87
Biochimica et Biophysica Acta, 4 1 1 ( 1 9 7 5 ) 8 7 - - 9 6 © E l s e v i e r S c i e n t i f i c P u b l i s h i n g C o m p a n y , A m s t e r d a m - - P r i n t e d in T h e N e t h e r l a n d s -
BBA 27748
C E L L U L A R RETINOL-BINDING PROTEIN
M A R K M. B A S H O R *
and FRANK CHYTIL**
Departments of Biochemistry and Medicine, Vanderbilt University, School of Medicine Nashville, Tenn. 37232 (U.S.A.) (Received April 10th, 1975)
Summary 1. A protein which binds retinol in vitro with high affinity and specificity was detected by sucrose gradient centrifugation or by gel filtration after preincubating rat tissue cytosols with all-trans-[ 3 H] retinol. This protein sediments in the 2 S region of sucrose gradients. Molecular size determination b y gel filtration indicates a molecular weight of 16 000. 2. Competition studies revealed that only all-trans-retinol, n o t retinal or retinoic acid, competes for binding. The binding of radioactive retinol is reversible. 3. This protein was detected in cytosols of rat liver, lung, spleen, brain, testis, ovaries, uterus and intestinal mucosa whereas heart or gastrocnemius muscle seem to lack this protein. 4. The cellular retinol binding protein was found in fetuses as early as day 12 of the gestation period and possessed the same specifity for the ligand as the one in adult tissues. 5. This binding c o m p o n e n t was not detected in cytosols prepared from Novikoff hepatoma, ascites hepatoma AS-30D, mouse Ehrlich ascites t u m o r and mouse pituitary t u m o r cell line AtT 20. 6. The cellular retinol binding protein seems to be different from that described to be present in the serum as suggested by difference in size and b y the inability of the antisera against the serum retinol binding protein to remove the cellular bindh~g protein from the cytosol preparations.
Introduction In a recent report we presented preliminary evidence that cytosols prepared by high speed centrifugation of several rat tissue homogenates (lung, * Present address: Department of Biology, Umverslty of California San Diego, La Jolla, Cahf., U.S.A. *~ To whom rep~lnt requests sho,lld be addressed.
88 kidney, intestinal mucosa, liver and testis) contained a protein capable of binding [3 H]retinol in vitro with high specifity [1]. This protein has an approximate molecular weight of 16 000 as estimated by gel filtration and shows a sedimentation coefficient of 2 S on sucrose gradients. In h u m a n [1] as well as in rabbit lung [2] proteins of comparable properties have been detected. Furthermore a partial purification of the retinol binding protein from rabbit lung has been achieved in this laboratory [2]. Heptane extracts o f this material have a fluorescence spectrum identical to that of authentic all-trans-retinol in heptane suggesting that retinol was in fact the naturally occurring ligand [2]. The in vitro binding of [3H] retinol in tissue cytosols could be easily distinguished from binding to a serum c o m p o n e n t which sedimented in the 4.6 S region of sucrose gradients and which had a molecular weight of approximately 67 000 [1]. The work to be reported here was undertaken to extend the above studies with regard to the following points of interest. First, a number of normal and neoplastic tissues were surveyed to determine the distribution of the cellular retinol binding protein. Second, fetal rats at various stages of development were studied in order to see whether the appearance of this protein corresponds to the time when the requirement for vitamin A becomes critical [3]. Third, the components in tissue, serum and amniotic fluid which bind retinol in vitro were further characterized in order to distinguish them from the serum transport protein complex [4,5] for vitamin A. Materials and Methods
Matertals. All-trans-[15-3H]retinol (1.25 Ci/mmol) was purchased from New England Nuclear Corporation. Unlabelled retinol, retinal, retinoic acid, retinyl acetate and palmitate (all in the all-trans-configuration) were purchased from Sigma. Rat serum albumin, Fraction V [6] was also from Sigma or prepared by the m e t h o d of Debro and coworkers [7]. Prior to use lyophylized preparations were dissolved in 0.05 M Tris • HC1 buffer, pH 7.5 to the desired protein concentration. Anti-rat serum albumin (rabbit-immunoglobulin) was obtained from Nutritional Biochemicals Corporation or Cappel Laboratories. Allyl alcohol, crotyl alcohol and ~-ionone were products of Eastman Kodak. Anti-rat serum retinol binding protein (sheep) was a generous gift of Dr D.S. Goodman. Specimens of Novikoff Hepatoma, mouse Ehrlich ascites cells and ascites h e p a t o m a AS-30D [8] were kindly provided by Dr L.S. Hnilica and the AtT-20 cells [9] by Dr R. Harrison. Animals. Adult Sprague-Dawley rats weighing 150--250 g (Holtzman, Madison, or Harlan Industries, Indianapolis) were fed Lab Chow (Ralston, Purina Co.) ad libitum. Fetuses were obtained at days 12, 14, and 16 of gestation. Female rats were left overnight with a male and the first day o f pregnancy was determined by a microscopic examination of a vaginal smear for the presence of sperm. The fetuses were delivered by Caesarean section. The animals were killed by decapitation. Preparation of cytosols. Tissues were homogenized in four volumes (w/v) of 0.05 M Tris • HC1 buffer, pH 7.5 in a glass-Teflon homogenizer. The homogenates were then centrifuged at 31 000 × g for 10 min and the resulting
89 supernatant centrifuged again at 105 000 × g for 60 min at 4°C. The highspeed supernatants (cytosols) were used immediately or after storage at --20 ° C. Preparation of serum and amniotic fluid. Blood was collected by cardiac puncture of animals anesthetized with ether and allowed to clot at 0°C for 15--20 min. Serum was then obtained as described earlier [1]. Amniotic fluid was collected from the amniotic sac by aspiration and clarified, when necessary, by centrifugation at 314300 × g for 10 min in a refrigerated centrifuge. Incubation with [3 H] retinol. In principle, conditions used were those described earlier [1]. Briefly, 1 ml aliquots of the appropriate samples were incubated in the dark at 4°C for 12--16 h with 4 . 1 0 - 8 M [3H]retinol. The isotope was added in 5 #l of ethanol. Unlabeled all-trans-retinol, retinal and retinoic acid were used for binding competition studies. The unlabelled competitors were added to the incubation mixtures in 5 pl of ethanol to a final concentration of 8 • 10 -6 M (200-fold excess over the labelled retinol). Control samples received ethanol alone. Sucrose gradient centrifugation. Binding of [3 H] retinol was determined by sucrose gradient centrifugation as previously described [1] with the exception t h a t linear 5--20% sucrose (w/v) gradients were prepared in 0.05 M Tris • HC1 buffer, pH 7.5. Determination of [3 H] retinol binding by gel filtration. In some experiments the a m o u n t of bound [3 HI retinol in serum or tissue cytosols was determined as follows. An aliquot (0.1 ml) of the incubation mixture was applied to a column (1.1 X 5.5 cm) of Sephadex G-25 (Fine) equilibrated with 0.05 M Tris • HCI buffer, pH 7.5. The sample was eluted with the same buffer at a flow rate o f about 25 ml/h; 0.5 ml fractions were collected. The fractions were then transferred to counting vials, mixed with 5 ml of the scintillation cocktail [ 1 ], and counted. [3 H] Retinol eluting from the columns in the void volume was considered " b o u n d " retinol and was clearly separated from the free [3 H] retinol appearing in later fractions. Molecular weight determination. The molecular weight of the [3 H] retinol binding components in serum and cytosols from various tissues was estimated either from the sedimentation data or by gel filtration as described previously
