Interactions of all-trans, 9-, 11-, and 13-cis-retinal, all-trans-retinyl acetate, and retinoic acid with human retinol-binding protein and prealbumin.

Abstract The reconstitution of analogs of the native chromophore retinol (vitamin A) with apo-retinol-binding protein (apo-RBP) from human plasma was studied by chromatographic and spectroscopic techniques. All-trans- and 9-, 11-, and 13-cis-retinal combined with apo-RBP in a 1:1 molar ratio while only about 0.85 mole of all-trans-retinoic acid and 0.75 mole of retinyl acetate combined with 1.0 mole of apo-RBP. The various chromophores all bound to the same site on retinol-binding protein and were competitive inhibitors of each other's binding. Upon reconstitution the absorption peak of the chromophore shifted some 9 to 12 nm to the red with the retinal chromophores, 5 nm to the red with retinyl acetate, and 15 nm to the blue with retinoic acid (at pH 9). All the chromophore-RBP complexes showed an induced (extrinsic) Cotton effect of the chromophore absorption band with a rotatory strength of the same order of magnitude as that of the retinol isomers-RBP. Similar to the behavior of the free chromophore in aqueous detergent solutions, the absorption and CD spectra of retinoic acid-RBP were pH dependent, with both the absorption and CD chromophore peak decreasing in magnitude and shifting to the red upon a decrease in pH. The reconstitution of apo-RBP with the various chromophores led to a marked (more than 90%) quenching of the fluorescence of the inherent protein chromophores. Increasing the ionic strength of a retinal-RBP complex led to a shift of the absorption to the blue (7 nm with the 9-cis-retinal), a small decrease in the size of the absorption band, and marked decrease (up to 39% with 11-cis-retinal) of the rotatory strength. Increasing the ionic strength of retinoic acid-RBP solution led to a shift of the absorption peak to the red of about 1 nm, a 2.4% increase in the absorption band, and an increase of about 1% in the rotatory strength. The same change in ionic strength had no effect on the absorption of the free chromophore in an aqueous nonionic detergent solution. Reduction of retinal-RBP complexes with NaBH4 showed that the retinal aldehyde function was free and was not covalently linked to the protein. This was also shown by illuminating frozen solutions of retinal-RBP complexes with linearly polarized light (photoselection) and measuring the resulting linear dichroism spectrum. The linear dichroism spectrum of illuminated retinal-RBP complexes was similar to that of the corresponding free retinal isomer in aqueous digitonin solution. None of the retinal isomers-RBP and retinyl acetate-RBP complexes was bound to prealbumin (thyroxine-binding protein) at physiological ionic strength, whereas retinoic acid-RBP was bound to prealbumin under the same conditions as judged by gel filtration chromatography. On the other hand, the addition of prealbumin to retinal-RBP complexes at low ionic strength resulted in a shift to the blue of the absorption peak (8 nm with 9-cis-retinal-RBP). Increasing the ionic strength of a retinal-RBP and prealbumin solution resulted in a further shift to the blue (5 nm with 9-cis-retinal-RBP) and an increase in the area of the chromophore absorption band. Adding salt to a retinoic acid-RBP solution containing prealbumin resulted in a shift of 1 nm to the red and an increase in the absorption band. It was concluded from these experiments that although the retinals, retinoic acid, and retinyl acetate bind to retinol-binding protein at the same site as the retinol isomers, the binding of these various chromophores resulted in a somewhat altered conformation of the reconstituted retinol-binding protein. This in turn made it impossible for the retinals- and retinyl acetate-RBP complexes to bind to prealbumin and led to the various subtle differences in spectroscopic behavior of the various chromophore-RBP complexes upon changes in ionic strength and interaction with prealbumin.

The reconstitution of analogs of the native chromophore retinol (vitamin A) with apo-retinol-binding protein (apo-RBP) from human plasma was studied by chromatographic and spectrosdopic techniques.
