In 1989, P. Philippov’s group from M.V. Lomonosov Moscow State University invented a method for purification of the visual G-protein transducin (Gt) and other G-proteins. The idea of the method was based on the ability of visual rhodopsin to bind and to release transducin in the absence and in the presence of GTP, respectively. For this aim, a column with delipidated visual rhodopsin immobilized on Concanavalin A Sepharose was used. Chromatography of a crude extract of bovine rod outer segments on the column allowed one to obtain a set of transducin subunits with a slight contamination of cGMP-phosphodiesterase. Also, an admixture of an unknown protein with an apparent molecular weight of 26 K was seen on the electrophoregram. The unknown protein attracted the attention of the group since the capability of binding to rhodopsin had been a characteristic feature of several key photoreceptor proteins, such as transducin, rhodopsin kinase, and arrestin. That is why the group decided to study this protein in more detail. The protein named “p26” was purified to a homogeneous state and used to raise specific antibodies. Screening of the retina and a number of other tissues for the presence of p26 with the use of the antibodies detected this protein only in the retina, in particular in the photoreceptor layer. It was also demonstrated that the amino acid sequence of p26 contained several calcium-binding sites of the EF-hand type, and the ability of p26 to bind Ca2+ was confirmed by experiments with calcium-45. In addition, p26 was suggested to be a Ca2+-specific regulator of photoreceptor guanylate cyclase, a key enzyme of photoreceptors recovery, and due to this ability it was rechristened as “recoverin.” Afterward, a 26 K protein named “S-modulin” was purified by S. Kawamura and colleagues from frog rod outer segments and shown to have a primary structure similar to bovine recoverin. Later it became clear that the binding of recoverin to rhodopsin is not quite specific as recoverin, due to its Ca2+-myristoyl switch, is capable of binding to hydrophobic substances, e.g., Phenyl Sepharose, in a Ca2+-dependent manner. That recoverin is capable of activating guanylate cyclase was, however, disproved in subsequent works. The mistake in the initial assignment of the recoverin function might be explained by the presence of endogeneous guanylate cyclase activator(s), GCAP1 and/or and GCAP2, in the recoverin preparations used in the preceding works. Nevertheless, recoverin continues to be considered as a participant of the photoreceptor recovery but currently as a Ca2+-sensor of rhodopsin kinase, the enzyme catalyzing phosphorylation and thus desensitization of the visual receptor rhodopsin (for reviews, see Senin et al. 2002, Philippov et al. 2006). After the discovery of recoverin, a number of homologous Ca2+-binding proteins were described, which formed a family of the neuronal calcium sensor (NCS) proteins. The expression of the NCS proteins is restricted within neurons and neuroendocrine cells, wherein these proteins provide a Ca2+-sensitivity to a number of protein targets (for a review, see Burgoyne 2007; Philippov et al. 2006).
Another line of the recoverin research was started in 1987, when an antigen with an apparent molecular weight of 23 K was found in sera of patients with cancer-associated retinopathy (CAR). This so-called CAR-antigen was then purified from bovine rod outer segments and shown to be identical to recoverin. This finding has sprung an intensive study of recoverin as a paraneoplastic antigen in cancer (for reviews, see Senin et al. 2002; Adamus 2006; Philippov et al. 2006). More recently, recoverin has become the first member of a new group of cancer-specific antigens designated as “cancer-retina antigens” which includes several key retinal proteins. In health, these proteins are highly specific for the retina, but in cancer they can be expressed in malignant tumors localized outside the retina (Bazhin et al. 2007).
Tissue and Cellular Distribution of Recoverin
Immunochemical analysis demonstrated the presence of recoverin in the adult retina of all investigated species. Among these are: man, bull, monkey, mouse, rat, rabbit, frog, chameleon, and newt. In the case of the chicken retina, contradictory data were obtained: Recoverin-positive immunoreaction was described in one case, but it was not found in another work. In addition to the retina, recoverin immunoreactivity was observed in the ocular ciliary epithelium, pinealocytes of the pineal organ, and rat olfactory epithelium. Within the retina, recoverin-positive reaction was found in photoreceptor cells as well as in higher order neurons (bipolar and ganglion cells) of a number of species and in amacrine cells of lamprey Lampreta fluviatilis. As already noted, recoverin is suggested to function in photoreceptors as a Ca2+-sensor of rhodopsin kinase, but its role in neurons different from photoreceptors and in tissues different from the retina remains unknown (for reviews, see Senin et al. 2002; Philippov et al. 2006).