[1]. Immunological procedures. An immunoglobulin fraction was isolated from control or immune sera by repeated a m m o n i u m sulfate precipitation
[10]. Immunodiffusion experiments were performed by Ouchterlony doublediffusion technique [ 11] using commercially prepared plates (Hyland, Travenol Laboratories). Immunoelectrophoresis was performed on cellulose acetate strips equilibrated with 0.0275 M barbiturate buffer pH 8.6 using a Beckman Model R 101 microzone apparatus. Electrophoresis was carried o u t for 40 min at 150 V. Treatment of serum with antibody. Rat serum was diluted 80-fold in 0.05 M Tris • HCI buffer pH 7.5 prior to incubation with anti-rat serum albumin immunoglobulin. Aliquots of the diluted serum were then incubated with 200 or 400 ~1 of anti-albumin or control immunoglobulins at 37°C for 60 min and then overnight at 4°C. The samples were then centrifuged at 105 000 × g for 60 min and 1.0 ml of each supernatant was incubated with 4 • 10 -8 M
90 [3 H] retinol overnight. Binding of [3 H] retinol was determined by sucrose gradient centrifugation. Treatment o f cytosols with antibody. A similar procedure was used when studying the effect of anti-rat retinol binding protein serum and anti-rat albumin immunoglobuhn on the cellular retinol binding protein. Aliquots of testis cytosol (1.0 ml) previously labelled with 4 • 10 -8 M [3 H] retinol were incubated with 200, 400, 600 or 800 pl of antiserum or control sheep serum for 1 h at 37°C followed by overnight incubation at 4°C. Samples were then assayed for [3 H] retinol binding by sucrose gradient centrifugatlon as described above. Protein assay. Protein was determined by the m e t h o d of Lowry et al. [12]. Results
T~me course o f [3 H] retmol binding m cytosols. In order to determine the optimum time for incubation with tissue cytosols, a sample of liver cytosol (15 mg protein/ml) was mixed with [3 H] retinol, 4 • 10 -8 M. At various times thereafter a 0.1 ml aliquot was removed and the a m o u n t of b o u n d [3 H]retinol determined by gel hltration; The results of this experiment are presented in Fig. 1. Binding was essentially completed in 4 h and remained unchanged for at least 20 h thereafter. In the following experiments an incubation time o f 12-16 h was used. Reversibility o f retinol binding. It was desirable to know if the binding of [3 H] retinol to the cellular retinol binding protein was reversible. This was done by incubating testis cytosol (usually 7--8 mg protein/ml) with [3 H] retinol to achieve m a x i m u m binding, followed by the addition of a 200-fold excess of
24
I0
o_ x
16
0 x
8
12
n (J
6
(.9
4
8
2
4 I
4
I
I
I
I
I
I
I
I
8 12 16 2 0 2 4 HOURS INCUBATION
I
2
5
HOURS
i
4
I
5
INCUBATION
F i g . 1. E f f e c t o f i n c u b a t i o n t i m e o n [ 3 H ] r e t m o l b i n d i n g . L : v e r c y t o s o l w a s i n c u b a t e d a t 4 ° C f o r v a r i o u s p e r i o d s o f t u n e f o l l o w i n g t h e a d d i t i o n o f 4 • 1 0 - 8 M [ 3 H ] r e t m o l . T h e b i n d i n g w a s a n a l y z e d a t e a c h tLrne p o i n t b y gel f i l t r a t i o n . F i g . 2. D i s p l a c e m e n t of bound [3H]retmol m testis cytosol foUowmg the add~tlon of excess uniabelled retmol. Testm eytosol was incubated overnight with 4.10 -8 M [3H] retmol. Then a 200-fold excess of uniabelled retmol was added. Ahquots of the mixture were removed at various times thereafter and assayed f o r [ 3 H ] r e t m o l b i n d i n g b y gel f i l t r a t i o n .
91 unlabelled retinol. As can be seen in Fig. 2, b o u n d [3 H] retinol could be displaced to some extent following the addition of unlabelled retinol. The binding of retinol to the cellular retinol binding protein is apparently n o t covalent as shown recently [ 2 ] . Molecular size of the retinol binding component. The apparent molecular weight of the cellular binding protein from a variety of tissue was estimated on a calibrated Sephadex G-100 column. In all cases the cellular binding protein eluted after a myoglobin standard, giving an approximate molecular size of 16 000. Purified serum retinol-binding protein has been reported to elute before myoglobin on a Sephadex G-100 column, giving an approximate molecular size of 19 000(13). A refinement of apparent molecular weight determined by gel filtration has been reported for cellular retinol binding protein purified 1000 fold from rat testes [14]. Using a standardized Sephadex G-75 column the partially purified natural elutes rear ribonuclease A, giving an even lower apparent molecular weight of 14 000. Tissue distribution of the binding component. We have previously described the binding of [3 H] retinol to a 2 S protein in cytosols from rat testis, liver, lung, kidney and intestinal mucosa [1]. The tissues examined in the present study, ovary, uterus, spleen and brain also contain the retinol binding protein as shown in a typical experiment in Fig. 3. On the other hand when cytosols from heart muscle ~ and gastrocnemius muscle were incubated with [3 H] retinol no 2 S peak was detected on sucrose gradient, as can be seen from Fig. 3. These tissues apparently do not contain this 2 S protein or the amounts present are below the sensitivity of the assay. The muscle cytosols were prepared in the same manner as the other tissue cytosols (homogenized in 4 volumes of buffer). Reducing the volume of buffer to 2 yielded the same results, suggesting that the retinol binding c o m p o n e n t was in fact absent. When labelled retinol was applied on sucrose gradients in the absence of any protein in 0.05 M Tris • HC1 buffer (pH 7.5) only, the radioactivity showed pattern n o t distinguishable from that shown for muscle in Fig. 3. [3 H] Retinol binding in tumor cytosols. Several tumors have also been examined for retinol binding component. No binding was observed in Novikoff Hepatoma. Likewise no 2 S binding was seen in a mouse pituitary t u m o r cell line (ART-20 cells) maintained in tissue culture, mouse Ehrlich ascites cell or a hepatoma strain AS-30D maintained as an ascites tumor. On the other hand [3 H] retinol binding in normal liver cytosol was previously observed [1]. Whether the binding c o m p o n e n t does exist in normal mouse pituitary remains to be shown. Ontogeny of the retinol-binding component. In light of the evidence concerning the requirement of vitamin A for fetal development [3] the fetal cytosols were examined at various stages of development for the presence of the 2 S binding component. The 2 S binding protein for [3 H] retinol was detected in the earliest stage tested, day 12 of gestation. No 2 S binding was observed in amniotic fluid preparations from days 12, 14 or 16 of gestation, although binding in the 4.6 S region of the gradient was found, as will be discussed below. Ligand specificity of the retinol binding component. The previous work has shown that binding of [3 HI retinol in cytosols from adult rat testis [1] or
92
i
:7 0 I'£9
z
o
I'-£.9
I0
LL
A
OVARY
6
IX LL
b
t~
x
x
I:L (D
I:L {.9 '~ 4 8 12 16 TOP FRACTION NUMBER
i
i
i
A
12
z - + RETINAL-
8
L+ RETINOIC
,~ 4
8
12
16
4
8
12
16
/
TOP
TOP FRACTION
F i g . 3. S u c r o s e g r a d i e n t c e n t r l f u g a t l o n of cytosols from ovary and muscle tion with 4 " 10 -8 M [3H] retmol. For details see Materials and Methods.
