Robust self-association is a common feature of mammalian visual arrestin-1

Abstract

Arrestin-1 binds light-activated phosphorhodopsin and ensures rapid signal termination. Its deficiency in humans and mice results in prolonged signaling and rod degeneration. However, most of the biochemical studies were performed on bovine arrestin-1, which was shown to self-associate forming dimers and tetramers, although only the monomer binds rhodopsin. It is unclear whether self-association is a property of arrestin-1 in all mammals, or a specific feature of bovine protein. To address this issue, we compared self-association parameters of purified human and mouse arrestin-1 with those of bovine counterpart using multi-angle light scattering. We found that mouse and human arrestin-1 also robustly self-associate, existing in monomer-dimer-tetramer equilibrium. Interestingly, the combination of dimerization and tetramerization constants in these three species is strikingly different. While tetramerization of bovine arrestin-1 is highly cooperative, with K_D,dim⁴ > K_D,tet, in mouse K_D,dim ~ K_D,tet, whereas in human K_D,dim ≪ K_D,tet. Importantly, in all three species at very high physiological concentrations of arrestin-1 in rod photoreceptors, most of it is predicted to exist in oligomeric form, with relatively low concentration of free monomer. Thus, it appears that maintenance of low levels of active monomer is the biological role of arrestin-1 self-association.

Arrestin-1⁴ binds light-activated rhodopsin phosphorylated by GRK1 (a.k.a. rhodopsin kinase) with high affinity (1), ensuring the termination of light-induced signaling with sub-second kinetics (2). Arrestin-1 knockout in mice dramatically slows the photoresponse shutoff in rod (3) and cone (4) photoreceptors. Arrestin-1 deficiency in humans results in Oguchi disease, a form of stationary night blindness (5). Arrestin-1 is expressed at very high levels in both photoreceptor types, being the second most abundant signaling protein after corresponding opsins (4, 6, 7). Considering that the rhodopsin concentration in rod outer segments (OS) is ~3 mM (8), the average cytoplasmic concentration of arrestin-1 (which is expressed at ~0.8:1 ratio to rhodopsin (6, 7, 9)) is expected to be >1 mM (1). Dark-adapted rods are used in most studies of the molecular mechanisms of rod signaling in genetically modified mice (reviewed in (10–12)). In the dark, ~85% of arrestin-1 resides in the inner segments, cell bodies, and synaptic terminals (6, 9, 13–15), which brings its concentration in these compartments to >2 mM (9). While the majority of the functional studies were performed in mice and humans, the biochemical properties of arrestin-1 were mostly studied using bovine protein. Bovine arrestin-1 robustly self-associates (16, 17), cooperatively forming dimers and tetramers (16, 18, 19). It is unclear whether this is a peculiar property of bovine protein, or a common feature of mammalian arrestin-1 species. In addition, since only monomeric arrestin-1 binds rhodopsin and quenches the signaling (18), the concentration of the monomer is an important functional parameter; it can be calculated based on total arrestin-1 if the self-association constant(s) are known. In view of therapeutic potential of “enhanced” mutants that can compensate for deficits of rhodopsin phosphorylation in vivo (20), characterization of human arrestin-1 is particularly important. Therefore, we explored oligomerization of purified mouse and human arrestin-1 and found that self-association is a common feature of mammalian arrestin-1. Interestingly, we found that while the values of dimerization and tetramerization constants of arrestin-1 from three species are very different, the underlying molecular interactions appear to be similar: the same point mutations render bovine and mouse arrestin-1 constitutively monomeric.

Methods

Materials

[γ-³²P]ATP, [¹⁴C]leucine, and [³H]leucine were purchased from DuPont NEN. All restriction enzymes were purchased from New England Biolabs. Sepharose 2B and all other chemicals were from sources previously described (21, 22). Rabbit reticulocyte lysate was purchased from Ambion, and SP6 RNA polymerase was prepared as previously described (23). 11-cis-retinal was generously supplied by Dr. R. K. Crouch and the National Eye Institute.

