Fish Physiology and Biochemistry vol. 8 no. 2 pp 111-120 (1990) Kugler Publications, Amsterdam/Berkeley

The purification and partial characterization of carp, Cyprinus carpio, vitellogenin Charles R. Tyler and John P. Sumpter Department of Biology and Biochemistry, Brunel University, Uxbridge, Middlesex, UB8 3PH, U.K.

Keywords: carp, vitellogenin, purification

Abstract A procedure is described for the isolation of intact vitellogenin (c.-VTG) from the carp, Cyprinus carpio. VTG was induced in juvenile females using oestradiol-170 and purified from the plasma using a combination of gel-filtration chromatography on Sepharose 6B and ion exchange chromatography on DEAE-cellulose. Purification procedures were conducted at low temperatures (below 9°C) in the presence of the proteolytic enzyme inhibitor aprotinin to prevent degradation. Intact c-VTG had an apparent molecular mass of 390,000 Daltons, but when extracted from plasma in the absence of aprotinin it underwent proteolysis into at least 2 protein fragments (apparent molecular masses of 230,000 and 96,000 Daltons), showing an instability of the native dimer. An amino acid analysis of c-VTG showed that its composition was almost identical to goldfish VTG, a species closely allied to the true carps and also similar to other oviparous vertebrate VTGs. Collectively, these data indicate that using these purification procedures VTG from carp, and probably other teleost species, can be isolated in an intact, highly purified form.

Introduction

mature females and males (Beams and Kessel 1973; Wallace 1978; de Vlaming et al. 1980; van Bohemen

Vitellogenin(s) (VTGs) are among the major yolk protein precursors in both oviparous vertebrate and invertebrate groups. In vertebrates they are large, female specific, glycolipophosphoproteins produced by the liver in response to oestrogens. After their secretion into the blood in maturing females, VTGs are transported to the ovary where they are selectively sequestered by the developing oocytes and subsequently proteolytically processed into the lipovitellin(s) and phosvitin(s), the predominant yolk proteins (Wallace 1978). Synthesis of VTGs may also be induced by oestrogen treatment of im-

et al. 1982).

Vitellogenin(s) have been demonstrated in a wide variety of fish (Le Menn 1978; Tata and Smith 1979; Nath and Sundararaj 1981; Campbell and Idler 1980; Wallace and Selman 1981) and in many cases their hepatic origin confirmed (Plack and Frazer 1971; Campbell and Idler 1976; De Vlaming et al. 1977; Le Menn and Lamy 1977; Idler and Campbell 1980). However, with a few exceptions (Fundulus heteroclitis, Wallace and Selman (1982); the medaka, Oryzias latipes, Hamazaki et al. (1987) and the spotted seatrout, Cynoxion nebulosus,

Correspondence to: Dr. C.R. Tyler, Dept. of Biology and Biochemistry, Brunel University, Uxbridge, Middx. UB8 3PH. Tel: 08985 74000 ext.2089.

112 Copeland and Thomas (1988) relatively few attempts have been made to isolate and characterise fish VTG, due at least in part to the apparent difficulties encountered in obtaining pure and intact forms. A number of techniques have been developed for the isolation of vertebrate VTG, but many produce a final preparation that is either partially degraded, contaminated with other plasma proteins (ultracentrifugation: Redshaw and Follett (1971)), or contains toxic chemicals such as dimethylformamide (Ansari et al. 1971). The latter clearly makes the VTG isolate unsuitable for further physiological studies either in vivo or in vitro. Successful methods developed for VTG isolation from other vertebrate groups often rely on the high surface charge of VTG and include EDTA-Mg + + (Ca + +) precipitation and/or ion exchange chromatography. Teleost VTG(s), however, are generally less phosphorylated than those of other vertebrate VTG(s) and therefore these methods either fail to precipitate VTG or, with few exceptions (Wallace and Selman 1982) are unable to separate it adequately from other plasma proteins. Furthermore, teleost VTG(s) appear susceptible to proteolysis (Wallace and Selman 1982) and the use of divalent cations may activate proteolytic enzymes (Norberg and Haux 1985), thereby increasing degradation. Even in the few studies where teleost VTG has been handled in the presence of proteolytic enzyme inhibitors, generally degradation of the molecule has not been eliminated. The greater susceptibility of teleost VTG to proteolysis compared to the VTGs of other oviparous species has been attributed to their differences in chemical composition and structure (Norberg and Haux, 1985). The family Cyprinidae contains the majority of Europe's freshwater fish. In this family the 'carp' species command the greatest commercial value; they are important both for sport-angling and, on a worldwide basis, they are singularly the most extensively farmed fish, their production rates far exceeding that of any other family of fish. However, quite surprisingly considering their commercial importance, little is known about the reproductive biology of carp, compared to other farmed fish, such as the salmonids. Very few publications deal

