate the plasma, which was frozen at –80°C until use. To collect fully grown vitellogenic follicles, ovaries were removed from gravid females and the follicles were isolated in ice-cold Ringers solution for mummichog oocytes (113 mM NaCl, 5 mM KCl, 2 mM CaCl2, 1 mM MgSO4, 1 mM Na2HPO4, 1 mM NaHCO3, 1 mM sodium pyruvate, 5 mM glucose, 5 mM Hepes, 0.001% phenol red, and 0.1 M NaOH pH 7.5 [43]). The fresh follicles were then homogenized in a 1.5-ml microfuge tube with ice cold 0.9% NaCl solution containing 0.005% phenylmethylsulfonylfluoride (PMSF) and 0.01% NaN3 at a concentration of 10% (wt/vol). After centrifugation at 10 000 x g for 15 min at 4°C, the supernatant was collected as the follicle extract (Fe) and stored at –80°C until use.
Chromatography
Hydroxylapatite column chromatography was performed using Bio-Gel HT (Bio-Rad, Hercules, CA). Approximately 20 cm3 of the gel was packed into a 3.0 x 3.0 cm column and equilibrated with 0.1 M potassium phosphate buffer at pH 7.3 (0.1 M KP). For separation of Vgs from plasma samples, the column was eluted stepwise with 0.2 M KP and then 1.0 M KP. Eluted fractions were collected at a volume of 1.6 ml per tube. Desired protein fractions were collected, dialyzed overnight at 4°C against the 0.9% NaCl solution mentioned above using Visking tubing (cutoff <10 kDa; Visking, London, U.K.), and then concentrated with a water pickup polymer (Ms. BTAURY-KN, ATTO, Tokyo, Japan).
As described by Matsubara and Sawano [11], Vgs and yolk proteins were purified by gel filtration using a prepacked column of Superose 6 HR 10/30 (1 x 30 cm; Amersham Biotech, Buckinghamshire, England). The native mass of Vgs and yolk proteins was estimated by chromatography on a Superose 6 column calibrated with standard proteins (HMW and LMW Gel Filtration Calibration Kit; Amersham Biotech).
High-performance anion exchange chromatography was carried out using a prepacked Mono Q HR 5/5 column (5 x 50 mm; Amersham Biotech). The column was equilibrated with 20 mM Tris-HCl pH 8.0 containing 150 mM NaCl, and protein preparations were eluted with a linear NaCl gradient (150–500 mM) at a flow rate of 1.0 ml/min.
Analyses of Phosphorous and Lipid
Phosphoprotein phosphorus in the purified preparations of Vg and yolk proteins was measured as described by Gamst and Try [44]. The lipid composition of purified yolk proteins was assessed using enzymatic procedures with commercial kits (Phospholipid B-test, Triglyceride G-test, and Cholesterol E-test; Wako, Osaka, Japan).
Immunological Procedures
Polyvalent antisera were raised in rabbits against purified preparations of the 600 kDa Vg fraction of plasma proteins from E2-injected fish (a-600 Vg) and the 400 kDa yolk protein fraction from follicle extracts (a-400 Yp). To ensure specificity, the a-600 Vg and a-400 Yp were absorbed (ab) with an appropriate quantity (1 mg antigen/ml antiserum) of purified 400 kDa yolk protein (400 Yp) and 600 kDa Vg (600 Vg), respectively, and the absorbed antisera were designated as ab-a-600 Vg and ab-a-400 Yp. Immunoelectrophoresis was performed according to the method described by Grabar and Williams [45] in 1.2% agarose gels containing 25 mM barbital buffer pH 8.6. After electrophoresis, troughs in the gels were filled with 100 µl of ab-a-600 Vg or ab-a-400 Yp. The gels were then incubated in a moist chamber for 24 h at room temperature, washed with 2% NaCl, dried, and stained with Amido Black 10B. For further testing of the antigenicity of purified Vgs and yolk proteins, double immunodiffusion was performed in 1.0% agarose gels containing 0.9% NaCl as described by Ouchterlony [46].
Polyacrylamide Gel Electrophoresis and Transblotting
We performed SDS-PAGE on precast 5%–20% polyacrylamide gradient gels (ATTO). The molecular mass of individual protein bands was estimated on gels calibrated with Broad Molecular Marker proteins (Sigma, St. Louis, MO). After SDS-PAGE, semidry transblotting was performed using polyvinylidene difluoride (PVDF) membranes (Immobilon-PSQ; Millipore, Bedford, MA) and a semidry transfer apparatus (Trans-Blot SD; Bio-Rad). The transferred proteins were stained with Coomassie Brilliant Blue R 250, destained in 50% methanol and 5% acetic acid, washed with Milli Q (Millipore) water, and air-dried. Desired protein bands were then cut out from the membranes for N-terminal peptide sequencing.
N-Terminal and Inner Amino Acid Sequence Analysisof Yolk Proteins
N-terminal amino acid sequences of banded peptides generated from purified yolk proteins during SDS-PAGE were analyzed with a PPSQ-21 Protein Sequencer (Shimadzu, Kyoto, Japan) using the transblotted PVDF membrane chips as the sample. To obtain internal amino acid sequences of the 400 Yp, the purified 400 Yp preparation was digested for 30 min at 37°C with trypsin (Sigma) in 0.4 M phosphate buffer pH 7.0. The digested polypeptides were then subjected to SDS-PAGE, transblotting, and N-terminal peptide sequencing, as described above.