NUMBER
(gastrocnemlus)
after incuba-
F i g . 4. L l g a n d s p e c l f l t y o f c e l l u l a r r e t m o l b i n d i n g p r o t e i n m 1 6 - d a y - o l d fetuses Fetal cytosols were p r e p a r e d a n d i n c u b a t e d o v e r n i g h t w i t h 4 • 1 0 - 8 M [ 3 H ] r e t m o l in t h e a b s e n c e ( ~ ) a n d p r e s e n c e o f 2 0 0 - f o l d m o l a r e x c e s s o f A r e t m o l ( / ) , B r e t m o l ( u ) o r r e t m o l c a c i d (z~). S u c r o s e g r a d i e n t c e n t n f u g a t m n was p e r f o r m e d as d e s c r i b e d m M a t e r i a l s a n d M e t h o d s .
adult rabbit lung [2] show a high specifity. Of the compounds tested only all-trans-retinol and not retinoic acid added in a 200-fold excess would compete with the radioactive ligand. In most preparations retinal did not compete either. Similar specifity was observed in cytosols prepared from fetal rat at the day 16 of gestation as shown in Fig. 4 suggesting that the fetal tissue binding protein is not different from that detected in the adult animal. The binding specificity was further studied in adult rat testis cytosols using a 200-fold excess of retinyl acetate of palmitate. Both compounds competed as effective as retinol. The peak of radioactivity observed in the presence of an excess of nonradioactive retinol represents non specific binding of [3 H] retinol displaced by the competitor to other macromolecules present in the cytosols such as albumin. It was also of interest to test other compounds which do not have vitamin A activity, but which resemble portions of the retiaol molecule, for their ability to compete for tritiated retinol binding. Allyl alcohol, crotyl alcohol or ~-ionone added at 8 • 10 -6 M concentration were unable to compete. Bmding o f [3H] retmol to serum and amniotic fluid components. Serum and amniotic fluid show no evidence for 2 S binding component but both contain a binding component(s) for [3 H] retinol sedimenting at 4.6 S, the same position as observed for [3 H] retinol bound to authentic rat serum albumin (Fig. 5, testis cytosol included for comparison). If diluted serum was preincubated with rabbit anti-rat serum albumin immunoglobulin fraction, the 4.6 S binding peak for [3 H] retinol was abolished (Fig. 6). No antibodies directed against rat serum prealbumin were found using immunoelectrophoresis in the above anti-rat serum albumin immunoglobulins. These results suggest strongly that serum albumin is responsible for the in vitro binding of [3 H] retinol by serum. Immunodiffusion and immunoelectrophoresis of rat amniotic fluid and tissue cytosols against anti-rat serum albumin antiserum indicated the presence
93
TESTIS
~-AMNIOTIC FLUID
CYTOSOL 14 z o F--
~-,.,~ERUM ( I:lO} ~'~--ALBUMI N
U_
'o ×
6
z O t(D
•R
6
c3
n~ t~
13_ o
4
o_ x ,
4
8
12
16 20
FRACTION NUMBER TOP
D_ {.9
2
i
i
i
i
T~Op4 8 12 16 FRACTION NUMBER
F i g . 5. S u c r o s e gradient profiles o f [ 3 H ] r e t i n o l b i n d i n g b y s e r u m , ananiotic fluid, s e r u m a l b u m i n and testis c y t o s o l . R a t s e r u m ( d i l u t e d 1 : 1 0 w i t h 0 . 0 5 M T r i s • H C1 b u f f e r , p H 7 . 5 ) a m m o t i c fluid, s e r u m a l b u m i n (4 m g per m l o f the a b o v e b u f f e r ) w e r e i n c u b a t e d w i t h t n t i a t e d r e t i n o l and a n a l y z e d b y s u c r o s e grachent c e n t r i f u g a t i o n . F i g . 6 . S u c r o s e gradient profiles of [ 3 H ] r e t l n o l b i n d i n g in s e r u m t r e a t e d w i t h c o n t r o l or anti-rat s e r u m a l b u m i n i m m u n o g l o b u h n s . R a t s e r u m w a s d i l u t e d 8 0 - f o l d w i t h 0 . 0 5 M T n s • H C I b u f f e r , p H 7 . 5 prior t o t r e a t m e n t w i t h rabbit c o n t r o l n n m u n o g l o b u l i n s ( o ) or anti-rat s e r u m a l b u m i n i m m u n o g l o b u h n s (~). T h e s a m p l e s w e r e t h e n a n a l y z e d for [ 3 H ] r e t i n o l b y s u c r o s e g r a d m n t c e n t r i f u g a t i o n .
of albumin, suggesting the 4.6 S peak observed in amniotic fluid and occasionally in tissue cytosols also represents binding to albumin, due to serum contamination. If additional serum is added intentionally to testis cytosol no difference is observed in the 2 S peak, suggesting that the cellular retinol binding protein does not form complexes with serum components as the serum retinol binding protein does with prealbumin. Effect o f anti-rat serum retinol binding protein antiserum on [3H] retinol binding in testis cytosol. To examine the possibility that the binding of [3 H] retinol in tissue cytosols by a c o m p o n e n t sedimenting at 2 S was due to contaminating serum retinol binding protein described earlier [ 4 , 5 ] , testis cytosol was incubated with monospecific antiserum to rat serum retinol binding prorein. The results presented in Fig. 7 show that under conditions employed, which are similar to those employed for removing the serum albumin, antiserum against retinol binding protein does not affect the amount of [3 H]retinol bound in testis cytosol or the position of the peak on sucrose gradient to which the bound radioactive retinol sediments. While Fig. 7 presents the results obtained with 800 pl of antiserum, similar results were obtained when 200, 400, and 600 pl of antiserum was employed. The antiserum gave a positive double diffusion and immunoelectrophoretic tests with rat serum, but no precipitin bands were observed when testis cytosol was tested. Preincubation o f testis cytosol with rabbit anti-rat albumin immunoglobulins failed to remove the cellular retinol binding protein. The shoulder of radioactivity in the above figure originated probably by binding of radioactive retinol to albumin of the added antiserum. These results indicate that the cellular retinol binding protein is antigenically distinct from that found in the serum.
94 i
z
12
i
f
2s
I..ha:
8
b x
4
a_ i
T~P 4
i
i
i
8
[2
16
FRACTION NUMBER Fig.