Site-directed mutagenesis

Bovine, mouse, and human arrestin-1 cDNA were cloned into pGEM2-based plasmid with an "idealized" 5'-untranslated region (23) under control of a SP6 promoter. All mutations were introduced by PCR using an appropriate mutagenizing oligonucleotide as a forward primer and an oligonucleotide downstream from the far restriction site to be used for subcloning as a reverse primer. Resulting fragments of various lengths and an appropriate primer upstream of the near restriction site were then used as reverse and forward primers, respectively, for the second round of PCR. Resulting fragments were subcloned back, and all constructs were confirmed by dideoxynucleotide sequencing.

In vitro transcription, translation, and evaluation of mutants’ stability

Plasmids were linearized with Hind III before in vitro transcription to produce mRNAs encoding full-length arrestin proteins, as described (22, 24). All arrestin proteins were labeled by incorporation of [³H]leucine and [¹⁴C]leucine with the specific activity of the mix being 1.5–3 Ci/mmol, resulting in the specific activity of arrestin proteins within the range of 66-85 Ci/mmol (145-187 dpm/fmol). The translation of every protein produced a single labeled band with the expected mobility on SDS-PAGE. The relative stability of all mutants used in this study (evaluated as described in (25)) exceeds 90% of corresponding wild type arrestin.

Rhodopsin preparations

Urea-treated rod OS membranes were prepared, phosphorylated with rhodopsin kinase and regenerated with 11-cis-retinal as described (21). The stoichiometry of phosphorylation for the rhodopsin preparations used in these studies was 3.7 mol phosphate/mol rhodopsin.

Direct binding assays

Arrestin-1 binding to rhodopsin was performed, as described (24, 26). Briefly, in vitro translated tritiated arrestins (100 fmol) were incubated in 50 mM Tris-HCl, pH 7.5, 0.5 mM MgCl₂, 1.5 mM dithiothreitol, 50 mM potassium acetate with 7.5 pmol (0.3 μg) of the various functional forms of rhodopsin in a final volume of 50 μl for 5 min at 37°C either in the dark or under room light. The samples were immediately cooled on ice and loaded under dim red light onto 2 ml Sepharose 2B columns equilibrated with 10 mM Tris-HCl, pH 7.5, 100 mM NaCl. Bound arrestin eluted with the disc membranes in the void volume (between 0.5 – 1.1 ml). Nonspecific binding determined in the presence of 0.3 μg liposomes was subtracted. Arrestin-1 binding to microtubules (MT) (purified tubulin polymerized in the presence of taxol) was performed, as described (27). Briefly, 200 femtomoles of the indicated in vitro translated arrestins were incubated in 50 mM Tris-HCl, pH 7.4, 0.5 mM MgCl₂, 1.5 mM DTT, 1 mM EGTA, and 50 mM potassium acetate for 20 min at 25°C with 20 μg of pre-polymerized tubulin. MTs along with bound arrestin were pelleted. MT-arrestin pellets were not washed due to the low affinity (i.e. high off-rate) of the interaction. The pellet was dissolved in 0.1 ml of 1% SDS, 50 mM NaOH, and bound arrestin was quantified by liquid scintillation counting. Non-specific “binding” (arrestin pelleted without microtubules) was subtracted.

Arrestin-1 purification and analysis of its self-association

WT and mutant mouse, WT bovine, and WT human arrestin-1 were expressed in E. coli and purified, essentially as described (22). All light scattering measurements were made with a DAWN EOS detector coupled to an Optilab refractometer (Wyatt Technologies) following gel filtration on a 7.8 mm (ID) × 30.0 cm (L) silica-based column along with its guard column (Wyatt Technologies). The arrestin samples (100 μl) at different concentrations were incubated in fresh 5 mM DTT for 30 min at room temperature to disrupt covalent inter-arrestin disulfide bonds and injected onto the column at 25°C, at a flow rate of 0.8 ml/min in 50 mM MOPS, 100 mM NaCl, pH 7.2. The column used did not resolve oligomeric species, but simply acted as a filter to remove highly scattering particulates. Light scattering at 18 angles (15°–160°), absorbance at 280 nm, and refractive index (at 690 nm) for each sample were taken for a slice centered at the peak of the elution profile and of width approximately that of the profile at half maximum (18). The experimental weight-averaged molecular weight values were obtained from the protein concentration and light scattering data using ASTRA 5.3.4.16 software (Wyatt Technologies). The weight-averaged molecular weight data were analyzed using the two-step monomer-dimer-tetramer (MDT) model (16),

$$2\text{M}\stackrel{{\text{K}}_{dim}}{\leftrightarrow }\text{D}2\text{D}\stackrel{{\text{K}}_{\text{tet}}}{\leftrightarrow }\text{T}$$

where M, D, and T are monomer, dimer, and tetramer, respectively. Details of the analysis have been previously described (28). Except where noted, the equilibrium constants are given in terms of the corresponding dissociation constants, K_D,dimer (K_D,dim) and K_{D, tetramer} (K_D,tet). The errors in equilibrium constants were determined from least-squares fitting of the data to MDT model, taking into account an estimated error of ± 1 kDa is the computed values of the average molecular weight (28).

