All chemicals used in the work were purchased from Sigma-Aldrich (USA) and from Fluka, Merck, Serva, and Ferak (Germany). Treatment and care of animals were conducted according to the Sechenov Institute’s Rules and the ARVO Statement for the Use of Animals in Ophthalmic and Vision Research. In the experiments, the Grass frogs (Rana temporaria) were used. The procedures of isolation of ROSs as well as solubilization of rhodopsin with digitonin and SDS were described early [17]. Detergent/protein ratio in the sample extracts was 4:8 (in g/g).

The original protocols of BN-PAGE [19] and SDS-PAGE [20] were taken as a basis. In order to have other denaturing electrophoresis systems in addition to SDS-PAGE, the original protocol of BN-PAGE [19] was modified by using Coomassie Brilliant Blue G-250 (CBB) in conjuction with SDS (BSDS-PAGE) instead of original mild detergents (see details in [19]) or by adding of urea (2, 4, 6, 7 M) to solubilization buffer (BU-PAGE) instead of 6-aminohexanoic acid. Sometimes urea was added also to cathode buffer for final concentration 2 М. In some SDS-PAGE or BSDS-PAGE experiments the dry SDS was added to digitonin extracts for final concentration 0.5%, thus preparing mixed digitonin/SDS micelles of opsin.

The protocol of formation stable secondary amine, retinylopsin (RO), in which retinal is still bound to denatured opsin at the chromophore-binding site, was described in details early [17]. Briefly, all ROS obtained from 1 - 2 frogs were suspended in 500 μl of mixture of 0.25 M acetate buffer with 0.7 M NaBH₃CN, pH 5.0 [19], and incubated 1 h in the light at room temperature. After finishing of reaction, the suspension was sedimented at 20,000 g for 15 min and washed five times with 0.05 M imidazole buffer, pH 7.0, using the same conditions of centrifugation. The final pellet was immediately solubilized, the extract was loaded onto a gel, and after electrophoresis the fluorescence of RO on the gel was viewed and documented.

For all kinds of electrophoresis a vertical electrophoresis chamber MiniProtean 3 Cell (BioRad, USA) with 10 × 8 cm glasses, 1 mm thickness spacer, and 5.6 ml volume of gel were used. The power supply was a Power Pac HC (250 V/3 A/300 W; BioRad). A J5 Jule gradient former (Jule, USA) was used to prepare a linear gradient of polyacrylamide gel (8% - 26%) for BN-, BUor BSDSPAGEs. The aliquots (2 - 15 μl, 10 - 40 μg protein) were loaded under the cathode buffer onto the sample gel using a Guide ReadyGel Sample loading device (Bio-Rad, USA) and a syringe from Hamilton (USA). After sample loading onto the gel (5 - 10 μl, 10 - 35 μg of protein), the run parameters were kept according to original protocols [19,20]. After three kinds of electrophoresis the sample gel was cut off, and the rest of the gel, depending on the task, was subjected to: 1) Standard fixation followed by protein staining with CBB or AgNO₃ [21]; 2) Electro blotting of proteins onto PVDF membrane (Sigma) and subsequent staining of bands with reaction product formed by interaction rhodopsin-specific mAb (titer 1: 1000, mouse, clone 1D4; Sigma) with anti-mouse IgG (titer 1:30,000; Sigma), containing conjugated alkaline phosphatase [17]; 3) electro elution of protein bands from the gel in denaturing buffer.

The electro blotting apparatus and power supply were the Mini TransBlot Electrophoresis Transfer Cell and Power Pac HC (250 V, 3 A, 300 W), respectively (both from BioRad). The electro elution apparatus and power supply were the Mini Whole Gel Eluter equipped 14-channels chamber core and Power Pac HC (250 V, 3 A, 300 W), respectively (all from BioRad). The electro elution of digitonin or SDS solubilized protein bands from the gel slabs after BN-or BSDS-PAGEs was proceeded in denaturing buffer (43 mM Imidazole - 35 mM HEPES −1% SDS, pH = 7.4). Sometimes before electro elution a narrow strip of the gel was cut off from the slab and subjected fixation and staining to control the separation quality. Following equilibration in the denaturing buffer, the slab gel was removed from the buffer and placed on a glass plate. The template was used to trim the excess gel from the part of the gel containing the separated rhodopsin (opsin) oligomers. The Mini Whole Gel Eluter was filled with the same buffer. The gel was transferred to it and overlaid with soaked filter paper. Electro elution was performed under constant current 50 mA and voltage limited to 5 - 10 V. After 20 min elution, the reverse current was applied to Eluter for 30 sec to dislodge proteins bound to the cellophane membrane. Sometimes proteins after elution were concentrated by spin filtration on the dry Sephadex G-25 (Pharmacia) using a small column and the angle rotor N5404 of K280R centrifuge (Centurion, England).