Within photoreceptors, recoverin is detected in the outer and inner segments, cell bodies, and synaptic pedicles. Most of the protein is localized in rod inner segments, with approximately 12% present in the outer segments in the dark and less than 2% remaining in that compartment in the light (Strissel et al. 2005). Thus, light causes a reduction of an amount of recoverin in rod outer segments, accompanied by its redistribution toward rod synaptic terminals.
Recoverin is a compact (23.4 K, 201 amino acids) protein consisting of two globular N- and C-terminal domains, separated by a short linker. N-terminal glycine of recoverin is acylated predominantly with the myristic acid residue (C14:0) or, to a lesser extent, with one of the following fatty acid residues: C14:1 (5-cis), C14:2 (5-cis, 8-cis), or C12:0. Each of recoverin domains contains a pair of potential Ca2+-binding sites of the EF-hand type: In total, the recoverin molecule contains four potential Ca2+-binding sites which are disposed evenly along the amino acid chain of the protein. Of these, only two – EF-hand 2 and EF-hand 3 – are capable of binding calcium ions. Whereas EF-hand 1 and EF-hand 4 are inactive in this respect due to the following structural “defects”: (1) In the sequences of EF-hand 1 and EF-hand 4, residues of negatively charged amino acids critical for the coordination of Ca2+ are missing in the 1st and 3rd positions of the 12-mer Ca2+-binding loops; (2) EF-hand 1 cannot accept the conformation needed for the binding of calcium as P40 is present in the fourth position of the EF-hand 1 loop; (3) EF-hand 4 contains a salt bridge between the side chains of the K161 and G171 in the 2nd and 12th positions of the Ca2+-binding loop; and (4) the highly conserved glycine at position 6 of the EF-hand 4 loop is replaced by the aspartic acid residue (D165) (for reviews, see Senin et al. 2002; Philippov et al. 2006).
It should be added that recoverin, in addition to ten α-helices A-J normally present in other NCS proteins, contains an extra α-helix K which together with α-helix J forms the so-called C-terminal segment of the protein. The segment is structurally variable within the NCS family and is suggested to provide NCS proteins with unique functional features. Thus, the recoverin C-terminal segment is involved in regulating its Ca2+-binding properties, as well as in recognizing and regulating rhodopsin kinase (Weiergräber et al. 2006; Zernii et al. 2011).
Molecular Properties of Recoverin
Recoverin molecule is characterized by a set of key properties required for the signaling activity of the protein. Among them the most important are calcium binding, N-terminal myristoylation, and the ability to bind to phospholipid membranes. In recombinant nonmyristoylated recoverin, the binding of calcium to EF-hands 2 and 3 occurs independently with different affinities: KD = 6.9 and 0.11 μM, respectively. In contrast, the binding of calcium to recombinant myristoylated recoverin is a cooperative sequential process (Hill coefficient = 1.75), wherein EF-hand 3 is occupied first, facilitating the subsequent filling of EF-hand 2 (an apparent KD of the complex formed is equal to 17 μM). Thus, N-terminal myristoylation confers onto recoverin the cooperativity in calcium binding to EF-hands 2 and 3. Also, the myristoyl residue significantly stabilizes the conformation of the Ca2+-free protein during the stepwise transition toward the fully Ca2+-occupied state (for reviews, see Senin et al. 2002; Philippov et al. 2006).