Isolation of cDNA Clones
Total RNA from a sample of livers pooled from several E2-treated mosquitofish was prepared with ISOGEN (Nippon-GENE, Tokyo, Japan). Poly (A)+ RNA (mRNA) was isolated from the total RNA using oligo(dT)-latex beads (Oligotex dT; Takara Bio, Otsu, Japan). A liver cDNA library was then constructed in Lambda ZAP II using this mRNA and a cDNA synthesis kit (Invitrogen, Carlsbad, CA). This cDNA library contained 2.28 x 106 pfu/ml with <0.3% nonrecombinants. Figure 1 shows the strategy used to obtain complete cDNA sequences for the three different mosquitofish Vgs. For isolation of 600 VgA cDNA, the sequence of a specific polypeptide domain of the protein (VSKAAAAECRFIEDTLYTFNNKSY) was used to design degenerate PCR primers. This peptide sequence was deduced from N-terminal amino acid sequences of polypeptides derived from the 28 kDa yolk protein (Fe-21 and 19, numbers indicate kDa; Table 1). Degenerate oligonucleotides (600 VgA-F, 5'-AAGGCTGCTGCAGCTGAATGYAGRTT-3'; nested 600 VgA-F, 5'-CTSTACACATTCAACAACAAGAGYTA-3') corresponding to underlined positions of the targeted 600 VgA domain (shown above) were synthesized as forward PCR primers for rapid amplification of 3' cDNA ends (3'-RACE) using the SMART RACE cDNA Amplification Kit (Clontech, Palo Alto, CA) with mRNA from livers of E2-injected fish as the template. The resulting PCR product (approximately 800 base pairs; bp) was cloned using a TOPO TA cloning kit (Invitrogen), subcloned, sequenced in both directions, and compared with published piscine Vg cDNA sequences to verify that it encoded part of an A-type of Vg. A VgA gene-specific primer pair (sense 1, 5'-TGCTGCAGCTGAATGCAGAT-3'; antisense 2, 5'-CCTTTCATCCAGTCAGCAAT-3') was designed to target the partial VgA cDNA sequence and used in PCR with a digoxigenin (DIG) DNA labeling mixture (Boehringer-Mannheim, Mannheim, Germany) to amplify a DIG-labeled probe for screening the liver cDNA library. A total of 5 x 104 plaques were screened using this probe. After two rounds of screening, positive clones were selected, and the clone containing the longest cDNA insert was subcloned and sequenced in both directions by the primer-walking method using a BigDye terminator v3.0 cycle sequencing kit and ABI PRISM 310 Genetic Analyzer (Applied Biosystems, Foster City, CA). Because the 5' terminus of the insert in this clone was truncated, a second probe targeting a sequence near the 5' end of the insert was amplified from the liver cDNA library using a new primer pair (sense 3, 5'-AGCTYGTRGAACCTGAGCTC-3'; antisense 4, 5'-TATGCCAACCCKATGTCYTT-3') and the DIG DNA labeling mixture. The second probe was used to screen the liver cDNA library again and the clone containing the longest insert including a start codon (ATG) was selected after two rounds of screening, subcloned, and sequenced across its cloning site in both directions.
|
|
Isolation of cDNA clones encoding 600 VgB was carried out using a degenerate pair of oligonucleotide primers (600 VgB-F, 5'-AAGTTTGAGTATGCYAACGG-3'; 600 VgB-R, 5'-GCRTCAATGTGAACTCTGAG-3') designed to target nucleotide sequences showing high similarity between Fun VgII (GenBank accession number AAB17152 and Bar VgB but low similarity to Fun VgI (GenBank accession number T43141). PCR was carried out using the liver cDNA library as template and the PCR product (approximately 1400 bp) was cloned by the TOPO TA method, subcloned, sequenced in both directions, and verified to encode a portion of a B-type of piscine Vg. A DIG-labeled probe was then amplified from the liver cDNA library as described above, with the exception that a VgB gene specific primer set (sense 5, 5'-GTTTGAGTATGCCAACGGTGT-3'; antisense 6, 5'-GTAGCAGTGGCCTCTAAGAT-3') targeting nucleotide sequences of the partial VgB cDNA was used. A total of 1 x 105 plaques were screened using this probe. The two clones containing the longest inserts were selected after two rounds of screening and their inserts were sequenced in both directions. Because the 5' termini of these clones were truncated, 5'-RACE was performed using 5'-RACE System (Invitrogen) with the 600 VgB-R and nested 600 VgB-R (5'-CACCGTTGGCATACTCAAAC-3') primers, which were designed to target nucleotide sequences of inserts in the truncated clones. The resulting RACE product was then subcloned and sequenced, as described above.
To isolate cDNA clones encoding the 400 Vg, portions of two polypeptide sequences (TYGPLEKKGKIIYSFEDVDIN and PECSTTEAIFNVKAFAIXENQKPE) were used to design degenerate PCR primers. These peptide sequences were deduced from the N-terminal amino acid sequence of intact yolk proteins or peptides present in trypsin digests (t) of the 400 Yp molecule (tFe-80, Fe-33, 26; Table 1). Degenerate oligonucleotides (400 Vg-F, 5'-GARAARAARGGYAAAATTAT-3'; 400 Vg-R, 5'-GGYTTYTGRTTYTCAATAAT-3') corresponding to underlined positions of the targeted 400 Vg polypeptide domains (shown above) were synthesized and used as PCR primers with the liver cDNA library as template. The PCR product was cloned by the TOPO TA method, subcloned, sequenced in both directions, and verified to encode a portion of a Pvl (C-type) piscine Vg. A VgC gene specific primer set (sense 7, 5'-CTACTCGTTTGAGGATGTGG-3'; antisense 8, 5'-CTTGTTCACCAGAGAGCCAT-3') was designed to target nucleotide sequences of the partial VgC cDNA and used to amplify a DIG-labeled probe from the cDNA library, as described above. A total of 4 x 104 plaques were screened using this probe. The two clones containing the longest 400 Vg DNA inserts, one of which contained a start codon, were selected after two rounds of screening and their inserts were sequenced in both directions.