7. Sucrose
(~') a n d w i t h Methods.
gradients
ant:-serum
profiles
of
[3H]
retlnol
(;~) a g a i n s t r a t s e r u m
binding
retmol
m
testis
binding
cytosols
pretreated
protein
F o r d e t a i l s see M a t e r i a l s a n d
with
control
serum
Discussion The studies presented here indicate that numerous rat organs contain a macromolecule capable of binding [-~H] retinol in vitro with high speclfmity. The binding in tissue cytosols occurs fairly rapidly, with maximum occurring within four hours. That maximum binding does not occur sooner may be due to the presence of endogenous retinol bound to the tissue c o m p o n e n t which the added radioactive retinol must displace. The binding of retinol is apparently reversible and not covalent, as was demonstrated b y the ability of unlabelled retinol to displace previously b o u n d [3 H] retinol. Furthermore it was recently demonstrated that rabbit lung contains a retinol binding protein that is probably analogous to that described here for rat tissues [2]. The fluorescence spectrum of a heptane extract of this binding c o m p o n e n t was identical to all-trans-retinol, indicating that the binding c o m p o n e n t carries all-trans-retinol m vivo and that the binding is non-covalent. Partially purified rat testicular retinol binding protein also contains all-trans-retinol [ 14]. The binding specifity established by competitive assays :s remarkable. In cytosols retinoic acid never h a s b e e n found to compete, whereas retinal appeared to partially c o m p e t e in some preparations for [3 H] retinol binding [ 1]. However this may reflect the in vitro reduction of retinal to retinol, with the retinol thus formed being the competing ligand. Similarly the fact that retmyl acetate and retinyl palmitate c o m p e t e as effectively as retinol for [a H] retinol binding in testis cytosol may result from cleavage of these esters to yield retinol. In partially purified preparations retinal [2] or retinyl esters [14] do not compete. Furthermore the inability of other c o m p o u n d s representing portions of the retinol molecule to c o m p e t e for [3 H] retinol binding again emphasizes the highly specific nature of the binding site. The tissue retinol binding protein shows a rather wide distribution in different tissues. Heart and gastrocnemius muscle are the only negative tissues found at this time. The failure to detect the binding protein in these organs is quite interesting when one considers that to our knowledge muscle tissue is apparently unaffected by vitamin A deficiency [15]. The retinol binding c o m p o n e n t could not be detected in several tumors or
95 t u m o r cell lines. At present it is not possible to ascribe any functional significance to the apparent loss of the 2 S binding protein in the tumors. Nevertheless these findings might be of interest as there have been numerous reports indicating that the administration of excess vitamin A reduces the occurrence of dimethylbenz(a)anthracene or benzo(a)pyrene induced tumors of certain epithelial tissues [ 16--20]. The fetal requirement for vitamin A becomes critical around day 14 of gestation [3]. The cellular binding protein was detected in 12-day-old fetuses suggesting that its appearance precedes the critical period of vitamin A action. The fetal retinol binding protein could also be distinguished from [3 HI retinol binding in amniotic fluid. Fetal cytosols were prepared by homogenizing the whole fetuses. Possible developmental changes in different organs have not yet been determined. Retmol is transported in the serum of several vertebrate species b o u n d to a specific transport protein, which has come to be called "retinol-binding protein" or RBP [13,21]. This protein has a reported molecular weight of 20 000--22 000, S: 0,w of 2.06 S depending u p o n the species from which it is isolated, and the m e t h o d of determining the molecular weight. The ctrculating form of retinol-binding protein is as a 1 : 1 complex with thyroxine-binding prealbumin. This complex of retinol-bindmg protein and prealbumin (a protein of approximately 56 000 molecular weight) has a sedimentation coefficient of approximately 4.6 S, the same as has been observed for in vitro [3 H] retmol binding in serum m the experiments reported here. That the addition of anti-rat serum albumin irnmunoglobulins to the serum was able to abolish binding of tritiated retinol m the 4.6 S region strongly suggests that in vitro binding of [3H]retinol m serum is due to an interaction between [3 H] retinol and serum albumin rather than with serum retinol-binding protein. Since albumin is k n o w n to bind various long-chain fatty acids [22] it is not difficult to imagine that retinol might also be bound. However, since there is a specific transport protein for retinol in serum, the in vitro binding which we observe to albumin is probably w i t h o u t physiological significance. It also seems unlikely that the m vitro binding of [3 H] retmol in tissue cytosols as described here is due to labelling of serum retinol-bmding protein present because of serum contamination. No 2 S peak is observed when serum is labelled with [3 H] retinol in vitro. That the 2 S peak arises by means of dissociation of the serum retmol-binding protein-prealbumin complex d u n n g preparation of cytosols to give free serum retinol-bmding protein is unlikely because testis has apparently lower serum contamination than liver but a higher level of binder [13]. Muscle and t u m o r extracts which also have serum contamination, show no evidence for binder at all. The failure of anti-rat serum retinol-binding protein antiserum to affect the binding of [3 H] retinol in testis provides further evidence that the tissue binding c o m p o n e n t is distinct, at least immunologically, from serum retinolbinding protein. Moreover quite recently it was observed that the fluorescence excitation spectrum of the partially purified cellular retinol binding protein from rat testes, differs from that described for serum retinol binding protein [13,14]. We therefore suggest to call the 2 S binding protein, cellular retinol binding protein.
96 Acknowledgements This work was supported by U.S.P.H.S. grants AM-05441, HD-05384, HL-15341 and HL-14214. It has been presented by M.M. Bashor in partial fulfillment of the degree of Doctor of Philosophy, Vanderbilt University, Nashville, Tennessee. M.M. Bashor was a USPHS predoctoral trainee. The authors are indebted to Dr D.E. Ong for his help in preparation o f this manuscript. References 1 Bashor, M.M., Toft, n . o . and Chytfl, F. (1973) Proc. Natl. Acad. Sci. U.S. 70, 3483--3487 2 0 n g , D.E. and Chytfl, F. (1974) Blochem. Blophys.Res. Commun. 59, 221--229 3 Thompson, J.N,, Howell, J. McC. and Pitt, G.A.J. (1964) Proc. R. Soc. 159, 51{)--535 4 Raz, A. and Goodman, D.S. (1969) J. Biol. Chem. 244, 3230--3237 5 Peterson, P.A. (1971) J. Biol. Chem. 246, 44--49 6 Cohn, E.J., Strong, L.E., Hughes, W.L., Mulford, D.J., Mehn, M. and Taylor, H.L. (1946) J. Am. Chem. Soc. 68, 4 59--475 7 Debro, J.R., Taxwer, H. and Korner, A. (1957) J. Lab. Chn. Med. 50, 728--732 8 Smith, D.F., Nevi, G. and Wolborg, Jr, E.F. (1973) Biochemistry 12, 2111--2118 9 Buonasmssi, V., Sato, G. and Cohen, A.L. (1962) Proc. Natl. Acad. ScL U.S. 48, 1184--1190 10 Stelos, P.V. (1967) H a n d b o o k of Expertmental I m m u n o l o g y (Weir, D.M., ed.), pp. 3--18, Davts Company, Philadelphia 11 Ouchterlony, O. (1967) m H a n d b o o k of Experimental I m m u n o l o g y (Weir, D.M., ed.), pp. 655--706, Davis Company, Philadelphia 12 Lowry, O.H., Rosebrough, N.J., Farr, A.L. and Randall, R.J. (1951) J. Biol. Chem. 1 9 3 , 2 6 5 - - 2 7 5 13 Muto, Y. and Goodman, D.S. (1972) J. Biol. Chem. 247, 2533--2541 14 Ong, D.E. and Chytil, F. (1975) J. Biol. Chem. in press 15 Moore, T. (1967) m The Vitamins (Sebrell, Jr, W.H. and Harris, R.S., eds), Vol. I, pp. 245--266, Academic Press, New York 16 Chy, E.W. and Malmgren, R.A. (1965) Cancer Res. 25, 884 --895 17 McMichael, H. (1965) Cancer Res. 25, 947--955 18 Davies, R.E. (1967) Cancer Res. 27, 237--241 19 Saffiottl, V., Montesano, R., Sella Kumar, A.R. and Borg, S.A. (1967) Cancer Res. 20, 857--864 20 Crocker, T.T. and Sanders, L.L. (1970) Cancer Res. 30, 1312--1318 21 Glover, J. (1973) Vltam. Horm. 31, 1--42 22 Goodman, D.S. (1958) J. Am. Chem. Soc. 80, 3 8 9 2 - - 3 8 9 8
Biochimica et Biophysica Acta, 4 1 1 ( 1 9 7 5 ) 8 7 - - 9 6 © E l s e v i e r S c i e n t i f i c P u b l i s h i n g C o m p a n y , A m s t e r d a m - - P r i n t e d in T h e N e t h e r l a n d s -
BBA 27748
C E L L U L A R RETINOL-BINDING PROTEIN
M A R K M. B A S H O R *
and FRANK CHYTIL**
Departments of Biochemistry and Medicine, Vanderbilt University, School of Medicine Nashville, Tenn. 37232 (U.S.A.) (Received April 10th, 1975)
Summary 1. A protein which binds retinol in vitro with high affinity and specificity was detected by sucrose gradient centrifugation or by gel filtration after preincubating rat tissue cytosols with all-trans-[ 3 H] retinol. This protein sediments in the 2 S region of sucrose gradients. Molecular size determination b y gel filtration indicates a molecular weight of 16 000. 2. Competition studies revealed that only all-trans-retinol, n o t retinal or retinoic acid, competes for binding. The binding of radioactive retinol is reversible. 3. This protein was detected in cytosols of rat liver, lung, spleen, brain, testis, ovaries, uterus and intestinal mucosa whereas heart or gastrocnemius muscle seem to lack this protein. 4. The cellular retinol binding protein was found in fetuses as early as day 12 of the gestation period and possessed the same specifity for the ligand as the one in adult tissues. 5. This binding c o m p o n e n t was not detected in cytosols prepared from Novikoff hepatoma, ascites hepatoma AS-30D, mouse Ehrlich ascites t u m o r and mouse pituitary t u m o r cell line AtT 20. 6. The cellular retinol binding protein seems to be different from that described to be present in the serum as suggested by difference in size and b y the inability of the antisera against the serum retinol binding protein to remove the cellular bindh~g protein from the cytosol preparations.