with aspects relating to vitellogenesis and ovary growth. To enable studies of this nature to be conducted it is necessary to identify and isolate intact carp vitellogenin (c-VTG). With this in mind the objective of the present study was to develop a simple, reproducible method for obtaining a preparation of intact VTG, of high purity, from the carp, Cyprinus carpio. Both physical and chemical parameters were subsequently used to characterise the purified c-VTG isolate. The purified c-VTG obtained was used to develop a radioimmunoassay (Tyler and Sumpter 1989), which in turn will allow us to investigate the regulation of ovarian growth. Materials and methods Fish, stimulation of VTG productionand collection of blood One to 2-year old mirror carp, C. carpio and rainbow trout, Salmo gairdneri, were purchased from commercial fish farms and maintained in aquaria provided with running tap water at 15 + 5C. The fish were held on a 12-hour light/12-hour dark photoperiod and fed on commercial pelleted food (Mainstream, B.P. Nutrition). All hormone and isotope injections and blood sampling were carried out under anaesthesia using 2-phenoxyethanol (1: 2000; Sigma Chemical Company, Poole, Dorset). Induction of VTG synthesis was accomplished using oestradiol-17 at a concentration of 10 /g.g-' body weight. The hormone was dissolved in 100 l of 70% ethanol, mixed with an equal volume of 0.9% NaCI and injected into the peritoneal cavity. Twenty Curie of 3 H-leucine (37 Bq.ml-l; Amersham International plc, Amersham, Bucks), used as a marker for VTG, was diluted in 0.9% NaC to give an injection volume of 200 1al, and also injected into the peritoneal cavity. Blood samples were collected from the caudal sinus into chilled, heparinized syringes either in the presence (20 T.I.U. ml-' blood) or absence of the proteolytic enzyme inhibitor, aprotinin (Sigma Chemical Company). The blood was centrifuged immediately at 1500 x g and 4°C for 15 minutes and the resulting plasma withdrawn and either used immediately or deep frozen at -20 0 C.

113 Chromatographicseparations All chromatography and other separative procedures were carried out below 9°C in an attempt to minimise proteolysis.

Gel filtration Gel filtrations were performed on a 100 x 1.5 cm column of Sepharose 6B (Pharmacia, Milton Keynes, Bucks, U.K.) equilibrated with 0.1 M sodium phosphate buffer, pH 7.8. During the purification of c-VTG for chemical and physical analyses aprotinin was added to the column buffer at 100 T.I.U. -1. The columns were run at 10 ml.h- ' and the fractions eluted (2 ml) analysed for protein (absorbance at 276 nm) and radiolabel in a Packard B460 scintillation counter. The column was calibrated for molecular weight determinations using a high molecular weight calibration kit (Pharmacia) containing adolase (MW 158,000), catalase (MW 232,000), ferritin (MW 440,000) and thyroglobin (MW 669,000). Ion exchange DEAE-cellulose (Pharmacia) was packed into a 8 x 2.5 cm column (40 ml bed volume) and equilibrated with 5 volumes of starting buffer (10 mM K2 HPO4 , 2.5 mM KH 2PO4 , pH 7.6). The partially purified carp VTG from the gel-filtration on Sepharose 6B was equilibrated with the starting buffer (as described below) and applied to the column. Unbound proteins were eluted with one bed volume of starting buffer. To separate proteins bound to the gel a non-linear potassium phosphate gradient was used (10 mM K2HPO4, 2.5 mM KH 2PO 4 to 300 mM K2HP0 4, 75 mM KH 2PO 4) at a flow rate of 40 ml.h - l , and 8 ml fractions were collected. The eluent was monitored for protein and radiolabel, and those fractions containing c-VTG retained.

fractions containing c-VTG had to be equilibrated to a new buffer before their application to the next column. Furthermore, the c-VTG often eluted in a relatively large volume which had to be reduced considerably before the next purification step. Equilibration to column buffers and concentration of c-VTG were performed using an Amicon ultrafiltration apparatus fitted with an XM50 (50,000 Daltons molecular weight pore size) membrane.

Isolation of c- VTG The steps involved in the isolation of intact c-VTG are shown in Fig. 1.