Myristoylated recoverin is capable of binding to hydrophobic surfaces, such as the photoreceptor and artificial lipid membranes. Depending on calcium concentration, compartmentalization of recoverin is reversibly changed from a soluble Ca2+-free form to a membrane-bound Ca2+-containing form. This process is due to the mechanism of the Ca2+-myristoyl switch operating in recoverin: After EF-hand 3 is filled by calcium, EF-hand 2 is subsequently filled, which triggers the exposition of the myristoyl group attaching recoverin to the membrane. Solid-state NMR studies revealed that the Ca2+-bound protein is positioned on the membrane surface so that its long molecular axis is oriented 45° with respect to the normal membrane. The myristoyl group is buried inside the membrane, whereas the N-terminal region of recoverin points toward the membrane surface, with close contacts formed by basic residues K5, K11, K22, K37, R43, and K84. This orientation of the membrane-bound protein allows an exposed hydrophobic crevice, near the membrane surface, to serve as a binding site for the target protein, rhodopsin kinase (for review, see Ames and Lim 2012). The half-maximal binding of recoverin to photoreceptor membranes in vitro occurs at 2.5 μM of a free calcium concentration ([Ca2+]f), which is slightly out of the physiological range of cytoplasmic [Ca2+]f. However, extrapolation to in vivo conditions in rod outer segments, bearing stacks of densely packed membranes, reveals that the apparent affinity of recoverin to calcium is in the submicromolar (i.e., physiological) range of [Ca2+]f (for reviews, see Senin et al. 2002; Philippov et al. 2006). The binding of recoverin to membranes depends on their lipid composition: It is enhanced with the elevation of the content of phosphatidylserine (Senin et al. 2007), polyunsaturated phospholipids (Calvez et al. 2011), and most notably cholesterol. High cholesterol content in photoreceptor disk membranes found at the base of rod outer segments might favor the affinity of recoverin to the membranes and shift its binding to the physiological range of [Ca2+]f (for a review, see Philippov et al. 2006). Similar effect is produced by caveolin-1, a transmembrane scaffold protein found in cholesterol-enriched photoreceptor rafts (for a review, see Philippov et al. 2006). Apparently, caveolin-1 is capable of forming a complex with recoverin thereby increasing Ca2+ affinity of the latter (Zernii et al. 2013). The Ca2+ dependence of the recoverin binding to photoreceptor or artificial lipid membranes is also regulated by the C-terminal segment of recoverin, which serves as an internal modulator of its Ca2+ sensitivity and functional activity (Weiergräber et al. 2006; Senin et al. 2007; Zernii et al. 2011).
Along with Ca2+ binding, which is a key molecular property of recoverin, the protein can bind Zn2+ with stoichiometry of 1:1 and apparent KD of 30 and 7.1 μM for apo- and Ca2+-loaded protein forms, respectively (Permyakov et al. 2003). Also, recoverin molecules are able to form a disulfide dimer and thiol-oxidized monomer under mild oxidizing conditions, involving unique C39 highly conserved within NCS family (Permyakov et al. 2007). The functional properties of thiol-oxidized monomer were examined using recoverin mutant, containing oxidation mimicking substitution C39D. Although the global structure of the mutant displays no obvious differences from that of the wild type protein, the introduced substitution suppresses membrane association of recoverin and disrupted its binding to rhodopsin kinase (Permyakov et al. 2012; Ranaghan et al. 2013). Disulfide dimerization of recoverin was shown to occur after prolonged intense illumination of mammalian eyes under ex vivo and in vivo conditions. The illumination triggers time-dependent accumulation of disulfide homodimers of recoverin and its higher order disulfide cross-linked species, including a minor fraction of mixed disulfides with some other intracellular proteins; moreover, the disulfide dimer of recoverin demonstrates an increased propensity to multimerization and aggregation. It is suggested that the disulfide dimerization of recoverin is involved in the light-induced damage of photoreceptor cells and progression of age-related macular degeneration (Zernii et al. 2015).
