Injection of estrogen into male Xenopus laevis induced the appearance of retinals (retinal and 3-dehydroretinal) and a considerable increase in the amount of retinols (retinol and 3-dehydroretinol) in the blood plasma. These retinoids were mainly in the all-trans form. Without estrogen injection, retinols were normally found in the blood plasma of both males and females, but only trace amounts of retinals were detected and these were restricted to the plasma of females.

The proteins in the blood plasma of estrogen-injected males were separated into two fractions. One fraction included vitellogenin, the precursor of egg yolk proteins, and the other contained some plasma proteins other than vitellogenin. Retinals were detected in the former and retinols in the latter. It is suggested that retinals are bound to vitellogenin and are taken up into oocytes in the process of vitellogenesis.

Previous papers (Seki et al. 1987; Azuma et al. 1988, 1990) have described our studies on retinoids in the eggs of toads and frogs. All-trans retinals were present but other retinoids and carotinoids were hardly detectable. In Xenopus laevis, a decrease in the level of egg retinals (all-trans retinal and 3-dehydroretinal) during the period of development before the larvae began to feed was accompanied by the formation of retinols (all-trans retinol and 3-dehydroretinol), retinyl esters (all-trans retinyl ester and 3-dehydroretinyl ester) and 11-cis retinals (retinal and 3-dehydroretinal). These results suggest that egg retinals are the source of the vitamin A necessary for larval development in Xenopus laevis.

Recent experiments have confirmed that retinals in the egg are bound to a yolk protein, lipovitellin 1, through a Schiff base linkage (Irie et al. 1991). It is known that egg yolk proteins are derived from a maternal blood plasma protein, vitellogenin, and that vitellogenin is synthesized in the liver, secreted into the blood circulation and sequestered by oocytes (Wallace and Jared, 1968, 1969; Wallace and Bergink, 1974; Wallace, 1985). If retinals are bound to vitellogenin, then retinals should also be produced in the liver, bound to vitellogenin and taken up into oocytes. Vitellogenin synthesis is normally controlled in females, and can be initiated in males by exogenous estrogen (Follett et al. 1968; Munday et al. 1968; Wallace and Dumont, 1968; Wallace and Jared, 1968; Dolphin et al. 1971). The concentration of vitellogenin in the blood plasma of an estrogen-injected male is higher than that of a normal female (Wallace and Jared, 1968; Selman and Wallace, 1983; Wallace, 1985). This study shows that retinals can be detected in the blood plasma of females and estrogen-injected males and that the amount of retinals in the former is smaller than that in the latter. Furthermore, the quantity of blood plasma retinol is also increased in estrogen-injected males. The fractionation of blood plasma proteins shows that the vitellogenin fraction contains retinals but little retinol. The results suggest that retinals are bound to vitellogenin but that retinols are not.

Animals and blood plasma

Adult male and female Xenopus laevis were purchased from Hamamatsu Seibutsu Kyozai Co., Ltd (Japan). They were reared at approximately 23°C and fed twice weekly with toad food supplied by the same company. Animals were injected via the dorsal lymph sac with 1mg of estradiol-17 β dissolved in 0.5ml of olive oil (Dolphin et al. 1971). Control animals received injections of the medium only. Several days or weeks after the injection, the toads were anesthetized by hypothermia and subjected to ventrotomy. Blood was collected from the heart using a syringe filled with 50 % Dulbecco’s Ca2+-free saline with 0.07mol l−1 sodium citrate (Wiley et al. 1979). The blood sample was centrifuged (1400 g, 10min) and the plasma was obtained as the supernatant.

