Crystal structure of the N-terminal domain of ryanodine receptor from the honeybee, Apis mellifera.

Ryanodine receptors (RyRs) are the molecular target of diamides, a new chemical class of insecticides. Diamide insecticides are used to control lepidopteran pests and were considered relatively safe for mammals and non-targeted beneficial insects, including honey bees. However, recent studies showed that exposure to diamides could cause long-lasting locomotor deficits of bees. Here we report the crystal structure of RyR N-terminal domain A (NTD-A) from the honeybee, Apis mellifera, at 2.5 Å resolution. It shows a similar overall fold as the RyR NTD-A from mammals and the diamondback moth (DBM), Plutella xylostella, and still several loops located at the inter-domain interfaces show insect-specific or bee-specific structural features. A potential insecticide-binding pocket formed by loop9 and loop13 is conserved in lepidopteran but different in both mammals and bees, making it a good candidate targeting site for the development of pest-selective insecticides. Furthermore, a conserved intra-domain disulfide bond was observed in both DBM and bee RyR NTD-A crystal structures, which explains their higher thermal stability compared to mammalian RyR NTD-A. This work provides a basis for the development of novel insecticides with better selectivity between pests and bees by targeting a distinct site on pest RyRs, which would be a promising strategy to overcome the current toxicity problem.

Ultrastructural Analysis of Self-Associated RyR2s.

In heart, type-2 ryanodine receptor (RyR2) forms discrete supramolecular clusters in the sarcoplasmic reticulum known as calcium release units (CRUs), which are responsible for most of the Ca(2+) released for muscle contraction. To learn about the substructure of the CRU, we sought to determine whether RyR2s have the ability to self-associate in the absence of other factors and if so, whether they do it in a specific manner. Purified RyR2 was negatively stained and imaged on the transmission electron microscope, and RyR2 particles closely associated were further analyzed using bias-free multivariate statistical analysis and classification. The resulting two-dimensional averages show that RyR2s can interact in two rigid, reproducible configurations: "adjoining", with two RyR2s alongside each other, and "oblique", with two partially overlapped RyR2s forming an angle of 12°. The two configurations are nearly identical under two extreme physiological Ca(2+) concentrations. Pseudo-atomic models for these two interactions indicate that the adjoining interaction involves contacts between the P1, SPRY1 and the helical domains. The oblique interaction is mediated by extensive contacts between the SPRY1 domains (domains 9) and P1 domains (domains 10) of both RyR2s and not through domain 6 as previously thought; in addition its asymmetric interface imposes steric constrains that inhibit the growth of RyR2 as a checkerboard, which is the configuration usually assumed, and generates new configurations, i.e., "branched" and "interlocked". This first, to our knowledge, structural detailed analysis of the inter-RyR2 interactions helps to understand important morphological and functional aspects of the CRU in the context of cardiac EC coupling.

Human cardiac ryanodine receptor: preparation, crystallization and preliminary X-ray ANALysis of the N-terminal region.

Human ryanodine receptor 2 (hRyR2) is a calcium ion channel present in the membrane of the sarcoplasmic reticulum of cardiac myocytes that mediates release of calcium ions from the sarcoplasmic reticulum stores during excitation- contraction coupling. Disease-causing mutations of hRyR2 are clustered into N-terminal (amino acids 1-600), central (amino acids 2100-2500) and C-terminal (amino acids 3900-5000) regions. These regions are believed to be involved in regulation of channel gating. The N-terminal region of hRyR2 has been implicated in regulating basal channel activity by interaction with the central hRyR2 region. This paper reports preparation, crystallization and preliminary X-ray analysis of recombinant hRyR2(1-606) N-terminal fragment. Soluble hRyR2(1-606) was expressed in Escherichia coli. Purification conditions were optimized using thermal shift assay. The quality and stability of the sample was probed by dynamic light scattering. A monomeric protein showing over 95% purity was obtained. The protein was crystallized by the hanging drop vapor-diffusion method. Diffraction data with resolution 2.39 Å were collected and processed.

Both basic and acidic amino acid residues of IpTx(a) are involved in triggering substate of RyR1.