Introduction In a recent report we presented preliminary evidence that cytosols prepared by high speed centrifugation of several rat tissue homogenates (lung, * Present address: Department of Biology, Umverslty of California San Diego, La Jolla, Cahf., U.S.A. *~ To whom rep~lnt requests sho,lld be addressed.
88 kidney, intestinal mucosa, liver and testis) contained a protein capable of binding [3 H]retinol in vitro with high specifity [1]. This protein has an approximate molecular weight of 16 000 as estimated by gel filtration and shows a sedimentation coefficient of 2 S on sucrose gradients. In h u m a n [1] as well as in rabbit lung [2] proteins of comparable properties have been detected. Furthermore a partial purification of the retinol binding protein from rabbit lung has been achieved in this laboratory [2]. Heptane extracts o f this material have a fluorescence spectrum identical to that of authentic all-trans-retinol in heptane suggesting that retinol was in fact the naturally occurring ligand [2]. The in vitro binding of [3H] retinol in tissue cytosols could be easily distinguished from binding to a serum c o m p o n e n t which sedimented in the 4.6 S region of sucrose gradients and which had a molecular weight of approximately 67 000 [1]. The work to be reported here was undertaken to extend the above studies with regard to the following points of interest. First, a number of normal and neoplastic tissues were surveyed to determine the distribution of the cellular retinol binding protein. Second, fetal rats at various stages of development were studied in order to see whether the appearance of this protein corresponds to the time when the requirement for vitamin A becomes critical [3]. Third, the components in tissue, serum and amniotic fluid which bind retinol in vitro were further characterized in order to distinguish them from the serum transport protein complex [4,5] for vitamin A. Materials and Methods
Matertals. All-trans-[15-3H]retinol (1.25 Ci/mmol) was purchased from New England Nuclear Corporation. Unlabelled retinol, retinal, retinoic acid, retinyl acetate and palmitate (all in the all-trans-configuration) were purchased from Sigma. Rat serum albumin, Fraction V [6] was also from Sigma or prepared by the m e t h o d of Debro and coworkers [7]. Prior to use lyophylized preparations were dissolved in 0.05 M Tris • HC1 buffer, pH 7.5 to the desired protein concentration. Anti-rat serum albumin (rabbit-immunoglobulin) was obtained from Nutritional Biochemicals Corporation or Cappel Laboratories. Allyl alcohol, crotyl alcohol and ~-ionone were products of Eastman Kodak. Anti-rat serum retinol binding protein (sheep) was a generous gift of Dr D.S. Goodman. Specimens of Novikoff Hepatoma, mouse Ehrlich ascites cells and ascites h e p a t o m a AS-30D [8] were kindly provided by Dr L.S. Hnilica and the AtT-20 cells [9] by Dr R. Harrison. Animals. Adult Sprague-Dawley rats weighing 150--250 g (Holtzman, Madison, or Harlan Industries, Indianapolis) were fed Lab Chow (Ralston, Purina Co.) ad libitum. Fetuses were obtained at days 12, 14, and 16 of gestation. Female rats were left overnight with a male and the first day o f pregnancy was determined by a microscopic examination of a vaginal smear for the presence of sperm. The fetuses were delivered by Caesarean section. The animals were killed by decapitation. Preparation of cytosols. Tissues were homogenized in four volumes (w/v) of 0.05 M Tris • HC1 buffer, pH 7.5 in a glass-Teflon homogenizer. The homogenates were then centrifuged at 31 000 × g for 10 min and the resulting
89 supernatant centrifuged again at 105 000 × g for 60 min at 4°C. The highspeed supernatants (cytosols) were used immediately or after storage at --20 ° C. Preparation of serum and amniotic fluid. Blood was collected by cardiac puncture of animals anesthetized with ether and allowed to clot at 0°C for 15--20 min. Serum was then obtained as described earlier [1]. Amniotic fluid was collected from the amniotic sac by aspiration and clarified, when necessary, by centrifugation at 314300 × g for 10 min in a refrigerated centrifuge. Incubation with [3 H] retinol. In principle, conditions used were those described earlier [1]. Briefly, 1 ml aliquots of the appropriate samples were incubated in the dark at 4°C for 12--16 h with 4 . 1 0 - 8 M [3H]retinol. The isotope was added in 5 #l of ethanol. Unlabeled all-trans-retinol, retinal and retinoic acid were used for binding competition studies. The unlabelled competitors were added to the incubation mixtures in 5 pl of ethanol to a final concentration of 8 • 10 -6 M (200-fold excess over the labelled retinol). Control samples received ethanol alone. Sucrose gradient centrifugation. Binding of [3 H] retinol was determined by sucrose gradient centrifugation as previously described [1] with the exception t h a t linear 5--20% sucrose (w/v) gradients were prepared in 0.05 M Tris • HC1 buffer, pH 7.5. Determination of [3 H] retinol binding by gel filtration. In some experiments the a m o u n t of bound [3 HI retinol in serum or tissue cytosols was determined as follows. An aliquot (0.1 ml) of the incubation mixture was applied to a column (1.1 X 5.5 cm) of Sephadex G-25 (Fine) equilibrated with 0.05 M Tris • HCI buffer, pH 7.5. The sample was eluted with the same buffer at a flow rate o f about 25 ml/h; 0.5 ml fractions were collected. The fractions were then transferred to counting vials, mixed with 5 ml of the scintillation cocktail [ 1 ], and counted. [3 H] Retinol eluting from the columns in the void volume was considered " b o u n d " retinol and was clearly separated from the free [3 H] retinol appearing in later fractions. Molecular weight determination. The molecular weight of the [3 H] retinol binding components in serum and cytosols from various tissues was estimated either from the sedimentation data or by gel filtration as described previously
[1]. Immunological procedures. An immunoglobulin fraction was isolated from control or immune sera by repeated a m m o n i u m sulfate precipitation
[10]. Immunodiffusion experiments were performed by Ouchterlony doublediffusion technique [ 11] using commercially prepared plates (Hyland, Travenol Laboratories). Immunoelectrophoresis was performed on cellulose acetate strips equilibrated with 0.0275 M barbiturate buffer pH 8.6 using a Beckman Model R 101 microzone apparatus. Electrophoresis was carried o u t for 40 min at 150 V. Treatment of serum with antibody. Rat serum was diluted 80-fold in 0.05 M Tris • HCI buffer pH 7.5 prior to incubation with anti-rat serum albumin immunoglobulin. Aliquots of the diluted serum were then incubated with 200 or 400 ~1 of anti-albumin or control immunoglobulins at 37°C for 60 min and then overnight at 4°C. The samples were then centrifuged at 105 000 × g for 60 min and 1.0 ml of each supernatant was incubated with 4 • 10 -8 M