Sequence Analyses
Analyses of the deduced Vg amino acid sequences were performed using the FASTA homology search tool at the DNA Data Bank of Japan (available at http://www.ddbj.nig.ac.jp/E-mail/homology.html). Alignments of amino acid sequences of 600 VgA, 600 VgB, and 400 Vg, and assessments of identity and similarity between these three mosquitofish Vgs and other piscine Vgs were performed using CLUSTAL W [47]. Alignment of putative receptor binding regions in the mosquitofish Vgs and those of other teleosts was manually conducted using MacVector software version 7.0 (Oxford Molecular Ltd., Madison, WI). The deduced amino acid sequence of this region identified in tilapia (O. aureus) Vg1 [48] (GenBank accession number AAD01615 was used as a template to perform a multiple CLUSTAL W alignment with the deduced amino acid sequence of the three mosquitofish Vgs and zebrafish vg3 [28] (GenBank accession number AAG30407.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
To detect circulating female-specific proteins whose production was induced by estrogen in mosquitofish, samples of blood plasma from males (MP), vitellogenic females (FP) and E2-treated females (EP) were subjected to gel filtration on a column of Superose 6 (Fig. 2). The elution profile of EP, but not MP, showed a major peak of UV absorption (280 nm) at an elution volume of 14.1 ml with a shoulder eluting at 15.4 ml (Fig. 2, A and C). Molecular masses of one or more proteins in the estrogen-induced peak and shoulder were estimated to be approximately 600 kDa and 400 kDa, respectively. The 600 kDa peak also was observed in the chromatogram of FP (Fig. 2B), indicating that it represents a plasma protein constituent of vitellogenic females but not of males. On the basis of these observations, the 600 kDa and 400 kDa protein fractions were considered to be crude preparations of mosquitofish Vgs. To identify Vg-derived yolk proteins in mosquitofish oocytes, Fe prepared from vitellogenic females was subjected to chromatography on Superose 6 (Fig. 2D). Three peaks of UV absorption, corresponding to approximate molecular masses of 560 kDa, 400 kDa, and 28 kDa (elution volumes 14.3 ml, 15.4 ml, and 18.3 ml, respectively) were observed. The three peak fractions (shaded areas in Fig. 2D) were collected and designated as yolk protein preparations 560 Yp, 400 Yp, and 28 Yp, respectively. In addition, a phosphorus-rich peak of Fe protein, which did not correspond in elution position to any of the three peaks of UV absorption, was detected (data not shown). The estimated mass of phosphoproteins in this fraction was approximately 32 kDa.
|
To confirm that the E2-induced 600 kDa and 400 kDa plasma proteins were mosquitofish Vgs, these proteins were further purified and characterized. The 600 kDa protein was purified from EP by two steps of column chromatography, on hydroxylapatite and Superose 6, respectively (Fig. 3A). The elution pattern of 600 kDa protein from Superose 6 showed a single symmetric peak (data not shown) and the peak fractions were pooled, concentrated, and designated as the purified 600 Vg preparation. Because its concentration was highest in Fe (compare Fig. 2, C and D), the 400 kDa protein was purified from Fe instead of from EP or FP. A diagram of procedures for purification of yolk proteins from Fe is shown in Figure 3B. To purify 400 Yp, the 400 kDa peak fraction from gel filtration was subjected to ion-exchange chromatography on Mono Q. Pass-through fractions eluted by 20 mM Tris-HCl buffer pH 8.0 containing 150 mM NaCl were collected and concentrated. The purity and molecular mass of protein in these fractions was confirmed by rechromatography on the Superose 6 column. Proteins in the resulting peak fractions were pooled and designated as purified 400 Yp. The purified 600 Vg and 400 Yp were then used to raise antisera against each protein (a-600 Vg and a-400 Yp) in rabbits, and the antisera were cross-absorbed as described above to ensure specificity of the final products (ab-a-600 Vg and ab-a-400 Yp).
|
Identification of Three Forms of Vg and Their Derivative Yolk Proteins
Figure 4A shows the results of immunoelectrophoresis conducted to test the antigenicity of MP, FP, EP, purified 600 Vg and 400 Yp, and Fe against the ab-a-600 Vg and ab-a-400 Yp antisera. Strong cross-reactions to ab-a-600 Vg were observed for FP, EP, and purified 600 Vg, but not for MP and purified 400 Yp. Two precipitin arcs were formed at an anodic position near the sample well by FP, EP, and purified 600 Vg. The arcs produced by the EP and 600 Vg preparations clearly crossed each other. On the basis of these observations, it was concluded that the 600 Vg preparation contained two different ab-a-600 Vg-reactive proteins (600 Vgs) with distinct antigenicities. Two precipitin arcs were also formed when Fe was subjected to immunoelectrophoresis using the ab-a-600 Vg antiserum, suggesting that the two 600 Vgs give rise to two distinct types of product yolk proteins. The ab-a-400 Yp antiserum reacted strongly to FP, EP, and purified 400 Yp during immunoelectrophoresis, but it did not cross-react to MP and purified 600 Vg. As the reactive protein produced only a single precipitin line against ab-a-400 Yp, the 400 Yp appeared to contain only a single immunoreactive species of protein, one derived from 400 Vg. To examine mutual antigenic relationships between the two 600 Vgs and 400 Yp, double-immunodiffusion was performed (Fig. 4B). Precipitin lines formed against ab-a-600 Vg by purified 600 Vg and EP were fused to each other, as expected, because 600 Vg was purified from EP. The 400 Yp and EP each formed a single precipitin line against ab-a-400 Vg, and these two lines fused together. Collectively, with the results of immunoelectrophoresis, these observations indicate that Fe from vitellogenic females contains two distinct yolk proteins derived from 600 Vg and that the single 400 Yp in oocytes is identical in antigenicity and, therefore, derived from the 400 kDa protein (400 Vg) in EP. Moreover, the precipitin lines formed by EP against ab-a-600 Vg and ab-a-400 Yp crossed one another, confirming that the 400 Vg differs immunologically from the 600 Vgs.
|
Banding patterns in SDS-PAGE of plasma proteins, Fe, and Vgs isolated by gel filtration on Superose 6 (Fig. 2, shaded area) are shown in Figure 5. The EP and FP, but not MP, produced two major bands of 195 and 142 kDa in SDS-PAGE. After gel filtration of EP on Superose 6, proteins generating the 195 and 142 kDa bands were detected mainly in the 600 kDa peak fraction and the 400 kDa shoulder fraction, respectively. The Fe generated 11 yolk protein bands in SDS-PAGE, which were named for their apparent molecular mass as Fe-195, Fe-175, Fe-142, Fe-126, Fe-112, Fe-66, Fe-55, Fe-33, Fe-26, Fe-21, and Fe-19 (Fig. 5). When the main UV-absorbing peaks of yolk proteins eluting from the Superose 6 column (560 Yp, 400 Yp, and 28 Yp) were subjected to SDS-PAGE, the 560 Yp generated Fe-195, Fe-175, Fe-126, Fe-66, and Fe-55. The 400 Yp was resolved into Fe-142, Fe-126, Fe-112, Fe-33, and Fe-26. The Fe-21 and Fe-19 were generated only by the 28 Yp fraction, and they were the only bands produced by this fraction.