Amino acid analyses Purified c-VTG and rainbow trout VTG (rt-VTG) were subjected to amino acid analyses for further characterisation. The analyses were performed on a custom-built, automatic amino acid analyser. The system was controlled by a Nascom III computer. Briefly, the proteins were initially hydrolysed in constant boiling HCI and 100°C for 24h, degassed to remove oxygen, and then treated with 10% sulphosalicyclic acid to precipitate protein that would otherwise irreversible bind to the ion exchange resin. The amino acid sample was then injected onto a column of strong action exchange resin and amino acids eluted by a stepwise gradient of increasing pH and ionic strength. The protein hydrolysates were then eluted by a series of lithium citrate buffers. The detection methodology adopted was colorimetric detection by ninhydrin (detailed methodology in Macromolecular Analysis, University of Birmingham, Birmingham, England). Tryptophan, which is often totally destroyed by HCI hydrolysis, was not measured.

Ultrafiltration

Results

The buffers used in the gel-filtration and ion-exchange chromatography differed and therefore

Figure 2 depicts the separation on Sepharose 6B of plasma proteins from a juvenile carp 2 days after an

114 Procedure for the isolation of carp vitellogenin. Time (days) - 1. 0 Intraperitoneal injection of oestradiol-1 7(1 0pg.g body wt.) 10

Boost with oestradiol-1 7(1 Og. g-1 .body wt.)

17

I

Blood collected into chilled syringes containing 20 T.I.U. ' aprotinin.ml . Blood centrifuged and plasma withdrawn.

18

Plasma run on Sepharose-6B with aprotinin under refrigerated conditions and eluents analysed for protein and radiolabel marker (3H-leucine). Vitellogenic fractions pooled. I

19

Partially purified c-VTG pool equilibrated with ion-exchange buffer and concentrated by ultrafiltration, then applied to DEAE-cellulose column. Proteins eluted with potassium phosphate gradient. 1

20

Carp vitellogenin pool from DEAE -cellulose column equilibrated and concentrated by ultrafiltration, then gel-filtered on Sepharose-6B. PURIFIED CARP VITELLOGENIN.

Fig. 1. Flow diagram showing the procedure for the isolation of intact carp vitellogenin.

intraperitoneal injection of 3 H-leucine (but not oestradiol). The radiolabel trace clearly shows that little, if any, of the 3H-leucine had become incorporated into the plasma proteins, with over 95 % of the isotope in the plasma eluting at te VT. When a fish was also injected with oestradiol-173 there was

a dramatic change in the elution pattern of the label, over 95% now eluted with the major plasma proteins in a discrete peak centred around fractions 65-70 (Fig. 3a). This peak of radiolabel indicated the synthesis of a new blood protein induced by oestrogen which, together with its apparent molecular mass of approximately 390,000 Daltons, tentatively identified it as carp vitellogenin. In these initial observations the c-VTG appeared to elute as a single, discrete peak. As mentioned in the introduction, however, there are a number of reports in the literature emphasizing the instability of teleost VTG and the difficulties involved in obtaining intact forms. An investigation was therefore made into the stability of VTG isolated from both carp and rainbow trout plasma (using 3 H-leucine as a tracer; leucine is a major amino acid of both carp and rainbow trout VTG, see Table 1) in the presence and absence of the proteolytic enzyme inhibitor aprotinin. When aprotinin was included it was present both in the syringe into which the blood was collected and in the chromatography buffer. When vitellogenic carp plasma was collected in the absence of aprotinin the 3 H-leucine eluted as two peaks of lower molecular mass rather than a single peak (compare Figs. 3a and 3b). In contrast to c-VTG, when rainbow trout vitellogenic plasma was subjected to gel filtration in the

C

_ E

C

0

I.,-

. r_

In q

0

E

C C

oC

0 VI .0

a .)

co

20

40

60

80

100

Fraction number (2ml) 3 Fig. 2. Gel-filtration on Sepharose 6B of 1 ml plasma treated with aprotinin from a juvenile carp previously injected with H-leucine. 3 ) and for H-leucine (- * -) Fractions were monitored for absorbance at 276 nm (uninterrupted line:

115 (a)

C

E E C 0. C 0

Cu R Cu

0 2 o un M

co

0

20

40

60

80

100

120

(b) C 0

E C CD Cl-

E

c G) O C

0. 0'

E

8

.0 0 In .0

co

0

20

40

60

80

100

120

Fraction number (2ml) Fig. 3. Gel-filtration on Sepharose 6B of I ml plasma samples containing labelled VTG from carp. Blood samples were collected and columns run in buffers with (a) or without (b) aprotinin. Fractions were monitored for absorbance at 276 nm (uninterrupted line: ) and for 3 H-leucine (-* -).