90 [3 H] retinol overnight. Binding of [3 H] retinol was determined by sucrose gradient centrifugation. Treatment o f cytosols with antibody. A similar procedure was used when studying the effect of anti-rat retinol binding protein serum and anti-rat albumin immunoglobuhn on the cellular retinol binding protein. Aliquots of testis cytosol (1.0 ml) previously labelled with 4 • 10 -8 M [3 H] retinol were incubated with 200, 400, 600 or 800 pl of antiserum or control sheep serum for 1 h at 37°C followed by overnight incubation at 4°C. Samples were then assayed for [3 H] retinol binding by sucrose gradient centrifugatlon as described above. Protein assay. Protein was determined by the m e t h o d of Lowry et al. [12]. Results
T~me course o f [3 H] retmol binding m cytosols. In order to determine the optimum time for incubation with tissue cytosols, a sample of liver cytosol (15 mg protein/ml) was mixed with [3 H] retinol, 4 • 10 -8 M. At various times thereafter a 0.1 ml aliquot was removed and the a m o u n t of b o u n d [3 H]retinol determined by gel hltration; The results of this experiment are presented in Fig. 1. Binding was essentially completed in 4 h and remained unchanged for at least 20 h thereafter. In the following experiments an incubation time o f 12-16 h was used. Reversibility o f retinol binding. It was desirable to know if the binding of [3 H] retinol to the cellular retinol binding protein was reversible. This was done by incubating testis cytosol (usually 7--8 mg protein/ml) with [3 H] retinol to achieve m a x i m u m binding, followed by the addition of a 200-fold excess of
24
I0
o_ x
16
0 x
8
12
n (J
6
(.9
4
8
2
4 I
4
I
I
I
I
I
I
I
I
8 12 16 2 0 2 4 HOURS INCUBATION
I
2
5
HOURS
i
4
I
5
INCUBATION
F i g . 1. E f f e c t o f i n c u b a t i o n t i m e o n [ 3 H ] r e t m o l b i n d i n g . L : v e r c y t o s o l w a s i n c u b a t e d a t 4 ° C f o r v a r i o u s p e r i o d s o f t u n e f o l l o w i n g t h e a d d i t i o n o f 4 • 1 0 - 8 M [ 3 H ] r e t m o l . T h e b i n d i n g w a s a n a l y z e d a t e a c h tLrne p o i n t b y gel f i l t r a t i o n . F i g . 2. D i s p l a c e m e n t of bound [3H]retmol m testis cytosol foUowmg the add~tlon of excess uniabelled retmol. Testm eytosol was incubated overnight with 4.10 -8 M [3H] retmol. Then a 200-fold excess of uniabelled retmol was added. Ahquots of the mixture were removed at various times thereafter and assayed f o r [ 3 H ] r e t m o l b i n d i n g b y gel f i l t r a t i o n .
91 unlabelled retinol. As can be seen in Fig. 2, b o u n d [3 H] retinol could be displaced to some extent following the addition of unlabelled retinol. The binding of retinol to the cellular retinol binding protein is apparently n o t covalent as shown recently [ 2 ] . Molecular size of the retinol binding component. The apparent molecular weight of the cellular binding protein from a variety of tissue was estimated on a calibrated Sephadex G-100 column. In all cases the cellular binding protein eluted after a myoglobin standard, giving an approximate molecular size of 16 000. Purified serum retinol-binding protein has been reported to elute before myoglobin on a Sephadex G-100 column, giving an approximate molecular size of 19 000(13). A refinement of apparent molecular weight determined by gel filtration has been reported for cellular retinol binding protein purified 1000 fold from rat testes [14]. Using a standardized Sephadex G-75 column the partially purified natural elutes rear ribonuclease A, giving an even lower apparent molecular weight of 14 000. Tissue distribution of the binding component. We have previously described the binding of [3 H] retinol to a 2 S protein in cytosols from rat testis, liver, lung, kidney and intestinal mucosa [1]. The tissues examined in the present study, ovary, uterus, spleen and brain also contain the retinol binding protein as shown in a typical experiment in Fig. 3. On the other hand when cytosols from heart muscle ~ and gastrocnemius muscle were incubated with [3 H] retinol no 2 S peak was detected on sucrose gradient, as can be seen from Fig. 3. These tissues apparently do not contain this 2 S protein or the amounts present are below the sensitivity of the assay. The muscle cytosols were prepared in the same manner as the other tissue cytosols (homogenized in 4 volumes of buffer). Reducing the volume of buffer to 2 yielded the same results, suggesting that the retinol binding c o m p o n e n t was in fact absent. When labelled retinol was applied on sucrose gradients in the absence of any protein in 0.05 M Tris • HC1 buffer (pH 7.5) only, the radioactivity showed pattern n o t distinguishable from that shown for muscle in Fig. 3. [3 H] Retinol binding in tumor cytosols. Several tumors have also been examined for retinol binding component. No binding was observed in Novikoff Hepatoma. Likewise no 2 S binding was seen in a mouse pituitary t u m o r cell line (ART-20 cells) maintained in tissue culture, mouse Ehrlich ascites cell or a hepatoma strain AS-30D maintained as an ascites tumor. On the other hand [3 H] retinol binding in normal liver cytosol was previously observed [1]. Whether the binding c o m p o n e n t does exist in normal mouse pituitary remains to be shown. Ontogeny of the retinol-binding component. In light of the evidence concerning the requirement of vitamin A for fetal development [3] the fetal cytosols were examined at various stages of development for the presence of the 2 S binding component. The 2 S binding protein for [3 H] retinol was detected in the earliest stage tested, day 12 of gestation. No 2 S binding was observed in amniotic fluid preparations from days 12, 14 or 16 of gestation, although binding in the 4.6 S region of the gradient was found, as will be discussed below. Ligand specificity of the retinol binding component. The previous work has shown that binding of [3 HI retinol in cytosols from adult rat testis [1] or
92
i
:7 0 I'£9
z
o
I'-£.9
I0
LL
A
OVARY
6
IX LL
b
t~
x
x
I:L (D
I:L {.9 '~ 4 8 12 16 TOP FRACTION NUMBER
i
i
i
A
12
z - + RETINAL-
8
L+ RETINOIC
,~ 4
8
12
16
4
8
12
16
/
TOP
TOP FRACTION
F i g . 3. S u c r o s e g r a d i e n t c e n t r l f u g a t l o n of cytosols from ovary and muscle tion with 4 " 10 -8 M [3H] retmol. For details see Materials and Methods.