All-trans-and 9-, ll-, and 13-cis-retinal combined with apo-RBP in a 1: 1 molar ratio while only about 0.85 mole of all-trans-retinoic acid and 0.75 mole of retinyl acetate combined with 1.0 mole of apo-RBP. The various chromophores all bound to the same site on retinol-binding protein and were competitive inhibitors of each other's binding.
Upon reconstitution the absorption peak of the chromophore shifted some 9 to 12 nm to the red with the retinal chromophores, 5 nm to the red with retinyl acetate, and 15 nm to the blue with retinoic acid (at pH 9). AU the chromophore-RBP complexes showed an induced (extrinsic) Cotton effect of the chromophore absorption band with a rotatory strength of the same order of magnitude as that of the retinol isomers-RBP.
Similar to the behavior of the free chromophore in aqueous detergent solutions, the absorption and CD spectra of retinoic acid-RBP were pH dependent, with both the absorption and CD chromophore peak decreasing in magnitude and shifting to the red upon a decrease in PH. The reconstitution of apo-RBP with the various chromophores led to a marked (more than 90%) quenching of the fluorescence of the inherent protein chromophores.
Increasing the ionic strength of a retinal-RBP complex led to a shift of the absorption to the blue ('7 nm with the 9-cis-retinal), a small decrease in the size of the absorption band, and marked decrease (up to 39% with ll-cisretinal) of the rotatory strength.
Increasing the ionic strength of retinoic acid-RBP solution led to a shift of the absorption peak to the red of about 1 nm, a 2.4% increase in the absorption band, and an increase of about 1% in the rotatory strength.
The same change in ionic strength had no effect on the absorption of the free chromophore in an aqueous nonionic detergent solution.
Reduction of retinal-RBP complexes with NaBH4 showed that the retinal aldehyde function was free and was not covalently linked to the protein.
This was also shown by illuminating frozen solutions of retinal-RBP complexes with linearly polarized light (photoselection) and measuring the * This study was supported by Grants EY 00331 and EYO0704 from the National Inst,itutes of Health.
resulting linear dichroism spectrum.
The linear dichroism spectrum of illuminated retinal-RBP complexes was similar to that of the corresponding free retinal isomer in aqueous digitonin solution.
None of the retinal isomers-RBP and retinyl acetate-RBP complexes was bound to prealbumin (thyroxine-binding protein) at physiological ionic strength, whereas retinoic acid-RBP was bound to prealbumin under the same conditions as judged by gel filtration chromatography.
On the other hand, the addition of prealbumin to retinal-RBP complexes at low ionic strength resulted in a shfit to the blue of the absorption peak (8 mn with 9-cis-retinal-RBP).
Increasing the ionic strength of a retinal-RBP and prealbumin solution resulted in a further shift to the blue (5 nm with 9-cis-retinal-RBP) and an increase in the area of the chromophore absorption band.
Adding salt to a retinoic acid-RBP solution containing prealbumin resulted in a shift of I nm to the red and an increase in the absorption band.
It was concluded from these experiments that although the retinals, retinoic acid, and retinyl acetate bind to retinolbinding protein at the same site as the retinol isomers, the binding of these various chromophores resulted in a somewhat altered conformation of the reconstituted retinolbinding protein.
This in turn made it impossible for the retinals-and retinyl acetate-RBP complexes to bind to prealbumin and led to the various subtle differences in spectroscopic behavior of the various chromophore-RBP complexes upon changes in ionic strength and interaction with prealbumin.
In the preceding paper of this series (I) we described the reconstitution of apo-retinol-binding protein from human plasma with four geometric isomers of retinol (vitamin A). The alltruns-and the 9-, II-, and 13-cis-retinols give about the same degree of reconstitution with apo-RBP,l namely, 0.9 to 0.95 molecule per molecule of protein.