Protein bands on the gels and PVDF membranes (Sigma) were viewed in scattered white light (staining by CBB or antibodies) or 365 nm UV light (RO fluorescence) emitted by the TFX-L/WL transilluminator (Vilber Lourmat, France). Because CBB completely quenched a shortlived and weak fluorescence of RO, only the gels after SDS-PAGE could be used to register fluorescence of this protein. Protein in the samples was determined by the method of Lowry et al. [22] after its precipitation by 3% CCl₃COOH, washing by H₂O and solution in 1 N NaOH. The standard protein was BSA.

After each experiment, the fresh or treated (fixed, stained and washed) gels were taken for the pictures by the PowerShot A3100 IS camera (Canon, Japan). The images as jpeg files were collected in computer hard disk and then they were transformed to tif format using Adobe Photoshop CS3 software. Each experiment was repeated 3 - 4 times and mostly typical and characteristic images were choice to illustrate the results of the experiments.

BN-PAGE and BSDS-PAGE of digitonin and SDS extracts of rhodopsin (opsin) followed by staining with CBB or mAb have confirmed our previous data [16,17] and have shown that protein pattern of extracts in the gradient gel are presented by dimer with a slight content of higher oligomers (Figure 1, lines 1-4).

No opsin monomer and protein impurities were detected in the gels. It seems also that the content of the higher oligomers in SDS extracts is more pronounced than in digitonin one. In spite of SDS is a most harsh detergent among well-known, the dissociation of opsin oligomers to monomer in SDS extracts under BSDSPAGE does not occur. This means that the used electrophoresis system together with CBB preserve the native no covalent interactions between monomers in opsin oligomers. Furthermore, the equal presence of opsin dimers and oligomers in both detergent extracts supports conclusion that native rhodopsin forms dimeric and higher order structures in the very photoreceptor membranes [23].

All oligomers in SDS or in digitonin/SDS mixed micelles under BSDS-PAGE, as compared to oligomers in digitonin micelles under BN-PAGE, showed a more rapid mobility toward the anode (Figures 1 and 2, lines 3, 4 compared to lines 1, 2). Analogical behavior of rhodopsin in other detergents was observed in the previous papers [16,17]. In the current work, we explain it by the graded degree of unfolding of rhodopsin in mild and harsh detergents, thus resulting in opsin structure the different changes of the compactness [16]. Indeed, references [24,25], using the biophysical methods, have recently shown that denatured states opsin in SDS micelles (0.3% - 15% SDS) are characterized a relatively greater burial of cystein and tryptophan residues and are

more compact as compared to the states observed in mild detergents or dodecylmaltozide/SDS mixed micelles. Thus, we concluded that opsin in SDS, as compared with rhodopsin in digitonin, has a distinctive shape. The behavior of rhodopsin (opsin) under BNand BSDSPAGEs supported the conclusion that not only the molecular mass of a protein but also its shape appears to be an important parameter determining its position in the gel [26]. From this, it follows that the use of BN-PAGE or BSDS-PAGE coupled with mild and harsh detergents for the determination of molecular mass of various proteins looks very doubtful.

Taking in account the entropic nature of hydrophobic interactions (the gain with increasing temperature and, conversely, the decrease when the temperature is lowered), seems likely that the conformational transitions in opsin dimer under the influence of SDS should be dependable from the temperature. Indeed, our experiments have shown that in contrast to digitonin extracts, where rhodopsin is in the stable folded state (Figure 2, lines 1 and 2), opsin in mixed SDS/digitonin extracts at low temperature showed ability to conformational transition from unfolded dimer state back to the initial folded dimer conformation (Figure 2, lines 3 and 4).

Introduction of urea in solubilizing and cathode buffers under BU-PAGE, resulted in both digitonin and SDS extracts roughly dose-dependent dissociation of dimer to monomer and inverse association of monomer to oligomers. This was manifested in decrease the proportion of major band of dimer and increase in the share of monomer and higher oligomers of opsin, respectively (Figure 3, lines 2 - 5, 7 - 10).