Imperatoxin A (IpTx(a)) is known to modify the gating of skeletal ryanodine receptor (RyR1). In this paper, the ability of charged aa residues of IpTx(a) to induce substate of native RyR1 in HSR was examined. Our results show that the basic residues (e.g., Lys¹9, Lys²°, Lys²², Arg²³, and Arg²4) are important for producing substate of RyR1. In addition, other basic residues (e.g., Lys³°, Arg³¹, and Arg³³ near the C-terminus and some acidic residues (e.g., Glu²9, Asp¹³, and Asp²) are also involved in the generation of substate. Residues such as Lys8 and Thr²6 may be involved in the self-regulation of substate of RyR1, since alanine substitution of the aa residues led to a drastic conversion to the substate. The modifications of the channel gating by the wild-type and mutant toxins were similar in purified RyR1. Taken together, the specific charge distributions on the surface of IpTx(a) are essential for regulation of the channel gating of RyR1.

CLIC2-RyR1 interaction and structural characterization by cryo-electron microscopy.

Chloride intracellular channel 2 (CLIC2), a newly discovered small protein distantly related to the glutathione transferase (GST) structural family, is highly expressed in cardiac and skeletal muscle, although its physiological function in these tissues has not been established. In the present study, [3H]ryanodine binding, Ca2+ efflux from skeletal sarcoplasmic reticulum (SR) vesicles, single channel recording, and cryo-electron microscopy were employed to investigate whether CLIC2 can interact with skeletal ryanodine receptor (RyR1) and modulate its channel activity. We found that: (1) CLIC2 facilitated [3H]ryanodine binding to skeletal SR and purified RyR1, by increasing the binding affinity of ryanodine for its receptor without significantly changing the apparent maximal binding capacity; (2) CLIC2 reduced the maximal Ca2+ efflux rate from skeletal SR vesicles; (3) CLIC2 decreased the open probability of RyR1 channel, through increasing the mean closed time of the channel; (4) CLIC2 bound to a region between domains 5 and 6 in the clamp-shaped region of RyR1; (5) and in the same clamp region, domains 9 and 10 became separated after CLIC2 binding, indicating CLIC2 induced a conformational change of RyR1. These data suggest that CLIC2 can interact with RyR1 and modulate its channel activity. We propose that CLIC2 functions as an intrinsic stabilizer of the closed state of RyR channels.

Ryanodine receptor type-1 (RyR1) expression and protein S-nitrosylation pattern in human soleus myofibres following bed rest and exercise countermeasure.

The ryanodine receptor type-I (RyR1) is one key player of the excitation-contraction coupling (E-CC) machinery. However, RyR1 expression in human skeletal muscle disuse and plasticity changes are not well documented. We studied the expression and the functional modifications of RyR1 following prolonged bed rest (BR) without and with exercise countermeasure (Resistive Vibration Exercise, RVE). Soleus biopsies were taken from a non-trained control (BR-CTRL) and trained (BR-RVE) group (each n = 10) before and after BR. In BR-CTRL group, a fibre type-specific immunopattern of RyR1 (type-I < type-II) was documented, and RyR1 immunofluorescence intensity and protein expression together with [(3)H]ryanodine binding were decreased after BR. In BR-RVE group, RyR1 immunosignals were increased and fiber type specificity was no longer present. RyR1 protein expression was unchanged, whereas [(3)H]ryanodine binding increased after BR. Confocal and biochemical analysis confirmed subcellular co-localisation and protein-protein interaction of RyR1 with nitric oxide (NO)-synthase type-1 (NOS1). S-nitrosylation of RyR1 was increased in BR-CTRLpost only, suggesting a reduction of RyR1 open channel probability by nitrosylation mechanisms following prolonged disuse. We conclude that following extended body deconditioning in bed rest, RVE countermeasure maintained normal RyR1 expression and nitrosylation patterns required for adequate E-CC in human performance control.

Overexpression and purification of the RyR1 pore-forming region.