All occupy the same binding site, give an enhanced quantum yield of fluorescence as compared to the free retinal, show a 3-to 5-nm shift to the red of the 6318 absorption maximum, and exhibit optical activity of the chromophore absorption band. Moreover, while apo-RBP is completely dissociated from prealbumin in buffers of physiological ionic strength, the reconstituted retinol isomers-RBP are tightly bound to prealbumin under the same conditions. In addition, the retinol isomer-RBP complexes show a hyperchromic effect of the chromophore absorption band when salt is added to a low ionic strength buffer in the presence of prealbumin.
A comparison of the retinol isomers-RBP complexes has shown that the native retinal-RBP complex isolated from human plasma is the all-trans isomer (1).
Goodman and Raz (2) recently have shown that apo-RBP can also combine with retinals and retinoic acid. This paper reports on the spectroscopic properties of reconstituted retinal isomers-RBP, retinyl acetate-RBP, and retinoic acid-RBP.
We also report on the interaction of these complexes with prealbumin.
Preparation of All-tmns-ret&o&Preparation was described in the preceding paper (1).
Incubation of Retinals, Retinyl Acetate, and Retinoic Acid with Apo-RBP-Incubations were performed as described in the preceding paper for incubation of retinols with apo-RBP (1).
Binding of Reconstituted Retinal-binding Protein to Prealbumin-Binding studies were performed by gel filtration chromatogra,phy with Sephades G-100 in 0.033 M phosphate buffer, pH 7, containing 0.1 M NaCi as described in the preceding paper (1).
Spectroscopic &%.&es-Absorption, fluorescence, and circular dichroism spectra were measured as described in the preceding paper (1). Photoselection and measurement of linear dichroism of retinal isomers at 77" K was performed exactly as described in Reference 4. Protein Concentrations-Protein concentrations were determined from the absorption at 280 nm using an ~80 value of 46,000 for retinal-RBP (1, 5), 40,400 for apo-RBP (I), and 76,300 for prealbumin (6, 7). Retinals, Retinoic Acid, and Retinyl Acetate Concentrations-Concentrations were determined from the absorption at the peak of the long wave length absorption band.
Nomenclature-As in the case of retinol-RBI' complexes, the reconstituted retinals, retinyl acetate, and retinoic acid are explicitly named.
The reconstitution of all-trans.retinyl acetate with ago-RBP leads to the formation of all-trans-retinyl acetate-RBP complex.

Reconstitution of Apo-RBP
with Retinal Isomers and Retinoic Acid-Similar to the findings of Goodman and Rae (2) we also found that apo-RBP combines not only w&h retinols, but also with several retinal isomers and with retinoic acid (Table I).
Similarly to retinol isomers, all the retinal isomers combined with apo-RBP to form a 1:l molar ratio complex.
We have prepared retinal-RBP complexes several times with different preparations of apo-RBP. In all cases the percentage reconstitution was about the same as reported in Table I, t,he extreme variations being between about 85 and 100%. When the incubation of retinal with apo-RBP was performed at 4", even after 18 hours only about 50% reconstitution took place, whereas at 23" the process was complete in less than 1 hour. Retinoie acid and retinyl acetate gave a somewhat lower percentage reconstitution (Table I). It is not clear whether this lower percentage reconstitution was real or whether it was due to uncertainties in the molar absorptivity values for these compounds (see below, "Effects of pH on the Absorption and CD Spectra of Retinoic Acid-RBP Comples").
The combination of retinal isomers with apo-RBP led in all cases to a considerable red shift in the absorption peak of t,he chromophore, ranging from 9 to 12 nm (as compared to the absorption of free retinal in ethanol, Table I). Most interestingly, the recombination of all-trans-retinoic acid with retinolbinding protein shifted its absorption peak to the blue some 15 nm (at pH 9). Similarly to the retinols, the retinol ester retinyl acetate showed a 5-urn shift to the red of the chromophore absorption maximum.