Preparation of vitellogenin from blood plasma

Vitellogenin was prepared from the blood plasma according to Wiley et al. (1979). Briefly, 5ml of the plasma sample was mixed with 20ml of 20mmol l-1 EDTA solution and 1.6ml of 0.5mol l−1 MgCl2 was added. This was further mixed gently and centrifuged (13000 g, 10min) to obtain a supernatant (sample A) and a precipitate. The precipitate was dissolved in 3ml of 1mol l−1 NaCl, 50mmol l-1 TrisCl buffer (pH7.5) and centrifuged (13000 g, 10min). The supernatant obtained in this way was mixed with 25ml of distilled water, resulting in the precipitation of vitellogenin. The precipitate was redissolved in 3ml of 1mol l−1 NaCl, 50mmol l-1 TrisCl buffer (pH7.5) and dialyzed against 1mol l−1 NaCl and then twice against 50mmol l−1 TrisCl buffer (pH7.5). The sample obtained (sample B) included vitellogenin.

Detection of retinoids in the samples

Retinoids in the blood plasma or fractionated samples (samples A and B) were extracted by the oxime method and analysed by high-performance liquid chromatography (HPLC) as reported previously (Azuma et al. 1988, 1990; Irie et al. 1991). An HPLC system (JASCO, Japan) equipped with a 4.6 mm × 250 mm column of YMC-Pack A-003-3 SIL (Yamamura Chemical Laboratories Co. Ltd, Japan) was used and a mixture of n-hexane, diethylether and ethanol (90:10:0.1 v/v) was eluted at a flow rate of 1.3 mlmin-1. The absorbances of fractions at 350nm and 330nm were measured with JASCO 875UV detectors and peak areas were determined by integrating with a Chromatopac C-R4A (Shimadzu, Japan). Sometimes the fluorescence, excited at 330 nm, was monitored at 470nm using a fluorescence spectrophotometer F1000 (Hitachi, Japan). Quantities of several retinoid isomers were estimated from their absorption coefficients and the peak areas of known amounts of standard retinoids (Azuma et al. 1988, 1990; Irie et al. 1991).

SDS–polyacrylamide gel electrophoresis

SDS–polyacrylamide gel electrophoresis (SDS–PAGE) was performed according to Laemmli (1970) with 7.5% acrylamide gel as the separating gel. The quantification of proteins in the samples was carried out by the Lowry method (Lowry et al. 1951) and by a method measuring the fluorescence excited at 280nm using bovine serum albumin as the standard protein. The detailed procedure has been described previously (Irie et al. 1991).

Estrogen induction of retinals and retinols in the blood plasma

Fig. 1 shows HPLC chromatograms of retinoids extracted from female plasma (Fig. 1B) and from estrogen-injected male plasma (Fig. 1C) together with a chromatogram of standard retinaloximes and retinols (Fig. 1A). The volume of blood plasma for the extraction of retinoids was 0.3ml. In the case of Fig. 1C, the blood was collected 10 days after estrogen injection. Retinals in the samples are extracted as retinaloximes, so that the peaks of retinaloximes in Fig. 1B,C correspond to retinals in the plasma of a female and of an estrogen-injected male. Small syn-peaks of all-trans retinaloximes (peaks 1 and 2) in Fig. 1B indicate that small amounts of all-trans retinal and 3-dehydroretinal (retinal1 and retinal2) are present in female plasma. However, large peaks of all-trans retinaloximes are observed in Fig. 1C, indicating the presence of large amounts of all-trans retinal and 3-dehydroretinal. Although not shown in this figure, retinals were barely detectable in the plasma of males without estrogen injection. The peaks due to all-trans retinol and 3-dehydroretinol (retinol1 and retinol2) (peaks 5 and 6) are also shown in Fig. 1B,C and are much larger in the latter. The unnumbered peaks have not been identified. The amounts of retinals and retinols in the blood plasma were calculated from such HPLC data.