presence (Fig. 4a) or absence (Fig. 4b) of aprotinin no proteolysis was observed; in both cases rt 3 HVTG eluted as a discrete peak centred around fraction 60, giving it a molecular weight of approximately 440,000 Daltons. These data indicated that proteolysis inhibitors had to be employed to avoid protein degradation during the isolation and purification of c-VTG. During the major purification of c-VTG, gel filtration of vitellogenic carp plasma in the presence of aprotinin produced protein and radiolabel profiles almost identical to those shown in Fig. 3a, and therefore these profiles are not shown. From this gel-filtration run fractions 63-70 inclusive were pooled, equilibrated with ion-exchange starting

buffer, concentrated to a final volume of 20 ml, and loaded onto the DEAE-cellulose column. Fractionation of the partially purified c-VTG preparation on DEAE-cellulose with a gradient of potassium phosphate of increasing ionic strength resulted in the elution profile shown in Fig. 5. Approximately 50% of the protein eluted from the column in the first 10 fractions, showing little, if any, binding to the DEAE cellulose. This protein contained less than 1007o of the 3 H-leucine. A second, much smaller peak of protein eluted early in the salt gradient, between fractions 12-15. It is not possible to positively identify this protein, but by analogy with a similar study on Xenopus plasma proteins it may represent albumin (Wiley et al. 1979). The remaining protein eluted from the DEAE-cellulose column was associated with over 9007o of the 3 H-leucine and resolved as a single peak centred around fraction 20. This peak, which eluted between 170 and 240 mM, making it the most highly charged protein, was presumed to be c-VTG. Fractions 16-24 inclusive from the DEAE-column, containing the cVTG, were pooled, equilibrated with gel-filtration column buffer and concentrated to a final volume of 4 ml. The chromatogram of a 2 ml aliquot of this purified c-VTG run on Sepharose 6B is shown in Fig. 6. Both the protein and radiolabel eluted as sharp, single, coincident peaks centred around fraction 67, identical in position to where it eluted from the initial Sepharose 6B column. This peak represented purified, intact c-VTG. Table 1 provides the results of the amino acid analyses of c-VTG and rt-VTG purified during this study. In addition the amino acid contents of goldfish (Carassiusauratus) and South African clawed toad (Xenopus laevis) VTGs (de Vlaming et al. 1980) are provided. This study is primarily concerned with the c-VTG; comparisons with the other two fish species and the toad VTG are discussed later. The c-VTG shows a predominance of non-polar, aliphatic amino acids, constituted primarily by alanine and leucine. The carboxylic amino acids are well represented by aspartic acid and glutamine acid. The hydroxylic amino acids are represented largely by threonine and especially serine. The lowest proportion of amino acids found in c-VTG was the sulphurcontaining group, notably cysteine and methionine.

116 Table 1. The amino acid compositions of various vitellogenins.

(a) c

Percentage of total amino acids

Amino acid Carp

Goldfish

Rainbow trout

Xenopus

9.2 6.2 7.3 4.6 11.1 5.1 11.5 1.0 6.6 2.0 5.0 9.5 2.9 4.0 7.1 2.9 4.5

8.7 5.2 10.1 4.9 13.5 4.9 8.0 6.0 2.5 4.8 8.3 3.0 3.6 7.8 3.5 5.1

E

E o

N

aspartic acid threonine serine proline glutamic acid glycine alanine cysteine valine methionine isoleucine leucine tyrosine phenylalanine lysine histidine arginine

6.7 5.4 7.6 5.9 11.8 5.1 12.6 0.12 6.3 1.9 5.4 10.5 2.8 2.8 6.3 3.4 5.0

6.5 5.5 6.9 5.5 11.9 4.6 12.8 6.9 2.0 6.6 10.8 2.6 2.9 7.0 2.3 4.9

The results for carp and rainbow trout are taken from this study, and the goldfish and Xenopus from de Vlaming et al. (1980).

Discussion A number of procedures have been developed for the isolation of intact VTG from amphibian and avian species (Ansari et al. 1979; Redshaw and Follett 1971; Wallace 1965; Wiley et al. 1979). However, these methods have proved less successful in their application to the isolation of teleost VTG(s). In many fish species VTG appears to be less easily resolved from other blood proteins and more heterogenous in nature than in higher oviparous vertebrates. VTG is a molecule of high complexity and its intricate structure, together with the fact that structurally distinct VTGs exist, at least in amphibians (reviewed in Wahli et al. 1981) and birds (Wang et al. 1983) have made it difficult to thoroughly characterize even in those animals that have received extensive study. Clearly, one of the prerequisite for both physiochemical characterisation and physiological studies using purified teleost VTG (for example in the development and application of a specific VTG radioimmunoassay to study aspects of ovarian development) is the isolation of intact VTG.