Fluorescence af Retinals and Retinoic Acid-RBP Complexes-Retinal isomers are not fluorescent at room temperature.
It was not surprising therefore to find that the various retinal-RBP complexes were nonfluorescent when excitated between 330 and 450 nm and measured between 380 and 600 nm. The only indication that a complex between retinal and RRI' was formed was the marked quenching of the protein fluorescence (Table II). A control experiment showed that the addition of free retinal to retinol-RBP complex did not cause any quenching. When all- c The excitatio$wave length was 330 nm. Fro. 1. CD spectra of retinal isomers-RBP. All compounds (final concentration, 30 pM) were dissolved in 2 mM Tris-HCl, pH 9.0.
The path length was 10 mm. Each spectrum represents the average of 16 scans.
tmlzs-retinal was added to a retinal-RBP complex there was very little increase in fluorescence (Table IV,

below).
This was interpreted to mean that the retinals occupied the same site as the retinols in retinol-binding protein and that this site was already fully saturated by retinals.
Similar results were also obtained with all-trans-retinoic acid. Circular Dichroism of Retinal Isomers-,

Retinyl
Acetate-, a,nd Retinoic Acid-RBP Complexes-Whereas free retinols and retinals in solution are devoid of optical activity, retinol isomers-RlfP complexes show an extrinsic Cotton effect of the chromophore absorption band (1). Similarly, the retinal isomers-RRP complexes show a CD band at their chromophore absorption band (Fig. 1). The CD bands of the retinal-RBP complexes peak at about the same wave length as the absorption peak of the complex (compare Table I  complexes was about the same as that of the retinol-RBP complexes (Table III; compare Table  IV in Reference 1). The all-trans-retinal-RBP complex possessed the largest CD band similar to that of the all-trans isomer among retinol-REP complexes. Roth the all-truns-ret.inyl acetate-and the all-frans-ret'inoic acid-RBP showed an induced CD band, peaking at about the same wave length as the chromophore absorption band and again showing a rotatory strength similar to that of the retinoland retinal-RRP complexes (Table III). In a previous paper (3) we have shown that all-trans-retinol combines with bovine serum albumin to give a relatively stable complex that can be recognized by gel filtration chromatography and by enhancement of the retinol fluorescence quantum yield. All-lrans-retinal also combines with bovine serum albumin to give a complex with an Azgo:A395 ratio of 0.25, after gel filtration chromatography to remove escess retinal. Neither the alltrans-retinal nor the all-trans-retinol-bovine serum albumin complex showed any optical activity of the bound chromophore absorption band. At the level of sensitivity used for the CD measurements both the all-trccns-retinol-and the retinal-bovine serum albumin complexes had an induced optical activity band at least 100 times smaller than that of the corresponding retinolbinding protein complex-if they had any optical activity at all. We interpreted these findings to show that the induced optical activity of the retinols-, retinals-, and retinoic acid-RBP complexes was not merely an expression of binding, as such, but of a highly specific binding which results in an asymmetric chromophore-protein complex. ,632O Acid-RBP Complex-When free all-trans-retinoic acid was dissolved in aqueous 1% Emulphogene (a nonionic detergent) both the peak height and the peak position were pH dependent (Fig.  2%). Thus, while the peak position at pH 3.5 was at 355 nm, the peak position at pH 9.3 was at 341 nm and there was a concomitant 17% increase in peak height.
There was no change The path length was 10 mm.
in area under the chromophore absorption band (270 to 440 nm) upon changes in pH. It is clear that the ionization of the carboxyl group of retinoic acid has a marked effect on both the peak position and the peak height, the charged form having a higher peak and showing a blue shift.
When the same series of experiments was performed with an all-trans-retinoic acid-RBP complex, similar changes were observed. The chromophore peak position shifted from about 335 nm at pH 9 and pH 8 to about 355 nm at pH 3.65 (Fig. 2B).