|
To further characterize and identify proteins in the 560 Yp, 400 Yp, and 28 Yp gel filtration fractions, proteins in these fractions were separated by SDS-PAGE and transferred to PVDF membranes for N-terminal amino acid sequencing (Table 1). Although Fe-195 and Fe-175 produced by 560 Yp could not be sequenced and were possibly blocked on their N-terminus, some sequence data were obtained for Fe-126 (6 residues), Fe-66 (6 residues), and Fe-55 (5 residues). The Fe-126 sequence, TQVNYA, had high identity (4 of 6 residues) and similarity (6 of 6 residues) with residue 16–21 in the LvH domain of the deduced amino acid sequence of Fun VgII (GenBank accession number U70826). The N-terminal amino acid sequences of Fe-66 and Fe-55, DKILSS and AARSN, respectively, yielded a good fit to amino acid positions 1072 to 1077 (5 of 6 residues identical; 6 of 6 residues similar) and 1117 to 1121 (4 of 5 residues identical; 5 of 5 residues similar) of the Pv domain of Fun VgI (GenBank accession number T43141). The N-terminal amino acid sequences of Fe-21 and Fe-19 from 28 Yp (24 residues, VSKAAAAECRFIEDTLYTFNNKSY) were identical to one another and highly similar to that of the ß'c domain of Bar VgA (identity, 62.5%; similarity, 75.0%) [22] and to the homologous sequence (residues 1227 to 1250) of Fun VgI (identity, 87.5%; similarity, 95.8%).
No sequence could be obtained for the major band generated by 400 Yp, Fe-112, possibly because its N-terminus was blocked. However a sequence of 24 residues was obtained for the N-termini of Fe-33 and Fe-26, and these two sequences were identical (PEXSTTEAIFNVKAFAIXENQKPE). To obtain additional internal sequence data for 400 Yp, a purified preparation of the protein was digested with trypsin at pH 7.0 and the resulting digests were subjected to SDS-PAGE and blotted onto PVDF membranes for N-terminal sequencing. In addition to Fe-112, the digested 400 Yp gave rise to two bands having apparent molecular masses of 80 kDa (tFe-80) and 30 kDa (tFe-30) (Fig. 5). N-terminal sequencing of tFe-80 revealed 21 residues (TYGPLEKKGKIIYSFEDVDIN), but no sequence data could be obtained for tFe-30 (Table 1). Identities and similarities of the sequences of the tFe-80 and Fe-33 polypeptides derived from Fe-112 with zebrafish (Zeb) vg3 (GenBank accession number AAG30407 were only 14.3% and 38.1% (position 275 to 295) and 25.0% and 54.2% (position 1053 to 1070), respectively. Slightly higher corresponding identities to goby (Gob) Vg-320 (GenBank accession number AB088474) of 23.8% and 38.1% (position 276 to 296) and 25.0% and 45.6% (position 1064 to 1079), respectively, were observed.
When protein-bound phosphorus in the purified 600 Vg and 400 Yp preparations was measured, these preparations had phosphorus contents of approximately 0.44% and 0.05% (wt/wt), respectively (Table 2). The percentage of total lipid by weight and the lipid class composition (percent phospholipid, triglyceride, and cholesterol) of the 600 Vg and 400 Yp also were assessed. The total lipid content by weight of 600 Vg and 400 Yp was 16.2% and 13.7%, respectively, indicating that both were lipoproteins. The lipid class analysis revealed a predominance of phospholipid, which accounted for 84% and 75% of total lipids in 600 Vg and 400 Yp, respectively. As to the remaining lipid classes, 600 Vg had a slightly lower content of triglycerides (10% versus 14%) and higher content of cholesterol (16% versus 11%) than the 400 Yp.
|
Cloning and Sequence Analyses of Three Formsof Vg cDNA
Full-length cDNA sequences encoding the three forms of mosquitofish Vg, the two 600 Vgs tentatively designated as 600 VgA and 600 VgB and the 400 Vg (also 400 Yp), were obtained from a liver cDNA library of E2-treated female mosquitofish as described in Materials and Methods (see Isolation of cDNA Clones) and illustrated in Figure 1. The longest 600 VgA cDNA clone (approximately 5.1 kbp), which included a start codon, stop codon, and polyadenylation signal (AATAAA), was sequenced completely. The insert in this clone contained an open reading frame of 5085 bp encoding 1695 amino acid residues including 15 residues of putative signal peptide (Fig. 6A, GenBank accession number AB181835). Figure 6A shows an alignment of the deduced amino acid sequence of the 600 VgA with that of Fun VgI. These two Vgs showed maximum identity (76%) when the 600 VgA sequence was compared with Vg sequences available for several other teleost species. The molecular mass of the deduced amino acid sequence of the 600 VgA without signal peptide was 187 kDa, similar to that of the 600 VgA protein as assessed by SDS-PAGE (195 kDa). The full-length nucleotide sequence encoding the 600 VgB was obtained by aligning sequences of a partial 600 VgB cDNA clone and a 5'-RACE fragment (see Materials and Methods). The composite sequence contained an open reading frame of 5025 bp encoding 1675 amino acid residues, including 15 residues of signal peptide (Fig. 6B; GenBank accession number AB181836). An alignment of the deduced amino acid sequence of the 600 VgB with that of Fun VgII is shown in Figure 6B. These two Vgs showed maximum homology (78%) when the 600 VgB sequence was compared with that of Vg from other teleosts. The molecular mass of the deduced amino acid sequence of the 600 VgB without signal peptide was 185 kDa, only slightly less than the 195 kDa mass of the 600 Vg polypeptide estimated by SDS-PAGE. The sequence of the full-length 400 Vg cDNA clone contained a 3726 bp open reading frame encoding 1242 amino acid residues including 15 residues of putative signal peptide (Fig. 6C; GenBank accession number AB181837). Figure 6C shows an alignment of the deduced 400 Vg amino acid sequence with that of the phosvitinless zebrafish Vg, Zeb vg3, with which it had high homology compared with other teleost Vg sequences. The molecular mass of the deduced amino acid sequence of the 400 Vg without signal peptide was calculated to be 139 kDa, a value similar to that of the 400 Vg polypeptide determined by SDS-PAGE (142 kDa).