Results

The level of arrestin-1 in photoreceptors (4, 6, 7, 9) is three orders of magnitude higher than the level of non-visual arrestins in other neurons (29, 30). Another characteristic feature of arrestin-1 is robust self-association at physiological concentrations (16–18). Since all previous studies of arrestin-1 self-association were performed with bovine protein, we tested whether WT mouse arrestin-1 uses the same oligomerization mechanism. To this end, we expressed mouse arrestin-1 in E. coli, purified it, and tested its self-association by measuring the dependence of the average molecular weight on its concentration by multi-angle light scattering. We found that mouse arrestin-1 forms dimers and tetramers (Fig. 1A) like its bovine homolog (18). Although both dimerization (K_D,dim=57.5±0.6 μM) and especially tetramerization (K_D,tet=63.1±2.6 μM) constants are higher than corresponding values for bovine arrestin-1 (37.2±0.2 μM and 7.4±0.1 μM, respectively (18)), at physiological concentrations of >2 mM (1, 9) only a small fraction of WT mouse arrestin-1 would be monomeric (Fig. 1B). Interestingly, while tetramerization of bovine arrestin-1 was invariably found to be cooperative (16, 18, 19), i.e., K_D,dim > K_D,tet, the mouse protein has virtually equal K_D,dim and K_D,tet, indicating a lack of cooperativity in self-association (Table 1).

Fig. 1. Mouse and human arrestin-1 form dimers and tetramers at physiological concentrations.

A. The average molecular weight of wild type mouse arrestin-1 as a function of total concentration (black circles) was determined from the light scattering data as described in the Methods. The solid curve is a least-squares fit of the data to the MDT model with K_D,dim = 57.5±0.6 μM and K_D,tet = 63.1±2.6 μM. The data for bovine arrestin-1 (18) are shown as squares for comparison. B. The percentage of mouse arrestin-1 molecules in monomer (M, straight line), dimer (D, dashed line), and tetramer (T, dotted line) as a function of total arrestin-1 concentration computed for the MDT model and the data in panel A. C. The average molecular weight of wild type human arrestin-1 as a function of total concentration (black circles) was determined from the light scattering data. The solid curve is a least-squares fit of the data to the MDT model with K_D,dim = 2.95±0.02 μM and K_D,tet =224±5 μM. D. The percentage of human arrestin-1 molecules in monomer (M, straight line), dimer (D, dashed line), and tetramer (T, dotted line) as a function of total arrestin-1 concentration computed for the MDT model and the data in panel C. Vertical lines in B and D correspond to arrestin-1 concentrations in the outer segment (300 μM, black) and cell body (2,000 μM, gray) of dark-adapted rod.

Table 1.

Equilibrium constants characterizing self-association of WT and mutant mouse, human, and bovine arrestin-1.

Protein	log Kdima	log Kteta	KD,dim, μM	KD,tet, μM
Mouse arrestin-1	4.24±0.04	4.20±0.17	57.5±0.6	63.1±2.6
Mouse arrestin-1-(F86A,F198A)	3.27±0.05	-	537±9	-
Mouse arrestin-1-(F86A,F198A, A349V)	3.14±0.11	-	724±26	-
Human arrestin-1	5.53±0.03	3.65±0.08	2.95±0.02	224±5
Bovine arrestin-1	4.43±0.02	5.13±0.03	37.2±0.2	7.4±0.1
Bovine arrestin-1-(F85A,F197A)	3.28±0.10	-	525±16	-

^aK_dim and K_tet are the association constants determined from light scattering analysis.