NUMBER
(gastrocnemlus)
after incuba-
F i g . 4. L l g a n d s p e c l f l t y o f c e l l u l a r r e t m o l b i n d i n g p r o t e i n m 1 6 - d a y - o l d fetuses Fetal cytosols were p r e p a r e d a n d i n c u b a t e d o v e r n i g h t w i t h 4 • 1 0 - 8 M [ 3 H ] r e t m o l in t h e a b s e n c e ( ~ ) a n d p r e s e n c e o f 2 0 0 - f o l d m o l a r e x c e s s o f A r e t m o l ( / ) , B r e t m o l ( u ) o r r e t m o l c a c i d (z~). S u c r o s e g r a d i e n t c e n t n f u g a t m n was p e r f o r m e d as d e s c r i b e d m M a t e r i a l s a n d M e t h o d s .
adult rabbit lung [2] show a high specifity. Of the compounds tested only all-trans-retinol and not retinoic acid added in a 200-fold excess would compete with the radioactive ligand. In most preparations retinal did not compete either. Similar specifity was observed in cytosols prepared from fetal rat at the day 16 of gestation as shown in Fig. 4 suggesting that the fetal tissue binding protein is not different from that detected in the adult animal. The binding specificity was further studied in adult rat testis cytosols using a 200-fold excess of retinyl acetate of palmitate. Both compounds competed as effective as retinol. The peak of radioactivity observed in the presence of an excess of nonradioactive retinol represents non specific binding of [3 H] retinol displaced by the competitor to other macromolecules present in the cytosols such as albumin. It was also of interest to test other compounds which do not have vitamin A activity, but which resemble portions of the retiaol molecule, for their ability to compete for tritiated retinol binding. Allyl alcohol, crotyl alcohol or ~-ionone added at 8 • 10 -6 M concentration were unable to compete. Bmding o f [3H] retmol to serum and amniotic fluid components. Serum and amniotic fluid show no evidence for 2 S binding component but both contain a binding component(s) for [3 H] retinol sedimenting at 4.6 S, the same position as observed for [3 H] retinol bound to authentic rat serum albumin (Fig. 5, testis cytosol included for comparison). If diluted serum was preincubated with rabbit anti-rat serum albumin immunoglobulin fraction, the 4.6 S binding peak for [3 H] retinol was abolished (Fig. 6). No antibodies directed against rat serum prealbumin were found using immunoelectrophoresis in the above anti-rat serum albumin immunoglobulins. These results suggest strongly that serum albumin is responsible for the in vitro binding of [3 H] retinol by serum. Immunodiffusion and immunoelectrophoresis of rat amniotic fluid and tissue cytosols against anti-rat serum albumin antiserum indicated the presence
93
TESTIS
~-AMNIOTIC FLUID
CYTOSOL 14 z o F--
~-,.,~ERUM ( I:lO} ~'~--ALBUMI N
U_
'o ×
6
z O t(D
•R
6
c3
n~ t~
13_ o
4
o_ x ,
4
8
12
16 20
FRACTION NUMBER TOP
D_ {.9
2
i
i
i
i
T~Op4 8 12 16 FRACTION NUMBER
F i g . 5. S u c r o s e gradient profiles o f [ 3 H ] r e t i n o l b i n d i n g b y s e r u m , ananiotic fluid, s e r u m a l b u m i n and testis c y t o s o l . R a t s e r u m ( d i l u t e d 1 : 1 0 w i t h 0 . 0 5 M T r i s • H C1 b u f f e r , p H 7 . 5 ) a m m o t i c fluid, s e r u m a l b u m i n (4 m g per m l o f the a b o v e b u f f e r ) w e r e i n c u b a t e d w i t h t n t i a t e d r e t i n o l and a n a l y z e d b y s u c r o s e grachent c e n t r i f u g a t i o n . F i g . 6 . S u c r o s e gradient profiles of [ 3 H ] r e t l n o l b i n d i n g in s e r u m t r e a t e d w i t h c o n t r o l or anti-rat s e r u m a l b u m i n i m m u n o g l o b u h n s . R a t s e r u m w a s d i l u t e d 8 0 - f o l d w i t h 0 . 0 5 M T n s • H C I b u f f e r , p H 7 . 5 prior t o t r e a t m e n t w i t h rabbit c o n t r o l n n m u n o g l o b u l i n s ( o ) or anti-rat s e r u m a l b u m i n i m m u n o g l o b u h n s (~). T h e s a m p l e s w e r e t h e n a n a l y z e d for [ 3 H ] r e t i n o l b y s u c r o s e g r a d m n t c e n t r i f u g a t i o n .
of albumin, suggesting the 4.6 S peak observed in amniotic fluid and occasionally in tissue cytosols also represents binding to albumin, due to serum contamination. If additional serum is added intentionally to testis cytosol no difference is observed in the 2 S peak, suggesting that the cellular retinol binding protein does not form complexes with serum components as the serum retinol binding protein does with prealbumin. Effect o f anti-rat serum retinol binding protein antiserum on [3H] retinol binding in testis cytosol. To examine the possibility that the binding of [3 H] retinol in tissue cytosols by a c o m p o n e n t sedimenting at 2 S was due to contaminating serum retinol binding protein described earlier [ 4 , 5 ] , testis cytosol was incubated with monospecific antiserum to rat serum retinol binding prorein. The results presented in Fig. 7 show that under conditions employed, which are similar to those employed for removing the serum albumin, antiserum against retinol binding protein does not affect the amount of [3 H]retinol bound in testis cytosol or the position of the peak on sucrose gradient to which the bound radioactive retinol sediments. While Fig. 7 presents the results obtained with 800 pl of antiserum, similar results were obtained when 200, 400, and 600 pl of antiserum was employed. The antiserum gave a positive double diffusion and immunoelectrophoretic tests with rat serum, but no precipitin bands were observed when testis cytosol was tested. Preincubation o f testis cytosol with rabbit anti-rat albumin immunoglobulins failed to remove the cellular retinol binding protein. The shoulder of radioactivity in the above figure originated probably by binding of radioactive retinol to albumin of the added antiserum. These results indicate that the cellular retinol binding protein is antigenically distinct from that found in the serum.
94 i
z
12
i
f
2s
I..ha:
8
b x
4
a_ i
T~P 4
i
i
i
8
[2
16
FRACTION NUMBER Fig.