Targets and Functions of Recoverin
A number of the in vitro data suggested that a major intracellular target of recoverin in rod outer segments is rhodopsin kinase (GRK-1) (Gorodovikova and Philippov 1993; Kawamura 1993; for a review, see Philippov et al. 2006). According to these data, at high calcium, corresponding to a dark state of photoreceptor cells, rhodopsin kinase forms a complex with recoverin and becomes inactive; at low calcium, corresponding to the bleached state of photoreceptor cells, the complex dissociates allowing activation of the enzyme. However, the actual in vivo function of recoverin has been a subject of discussion over many years. The minimal hypothesis was that recoverin plays a role of an intracellular Ca2+ buffer that is in agreement with faster Ca2+ dynamics observed in recoverin knockout rods (Makino et al. 2004). Yet, the growing evidence confirms that besides Ca2+ buffering the protein operates as a Ca2+-dependent modulator of rhodopsin kinase. Indeed, the genetic deletion of recoverin, while not affecting amplitude of photoreceptor light response, accelerates time course of its decay, and similar effect is produced by overexpression of rhodopsin kinase indicating that these proteins act in concert (Makino et al. 2004; Chen et al. 2012; Chen et al. 2015). These data suggest that either absence of the rhodopsin kinase inhibition by Ca2+/recoverin or an increased amount of the enzyme would shorten catalytic activity of photoexcited rhodopsin thereby speeding up photoresponse recovery. Consistently, overexpression of rhodopsin kinase increases light-dependent phosphorylation of rhodopsin in vivo (Chen et al. 2012). It was shown, however, that rhodopsin kinase and recoverin, in addition to their role in rhodopsin regulation, can modulate the decay of activated cGMP-phosphodiesterase, and this ability may contribute to the light adaptation of photoreceptor cells which is known to be mediated by Ca2+/recoverin. In this case, rhodopsin kinase under Ca2+-dependent control of recoverin might phosphorylate additional protein substrates, which were speculated to be cGMP-phosphodiesterase itself, GTPase-activating proteins (GAPs), or transducin (Chen et al. 2015). One may add that another Ca2+-sensor, calmodulin, is also capable of inhibiting rhodopsin kinase in vitro. Calmodulin and recoverin were shown to recognize separate sites in the enzyme molecule thereby providing synergetic inhibition of its activity (Grigoriev et al. 2012).
Molecular interaction between recoverin and rhodopsin kinase has been investigated in detail. The filling of EF-hand 2 with calcium (in myristoylated recoverin the filling of EF-hand 2 occurs only after EF-hand 3 is already filled) results in the exposition of a cluster of the hydrophobic amino acids, providing recoverin with an ability to interact with rhodopsin kinase and thus inhibiting the activity of the enzyme. According to the surface plasmon resonance studies, the half-maximal binding of rhodopsin kinase to immobilized recoverin occurs at approximately 0.51 μM of rhodopsin kinase. Myristoylation has a little effect on the binding of recoverin to the kinase, but it shifts the half-maximal effect of calcium on the binding from 150 nM for nonacylated recoverin to 400 nM for myristoylated recoverin (for reviews, see Senin et al. 2002; Philippov et al. 2006). Recoverin binds to a region of residues 1–15 at the N-terminus of rhodopsin kinase (Higgins et al. 2006). Nuclear magnetic resonance studies of the complex between Ca2+-bound recoverin and a N-terminal fragment of rhodopsin kinase, residues 1–25 (RK1–25), revealed that the hydrophobic face of the RK1–25 helix (L6, V9, V10, A11, A14, and F15) interacts with an exposed hydrophobic groove on the surface of recoverin, lined by side-chains of the residues W31, F35, F49, I52, Y53, F56, F57, Y86, and L90. In that structure, the first eight residues of recoverin at the N terminus are solvent-exposed, enabling the N-terminal myristoyl group to interact with target membranes (for a review, see Ames and Lim 2012). The half-maximal inhibition of rhodopsin kinase by recoverin is observed at 2–3 μM and 1.5–1.7 μM of calcium in the case of nonacylated and myristoylated recoverin, respectively. At saturating calcium concentrations, the half-maximal inhibition of rhodopsin kinase occurs at 6.5–8 μM of nonmyristoylated recoverin and at 0.8–3 μM of myristoylated recoverin, suggesting that photoreceptor membranes enhance inhibitory effect of recoverin upon rhodopsin kinase. The inhibition of rhodopsin kinase by recoverin is facilitated when the cholesterol content of membranes is increased. As the cholesterol content in photoreceptor disk membranes changes along the axis of rod outer segment from 5% at the tip to 30% at the base, the above-mentioned effect of cholesterol might be of physiological importance (for reviews, see Senin et al. 2002; Philippov et al. 2006). The activity of recoverin as a Ca2+-sensor of rhodopsin kinase is also regulated by the C-terminal segment of recoverin, the internal modulator of its Ca2+-sensitivity, which plays an important role in targeting of rhodopsin kinase (Weiergräber et al. 2006; Zernii et al. 2011).