Table 1 shows the amounts of retinoids in the blood plasma of adult females and males and estrogen-injected males. Blood samples were collected from estrogen-injected males 1, 2 and 3 weeks after injection. Only the amounts of all-trans retinals (retinal and 3-dehydroretinal) and all-trans retinols (retinol and 3-dehydroretinol) were calculated from HPLC data. As shown in Table 1, all-trans retinals were already present in the male plasma 1 week after estrogen injection and the levels increased with time after the injection. Maximum amounts of retinals were found 2–3 weeks after injection in this experiment. The percentage of 3-dehydroretinal in all-trans retinals (retinal and 3-dehydroretinal) was about 50% in every case. In contrast, all-trans retinols (retinol and 3-dehydroretinol) were present in the plasma of females and males without estrogen injection (about 100pmolml-1). However, injection of estrogen increased the amount of all-trans retinols by a factor of 20–40 (Table 1). The percentage of 3-dehydroretinol in all-trans retinols (retinol and 3-dehydroretinol) is about 20% and is smaller than that of 3-dehydroretinal in all-trans retinals.

Retinals and retinols in the fractions of blood plasma

Blood plasma was prepared from estrogen-injected males 2 weeks after injection. After adding MgCl2 and EDTA to the plasma, the supernatant (sample A) and the precipitate (sample B) were obtained as described in Materials and methods. Sample A, sample B and the original blood plasma were assayed for retinoids by HPLC and for proteins by SDS–PAGE.

Fig. 2 shows HPLC chromatograms of retinoids extracted from sample A (Fig. 2A), sample B (Fig. 2B) and the blood plasma (Fig. 2C). Each sample was prepared from an aliquot of the same plasma solution. The elution peaks were identified using standard retinaloximes and retinols, as in Fig. 1. Fig. 2C shows the presence of both retinals and retinols in the blood plasma. In this case, the peaks due to all-trans retinal (1 and l), all trans 3-dehydroretinal (2 and 2), 13-cis retinal (3 and 3) and 13-cis 3-dehydroretinal (4 and 4) are clear. The two small peaks eluting just before peak 1 and two further peaks between peaks 4 and l are not numbered in this figure but are assumed to be due to 11-cis retinal and 3-dehydroretinal (M. Azuma, unpublished observation). Large peaks (5 and 6) due to all-trans retinols (retinol and 3-dehydroretinol) are also shown in Fig. 2C. As indicated in Fig. 2A, sample A has mostly all-trans retinol and 3-dehydroretinol (peaks 5 and 6), in amounts similar to those found in the blood plasma (Fig. 2C), and low levels of retinals, the peaks of which are not numbered. In contrast, sample B, shown in Fig. 2B, has large amounts of all-trans retinal (peaks 1 and l) and all-trans 3-dehydroretinal (peaks 2 and 2) but small amounts of retinols (peaks 5 and 6). The amount of retinals in sample B seems to be equal to that in the original blood plasma shown in Fig. 2C.

The separation of proteins in the fractionated samples (sample A and sample B) and in the blood plasma by SDS–PAGE is shown in Fig. 3. The amount of protein in sample A (Fig. 3A) was much greater than the amounts in sample B (Fig. 3B) and in the blood plasma (Fig. 3C). The pattern of sample B is very similar to that reported for vitellogenin in the blood plasma of Xenopus laevis (Wiley et al. 1979; Wiley and Wallace, 1981), indicating that the main protein in sample B is vitellogenin (arrowed in Fig. 3). The pattern of sample A shows that the main proteins have relative molecular masses smaller than that of vitellogenin. The results shown in Figs 2 and 3 indicate that the fraction including vitellogenin (sample B) has retinals but little retinol while the other fraction (sample A) contains retinols and some proteins other than vitellogenin

This study shows that an estrogen injection induces the appearance of retinals (mainly all-trans retinal and 3-dehydroretinal) and vitellogenin in the blood plasma of male Xenopus laevis. A method of purifying vitellogenin (Wiley et al. 1979) produced the fraction that precipitated in the presence of MgCl2 and EDTA and included vitellogenin and retinals (sample B). To examine the possibility that retinals in the blood plasma were bound to vitellogenin by the Schiff base linkage, as in the egg (Irie et al. 1991), the effects of sodium borohydride (NaBH4) on the retinals in the blood plasma and in the fractionated sample (sample B) were tested (results not shown). Hardly any retinals could be extracted from either the blood plasma or sample B after the addition of NABH4 but retinols could be extracted from the blood plasma. This is very similar to results obtained with egg retinals (Seki et al. 1987; Irie et al. 1991), and supports the idea that the retinals in the blood plasma are bound to vitellogenin by the Schiff base linkage. Small amounts of retinals were detected in the blood plasma of female Xenopus laevis without estrogen injection. These retinals also seemed to be bound to vitellogenin. The vitellogenin–retinal complex in the maternal blood plasma may be taken up into oocytes and converted into a lipovitellin–retinal complex during the process of vitellogenesis.