The significant difference between self-association constants of bovine and mouse arrestin-1 prompted us to analyze the human homologue (Fig. 1C,D). We found that purified human arrestin-1 also self-associates, with remarkably low K_D,dim=2.95±0.02 μM and relatively high K_D,tet=224±5 μM (Table 1). Interestingly, these disparate sets of constants in the three mammalian species yield predicted levels of the monomer in the cell body of dark-adapted rod at ~2 mM total arrestin-1 concentration (extrapolating measured mouse values (8, 9) to other species) in the relatively narrow range of 30–90 μM (Table 2). Total tetramer concentrations vary by only 30% (Table 2). The main difference is in the resulting dimer levels, which varies almost 5-fold, from predicted 59 μM in bovine to 281 μM in human rod (Table 2).

Table 2.

Predicted concentrations of monomer, dimer, and tetramer of mouse, human, and bovine arrestin-1 at concentrations in the outer segment (300 μM) and cell body (2,000 μM) of dark-adapted rods.

Arrestin-1	Total, μM	Monomer, μM (%)	Dimer, μM (%)	Tetramer, μM (%)
Bovine	300	27.6 (9.2%)	20.8 (13.9%)	57.7 (76.9%)
Mouse	300	52.8 (17.6%)	48.8 (32.5%)	37.4 (49.9%)
Human	300	15.5 (5.2%)	82.1 (54.7%)	30.1 (40.1%)
Bovine	2,000	46 (2.3%)	59 (5.9%)	459 (91.8%)
Mouse	2,000	95 (4.7%)	159 (15.9%)	397 (79.4%)
Human	2,000	29 (1.5%)	281 (28.1%)	352 (70.4%)

The value of measured K_D,dim between human and mouse arrestin-1 shows ~20-fold difference, and the value of K_D,tet between bovine and human proteins differs by ~30-fold. These dramatic differences raise the possibility that the three mammalian arrestin-1 species could use distinct interaction interfaces, so that common self-association phenotype could represent convergent evolution, rather than direct conservation. It has been recently shown that the structure of the solution tetramer of bovine arrestin-1 is different from that of the crystal tetramer (18). Extensive investigation yielded a model of the solution tetramer, where receptor-binding surfaces are shielded by “sister” subunits, which explains demonstarted inability of the oligomers to bind rhodopsin (19). Based on this model, a modified bovine arrestin-1 was constructed, where two (F85A,F197A) mutations predicted to disrupt NN (F85A) and CC (F197A) self-association interfaces were introduced (Fig. 2A,B,C) (19). Indeed, this mutant was shown to be essentially monomeric, with K_D,dim=525 μM and no detectable tetramer formation (19), independently confirming the model. To test whether the same subunit arrangement is present in the tetramer of mouse arrestin-1, we introduced homologous mutations (F86A,F198A), expressed this protein, and measured its self-association. We found that this mutation in mouse arrestin-1 yields the same non-self-associating phenotype as in its bovine counterpart, demonstrating K_D,dim=537 μM and no detectable tetramerization (Fig. 2D). Further disruption of the CC interface by the addition of an A349V mutation brought K_D,dim to 724 μM (Fig. 2D; Table 1). Thus, homologous mutations in bovine and mouse arrestin-1 affect their self-association in a similar manner, suggesting that the same interfaces are involved in oligomerization of both proteins, and the difference in constants reflects the relative energy of interactions between the subunits, rather than a global difference in the structure of the solution tetramer.

Fig. 2. The targeted disruption of arrestin-1 self-association.

A. Solution tetramer structure of bovine arrestin-1 (19). Areas enlarged to show inter-subunit contacts are indicated. B. CC interface (between the two C-domains) showing the positions of residues F197 (F198 in mouse) and A348 (A349 in mouse). C. NN interface (between the two N-domains) showing the position of F85 (F86 in mouse). D. The average molecular weight of the F86A,F198A (red circles) and F86A,F198A,A349V (blue circles) mouse arrestin-1 mutants as a function of total arrestin concentration were determined from the light scattering data (symbols). The fit of the data to the MDT model (solid lines) was obtained, as described (18). Note that neither mutant showed detectable tetramerization, so that the resulting fit describes monomer-dimer equilibrium. The wild type (WT) mouse arrestin-1 data (black circles) are shown for comparison.