|
|
|
Vg Peptide Sequence Homologies
When the deduced amino acid sequences of the three mosquitofish Vgs were aligned and compared (Table 3), identity and similarity between 600 VgA and 600 VgB were 46% and 67%, respectively. However, much lower identities and similarities were observed between 400 Vg and 600 VgA (identity, 17%; similarity, 36%) or 400 Vg and 600 VgB (identity, 21%; similarity, 37%). Results of similar comparisons of mosquitofish Vg sequences with those available for other teleosts are shown in Table 3. The 600 VgA exhibited considerably higher identities to Fun VgI (76%), Bar VgA (57%) and Had VgA (54%, GenBank accession number AAK15158 than to the VgII or VgBs of these species. Conversely, the 600 VgB showed higher identities to Fun VgII (78%), Bar VgB (62%), and Had VgB (59%, GenBank accession number AAK15157. The deduced amino acid sequence of the 400 Vg showed 44% identity to that of Zeb vg3 and 47% identity to Gob Vg-320. Alignments of the 400 Vg sequence to those available for other teleost Vgs revealed less than 26% identity in all cases.
|
Putative Vg Receptor-Binding Sites
Alignment of the deduced amino acid sequence of the putative receptor-binding site of tilapia Vg1 [48] (GenBank accession number AAD01615 with that of the three mosquitofish Vgs and zebrafish vg3 [28] (GenBank accession number AAG30407 revealed that this domain of mosquitofish Vgs is rich in basic amino acid residues, which are believed to mediate receptor binding (Fig. 7; see Discussion). However, the mosquitofish VgC exhibited a unique substitution of glutamine for lysine at position 185 of the aligned tilapia sequence, which is a residue believed to be of critical importance to ligand recognition by the Vg receptor [48].
|
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Several criteria for identification of teleost Vgs have been proposed based on the following protein characteristics: female specificity; inductivity by estrogen; high molecular mass; calcium and iron binding ability; carbohydrate, lipid, and phosphorous content; and immunological identity with the major yolk proteins [49, 50]. In mosquitofish, a single E2 injection induced two protein peaks of 600 and 400 kDa that were apparent when the blood plasma was subjected to gel filtration. Previously reported native molecular masses of teleost Vgs range from 300 to more than 600 kDa [1, 51]. The molecular masses of estrogen-induced mosquitofish proteins purified in this study fall within this range. Considering the inequality in relative contents of 600 and 400 kDa proteins in the gel filtration chromatogram of plasma from E2-treated fish, the mosquitofish seems to produce 600 kDa protein as the major form or forms of Vg (600 Vg). A peak of 600 kDa protein also was evident after gel filtration of plasma from naturally vitellogenic females, but not after similar chromatography of male plasma. Heterogeneity of the 600 Vg, in which two distinct antigens were recognized, was clearly verified by our immunological analyses. Duality in antigenicity of the 560 kDa major yolk protein (560 Yp) present in Fe also was confirmed, because two precipitin lines were formed by Fe when it was reacted with ab-a-600 Vg. Collectively, these findings indicate that two major yolk proteins in 560 Yp are derived from the two different Vgs present in the 600 Vg preparation. In addition, results from analysis of the lipid content of the 600 Vg (16.2% by weight) and the phosphorus content of the protein (0.44% by weight) were consistent with values previously reported for complete Vgs from other teleosts [52, 53]. Thus, on the basis of their molecular mass, female specificity, inducibility by estrogen, antigenic relationship to the major yolk proteins, and identity as lipophosphoproteins, the two proteins included within the 600 Vg preparation satisfied several criteria for classification as mosquitofish Vgs.
In mosquitofish, a small Vg-like protein having an apparent molecular mass of 400 kDa in native form was observed not only in plasma from estrogen-treated fish (400 Vg) but also in extracts of vitellogenic follicles (400 Yp). The native molecular mass of this small putative Vg appeared to remain unchanged after its incorporation into oocytes, which also was reported for an incomplete form of Vg (Vg-320) in the Japanese common goby [29]. The small mosquitofish Vg was purified from Fe instead of blood plasma from vitellogenic fish because the 400 Yp was abundant in the follicle extracts and could easily be separated from the other yolk proteins. The purified 400 Yp was a lipophosphoprotein with a slightly lower lipid content (13.7% by weight) and substantially lower phosphorus content (0.05% by weight) than the 600 Vgs. Low phosphorus content, indicative of a reduced Pv domain, also is a characteristic of a small Vg variant in the tilapia [25] and the Japanese medaka [27]. The reactivity of the antiserum against purified 400 Yp to the 400 kDa protein in EP (400 Vg) but not to purified 600 Vg indicates that the 400 Vg gives rise to the 400 Yp and is not a degradation product of 600 Vgs. These immunobiochemical observations confirm that three forms of Vg are present in the blood plasma of E2-treated mosquitofish.
Further verification of precursor-product relationships between the mosquitofish Vgs and yolk proteins was based on our analyses of the N-terminal amino acid sequences of yolk polypeptides. In gel filtration, homogenates of follicles from vitellogenic mosquitofish gave rise to three obvious protein peaks corresponding to 560 kDa, 400 kDa, and 28 kDa. The major 560 kDa yolk protein peak, which we considered to represent Lv of this species, was resolved into three bands of 126 kDa, 66 kDa, and 55 kDa in SDS-PAGE. In teleosts, as in other vertebrates, Lv is composed of LvH and LvL [2], and the 126 kDa band was presumed to represent mosquitofish LvH. Heterogeneous forms of LvH have been reported to exist in vitellogenic oocytes of barfin flounder, in which two different LvHs (vLvHA, 107 kDa; vLvHB, 94 kDa) are present and can be separated by SDS-PAGE. Each form of barfin flounder LvH is derived from a distinct Vg precursor molecule, either VgA or VgB [22]. However, in haddock [21] and white perch [13], and apparently in mosquitofish, the LvHs derived from different Vgs exhibit the same molecular mass and cannot be separated by SDS-PAGE.