Ryanodine receptor 1 (RyR1) is a large homotetrameric calcium channel that plays a pivotal role in skeletal muscle contraction. Sequence comparison and mutagenesis studies indicate that the pore architecture of RyR1, including the last two transmembrane helices and the luminal loop linking them, is similar to that of the bacterial KcsA K(+) channel. Here, we describe the overexpression and purification of the C-terminal polyhistidine-tagged RyR1 pore-forming region. The nonionic detergent lauryldimethylamine oxide (LDAO) was selected for solubilization of the protein based on its ability to extract the protein from the membrane and to maintain it in a monodisperse state. The protein was then purified using nickel-affinity chromatography and gel filtration. Gel filtration analysis confirmed that the RyR1 fragment containing the pore-forming region (amino acids 4829-5037) is sufficient to form a tetramer.

Maurocalcine interacts with the cardiac ryanodine receptor without inducing channel modification.

We have previously shown that MCa (maurocalcine), a toxin from the venom of the scorpion Maurus palmatus, binds to RyR1 (type 1 ryanodine receptor) and induces strong modifications of its gating behaviour. In the present study, we investigated the ability of MCa to bind to and modify the gating process of cardiac RyR2. By performing pull-down experiments we show that MCa interacts directly with RyR2 with an apparent affinity of 150 nM. By expressing different domains of RyR2 in vitro, we show that MCa binds to two domains of RyR2, which are homologous with those previously identified on RyR1. The effect of MCa binding to RyR2 was then evaluated by three different approaches: (i) [(3)H]ryanodine binding experiments, showing a very weak effect of MCa (up to 1 muM), (ii) Ca(2+) release measurements from cardiac sarcoplasmic reticulum vesicles, showing that MCa up to 1 muM is unable to induce Ca(2+) release, and (iii) single-channel recordings, showing that MCa has no effect on the open probability or on the RyR2 channel conductance level. Long-lasting opening events of RyR2 were observed in the presence of MCa only when the ionic current direction was opposite to the physiological direction, i.e. from the cytoplasmic face of RyR2 to its luminal face. Therefore, despite the conserved MCa binding ability of RyR1 and RyR2, functional studies show that, in contrast with what is observed with RyR1, MCa does not affect the gating properties of RyR2. These results highlight a different role of the MCa-binding domains in the gating process of RyR1 and RyR2.

Postulated role of interdomain interaction between regions 1 and 2 within type 1 ryanodine receptor in the pathogenesis of porcine malignant hyperthermia.

We have demonstrated recently that CICR (Ca2+-induced Ca2+ release) activity of RyR1 (ryanodine receptor 1) is held to a low level in mammalian skeletal muscle ('suppression' of the channel) and that this is largely caused by the interdomain interaction within RyR1 [Murayama, Oba, Kobayashi, Ikemoto and Ogawa (2005) Am. J. Physiol. Cell Physiol. 288, C1222-C1230]. To test the hypothesis that aberration of this suppression mechanism is involved in the development of channel dysfunctions in MH (malignant hyperthermia), we investigated properties of the RyR1 channels from normal and MHS (MH-susceptible) pig skeletal muscles with an Arg615-->Cys mutation using [3H]ryanodine binding, single-channel recordings and SR (sarcoplasmic reticulum) Ca2+ release. The RyR1 channels from MHS muscle (RyR1MHS) showed enhanced CICR activity compared with those from the normal muscle (RyR1N), although there was little or no difference in the sensitivity to several ligands tested (Ca2+, Mg2+ and adenine nucleotide), nor in the FKBP12 (FK506-binding protein 12) regulation. DP4, a domain peptide matching the Leu2442-Pro2477 region of RyR1 which was reported to activate the Ca2+ channel by weakening the interdomain interaction, activated the RyR1N channel in a concentration-dependent manner, and the highest activity of the affected channel reached a level comparable with that of the RyR1MHS channel with no added peptide. The addition of DP4 to the RyR1MHS channel produced virtually no further effect on the channel activity. These results suggest that stimulation of the RyR1MHS channel caused by affected inter-domain interaction between regions 1 and 2 is an underlying mechanism for dysfunction of Ca2+ homoeostasis seen in the MH phenotype.