It is evidently, that small but distinct increase in a share of the higher oligomers in the presence of urea is caused by a secondary self-association of the formed monomer (or oligomers). It is interesting, that in the condition of BU-PAGE, opsin monomer extracted by digitonin has a tendency to be divided on two sub bands (Figure 3, lines 2 - 5). This tendency is much less pronounced in SDS extracts (Figure 3, lines 7 - 10). Note that the relative functional integrity of frog rhodopsin solubilized with digitonin in the presence 2 and 4 M urea was still demonstrated by its characteristic UV-visible absorption spectrum. However, 6 - 7 M urea caused the deep destructive changes in rhodopsin that resulted in

detachment of 11-cis-retinal from opsin and respectively, in bleaching of rhodopsin even in digitonin extracts (data not shown).

As showed SDS-PAGE, retinyl-opsin (RO) formed directly in photoreceptor membrane, was presented in gel by fluorescent monomer, dimer and higher oligomers, regardless SDS or digitonin were initially used for solubilization (Figure 4, lines 1-7).

SDS extracts, as compared to digitonin extracts, demonstrated the larger number of protein impurities (molecular mass ~36, 50, 66 kDa), thus indicating the selectivity of digitonin for RO extraction. Similar to opsin monomer devoided of the chromophore under BU-PAGE (Figure 3, lines 2-5), RO monomer with “sewn” chromophore under SDS-PAGE (Figure 4, lines 1 - 4, 5 - 8) also demonstrated its subdivision on two sub bands. The phenomena of “splitting” (or vice versa “doubling”) which was detected under denaturing electrophoresis in a single-chain protein, opsin, cannot be interpreted now and requires further study.

The presence of opsin aggregates in SDS as well as in SDS + urea extracts, as revealed by BSDS-PAGE, BU-PAGE and SDS-PAGEs, contradicts to the conclusions of other authors obtained after use of biophysical methods [24,25]. They are as follows: 1) “SDS is able to prevent the exposure of large hydrophobic regions present in rhodopsin which otherwise leads to aggregation”, 2) “3% SDS + 8 M urea are the denaturing conditions of choice to study maximally unfolded rhodopsin without aggregation”. The reasons of such discrepancy are not clear and cannot be commented on now.

Can the various opsin oligomers obtained by BN-PAGE or BSDS-PAGE entirely dissociate to monomer?

In our experiments we showed that electro elution in denaturing conditions followed by SDS-PAGE induced complete dissociation of opsin dimers visible in the Figures 5(a), 6(a) to opsin monomers with molecular mass close to 40 kDa, regardless of BN-PAGE (run native rhodopsin) or BSDS-PAGE (run denatured opsin) were initially used for their isolation (Figure 5(b), lines 4-7; Figure 6(b), lines 9-12). Because of the proportion of higher rhodopsin oligomers (trimer and >) in digitonin extracts is sufficiently low (Figure 5(a)), it is difficult to see and understand on what products they dissociate (Figure 5(b)). Surprisingly, but clearly visible after BSDS-PAGE opsin trimer (Figure 6(a)), under similar conditions dissociated to a small fraction of monomer and a predominant share of dimer (Figure 6(b), lines 7,8), thus clearly indicating that dissociation in these conditions does not necessarily leads to the formation only monomers. The products of dissociation of opsin

tetramer and pentamer are difficult to identify, but they are evidently neither the monomer nor the dimer (Figure 6(b), lines 4 - 6). The data ambiguities obtained under dissociation of Op oligomers require further experiments.

As seen from the Figures 5(b), 6(b), all fractions of oligomers contain at least 2 unknown proteins with molecular mass close to 36 and 66 kDa. They first were noticed in our experiments under SDS-PAGE of RO and, as compared to digitonin extracts, were prevailing in SDS extracts (Figure 4). It means that both BN-PAGE and BSDS-PAGE do not provide a complete separation of complexes of opsin oligomers/concomitant proteins or opsin oligomers/intrinsic monomers, thus demonstrating that mAb-positive bands are non-homogenous. Complete homogeneity of Op oligomers is ensured only after use SDS-PAGE. The physiological function these concomitant proteins remain to be obscure.