The unusually weak signal obtained from N-terminal amino acid sequencing of the 126 kDa putative LvH band (Fe-126), which was the major yolk polypeptide in mosquitofish, led us to surmise that this band consists of at least two forms of LvH and that the N-terminus of the major form of LvH is blocked. The N-terminus of LvH (or Vg) is blocked in several other teleost species, presumably by pyroglutamyl or acetyl groups [54]. A homology search revealed that the limited N-terminal sequence that we obtained for the Fe-126 band, TQVNYA, was most identical to residues 16 to 21 of the deduced amino acid sequence of Fun VgII and of Bar VgB (4 of 6 residues). These residues are located in the LvH domain of the respective mummichog and barfin flounder Vgs. The N-terminal amino acid sequences of the 66 kDa band (Fe-66) and the 55 kDa band (Fe-55), DKILSS and AARSN, respectively, were most identical to amino acid positions 1072 to 1077, and 1117 to 1121 of Fun VgI. These positions are located just upstream from polyserine regions in the Pv domain of Fun VgI. This localization suggests that Fe-66 and Fe-55 are proteolytic variants of Pv-LvL complexes bound to LvH in the 560 Yp molecule. Such Pv-LvL complexes, which are substituted for LvL in the native Lv yolk protein, also have been detected in barfin flounder LvA [22].
The 28 kDa protein peak isolated from mosquitofish Fe by gel filtration was resolved into a 21 kDa band (Fe-21) and a 19 kDa band (Fe-19) in SDS-PAGE after reduction of the protein with 2-mercaptoethanol. The N-terminal amino acid sequences obtained for these two bands were identical to one another and also were similar to the N-terminus of barfin flounder ß'c [22]. These results suggest that the Fe-19 is a proteolytic variant of the N-terminus of a mosquitofish ß'c (Fe-21). Homology analyses of the Fe-21 sequence aligned to Vg sequences of other fishes showed highest identity when Fe-21 was compared with Fun VgI (75%), Had VgA (54%) and Bar VgA (59%). These observations indicate that Fe-21 represents the N-terminus of a mosquitofish ß'c derived from an A-type (or type I) Vg [23].
Unequal accumulation of the two 600 Vgs in mosquitofish oocytes was suggested by the results of N-terminal and internal amino acid sequencing. With the single exception of the one weak sequencing signal obtained for Fe-126, most sequences of the yolk peptides had higher homology to A-group teleost Vgs than to B-group Vgs. Accordingly, the major and minor forms of Vg in the 600 Vg preparation may belong to VgA (or VgI) group and the VgB (or Vg II) group, respectively, and these were tentatively designated as mosquitofish 600 VgA and 600 VgB.
As noted, the native molecular mass of the mosquitofish 400 Vg did not appear to change after its incorporation into oocytes, as both plasma from vitellogenic fish and vitellogenic Fe contained the same 400 kDa protein, based on tests of its antigenicity. However, purified 400 Vg gave rise to a 142 kDa band in SDS-PAGE under reducing conditions, whereas the 400 Yp yielded bands corresponding to 112 kDa (Fe-112), 33 kDa (Fe-33), and 26 kDa (Fe-26). These observations indicate that the native 400 Vg molecule was nicked by one or more lysosomal enzymes after uptake into growing oocytes, as typically occurs during formation of LvH and LvL [2]. The N-termini of the Fe-142 and Fe-112 bands were blocked, and these peptides could not be sequenced. However, N-terminal amino acid sequences obtained for the Fe-33 and Fe-26 bands were identical, suggesting that these two peptides arise from variant proteolytic processing of the parent 400 Vg molecule. Although the identity of the enzyme responsible for proteolytic processing of Vg into yolk proteins has not been confirmed in mosquitofish, an aspartyl endopeptidase, cathepsin D, is believed to perform this function in most vertebrates [55, 56], including several teleost species examined to date (reviewed in [3, 41]).
We confirmed the presence of three Vgs (600 VgA, 600 VgB, and 400 Vg) in mosquitofish by cloning and sequencing the three Vg cDNAs. The amino acid sequences of 600 VgA and 600 VgB deduced from their cDNAs contained the N-termini of the 600 VgA- and 600 VgB-derived yolk polypeptides, respectively, which verified the existence and identity of each type of 600 Vg. Lengths of both cDNAs encoding mosquitofish 600 Vg (5025–5088 bp) were similar to those reported for other teleost fishes, including the mummichog (VgI, 5112 bp; VgII, 5061 bp), rainbow trout (5100 bp) [57], and haddock (VgA, 5189 bp, GenBank accession number AF284035; VgB, 5103 bp, GenBank accession number AF284034). Moreover, the domain structure of the mosquitofish 600 Vgs (NH2-signal peptide-LvH-Pv-LvL-ß'c-C-terminal-COOH) was identical to that of Vgs in the species listed above. Based on the similarities in Vg size and domain structure, and considering the results of our sequence homology analyses (see Table 3), we categorized the mosquitofish 600 VgA and 600 VgB as complete teleost type A and type B Vgs, respectively. In contrast, the mosquitofish 400 Vg appears to be an incomplete form of Vg. The amino acid sequence of the 400 Vg deduced from cDNA contained N-terminal amino acid sequences of the constituent peptides of 400 Yp, verifying its identity. The size of the open reading frame of the cDNA encoding the 400 Vg (3726 bp) is smaller than that of complete (A and B type) Vgs and is similar to that of cDNAs encoding the zebrafish vg3 (3758 bp, GenBank accession number AF254638) and the Japanese common goby Vg-320 (3993 bp). As is the case with these incomplete zebrafish and goby Vgs, the 400 Vg has no obvious Pv domain and had an otherwise shortened coding region downstream from its LvH domain.