Aldolase potentiates DIDS activation of the ryanodine receptor in rabbit skeletal sarcoplasmic reticulum.

DIDS (4,4'-di-isothiocyanostilbene-2,2'-disulfonate), an anion channel blocker, triggers Ca2+ release from skeletal muscle SR (sarcoplasmic reticulum). The present study characterized the effects of DIDS on rabbit skeletal single Ca2+-release channel/RyR1 (ryanodine receptor type 1) incorporated into a planar lipid bilayer. When junctional SR vesicles were used for channel incorporation (native RyR1), DIDS increased the mean P(o) (open probability) of RyR1 without affecting unitary conductance when Cs+ was used as the charge carrier. Lifetime analysis of single RyR1 activities showed that 10 microM DIDS induced reversible long-lived open events (P(o)=0.451+/-0.038) in the presence of 10 microM Ca2+, due mainly to a new third component for both open and closed time constants. However, when purified RyR1 was examined in the same condition, 10 microM DIDS became considerably less potent (P(o)=0.206+/-0.025), although the caffeine response was similar between native and purified RyR1. Hence we postulated that a DIDS-binding protein, essential for the DIDS sensitivity of RyR1, was lost during RyR1 purification. DIDS-affinity column chromatography of solubilized junctional SR, and MALDI-TOF (matrix-assisted laser-desorption ionization-time-of-flight) MS analysis of the affinity-column-associated proteins, identified four major DIDS-binding proteins in the SR fraction. Among them, aldolase was the only protein that greatly potentiated DIDS sensitivity. The association between RyR1 and aldolase was further confirmed by co-immunoprecipitation and aldolase-affinity batch-column chromatography. Taken together, we conclude that aldolase is physically associated with RyR1 and could confer a considerable potentiation of the DIDS effect on RyR1.

Molecular nature of sulfhydryl modification by hydrogen peroxide on type 1 ryanodine receptor.

AIM: To elucidate the molecular nature of sulfhydryl modification by hydrogen peroxide on type 1 ryanodine receptor (RyR1). METHODS: Rabbit skeletal muscle sarcoplasmic reticulum was treated with hydrogen peroxide, then RyR1 complex was isolated. The proteins in the complex were analysed by electrophoresis, Western blot and electron microscopy. RESULTS: (1) Hydrogen peroxide induces inter-subunit cross-linking within the tetrameric RyR1 molecule; (2) in parallel to inter-subunit cross-linking, the RyR1 molecule changes morphology; (3) the chemical and morphological changes are reversible: upon reduction by reducing agents, the RyR1 molecule regains its original state. CONCLUSION: These findings suggest that the molecular mechanism of RyR1 channel activity in sarcoplasmic reticulum regulated by hydrogen peroxide is through inter-subunit cross-linking within the tetrameric RyR1 molecule, which in turn induces structural changes of RyR1.

Two-dimensional crystallization of the ryanodine receptor Ca2+ release channel on lipid membranes.

The ryanodine receptor (RyR) is the largest known membrane protein with a total molecular mass of 2.3 x 10(3) kDa. Well ordered, two-dimensional (2D) crystals are an essential prerequisite to enable RyR structure determination by electron crystallography. Conventionally, the 2D crystallization of membrane proteins is based on a 'trial-and-error' strategy, which is both time-consuming and chance-directed. By adopting a new strategy that utilizes protein sequence information and predicted transmembrane topology, we successfully crystallized the RyR on positively charged lipid membranes. Image processing of negatively stained crystals reveals that they are well ordered, with diffraction spots of IQ < or = 4 extending to approximately 20 angstroms, the resolution attainable in negative stain. The RyR crystals obtained on the charged lipid membrane have characteristics consistent with 2D arrays that have been observed in native sarcoplasmic reticulum of muscle tissues. These crystals provide ideal materials to enable structural analysis of RyR by high-resolution electron crystallography. Moreover, the reconstituted native-like 2D array provides an ideal model system to gain structural insights into the mechanism of RyR-mediated Ca2+ signaling processes, in which the intrinsic ability of RyR oligomers to organize into a 2D array plays a crucial role.