E

0. .Z

Cu

(b) c Cu o

E

E In

C o

qE 0 ) C o

C o

.0

C ., 3

0

20

40

60

80

100

120

Fraction number (2ml) Fig. 4. Gel-filtration on Sepharose 6B of 1 ml plasma samples containing labelled VTG from rainbow trout. Blood samples were collected and columns run in buffers with (a) or without (b) aprotinin. Fractions were monitored for absorbance at 276 nm and for 3 H-leucine (-*-). (uninterrupted line: -)

This paper describes a successful technique for the isolation and purification of c-VTG. Several criteria provide evidence that the plasma protein, labelled with 3H leucine in this study, is cVTG. Primarily, the protein is inducible by oestrogen treatment (compare Figs. 2 and 3a). All groups of oviparous vertebrates, from hagfish (Yu et al. 1981), sharks (Craik 1978) and other fish (Wiegand 1982) through to turtles (Ho et al. 1981) and birds (Gibbins and Robinson 1982) have been shown to synthesise and secrete VTG in response to oestrogens. Secondly, both its chromatographic and amino acid properties, which are described more fully below, resemble those of VTGs purified from other

117 0

c

C

l_

E

E

CD 1' N

o

0.

cJ , a

0 o .0 o

C 0

(n

.0

4

o

0

20

10

30

40

50so

60

Fraction number (8ml)

.2o

Fig. 5. DEAE-Cellulose chromatography of a partially purified c-VTG preparation (pooled fractions 63-70 from the Sepharose 6B column). The c-VTG preparation was equilibrated to the DEAE column starting buffer (10 mM K2 HPO4 , 2.5 mM KH 2PO4) by ultra filtration and fractionated with a gradient of increasing ionic strength up to 300 mM K2HPO 4, 75 mM KH 2PO 4. Eight ml fractions were collected and monitored for absorbance at 276 nm (uninterrupted line: ), 3H-leucine (- * -) and salt concentration (- o -). Intact carp vitellogenin

10 10

0

7

E u 0

C c tO

e-

.c C

0

10

o rC to

00

Total volume

.0

00

0 CA m en

'0

.i

0

20

40

60

80

100

Fraction number (2ml) Fig. 6. Gel-filtration on Sepharose 6B of purified c-VTG. Fractions (16-24) containing c-VTG obtained from the DEAE-cellulose column were concentrated to 4 ml and half of this loaded onto the Sepharose 6B column. Fractions were monitored for absorbance at 276 nm (uninterrupted line: -- ) and for 3 H-leucine (- * -).

oviparous vertebrates. Thirdly, the purified protein is female specific; using a radioimmunoassay (RIA) developed against this purified preparation it can be detected in the plasma of juvenile and maturing female carp but is completely absent in males (Tyler and Sumpter 1989). Blood concentrations of this

protein in female mirror carp increase from figs. ml- l in juveniles to 1 mg.ml- l in maturing fish, paralleling the increase in the gonadosonatic index, and thus providing evidence for its role in ovarian growth. The Sepharose 6B column chromatograms of

118 plasma proteins from oestrogen-treated rainbow trout show that when handled at low temperature, even in the absence of proteolytic enzyme inhibitors, rt-VTG is less susceptible to degradation than has previously been reported for other teleost species, including the sea trout, S. trutta(Norberg and Haux 1985), the catfish, Ictalurus nebulosa (Hickey and Wallace 1974) and the goldfish, C. auratus (De Vlaming et al. 1980). This finding corroborates previous studies on the rainbow trout in which VTG preparations were reported to be intact (Hara and Hirai 1978; Campbell and Idler 1980). Similarly, in the presence of aprotinin c-VTG eluted as an intact molecule with an apparent molecular mass of approximately 390,000 Daltons, and there were no signs of proteolytic degradation. In contrast to rt-VTG, however, c-VTG showed clear signs of degradation in the absence of aprotinin, fractionating on Sepharose 6B into at least 2 smaller components. In the goldfish, a cyprinid species closely related to the mirror carp and with a similar reproductive cycle, VTG was also demonstrated to be structurally unstable (De Vlaming et al. 1980). However, in the studies of de Vlaming and colleagues it was reported that even in the presence of proteolytic enzyme inhibitors VTG degradation could not be completely eliminated, an observation at variance with our experiences on the mirror carp reported here. Although clearly possible, it seems unlikely that the stabilities of VTG differ between these very closely related species, and it seems more probable that this difference was due at least in part to the use of different enzyme inhibitors in the two studies. The studies on goldfish VTG employed phenylmethylsulphonyl fluoride (PMSF), whereas we used aprotinin. We chose not to use PMSF because it has only very limited solubility in aqueous solutions, is highly toxic, and in our experience it can cause severe haemolysis of fish blood, an event which would be expected to release lytic enzymes into the plasma. Furthermore, whereas aprotinin has been shown to be effective at preventing VTG proteolysis in the marine teleost F. heteroclitus, PMSF was ineffective (Wallace and Selman 1982). In this study, even after freezing the VTG present in plasma samples of carp collected with aprotinin, it remained intact throughout the subsequent purifi-