An estrogen injection induces not only the appearance of retinals but also an increase in the amount of retinols (all-trans retinol and 3-dehydroretinol) in the blood plasma. These retinols were barely detectable in the vitellogenin fraction (sample B) but were found mostly in the fraction (sample A) that was soluble in the presence of MgCl2 and EDTA. Sample A did not include vitellogenin but contained other proteins whose relative molecular masses were lower than those of vitellogenin polypeptides (Mr approximately 200000). Estrogen induction of serum retinal-binding protein (RBP) mRNA has been reported in several papers (McKearin et al. 1987; McKearin and Shapiro, 1988; Whitman et al. 1990), suggesting that sample A obtained in this study probably contains large amounts of RBPs binding retinols. Retinols were barely detectable in oocytes (M. Azuma, unpublished observation) or eggs (Seki et al. 1987; Azuma et al. 1990; Irie et al. 1991). If oocytes have vitellogenin receptors but not RBP receptors, retinols would not be taken up into oocytes. These retinols may function in the same way as vitamin A in cells other than oocytes.

Small amounts of 11-cis retinals (retinal and 3-dehydroretinal) were detected in the blood plasma of estrogen-injected males (Fig. 2C) but were practically absent from the egg (Seki et al. 1987; Azuma et al. 1990; Irie et al. 1991). Retinals present in the egg occur mainly in the all-trans form together with small amounts of the 13-cis form (about 10%) (Seki et al. 1987). These results are inconsistent with the idea that retinals binding lipovitellin in the egg are derived only from retinals binding vitellogenin in the blood plasma. Experiments to investigate binding sites for retinal on both vitellogenin and lipovitellin, and to examine a possible stoichiometry between retinal and each of these proteins, are now in progress. The results will offer more reliable information on the vitellogenin–retinal complex and its conversion to the lipovitellin– retinal complex.

The authors thank Dr H. Takasaki (Osaka Kyoiku University, Japan) for kindly supplying X. laevis and Dr T. Suzuki (Hyogo College of Medicine, Japan) for his gift of an equimolar mixture of standard oximes. We are also grateful to Dr K. Azuma (Osaka Medical College, Japan) and Professor T. P. Williams (Florida State University, USA) for their critical reading of the manuscript. This work was supported in part by a Grant-in-Aid for Scientific Research to M.A. (01540595) from the Japanese Ministry of Education, Science and Culture.