There was a concomitant decrease in peak height, with the peak at pH 8 being 43% higher than that at pH 3.65. There was a small decrease in peak height at 280 nm in going from pH 9 to pH 8, with only minor changes upon further lowering of the pH (Fig. 2B). These changes in the chromophore absorption were partially reversible.
When the pH was brought back from 3.65 to 9, the peak position shifted back to 335 nm, but there was a loss of about 12% in intensity, probably due to destruction of the chromophore-RBP complex at acid pH (Fig.   2B, Curve 6). Supporting this interpretation was a 10% decrease in the size of the chromophore absorption band (290 to 450 nm) in going from pH 9 to pH 3.65, and a further 6% decrease when the pH was brought back to 9.0. The same reversible changes in the absorption spectrum of retinoic acid-RBP complex as a function of pH were seen in low and physiological ionic strength buffers.
Changes in pH also produced a shift in the CD spectrum of all-truns-retinoic acid-RBP. The peak position of the CD spectrum shifted from about 330 nm at pH 9 and pH 8 to about 340 nm at pH 3.65 (Fig. 3), with a concomitant 14% decrease in rotatory strength.
The changes in the CD spectrum were also reversible, with the peak position shifting back to about 330 nm in going from pH 3.65 to pH 9.0 and a further 6% decrease in rotatory strength.
Like the decrease in area under the absorption band, the decrease in rotatory strength was probably due to progressive destruction of the retinoic acid-RPB complex at acid pH. The changes in the CD spectrum of retinoic acid-RBP complex as a function of pH were the same in low a.nd physiological ionic strength buffers.

WAVELENGTH
(nm) FIG. 3. Effect of pH on the CD spectrum of retinoic acid-RBP complex.
The sample was the same as that shown in Fig. 2B. The path length was 10 mm. Each curve represents the average of eight scans.
The pH values were 1, pH 8.0 and 9.0; 2, 7.0; 3, 5.95; and 4,3.65; 5 was determined after the pH was brought back to 9.0. Possible Covalent Linkage of Retinal to Retirwl-binding Protein-Unlike retinols, which are fairly unreactive alcohols, the reactive aldehyde function of retinal has the potential to react with several reactive sites on the protein, most probably with a primary amino group.
It should be recalled in this connection that in the analogous situation of visual pigments, the retinal chromophore is covalently linked to an e-amino group of a lysine residue in rhodopsin.
Two methods were used to establish whether retinal was present as the free aldehyde or was covalently linked as a Schiff base.
Reduction of the free aldehyde with sodium borohydride produces the corresponding alcohol, while reduction of the Schiff base leads to a new covalent linkage between retinal and the amino group in the form of a stable primary amino linkage (9). Reduction of all-trans-retinal-RBP with NaBH4 produced a species that had an absorption spectrum similar to that of alltrans-retinol-RBP and which showed optical activity of the chromophore absorption band. One extraction of this reduced compound with ethanol (1) removed about 80% of the chromophore as free retinol.
This was approximately the same amount of retinol that could be extracted from native retinol-RBP by one ethanol extraction (1). The facts that the reduced alltrans-retinal in retinol-binding protein could be extracted with ethanol and that it had absorption and CD spectra similar to that of native retinol-RBP were taken as evidence that retinal is not covalently linked to retinol-binding protein and that the aldehyde function can be reduced in situ to produce the corresponding retinol-RBP complex. The second approach was to use the technique of photoselection and linear dichroism that was recently developed for identification and quantitative measurement of retinal isomers (4). The LD spectra of retinal isomers are unique and different from those of Schiff bases and other compounds of retinal (10). Retinols do not show photoselection and linear dichroism (10). Using this technique, the retinal isomers-RBP gave LD spectra similar to that of the corresponding free retinal (Fig. 4). Although the general shape of t.he LD spectrum of each retinal isomer was unique (4), making it possible to identify unambiguously each retinal isomer, the environment in which the retinals were dissolved had some effect on the details of the LD spectrum.