The function(s) of recoverin in the rod compartments different from the outer segments remains unspecified, but it is possible that the following data could help to provide a clue to this issue. Being localized in the rod inner segments, recoverin was found to enhance signal transfer between rods and rod bipolar cells via an unknown mechanism (Sampath et al. 2005). In the ribbon synapse of photoreceptors, recoverin is colocalized with membrane palmitoylated protein-4 (MPP4), a retina-specific scaffold protein, which has been implicated in organizing presynaptic protein complexes. Western blot analysis of bovine retinal anti-recoverin precipitates detects coprecipitating MPP4, supporting an association between the MPP4-containing protein complex and recoverin in vivo. However, immunoprecipitation experiments do not show a direct interaction between recoverin and MPP4 in 293-BNA cells cotransfected with both proteins (Förster et al. 2009). More recently, pull-down assay and surface plasmon resonance studies have revealed a neuron-specific Ca2+-binding protein caldendrin as a potential target for recoverin in retinal bipolar cells and pineal gland. In particular, both proteins are colocalized in these structures and an increase of intracellular calcium facilitates the translocation of caldendrin to intracellular membranes, which is under control of the complex formation with recoverin (Fries et al. 2010). Recent in vivo study provides insights into a role of recoverin in cones representing second type of photoreceptor cells. In mammalians, these cells express the same forms of recoverin and GRK1 as found in rods and exhibit similar alterations in response to recoverin knockout. Yet, in cones recoverin possesses stronger inhibition of the enzyme as evidenced by reduced amplitude of single photon response and flash sensitivity after recoverin removal. Furthermore, recoverin-deficient cones display more dominant role of recoverin in phototransduction inactivation under dim light conditions as compared to rods (Sakurai et al. 2015).
Recoverin in Cancer
In health, the expression of recoverin is mainly restricted within the retina. In cancer, recoverin can also be a paraneoplastic (or onconeural) antigen which is aberrantly expressed in tumors localized outside the nervous system. It is suggested that aberrant methylation of the recoverin gene region, overlapping the promoter upstream of the first exon and the first exon itself, controls expression of recoverin in cancer cells (Bazhin et al. 2010). The aberrant expression of recoverin in malignant cells causes an autoimmune response in some cancer patients followed by the development of paraneoplastic retina degeneration or cancer-associated retinopathy, CAR. Autoantibodies against recoverin (AAR) are detected in patients with different kinds of cancer (for reviews, see Adamus 2006; Bazhin et al. 2007).
A model of antibody-induced apoptosis of photoreceptor cells, underlying the CAR syndrome, has been proposed (for a review, see Adamus 2006). Serum AAR should be in sufficiently high titers to enter the eye and cause retinopathy. Circulating AAR cross the blood-retinal barrier, penetrate into retinal layers, and enter retinal cells by an active process of endocytosis. Once in the cell, AAR block the recoverin function in phototransduction resulting in the enhancement of rhodopsin phosphorylation and an increase in the concentration of intracellular calcium ions. The high intracellular calcium activates the mitochondria-dependent and caspase-9-dependent activation of caspase 3, leading to DNA fragmentation and cell death. Massive death of photoreceptor cells leads to retinal dysfunction and degeneration.