7. Sucrose
(~') a n d w i t h Methods.
gradients
ant:-serum
profiles
of
[3H]
retlnol
(;~) a g a i n s t r a t s e r u m
binding
retmol
m
testis
binding
cytosols
pretreated
protein
F o r d e t a i l s see M a t e r i a l s a n d
with
control
serum
Discussion The studies presented here indicate that numerous rat organs contain a macromolecule capable of binding [-~H] retinol in vitro with high speclfmity. The binding in tissue cytosols occurs fairly rapidly, with maximum occurring within four hours. That maximum binding does not occur sooner may be due to the presence of endogenous retinol bound to the tissue c o m p o n e n t which the added radioactive retinol must displace. The binding of retinol is apparently reversible and not covalent, as was demonstrated b y the ability of unlabelled retinol to displace previously b o u n d [3 H] retinol. Furthermore it was recently demonstrated that rabbit lung contains a retinol binding protein that is probably analogous to that described here for rat tissues [2]. The fluorescence spectrum of a heptane extract of this binding c o m p o n e n t was identical to all-trans-retinol, indicating that the binding c o m p o n e n t carries all-trans-retinol m vivo and that the binding is non-covalent. Partially purified rat testicular retinol binding protein also contains all-trans-retinol [ 14]. The binding specifity established by competitive assays :s remarkable. In cytosols retinoic acid never h a s b e e n found to compete, whereas retinal appeared to partially c o m p e t e in some preparations for [3 H] retinol binding [ 1]. However this may reflect the in vitro reduction of retinal to retinol, with the retinol thus formed being the competing ligand. Similarly the fact that retmyl acetate and retinyl palmitate c o m p e t e as effectively as retinol for [a H] retinol binding in testis cytosol may result from cleavage of these esters to yield retinol. In partially purified preparations retinal [2] or retinyl esters [14] do not compete. Furthermore the inability of other c o m p o u n d s representing portions of the retinol molecule to c o m p e t e for [3 H] retinol binding again emphasizes the highly specific nature of the binding site. The tissue retinol binding protein shows a rather wide distribution in different tissues. Heart and gastrocnemius muscle are the only negative tissues found at this time. The failure to detect the binding protein in these organs is quite interesting when one considers that to our knowledge muscle tissue is apparently unaffected by vitamin A deficiency [15]. The retinol binding c o m p o n e n t could not be detected in several tumors or
95 t u m o r cell lines. At present it is not possible to ascribe any functional significance to the apparent loss of the 2 S binding protein in the tumors. Nevertheless these findings might be of interest as there have been numerous reports indicating that the administration of excess vitamin A reduces the occurrence of dimethylbenz(a)anthracene or benzo(a)pyrene induced tumors of certain epithelial tissues [ 16--20]. The fetal requirement for vitamin A becomes critical around day 14 of gestation [3]. The cellular binding protein was detected in 12-day-old fetuses suggesting that its appearance precedes the critical period of vitamin A action. The fetal retinol binding protein could also be distinguished from [3 HI retinol binding in amniotic fluid. Fetal cytosols were prepared by homogenizing the whole fetuses. Possible developmental changes in different organs have not yet been determined. Retmol is transported in the serum of several vertebrate species b o u n d to a specific transport protein, which has come to be called "retinol-binding protein" or RBP [13,21]. This protein has a reported molecular weight of 20 000--22 000, S: 0,w of 2.06 S depending u p o n the species from which it is isolated, and the m e t h o d of determining the molecular weight. The ctrculating form of retinol-binding protein is as a 1 : 1 complex with thyroxine-binding prealbumin. This complex of retinol-bindmg protein and prealbumin (a protein of approximately 56 000 molecular weight) has a sedimentation coefficient of approximately 4.6 S, the same as has been observed for in vitro [3 H] retmol binding in serum m the experiments reported here. That the addition of anti-rat serum albumin irnmunoglobulins to the serum was able to abolish binding of tritiated retinol m the 4.6 S region strongly suggests that in vitro binding of [3H]retinol m serum is due to an interaction between [3 H] retinol and serum albumin rather than with serum retinol-binding protein. Since albumin is k n o w n to bind various long-chain fatty acids [22] it is not difficult to imagine that retinol might also be bound. However, since there is a specific transport protein for retinol in serum, the in vitro binding which we observe to albumin is probably w i t h o u t physiological significance. It also seems unlikely that the m vitro binding of [3 H] retmol in tissue cytosols as described here is due to labelling of serum retinol-bmding protein present because of serum contamination. No 2 S peak is observed when serum is labelled with [3 H] retinol in vitro. That the 2 S peak arises by means of dissociation of the serum retmol-binding protein-prealbumin complex d u n n g preparation of cytosols to give free serum retinol-bmding protein is unlikely because testis has apparently lower serum contamination than liver but a higher level of binder [13]. Muscle and t u m o r extracts which also have serum contamination, show no evidence for binder at all. The failure of anti-rat serum retinol-binding protein antiserum to affect the binding of [3 H] retinol in testis provides further evidence that the tissue binding c o m p o n e n t is distinct, at least immunologically, from serum retinolbinding protein. Moreover quite recently it was observed that the fluorescence excitation spectrum of the partially purified cellular retinol binding protein from rat testes, differs from that described for serum retinol binding protein [13,14]. We therefore suggest to call the 2 S binding protein, cellular retinol binding protein.
96 Acknowledgements This work was supported by U.S.P.H.S. grants AM-05441, HD-05384, HL-15341 and HL-14214. It has been presented by M.M. Bashor in partial fulfillment of the degree of Doctor of Philosophy, Vanderbilt University, Nashville, Tennessee. M.M. Bashor was a USPHS predoctoral trainee. The authors are indebted to Dr D.E. Ong for his help in preparation o f this manuscript. References 1 Bashor, M.M., Toft, n . o . and Chytfl, F. (1973) Proc. Natl. Acad. Sci. U.S. 70, 3483--3487 2 0 n g , D.E. and Chytfl, F. (1974) Blochem. Blophys.Res. Commun. 59, 221--229 3 Thompson, J.N,, Howell, J. McC. and Pitt, G.A.J. (1964) Proc. R. Soc. 159, 51{)--535 4 Raz, A. and Goodman, D.S. (1969) J. Biol. Chem. 244, 3230--3237 5 Peterson, P.A. (1971) J. Biol. Chem. 246, 44--49 6 Cohn, E.J., Strong, L.E., Hughes, W.L., Mulford, D.J., Mehn, M. and Taylor, H.L. (1946) J. Am. Chem. Soc. 68, 4 59--475 7 Debro, J.R., Taxwer, H. and Korner, A. (1957) J. Lab. Chn. Med. 50, 728--732 8 Smith, D.F., Nevi, G. and Wolborg, Jr, E.F. (1973) Biochemistry 12, 2111--2118 9 Buonasmssi, V., Sato, G. and Cohen, A.L. (1962) Proc. Natl. Acad. ScL U.S. 48, 1184--1190 10 Stelos, P.V. (1967) H a n d b o o k of Expertmental I m m u n o l o g y (Weir, D.M., ed.), pp. 3--18, Davts Company, Philadelphia 11 Ouchterlony, O. (1967) m H a n d b o o k of Experimental I m m u n o l o g y (Weir, D.M., ed.), pp. 655--706, Davis Company, Philadelphia 12 Lowry, O.H., Rosebrough, N.J., Farr, A.L. and Randall, R.J. (1951) J. Biol. Chem. 1 9 3 , 2 6 5 - - 2 7 5 13 Muto, Y. and Goodman, D.S. (1972) J. Biol. Chem. 247, 2533--2541 14 Ong, D.E. and Chytil, F. (1975) J. Biol. Chem. in press 15 Moore, T. (1967) m The Vitamins (Sebrell, Jr, W.H. and Harris, R.S., eds), Vol. I, pp. 245--266, Academic Press, New York 16 Chy, E.W. and Malmgren, R.A. (1965) Cancer Res. 25, 884 --895 17 McMichael, H. (1965) Cancer Res. 25, 947--955 18 Davies, R.E. (1967) Cancer Res. 27, 237--241 19 Saffiottl, V., Montesano, R., Sella Kumar, A.R. and Borg, S.A. (1967) Cancer Res. 20, 857--864 20 Crocker, T.T. and Sanders, L.L. (1970) Cancer Res. 30, 1312--1318 21 Glover, J. (1973) Vltam. Horm. 31, 1--42 22 Goodman, D.S. (1958) J. Am. Chem. Soc. 80, 3 8 9 2 - - 3 8 9 8