As can be seen from Fig. 4, the LD spectrum of retinal isomers dissolved in ethanol-glycerol (1: 1) was somewhat different from that in 2% aqueous digitonin-glycerol (1: 1). All four retinol isomers-RBP gave LD spectra which closely resembled that in digitonin-glycerol.
We interpret these findings as showing that the retinal in retinol-binding protein was present in the free aldehyde form and that the binding site of retinol-binding protein resembled more closely the hydrophobic environment of digitonin micelles than that of the more hydrophilic environment of the ethanolglycerol solution.
Since under standardized conditions the magnitude of the LD signals of retinal isomers is linearly related to concentration, it is possible to correlate these two measurements (4). The magnitude of the LD signals which were obtained from a given amount of retinal-RBP complex (as determined by absorption spectroscopy) was the same as that obt.ained from the equivalent concentration of free retinal. We interpreted this finding to mean that all the retinal complement of a retinal-RBP complex is in the free aldehyde form, because any other linkage would give a smaller LD signal for a given concentration.
Eflects on Abscrption and CD Spectra of Increase in Ionic BL is the linear dichroism spectrum of the frozen sample (77" K) before illumination with linearly polarized light. 1, free 9-c&-retinal in 2yo aqueous digitonin-glycerol (l:l, v/v); 2, free g-&s-retinal in ethanol-glycerol (l:l, v/v); 3, Q-cis-retinal-RBP complex in 2 mM Tris, pH 9, with 0.15 M NaCl-glycerol (l:l, v/v). All samples were illuminated for 150 min with linearly polarized light at 3GO nm prior to the recording of the linear dichroism spectra. B, spect,ra of all-trans-retinal.
BL is as in A above. 1, free all-lrans-retinal in 2% aqueous digitonin-glycerol (l:l, v/v); 8, free all-trans-retinal in ethanol-glycerol (l:l, v/v); 3, all-Iruns-retinal-RBP complex in 2 mM Tris, pH 9, and 0.15 M NaCl-glycerol (l:l, v/v). Illumination conditions were as in A above. All spectra represent the average of four scans. For further details on linear dichroism measurement see Reference 4.

Strength-When
solid NaCl (final concentration 0.15 M) was added to solutions of retina.l-and retinoic acid-RBP complexes in 2 mM Tris buffer, pH 9, changes were observed in both the absorption and the CD spectra (Table IV).
In no case was there a change in the absorption spectrum at the 280.nm band (Fig. 5). In all cases, though, there was an effect on the chromophore absorption band, which was shifted to the blue, 1 to 7 nm, in the case of the retinal-RBP complexes, and about 1 nm to the red in the retinoic acid-RBP complex (Table IV and Fig.  5). Although the absorpt,ion spectrum did shift considerably, as for instance in the case of the 9-cis-retinaI-RBP, the area under the absorption curve decreased only by some 2.7% as compared to that in low ionic strength.
Accompanying the change in the absorption spectrum, there were also changes in the CD spectrum.
The peak of the CD band also shifted to the blue upon increase in ionic strength. The decrease in rotatory strength was considerable, amount.ing to some 39% in 11-cis-retinal-RBP complex (Table IV and Fig.  6). The all-trans-retinoic acid-RBP showed a small increase in rotatory strength.
There were also changes in the far ultraviolet CD spectra (ZOO to 240 nm) of the retinal-and retinoic acid-RBP complexes.
These changes were on the order of up to 10% in peak height, with small shifts in peak position (on the order of 2 nm).
peaks at the same position as that of apo-RBP and retinol-RRP in the absence of prealbumin.
Although the retinal-RBP complexes were completely dissociated from prealbumin, they still had all the retinal bound to retinol-binding protein.