cation procedures. Collectively, these observations suggest that PMSF is less adequate in preventing VTG proteolysis of teleost VTG than is aprotinin. However, more detailed quantitative studies are needed before it is possible to describe a standard protocol suitable for the purification of intact VTG from any species of fish. Our molecular mass estimate of c-VTG (390,000 Daltons) is similar to that of rainbow trout VTG (440,000 Daltons) and to the VTGs of other oviparous species (reviewed in Ng and Idler 1983) and almost identical to that reported for goldfish VTG (380,000 Daltons; De Vlaming et al. 1980). Of all the many different characteristics attributed to VTG in different oviparous vertebrates a large molecular weight is one of the commonest. The DEAE-cellulose ion exchange column chromatogram (Fig. 5) clearly shows that gel-filtration alone is insufficient to resolve c-VTG completely from other plama proteins; other blood proteins also appear to have high molecular weights and therefore they co-elute with VTG. However, the highly charged nature of c-VTG enabled it to be completely separated from these other plasma proteins by ion-exchange chromatography. The high charge of VTG relative to other plasma proteins is a characteristic feature of most vertebrate VTGs and is linked to the protein's high level of phosphorylation, which in turn is due to the large number of serine residues in the protein (see Table 1). The ion-exchange chromatography was effective in reproducibly resolving intact c-VTG from other blood proteins of similar molecular weight. However, it is worth reporting that when large quantities (many tens of mgs) of protein were applied to the DEAE-column the separation of proteins was less effective. It is therefore suggested that to obtain VTG preparations of high purity using our procedure, moderate levels of protein and/or large bed volumes of DEAE cellulose are used. Although we performed no quantitative analyses with c-VTG, Wiley et al. (1979), in a similar study on Xenopus VTG, suggested that VTG concentrations above 2.5 mg.ml- 1 gel should be avoided if Xenopus VTG was to be effectively resolved from other plasma proteins. The success of the protocol adopted in these stu-

119 dies to isolate intact c-VTG is clearly seen in the final Sepharose 6b chromatogram (Fig. 6), where the purified preparation eluted in a single discrete form at an identical position to where it eluted in the initial gel-filtration (compare Figs. 3a and 6). Evaluating the purity of the final c-VTG preparation without a sequence analysis is difficult; gel filtration and ion exchange chromatography do not allow proteins with subtle difference in molecular weight and charge, respectively, to be distinguished (although these techniques combined are likely to resolve all major blood proteins from one another). Although we cannot accurately assess the purity of our c-VTG, our subsequent studies reported in the accompanying paper (Tyler and Sumpter 1989) strongly suggest our VTG preparation was not contaminated to any significant extent with any other blood protein. In that study we show that an RIA based on our purified c-VTG does not detect anything in the plasma of male carp. If our c-VTG had been contaminated with one or more other blood proteins, then the RIA would not be expected to be specific to VTG and thus plasma from males (which does not contain VTG) would be expected to crossreact in the assay. Detailed studies on both the domestic fowl (Wang et al. 1983) and Xenopus (Wahli et al. 1981) have clearly shown that these species contain multiple VTGs, encoded by distinct genes. The situation in fish, however, is unknown; most reports speak of VTG without implying whether there are one or more proteins. Studies on the goldfish suggest that there might be more than one VTG (Hori et al. 1979; De Vlaming et al. 1980) although recent genetic studies on the rainbow trout, though not yet complete, provide no evidence for either more than one VTG gene or more than one VTG protein (Chen 1983; Vaillant et al. 1988). Furthermore, extensive analyses on VTG in the rainbow trout showed only one VTG (Babin 1987). Our chromatography does not have the resolving power to answer the question of VTG multiplicity in the carp, and it should therefore be realised that the purified preparation of c-VTG may contain either a single VTG or multiple VTGs with similar physical properties. An amino acid analysis of c-VTG showed that it