Hiramatsu et al. [23] recently categorized teleost Vgs into three major groups (A, B, and C) based on their primary structures, their known physiological functions, or both. During oocyte maturation, the LvH derived from A type Vgs is usually heavily degraded, generating FAA that drive oocyte hydration [21, 22], whereas the LvH derived from B type Vgs undergoes no or limited hydrolysis during this time. As noted, the dual complete Vgs of mummichog, barfin flounder, and haddock can be classified into the Vg A and B groups as follows: group A, Fun VgI–Bar VgA– Had VgA; group B, Fun VgII–Bar VgB–Had VgB (also see Table 3). Phylogenetic analyses revealed that the Vg encoded by the zebrafish vg3 gene (Zeb Vg3, GenBank accession number AF406784) does not belong to any distinct cluster of vertebrate Vgs and, judging from its lack of a polyserine domain, it appears to be intermediate between invertebrate Vgs and all known vertebrate Vgs, possibly representing a primitive form of Pvl vertebrate Vg [28]. Hiramatsu et al. [23] later proposed that such Pvl vertebrate Vgs be classified as C-type Vgs, which typically are a minor form of Vg lacking a Pv domain, or having a greatly shortened Pv domain, that are most homologous to zebrafish vg3, insect Vgs, or chicken VgIII. Structural and biochemical features of a VgC include a lower content of phosphorus or serine residues, considerably lower molecular weight than VgAs or VgBs, and elution in fractions at lower NaCl concentration than the other types of Vg during anion-exchange chromatography. Our sequence homology analyses indicate that mosquitofish 600 VgA and 600 VgB obviously belong to the VgA and VgB groups, respectively, rather than to the VgC group. In contrast, the Pvl mosquitofish 400 Vg clearly belongs to the VgC group based on the classification scheme discussed above. It should be noted that some complete teleost Vgs that are unrelated to C-type Vgs, such as those from zebrafish (e.g., Vg1, [28], GenBank accession number NM170767), fathead minnow (Pimephales promelas, [58], GenBank accession number AF130354), and rainbow trout (Oncorhynchus mykiss, [57], GenBank accession number S82450), are difficult to assign unambiguously to either the VgA or VgB group.
Tolar et al. [35] recently reported identification and purification of a single Vg with a native molecular mass of 429 kDa from the plasma of mosquitofish treated with the estrogen, 17
-ethynylestradiol. These authors proposed to use this Vg preparation as a biomarker for the detection or evaluation (or both) of environmental estrogens. However, it is difficult to identify the Vg group to which this 429 kDa mosquitofish protein belongs because of the lack of information on its primary sequence, proteolytic processing during oocyte maturation, and product yolk proteins. We concur with the concept that Vg is a useful biomarker for exposure of fish to environmental estrogens [41]. However, the findings of the present study illustrate the need for detailed information on the suite of Vgs present in any teleost species used to assess estrogenic contamination of the environment. If different investigators use different Vg preparations, the results of toxicology studies that depend on a single Vg or an uncharacterized group of Vgs may not be comparable to those of other studies, even of the same species. The apparent complexity of the Vg system in mosquitofish and many other teleosts warrants that researchers verify the number and quantity of distinct Vg proteins in the species under investigation, or at least identify the major form of Vg in that species. Different forms of Vg are known to exhibit distinct sensitivities to induction by estrogen in terms of dose-response kinetics and maximal Vg levels produced. Specifically, it appears that the Pvl (C-type) Vgs may be less sensitive to induction by estrogen than complete (A- and B-type) Vgs [25, 26, 29]. Results of our analyses of the deduced amino acid sequences of the three mosquitofish Vgs as compared with those of tilapia Vg1 [48] (GenBank accession number AAD01615 and zebrafish vg3 suggest that some C-type Vgs also may have altered affinity for the Vg receptor (VgR) resulting from amino acid substitutions at key positions in their receptor binding site.
The site on Vg that binds to the VgR of teleosts is located in the Lv domain of the Vg molecule [5, 13]. The VgR belongs to a superfamily of lipoprotein receptors that include the very low density lipoprotein receptor and low density lipoprotein receptor (LDLR). Functional models of these types of lipoprotein receptors involve binding of lysine and arginine residues in the ligand to clusters of acidic residues in the seven or eight ligand-binding repeat domains of the receptor [59]. A recent study using the yeast two-hybrid system to study interactions of tilapia Vg with its receptor demonstrated that an 85-amino acid fragment (referred to as VtgSE) located in the LvH domain of Vg mediates receptor binding [48]. Alignment of the tilapia VtgSE sequence with that of the LDLR-binding sites on apolipoprotein (apo) B and apo E revealed a conserved short motif (HLTKTKDL) rich in basic amino acid residues. Site-directed mutagenesis of this motif indicated that its first lysine residue (181lysine) likely plays a critical role in receptor binding. When deduced amino acid sequences of Vgs from 27 species were aligned, lysine in this position was completely conserved among 26 species, being substituted for another basic residue, arginine, in the remaining one [48]. The zebrafish VgC (vg3) has this same substitution of arginine for lysine. However, the 181lysine, which is conserved in mosquitofish VgA and VgB, is substituted for glutamine in mosquitofish VgC (see Fig. 7). Our recent molecular cloning of two distinct Vgs in Japanese common goby (A. flavimanus) revealed that this species has the same substitution (glutamine for lysine) as mosquitofish at this residue in its C-type Vg (GenBank accession number BAC06191, but not in its A-type Vg (GenBank accession number BAC06190. The impact on receptor binding affinity of such nonconservative substitutions at a key functional residue in the ligand warrants investigation. Nothing is presently known about these novel C-type Vgs with regard to their receptor binding characteristics.