Azuma
,
M.
,
Seki
,
T.
and
Fujishita
,
S.
(
1988
).
Formation of visual pigment chromophores during the development of Xenopuslaevis
.
VisionRes
.
28
,
959
964
.
Azuma
,
M.
,
Seki
,
T.
and
Fujishita
,
S.
(
1990
).
Changes of egg retinoids during the development of Xenopus laevis
.
Vision Res.
30
,
1395
1400
.
Dolphin
,
P. J.
,
Ansari
,
A. Q.
,
Lazier
,
C. B.
,
Munday
,
K. A.
and
Akhtar
,
M.
(
1971
).
Studies on the induction and biosynthesis of vitellogenin, an oestrogen-induced glycolipophospho-protein
.
Biochem. J.
124
,
751
758
.
Follett
,
B. K.
,
Nicholls
,
T. J.
and
Redshaw
,
M. R.
(
1968
).
The vitellogenic response in the South African clawed toad (Xenopuslaevis) Daudin
.
J. Cell Physiol.
(Suppl.)
72
,
91
102
.
Irie
,
T.
,
Azuma
,
M.
and
Seki
,
T.
(
1991
).
The retinal and 3-dehydroretinal in Xenopus laevis eggs are bound to lipovitellin 1 by a Schiff base linkage
.
Zool. Sci.
8
,
855
863
.
Laemmli
,
U. K.
(
1970
).
Cleavage of structural proteins during the assembly of the head of bacteriophage T4
.
Nature
227
,
680
685
.
Lowry
,
O. H.
,
Rosebrough
,
N. J.
,
Farr
,
A. L.
and
Randall
,
R. J.
(
1951
).
Protein measurement with Folin phenol reagent
.
J.biol. Chem.
193
,
265
275
.
Mckearin
,
D. M.
,
Barton
,
M. C.
,
Keller
,
M. J.
and
Shapiro
,
D. J.
(
1987
).
Estrogen induces transcription of the Xenopus laevis serum retinol-binding protein gene
.
J. biol. Chem.
262
,
4939
4942
.
Mckearin
,
D. M.
and
Shapiro
,
D. J.
(
1988
).
Persistent estrogen induction of hepatic Xenopus laevis serum retinol binding protein mRNA
.
J. biol. Chem
.
263
,
3261
3265
.
Munday
,
K. A.
,
Ansari
,
A. Q.
,
Oldroyd
,
D.
and
Akhtar
,
M.
(
1968
).
Oestrogen-induced calcium-binding protein in Xenopus laevis
.
Biochim. biophys. Acta
166
,
748
751
.
Seki
,
T.
,
Fujishita
,
S.
,
Azuma
,
M.
and
Suzuki
,
T.
(
1987
).
Retinal and 3-dehydroretinal in the egg of the clawed toad, Xenopus laevis
.
Zool. Sci.
4
,
475
481
.
Selman
,
K.
and
Wallace
,
R. A.
(
1983
).
Oogenesis in Fundulusheteroclitus. III. Vitellogenesis
.
J. exp. Zool
.
226
,
441
457
.
Wallace
,
R. A.
(
1985
).
Vitellogenesis and oocyte growth in nonmammalian vertebrates
. In
Developmental Biology
, vol.
1
(ed.
L. W.
Browder
), pp.
127
177
. Philadelphia: Saunders.
Wallace
,
R. A.
and
Bergink
,
E. W.
(
1974
).
Amphibian vitellogenin: Properties, hormonal regulation of hepatic synthesis and ovarian uptake and conversion to yolk proteins
.
Am. Zool.
14
,
1159
1175
.
Wallace
,
R. A.
and
Dumont
,
J. N.
(
1968
).
The induced synthesis and transport of yolk proteins and their accumulation by the oocyte in Xenopus laevis
.
J. Cell Physiol.
(Suppl.)
72
,
73
89
.
Wallace
,
R. A.
and
Jared
,
D. W.
(
1968
).
Estrogen induces lipophosphoprotein in serum of male Xenopus laevis.
Science
160
,
91
92
.
Wallace
,
R. A.
and
Jared
,
D. W.
(
1969
).
Studies on amphibian yolk. VIII. The estrogen-induced hepatic synthesis of a serum lipophosphoprotein and its selective uptake by the ovary and transformation into yolk platelet proteins in Xenopus laevis
.
Devl Biol.
19
,
498
526
.
Whitman
,
M. M.
,
Harnish
,
D. C.
,
Soprano
,
K. J.
and
Soprano
,
D. R.
(
1990
).
Retinol-binding protein mRNA is induced by estrogen in the kidney but not in the liver
.
J. LipidRes
.
31
,
1483
1490
.
Wiley
,
H. S.
,
Opresko
,
L.
and
Wallace
,
R. A.
(
1979
).
New methods for the purification of vertebrate vitellogenin
.
Analyt.Biochem.
97
,
145
152
.
Wiley
,
H. S.
and
Wallace
,
R. A.
(
1981
).
The structure of vitellogenin. Multiple vitellogenins in Xenopus laevis give rise to multiple forms of the yolk proteins
.
J. biol. Chem.
256
,
8626
8634
.