had an almost identical composition to goldfish VTG (De Vlaming et al. 1980) which reinforced the close relationship between these two species. Further, comparisons of the c-VTG amino acid composition with that of rainbow trout VTG demonstrated strong similarities between cyprinid and salmonid VTG; the only amino acids to differ significantly were aspartic acid and cysteine, both having greater representation in trout VTG. Xenopus VTG also shows a similar amino acid composition to the teleosts (De Vlaming et al. 1980; Table 1), although it contains a lower proportion of the nonpolar amino acids (alanine, isoleucine and leucine) and a considerably higher level of serine (11.5% versus 7.3-7.6%). The lower level of serine in cVTG is consistent with observations on other teleost VTGs (reviewed in Ng and Idler 1983). From the results reported in this study it is concluded that purified c-VTG can be isolated intact from the plasma of oestrogenized carp, and that the molecule is stable enough for the c-VTG to be used in physiological studies (Tyler and Sumpter 1989).

References cited Ansari, A.O., Dolphin, P.J., Lazier, C.B., Munday, K.A. and Akhtar, M. 1971. Chemical composition of an oestrogeninduced calcium binding glycolipophosphoprotein in Xenopus laevis. Biochem. J. 122: 107-113. Babin, P.J. 1987. Apolipoproteins and the association of egg yolk proteins with plasma high density lipoproteins after ovulation and follicular atresia in the rainbow trout (Salmo gairdneri). .1. Biol. Chem. 262: 4290-4296. Beams, H.W. and Kessel, R.G. 1973. Oocyte structure and early vitellogenesis in the trout, Salmo gairdneri.Am. J. Anat. 136: 105-122. van Bohemen, Ch.G., Lambert, J.G.D. and van Oordt, P.G.W.J. 1982. Vitellogenin induction by estradiol in estrone-primed rainbow trout, Salmo gairdneri.Gen. Comp. Endocrinol. 46: 136-139. Campbell, C.M. and Idler, D.R. 1976. Hormonal control of vitellogenesis in hypophysectionised winter flounder (Pseudopleuronectes americanus Walbaum). Gen. Comp. Endocrinol. 28: 143-150. Campbell, C.M. and Idler, D.R. 1980. Characterization of an oestradiol-induced protein from rainbow trout serum as vitellogenin by the composition and radioimmunological cross reactivity to ovarian yolk-fractions. Biol. Reprod. 22: 605617.

120 Chen, T.T. 1983. Identification of and characterisation of oestrogen-responsive gene products in the liver of rainbow trout. Biochem. Cell Biol. 61: 802-810. Copeland, P.A. and Thomas, P. 1988. The measurement of plasma vitellogenin levels in the marine teleost, the spotted seatrout (Cynoscion nebulosus) by homologous radioimmunoassay. Comp. Biochem. Physiol. 91B: 17-23. Craik, J.C.A. 1978. The effects of oestrogen treatment on certain plasma constituents associated with vitellogenesis in the elasmobranch Scyliorhinus canicula L. Gen. Comp. Endocrinol. 35: 455-464. Gibbins, A.M.V. and Robinson, G.A. 1982. Comparative study of the physiology of vitellogenesis in Japanese quail. Comp. Biochem. Physiol. 72A: 149-152. Hamazaki, T.S., uch, I. and Yamagami, K. 1987. Purification and identification of vitellogenin and its immunohistochemical detection in growing oocytes of the teleost, Oryzias latipes. J. Exp. Zool. 242: 333-342. Hara, A. and Hirai, H. 1978. Comparative studies on immunochemical properties of female specific serum protein and egg yolk proteins in rainbow trout, Salmo gairdneri. Comp. Biochem. Physiol. 59B: 339-343. Hickey, E.D. and Wallace, R.A. 1974. A study of vitellogenic protein in the serum of oestrogen-treated Ictalurusnebulosus. Biol. Bull. 147: 481. Ho, S.M., Dawko, D. and Callard, I.P. 1981. Effects of exogenous 17-oestradiol on plasma vitellogenin levels in male and female Chrysemys and its modulation by testosterone and progesterone. Gen. Comp. Endocrinol. 43: 413-421. Idler, D.R. and Campbell, C.M. 1980. Gonadotropin stimulation of oestrogen and yolk precursor synthesis in juvenile rainbow trout. Gen. Comp. Endocrinol. 41: 384-391. Le Menn, F. 1978. Some aspects of vitellogenesis in a teleostean fish: Gobius niger L. Comp. Biochem. Physiol. 62A: 495500. Le Menn, F. and Lamy, M.M. 1977. Proteins vitellines seriques et hepatique chez un poisson teleostean, Gobius niger L., soumis a des traitements hormonaux. Bordeaux Medical 10: 213-216. Nath, P. and Sundararaj, B.I. 1981. Isolation and identification of female-specific serum lipophosphoprotein vitellogenin in the catfish Heteropnuestesfossilis Bloch. Gen. Comp. Endocrinol. 43: 184-190. Ng, T.B. and Idler, D.R. 1983. Yolk formation and differentiation in teleost fishes. In Fish Physiology, Vol. 9A, pp. 373-404. Edited by W.S. Hoar, D.J. Randall and E.M. Donaldson. Academic Press, New York. Norberg, B. and Haux, C. 1985. Induction, isolation and characterization of the lipid content of plasma vitellogenin from two Salmo species: Rainbow trout Salmo gairdneriand sea trout Salmo trutta. Comp. Biochem. Physiol. 81B, 869-876.