Most interestingly, while all-trans-retinol-RBP was bound to prealbumin at physiological ionic strength, the acetate ester, namely, all-trans-retinyl acetate-RBP, was completely dissociated. On the other hand all-frans-retinoic acid-RBP was bound firmly to prealbumin and appeared from the gel filtration column as one peak at the same position as that of retinol-RBP-prealbumin complexes.
Although the binding studies using the gel filtration technique showed that there was no binding between retinal-RRP and prealbumin, the interactions between the proteins were quite complex as judged by absorption and CD spectroscopy.
Addition of prealbumin to a solution of retinal isomers-RBP at low ionic strength shifted the absorption maximum to the blue some 8 nm with 9-cis-retinal-RBP and 5 nm with 11-cis-retinal-RBP (Table IV and Fig. 7). There were no apparent shifts with alltrans-retinal-, 13-cis-retinal-, and retinoic acid-RBP complexes. When solid NaCl was added to these solutions containing retinal-RBP and prealbumin at low ionic strength there was a further shift to the blue with all t.he retinal-RBP complexes and a small shift to the red with retinoic acid-RBP.
In all these cases, namely, both at low and at high ionic strength, there was no change at all in the absorption at 280 nm.
It is interesting to note that whereas increasing the ionic strength of a retinal-RBP solution alone led either to a small decrease in the chromophore absorption band or to no change, an increase in the ionic strength of a mixture of retinal-RBP and prealbumin led to an increase in the chromophore absorption band (Tables IV and V).
The analogous situation was seen with retinoic acid-RBP. There WR.S an increase of some 2.4% in the chromophore absorption when salt was added to retinoic acid-REP alone (Table  IV) and a further 2.3% iucrease, to a total of 4.7%, when salt was added to a retinoic acid-RRP and prealbumin mixture.  b In 2 mM Tris-HCI, pH 9, containing 0.15 M NaCI. c Calculated as the change in area under the absorption band between 300 and 600 nm. The area at low ionic strength was taken as lM)c/O.

DISCUSSIOX
The results reported in this paper further elaborate on the theme of the relatively low specificity as to nature and geometrical isomerism of the chromophores that bind to apo-RBP.
The analogs of all-trans.retinol, namely the alcohols (retinol isomers), aldehydes (retinals), acids (retinoic acid), and esters (retinyl acetate) bind with about the same high effectiveness to apo-RBP; all seem to be bound at the same site on retinol-binding protein and are thus competitive inhibitors.
This situation of low selectivity in the nature of the chromophore which is bound to retinol-binding protein stands in marked contrast to the visual pigment apoprotein, which binds only to the ll-and 9-c& retinal isomers. Yet, despite this low select.ivity with apo-RBP, it is clear that the binding of the various chromophores is not a case of nonspecific absorption and binding. All the chromophores studied, namely the retinols, retinals, retinoic acid, and retinyl acetate, show an extrinsic or induced optical activity of the chromophore absorption band with rotatory strength of about the same magnitude.
On the other hand, the binding of retinol and retinal to bovine serum albumin, a process which led to a considerable enhancement of the fluorescence of bound retinol similar to the enhancement upon binding to retinol-binding protein (3), does not lead to an induced optical activity in the chromophore absorption band.
It is interesting to note that t,he retinal isomers-RBP complexes all have a CD band in the 320-to 340~nm region (Fig. 1). The ll-&retinal visual pigment (rhodopsin) and the 9-&sretinal visual pigment (isorhodopsin) also show a prominent Cotton effect at about the same region-the so-called p band (11). The CD bands of the retinals-RBP and the visual pigments peak at about the same wave length, and they are quite strong despite the fact that there is no clear absorption band in this region in the retinal-RBP complexes (Fig. 7). It is an attractive hypothesis that the CD bands at about 320 to 340 nm in both the retinal-RBP complexes and visual pigments are de-6323 rived from the same inherent optical transition of the retinal chromophore. Retinals possess a reactive aldehgde function that under the proper conditions reacts rapidly with primary amino groups to give Schiff bases. It is interesting that even when large excesses of retinal w-ere incubated with retinol-binding protein none combined covalently with the protein.