CAR syndrome, similar to other paraneoplastic neurological syndromes, is a very rare event: Its occurrence is of the order of 1%. However, AAR might appear much more frequently (Bazhin et al. 2004). An important feature of the CAR syndrome, as well as other paraneoplastic syndromes, is that it can be manifested long before the clinical diagnosis of the underlying tumor (for a review, see Adamus 2006). Such a feature of the CAR syndrome and corresponding AAR could be useful to clinicians to predict the future development of a particular cancer.
Recoverin, initially named “p26,” is a Ca2+-binding protein with a predominantly retinal localization, which belongs to the neuronal calcium sensor (NCS) protein family. The recoverin molecule consists of 201 amino acid residues and contains four potential EF-hand Ca2+-binding sites, of which only two – EF-hands 2 and 3 – are capable of binding calcium. The N-terminus of recoverin is acylated, mainly myristoylated. Due to the mechanism of Ca2+-myristoyl switch, compartmentalization of recoverin is changed from a soluble Ca2+-free form to a membrane-bound Ca2+-containing form, and vice versa, depending on an external calcium concentration. In the Ca2+-free form, the N-terminal myristoyl moiety of recoverin is buried into the hydrophobic pocket of the protein; on calcium binding, the myristoylated N-terminus is exposed, providing membrane association of recoverin. Recoverin is suggested to operate as a Ca2+-sensor of rhodopsin kinase (G-protein-coupled receptor kinase 1, GRK-1), which catalyzes phosphorylation and thus desensitization of the visual receptor rhodopsin. In cancer, recoverin can also be a paraneoplastic (or onconeural) antigen, the aberrant expression of which in malignant tumors of some patients causes an autoimmune response and the development of paraneoplastic retina degeneration or cancer-associated retinopathy. An important feature of the CAR syndrome and underlying autoantibodies against recoverin is that they can be detected long before the clinical diagnosis of the corresponding tumor. Such a feature of the autoantibodies could be useful to clinicians to predict the future development of a particular cancer.
- Adamus G. The role of recoverin in autoimmunity. In: Philippov PP, Koch KW, editors. Neuronal calcium sensor proteins. New York: Nova Science Publishers; 2006. p. 181–99.Google Scholar
- Bazhin AV, Savchenko MS, Shifrina ON, Demoura SA, Chikina SY, Jaques G, Kogan EA, Chuchalin AG, Philippov PP. Recoverin as a paraneoplastic antigen in lung cancer: the occurrence of anti-recoverin autoantibodies in sera and recoverin in tumors. Lung Cancer. 2004;44:193–8.PubMedCrossRefGoogle Scholar
- Philippov PP, Senin II, Koch K-W. Recoverin: a calcium-dependent regulator of the visual transduction. In: Philippov PP, Koch KW, editors. Neuronal calcium sensor proteins. New York: Nova Science Publishers; 2006. p. 139–51.Google Scholar
- Zernii EY, Komolov KE, Permyakov SE, Kolpakova T, Dell’Orco D, Poetzsch A, Knyazeva KL, Grigoriev II, Permyakov EA, Senin II, Philippov PP, Koch K-W. Involvement of recoverin C-terminal segment in recognition of the target enzyme rhodopsin kinase. Biochem J. 2011;435:441–50.PubMedCrossRefGoogle Scholar
- Zernii EY, Zinchenko DV, Vladimirov VI, Grigoriev II, Skorikova EE, Baksheeva VE, Lipkin VM, Philippov PP, Senin II. Ca2+-dependent regulatory activity of recoverin in photoreceptor raft structures: the role of caveolin-1. Biol Membr. 2013;30:380–6.Google Scholar
- Zernii EY, Nazipova AA, Gancharova OS, Kazakov AS, Serebryakova MV, Zinchenko DV, Tikhomirova NK, Senin II, Philippov PP, Permyakov EA, Permyakov SE. Light-induced disulfide dimerization of recoverin under ex vivo and in vivo conditions. Free Radic Biol Med. 2015;83:283–95.PubMedCrossRefGoogle Scholar