Phosphorylated light-activated rhodopsin (P-Rh*) is the main binding target of arrestin-1 in the rod (31). The amount of rhodopsin present in the OS determines the amount of arrestin-1 that can translocated to this compartment (7), supporting the idea that rhodopsin binding holds arrestin-1 in the OS in the light (14). In contrast, in the dark most of arrestin-1 (the estimates for WT mouse range from ~85% (9) to 91% (6) - >95% (7, 14)) resides in other compartments of the cell (6, 13–15), where it is anchored via low-affinity binding to microtubules (14, 32, 33) abundant in the inner segment, perinuclear area, and synaptic terminals (34). Therefore, to test whether observed differences in self-association are the result of selective disruption of the interfaces involved, we compared the binding of WT and mutant forms of bovine and mouse arrestin-1 to P-Rh* and in vitro polymerized microtubules (Fig. 3). Using fully functional radiolabeled arrestins and standard direct binding assays (22, 27, 32, 35) we found that mutations that disrupt oligomerization in bovine and mouse arrestin-1 do not appreciably affect the binding to either partner (Fig. 3A,B). Arrestin-1 is a highly sensitive molecule, where even small conformational perturbations by mutagenesis result in dramatic changes of its binding to P-Rh* (20, 25, 26, 36–40) and microtubules (27, 32). Thus, virtually wild type binding to both rhodopsin and microtubules makes it highly unlikely that these mutations induce any global structural changes in the molecule, leaving targeted disruption of the self-association interfaces as the only plausible explanation of their phenotype.

Fig. 3. Mutations disrupting self-association do not affect arrestin-1 binding to P-Rh* and microtubules.

The binding of indicated radiolabeled arrestins to P-Rh* (100 fmol/assay) (A) and microtubules (200 fmol/assay) (B) was determined, as described in Methods. Means +/− SD of three experiments performed in duplicates are shown.

Discussion

Preferential binding of arrestin-1 to P-Rh* (41) and resulting quenching of rhodopsin signaling (42) were discovered in mid-1980s, and this remains its least controversial function to this day (1). The ability of bovine arrestin-1 (under the name of S-antigen) to self-associate was discovered a decade earlier (43), but no biological function was ascribed to this phenomenon. The interest in arrestin-1 self-association was revived when two crystal structures of bovine protein (44, 45) revealed virtually identical tetramers. Further studies confirmed its self-association in solution (16–18), although careful examination revealed that the structure of the solution tetramer that forms under much more physiological conditions is dramatically different from that found in crystal (18, 19). The oligomers were usually discussed as storage forms, an interpretation strongly supported by direct demonstration that only the monomer is capable of binding P-Rh* (18). Yet it remained unclear why out of all signaling proteins present at enormous (as compared to other neurons) concentrations in the rod (8), arrestin-1 is the only one that has a special apparently inactive storage form. Since all these studies were performed on bovine protein, it was not even clear whether self-association is a common feature of mammalian arrestin-1 species, which would be the case if it has physiologically relevant function.

Here we compared arrestin-1 from three mammals: (1) bovine, traditionally used for biochemical studies; (2) mouse, the best functionally characterized in vivo; and (3) human, the most therapeutically relevant. In each case we found that arrestin-1 robustly self-associates at physiological concentrations, which suggests that this feature is biologically important. Surprisingly, the thermodynamics of self-association in the three species are strikingly different. While tetramerization of bovine arrestin-1 is cooperative, in a sense that K_D,dim > K_D,tet, there is no such cooperativity in mouse (K_D,dim ~ K_D,tet), whereas in human K_D,dim ≪ K_D,tet, being 10- and 20-times lower than K_D,dim of bovine and mouse arrestin-1, respectively (Table 1). Despite these differences, the structures of the solution tetramers are likely similar, as judged by remarkably uniform effects of mutations destabilizing inter-subunit interfaces in bovine and mouse arrestin-1 (Fig. 2; Table 1). Importantly, these mutations do not appreciably affect arrestin-1 binding to P-Rh* and microtubules (Fig. 3), suggesting that the conservation of the interface residues reflects the need for self-association, rather than being a byproduct of conservation of other arrestin-1 functions. Most of the residues in the self-association interfaces identified in bovine arrestin-1 (19) and deduced by homology modeling in mouse and human proteins are conserved, with a few differences possibly responsible for distinct thermodynamics.