It is well known that most primary amino groups of proteins (c-amino of lysine residues) are "surface" groups and thus would be expected to be readily available.
We do not have an explanation at this time for the marked lack of reactivity of retinals toward the primary amino groups of retinol-binding protein.
We also have observed that retinals do not react readiIy with the amino groups of the visual pigment apoprotein (opsin) other than the group involved in the usual chromophore binding.2 As would be expected, the spectrum of free retinoic acid showed a marked dependence on pH.
It was somewhat surprising to find that the absorption spectrum of retinoic acid-RI3P was also pH dependent and that spectral changes due to pH change were reversible.
That this was not due to denaturation and release of free retinoic acid from retinol-binding protein was evident from the CD spectrum which also was reversibly pH dependent.
Thus we are led to conclude that the carbosyl group of retinoic acid-RRP complex is available to the solvent. We have shown also that the aldehyde function of retinal-RBP can be reduced in situ with NaBH4 to retinol-RBP, apparently without destruction of the complex.
The availability for interaction with solute of both the carboxyl and aldehyde groups of retinoic acid-and retinal-RBP complexes was distinct from that of retinol-RBP, in which the alcohol seems to be unavailable and protected from the solvent (3, 5).
Although retinols, retinals, retinoic acid, and retinyl acetate all bind to the same site on retinol-binding protein, the resulting conjugated chromophore-proteins differ in some notable ways. The most striking is the difference in binding to prealbumin at physiological ionic strength.
Only the retinol-and retinoic acid-RBP complexes were bound.
This striking difference could be due either to a direct interference in the binding between prealbumin and retinol-binding protein by the functional group of the polyene chromophore or to variation in conformation induced in retinol-binding protein by the different chromophores. Because of the experimental evidence showing that retinol is highly protected from interactions with various reagents when it is bound to retinol-binding protein (3, 5), it is probably deeply buried within the protein away from the surface of the protein.
This observation is difficult to reconcile with a theory that would implicate direct interference by the polyene functional group in the binding of prealbumin to retinol-binding protein.
Xoreover, it is not obvious from this hypothesis why the aldehyde function would interfere with binding to prealbumin and t.he acid function would not. It seems easier to assume that t.he different functional groups impart a somewhat different corrformation to the reconstituted retinol-binding protein and that this in turn impairs the binding between the two proteins.
A somewhat similar explanation for the effect of increased ionic strength on t.he absorption of retinal-and retinoic acid RBP (Table IV) seems reasonable.
Changes in ionic strength had absolutely no effect on the spectra of free retinals dissolved in an aqueous nonionic detergent solution.
The ionic strength effect was mediated then through electrostatic effects on the tertiary structure of the protein, which affected in turn the bind-ing site and interactions between retinals and retinol-binding protein.
The marked changes in the induced Cotton effect of the chromophore absorption band (Table IV) can be taken as further support for this hypothesis.
It is interesting that although retinal-REP complexes do not bind to prealbumin as judged by gel filtration chromatography, there were interactions between the proteins even at low ionic strength.
The addition of prealbumin to 9-and ll-cis-retinal-RBP at low ionic strength caused a shift of the chromophore absorption to the blue. Increasing the ionic strength caused a further shift to the blue (Table V). It is clear that the two proteins interacted even in low ionic strength buffer, although firm binding was not detected.
Recall that prealbumin did not change the chromophore peak position of retinol isomers-RBP either at low or at physiological ionic strength, although there was a hyperchromic effect on the absorption band when salt was added (1).
A more complete understanding of all these fascinating effects on the absorption, CD, and binding properties of the various chromophore-RBP complexes will have to await a better knowledge of the structural aspects of the polyene chromophore-binding site of retinol-binding protein.