Combining the results of our analyses of mosquitofish Vg cDNAs with those of our immunological and biochemical analyses of the corresponding Vgs and yolk proteins, we constructed a model describing molecular alterations of the three forms of mosquitofish Vg and their product yolk proteins during vitellogenesis and oocyte growth (Fig. 8). Within the oocyte, the 600 VgA and 600 VgB molecules are cleaved into two yolk proteins with molecular masses of 560 kDa (Lv-Pv) and 28 kDa (ß'c), respectively, with the major source of these yolk proteins being the 600 VgA. Considering the molecular masses of the various proteins, the 560 kDa yolk protein must be a dimeric molecule consisting of two sets of LvH and Pv-LvL complexes having apparent molecular masses of 126 kDa and 66 or 55 kDa, respectively. Small proportions of 195 kDa and 175 kDa polypeptides present in the 560 kDa yolk protein may arise from variation in Vg processing and are considered to consist of intact Vg A or B and LvH-Pv-LvL complexes, respectively. A minor protein present in extracts of vitellogenic follicles, presumably a yolk protein, contained phosphorus, showed no UV absorbance, and eluted at a position of 32 kDa after gel filtration on Superose 6. This protein is probably a Pv produced in minor quantities by complete cleavage of Vg (into Lv, ß'c, and Pv) as generally occurs in salmonids and other teleost species [2]. Although the 400 Vg is accumulated by oocytes without undergoing any apparent alteration to its native molecular mass, its product 400 Yp consists of subunits derived from LvH (112 kDa) and LvL (33 or 26 kDa). Combined with N-terminal sequencing of the yolk (Fe) peptides, our analyses of the homology of the deduced amino acid sequence of mosquitofish 400 Vg with those of mosquitofish VgA and VgB, as well as other teleost Vgs, confirmed that the Fe-112 and Fe-33 (or Fe-26) represent LvH and LvL domains of 400 Vg. In the Japanese common goby, the PvlVg (Vg-320) polypeptide (127 kDa) does not undergo any proteolytic alteration, yielding LvH and LvL subunits during the process of yolk accumulation [31]. The mosquitofish 400 Vg is the first example of a PvlVg that is processed into distinct LvH and LvL subunits during yolk formation.
|
In some teleost species, yolk proteins in oocytes and eggs remain in liquid form throughout vitellogenesis and oocyte maturation, which is in marked contrast to the yolk proteins of higher vertebrates, which form crystalline structures in the yolk granules or globules (see [60]). LaFleur et al. [20] speculated that the occurrence of liquid-phase yolk in the mummichog was due in part to small serine-enriched domains of VgI that are eventually processed into Pv, acquiring an increased hydrophilic potential that somehow prevents formation of insoluble complexes of crystalline yolk. Results from our analyses of the subunit structure of mosquitofish yolk proteins reveal another possible explanation of this phenomenon in mosquitofish. The Pv-LvL complexes associated with LvH likely endow the native, major 560 kDa yolk protein, with high solubility and inhibit its crystallization.
Accumulation of recent evidence for the presence of ß'c in various fish species, including the mosquitofish, suggests that the ß'c has a wide phylogenetic occurrence among teleosts [9–11, 13, 61]. In the chicken, a novel glycoprotein (YGP40) has been isolated from the yolk plasma and identified as the product of the C-terminal, cysteine-rich domain of chicken VgII [62]. Discovery of YGP40 in the chicken and ß'c in teleosts led to the expectation that products of the C-terminal domain of Vg, downstream from its LvL domain, are present as yolk proteins in other oviparous vertebrates as well. The predicted molecular mass of the C-terminal region of mosquitofish VgA, extending from the N-terminus of the ß'c domain to C-terminus of the protein, is 30 kDa, which is considerably larger than the mosquitofish ß'c monomers (21 or 19 kDa). This observation suggests that a 10 kDa yolk polypeptide derived from C-terminal end of mosquitofish 600 VgA still remains to be discovered. Very recently, such a fourth yolk protein product of Vg derived from its C-terminal coding region (Ct) was immunologically detected in postvitellogenic oocytes of barfin flounder using an antiserum raised against the recombinant Ct polypeptide [30].
Heretofore, in addition to the mosquitofish (Cyprinodontiformes, present study), three forms of Vg protein belonging to the VgA, VgB, and VgC groups have been characterized only in white perch (Perciformes) [23]. In the medaka (Beloniformes), two complete Vg cDNAs have been cloned and their deduced protein products, named Vg1 and VgII, appear to belong to VgA and VgB group, respectively [23]. An additional PvlVg-like protein [27] was later discovered in the ascites fluid of estrogen-treated medaka, extending the three Vg model to include this species as well. Thus, three different functional Vg genes are present at least one member of the teleost orders, Perciformes, Cyprinodontiformes, and Beloniformes. Reith et al. [21] suggested that the taxonomic distribution of VgA and VgB among haddock, mummichog, and barfin flounder results from the divergence of a Vg gene present in a common ancestor of the Paracanthopterygii and Acanthopterygii. All characterized Vgs or vitellins from invertebrates, such as insects and nematodes, lack a polyserine domain and, until the recent discovery of teleost VgCs, all characterized vertebrate Vgs contained the Pv domain [63]. Wang et al. [28] speculated that a primitive Vg in the ancestor of both invertebrates and vertebrates was likely a PvlVg and they suggested that the PvlVg in zebrafish was a conserved ancestral type of Vg. In the lamprey, a primitive jawless vertebrate, a Pv domain is present in Vg [64], suggesting that the acquisition of a Pv domain must have occurred quite early during vertebrate evolution, and possibly even before the appearance of vertebrates, as proposed by Wang et al. [28]. Discovery of PvlVg in several taxonomically diverse groups of teleosts supports the speculation that a functional PvlVg gene also is present in many other vertebrate species, where it remains to be identified. Further characterization of the Vg gene family in an array of fishes representing a full spectrum of different taxonomic groups and diverse reproductive life histories will provide new insights to the molecular evolution and function of the various typesof Vg.
Regarding potential functions of a multiple Vg system, we previously verified the disparate involvement of barfin flounder VgA and VgB, and their yolk protein products, in regulation of oocyte hydration during cytoplasmic maturation in addition to delivery of specific types of nutrients to developing embryos and larvae [22]. Differential processing of yolk proteins derived from two distinct Vgs during oocyte maturation has been confirmed in other teleosts inhabiting marine or brackish waters, such as the haddock [21] and the white perch [13]. Yolk degradation associated with massive oocyte hydration during maturation has been described in a number of teleost species (see for example [65–67]), although a dual Vg system has been verified to exist in relatively few of these. As noted, some teleosts that are viviparous or spawn demersal (or adhesive) eggs, such as mosquitofish or medaka, possess multiple Vgs, yet their oocytes exhibit no remarkable hydration as they mature. Additional information on the phylogenetic distribution of dual or multiple Vg systems and their association with different kinds of reproductive strategies is clearly needed. For those species in which multiple Vgs have been detected,