Rod photoreceptors function in dim light, saturating even at modest illumination levels (46). Thus, clues to possible biological functions of arrestin-1 self-association may be revealed through its effects on the state of arrestin-1 at concentrations found in the dark-adapted rod. Arrestin-1 distribution in dark-adapted rods was quantitatively measured only in mouse (6, 7, 9, 14). However, rod function in different mammals is similar, and arrestin-1/rhodopsin ratios in the OS of dark-adapted mouse (9) and frog (47) rods are remarkably close. Therefore, mouse expression levels were used for all species in the estimates that follow. In the dark the bulk (85–95% (6, 7, 9, 14)) of arrestin-1 is localized to the cell body, where expected concentration reaches ~2 mM, with much lower concentration of ~300 μM in the OS, where its main target rhodopsin is localized (9). Using these concentrations and measured values of K_D (Table 1), we estimated the expected equilibrium concentrations of arrestin-1 monomer, dimer, and tetramer in the dark-adapted rod (Table 2). In all three cases, arrestin-1 self-association makes the fraction of the monomer in the cell body relatively small, 1.5–4.7%. However, each species achieves this result in a somewhat different way (Tables 1, 2). In case of bovine protein, self-association is cooperative, with K_D,tet< K_D,dim, so that at 2 mM the bulk (>90%) of arrestin-1 is stored in the form of tetramer, while the fraction of the dimer is very small (<6%). In mouse, K_D,dim ~ K_D,tet, so that while the bulk (~80%) is still a tetramer, much larger fraction (~16%) exists as a dimer. Human arrestin-1 dimerizes more readily than the others, but shows less robust tetramerization, so that at the same concentration only 70% exists as a tetramer, whereas as much as 28% is a dimer. In all cases monomer concentration is <100 μM, with about 3-fold differences among species, dimer concentrations vary ~5-fold, whereas absolute tetramer levels vary by only ~30% (Table 2).

Light exposure induces massive arrestin-1 translocation from the inner segment and cell body to the OS, which takes up to 30–60 min (6, 13–15, 48, 49). Considering that the photoresponse, which is normally terminated with sub-second kinetics (2, 50–52), is greatly prolonged in arrestin-1 knockout animals (3), arrestin-1 present in the OS in the dark must be responsible for timely shutoff. Our analysis of arrestin-1 oligomerization suggests that level of the monomer, which is the only rhodopsin-binding form (18), is much higher in the OS (>50 μM) in nocturnal mice most reliant on rod vision, compared to <30 μM in bovine and just ~16 μM in human rods. Rod vision is less crucial for these latter diurnal species.

Using the second-order on-rate constant (k₁) recently measured for arrestin-1 binding to P-Rh* (~10⁶ M⁻¹s⁻¹ for P-Rh* in nanodiscs (53)) along with the monomer concentration allows one to estimate the expected pseudo-first order rate constant for an encounter of rhodopsin with arrestin-1 (k₁x[monomeric arrestin-1]), which in mouse yields ~50 s⁻¹. This estimate of the lower limit suggests that each rhodopsin encounters an arrestin-1 molecule on average once every ~20 ms. In other words, arrestin-1 “checks” the functional state of rhodopsin in this time interval and binds tightly when it encounters P-Rh* (1). Again, this yields an estimate of a lower limit to the active rhodopsin lifetime in mouse rods, which must be >20 ms because of additional time that GRK1 needs to attach three or more phosphates to light-activated rhodopsin necessary for the high-affinity arrestin-1 binding (21, 52). This number is consistent with the original experimental estimate of <60 ms (2) and more recent one of ~30 ms (51), as well as with the modeling of photoresponse dynamics based on these data (54, 55). By the same token, lower monomer concentrations in the OS of bovine and human rods (Table 2) suggest that the lifetime of the active rhodopsin in these species must be longer, >36 ms and >65 ms, respectively. Since in the OS arrestin-1 diffuses in the cytoplasm with complex geometry (56), while rhodopsin diffuses in two dimensions on the disc membrane, the actual on-rate could differ from these estimates, so these predictions need to be tested experimentally.