Plack, P.A. and Frazer, N.W. 1971. Incorporation of L-14C leucine into egg proteins by liver slices from cod. Biochem. J. 121: 857-862. Redshaw, M.R.and Follett, B.K. 1971. The crystalline yolk platelet proteins and their soluble plasma precursor in an amphibian, Xenopus laevis. Biochem. J. 124: 759-766. Tata, J.R. and Smith, D.F. 1979. Vitellogenesis: A versatile model for hormonal regulation of gene expression. Rec. Prog. Horm. Res. 35: 47-95. Tyler, C.R. and Sumpter, J.P. 1989. Development of a radioimmunoassay for carp. Cyprinus carpio, vitellogenin. Fish Physiol. Biochem.: this issue. Wahli, W., Dawid, I.B., Ryffel, G.U. and Weber, R. 1981. Vitellogenesis and the vitellogenin gene family. Science 212: 298-304. Wallace, R.A. 1965. Resolution and isolation of avian and amphibian yolk-granule proteins using DEAE-cellulose. Anal. Biochem. 11: 297-311. Wallace, R.A. 1978. Oocyte growth in nonmammalian vertebrates. In The Vertebrate Ovary. pp. 469-502. Edited by R.E. Jones. Plenum Press, New York. Wallace, R.A. and Selman, K. 1981. Cellular dynamic aspects of oocyte growth in teleosts. Am. Zool. 21: 325-343. Wallace, R.A. and Selman, K. 1982. A new procedure for the isolation of intact vitellogenin from teleosts. In Reproductive Physiology of Fish. p. 161. Edited by C.J.J. Richter and H.J.Th. Goos. Pudoc, Wageningen. Wang, S.Y., Smith, D.E. and Williams, D.L. 1983. Purification of avian vitellogenin III: Comparison with vitellogenins I and II. Biochemistry 22: 6206-6212. Wiegand, M.D. 1982. Vitellogenesis in fishes. In Reproductive Physiology of Fish. pp. 136-146. Edited by C.J.J. Richter and H.J.Th. Goos. Pudoc, Wageningen. Wiley, H.S., Opresko, L. and Wallace, R.A. 1979. New methods for the purification of vertebrate vitellogenin. Anal. Biochem. 97: 145-152. Yu, J.Y.L., Dickhoff, W.W., Swanson, P. and Gorbman, A. 1981. Vitellogenesis and its hormonal regulation in the Pacific hagfish, Eptatretus stouti L. Gen. Comp. Endocrinol. 43: 492-502. Vaillant, C., Le Guellec, C., Pakdel, F. and Valotoire, Y. 1988. Vitellogenin gene expression in primary culture of male rainbow trout hepatocytes. Gen. Comp. Endocrinol. 70: 284290. De Vlaming, V.L., Vodicnik, M.J., Bauer, T., Murphy, T. and Evans, D. 1977. 17/-oestradiol effects on lipid and carbohydrate metabolism and on the induction of a yolk precursor in goldfish, Carassiusauratus. Life Sci. 20: 1945-1952. De Vlaming, V.L., Wiley, H.S., Delahunty, G. and Wallace, R.A. 1980. Goldfish Carassius auratus vitellogenin: Induction, isolation, properties and relationships to yolk proteins. Comp. Biochem. Physiol. 67B: 613-623.