Several studies showed that light-dependent translocation of arrestin-1 to the OS (14), as well as transducin movement out of the OS (57), is energy-independent, largely driven by their interactions with non-moving partners (reviewed in (58)), whereas others suggested that active transport could be involved (59, 60). Arrestin-1 in photoreceptors is clearly at disequilibrium in the dark and in bright light (61). Thus, regardless of the mode of transportation, it must be “tethered” in the OS in the light and in the cell body in the dark by other proteins. Otherwise, the diffusion would quickly undo anything that active transport could achieve. Therefore, the concentrations of free monomer, dimer, and tetramer in the OS and cell body are likely equal (1). Preferential arrestin-1 localization outside of the OS in the dark was reported to be determined by its binding to microtubules (14, 33), abundant in the cell body and sparse in the OS (34). Other arrestin-1 binding partners, such as N-ethylmaleimide-sensitive factor (NSF, (62)) and/or enolase (63) could also serve as anchors in the cell body in the dark. However, the concentration of polymerized tubulin, where each αβ-dimer can bind arrestin-1 (27), by far exceeds all other putative anchors combined, suggesting that MTs likely serve as the main binding partner. Thus, the difference in the concentration of each form between these two compartments of the rod largely reflects its MT-associated fraction. The size of the rods and their OS are significantly different in different species. Unfortunately, the volume of the cytoplasm in the OS and other rod compartments, as well as the concentrations of rhodopsin and other signaling molecules, was carefully measured only in mice and frogs (64). However, arrestin-1/rhodopsin ratios in dark-adapted mouse OS (9) and ~20 times larger frog OS (47) are remarkably close. On the strength of these findings and the fact that the size of the OS is determined by the amount of rhodopsin present there (65), the estimates below are based on the assumption that arrestin-1 concentration in dark-adapted mammalian rods is similar. In this case, in all species the predicted levels of the monomer, dimer, and tetramer in the cell body exceed those in the OS by 1.7–1.9-fold, ~3-fold, and ~10-fold, respectively (Table 2). In bright light, much higher total arrestin-1 concentration in the OS was shown to be due to its binding to rhodopsin (6, 7, 14), as the amount of rhodopsin in the OS clearly limits molar amount of arrestin-1 that can translocate to this compartment (7). Similarly, in the dark the total concentration of each form of arrestin-1 in the cell body likely exceeds corresponding value in the OS by the amount that is bound to MTs. For example, in mouse rod the estimated total concentration of monomer is 52.8 μM in the OS and 95 μM in the cell body (Table 2). The latter exceeds the OS concentration by ~42 μM, suggesting that this amount (~44% of the total monomeric arrestin-1 in this compartment) is bound to MTs. Similarly calculated differences in the concentrations of monomer, dimer, and tetramer in the OS and the cell body suggest that the fractions of the individual MT-bound forms in the cell body are fairly close in the three mammalian species: monomer (40–47%), dimer (65–71%), and tetramer (87–91%). However, each of the three molecular forms of arrestin-1 apparently has distinct propensity to associate with MTs, with the order of potency being: tetramer ≫ dimer ≫ monomer.

To conclude, here we show that arrestin-1 self-association is conserved in three mammalian species, indicating that this phenomenon is biologically important. We present evidence that, although the dimerization and tetramerization equilibrium constants are very different in the three species examined, the overall structures of the solution tetramers are likely to be very similar. Despite the differences in thermodynamics of association, for each species the concentration of the active monomer is very low, while the bulk of arrestin-1 exists in the form of tetramer with the best ability to bind microtubules in the cell body. The concentrations of the arrestin-1 monomer in the OS, estimated from the oligomerization equilibrium constants, provide experimentally testable predictions regarding the lifetime of active rhodopsin in different species. An “enhanced” arrestin-1 mutant with increased affinity for light-activated unphosphorylated rhodopsin was recently shown to have therapeutic potential in genetic disorders with deficient rhodopsin phosphorylation (20). The biological importance of arrestin-1 self-association indicates that the changes in self-association, inadvertently produced by mutagenesis, may underlie reported deleterious effects of very high expression of this mutant (20). Thus, normal ability of any mutant form of arrestin-1 to oligomerize must be ascertained before it can be used for gene therapy.

Abbreviations

K_dim (=1/K_D,dim)

dimerization constant, i.e., equilibrium association constant of two arrestin-1 monomers

K_tet (=1/K_D,tet)

tetramerization constant, i.e., equilibrium association constant of two arrestin-1 dimers

GRK1

G protein-coupled receptor kinase 1, a.k.a. rhodopsin kinase

P-Rh*

phosphorylated light-activated rhodopsin