Voltage-sensitive equilibrium between two states within a ryanoid-modified conductance state of the ryanodine receptor channel.

We have investigated the influence of transmembrane holding potential on the kinetics of interaction of a cationic ryanoid, 8beta-amino-9alpha-hydroxyryanodine, with individual ryanodine receptor (RyR) channels and on the functional consequences of this interaction. In agreement with previous studies involving cationic, neutral, and anionic ryanoids, both rates of association and dissociation of the ligand are sensitive to transmembrane potential. A voltage-sensitive equilibrium between high- and low-affinity forms of the receptor underlies alterations in rates of association and dissociation of the ryanoid. The interaction of 8beta-amino-9alpha-hydroxyryanodine with RyR influences the rate of cation translocation through the channel. With this ryanoid bound, the channel fluctuates between two clearly resolved subconductance states (alpha and beta). We interpret this observation as indicating that with 8beta-amino-9alpha-hydroxyryanodine bound, the pore of the RyR channel exists in two essentially isoenergetic conformations with differing ion-handling properties. The equilibrium between the alpha- and beta-states of the RyR-8beta-amino-9alpha-hydroxyryanodine complex is sensitive to transmembrane potential. However, the mechanisms determining this equilibrium differ from those responsible for the voltage-sensitive equilibrium between high- and low-affinity forms of the receptor.

Caffeine sensitivity of native RyR channels from normal and malignant hyperthermic pigs: effects of a DHPR II-III loop peptide.

Enhanced sensitivity to caffeine is part of the standard tests for susceptibility to malignant hyperthermia (MH) in humans and pigs. The caffeine sensitivity of skeletal muscle contraction and Ca(2+) release from the sarcoplasmic reticulum is enhanced, but surprisingly, the caffeine sensitivity of purified porcine ryanodine receptor Ca(2+)-release channels (RyRs) is not affected by the MH mutation (Arg(615)Cys). In contrast, we show here that native malignant hyperthermic pig RyRs (incorporated into lipid bilayers with RyR-associated lipids and proteins) were activated by caffeine at 100- to 1000-fold lower concentrations than native normal pig RyRs. In addition, the results show that the mutant ryanodine receptor channels were less sensitive to high-affinity activation by a peptide (C(S)) that corresponds to a part of the II-III loop of the skeletal dihydropyridine receptor (DHPR). Furthermore, subactivating concentrations of peptide C(S) enhanced the response of normal pig and rabbit RyRs to caffeine. In contrast, the caffeine sensitivity of MH RyRs was not enhanced by the peptide. These novel results showed that in MH-susceptible pig muscles 1). the caffeine sensitivity of native RyRs was enhanced, 2). the sensitivity of RyRs to a skeletal II-III loop peptide was depressed, and 3). an interaction between the caffeine and peptide C(S) activation mechanisms seen in normal RyRs was lost.

Rapid determination of advanced glycation end products of proteins using MALDI-TOF-MS and PERL script peptide searching algorithm.

Advanced glycation end products (AGEs), which are composed of various glucose or carbohydrate adducts, are thought to be responsible for several diabetic and age-related complications. However, to date, specific sites on proteins that are modified by AGEs remain largely unknown. We report here the use of matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF-MS) to determine the type and localization of several AGEs formed in vitro on human beta-2-microglobulin (beta2M), and in vivo on type 2 ryanodine receptor calcium-release channel (RyR2), and sarco(endo)plasmic reticulum (SERCA2a). A PERL script algorithm, developed in-house, makes searching the relatively large amount of data generated by the MALDI-MS more manageable. The outstanding sensitivity of MALDI-TOF-MS coupled with the PERL script algorithm allows such an approach to be a very useful tool in detecting AGEs and other post-translational modifications. We believe that this method could be an important tool when searching for post-translational modifications on proteins.

Enrichment of triadic and terminal cisternae vesicles from rabbit skeletal muscle.

An enriched triad and terminal cisternae preparation was achieved from skeletal muscle through alterations of the differential centrifugation and muscle homogenization protocols. Both yield and specific activity (pmoles of radioligand binding per mg protein) were optimized for (3)H-PN200-110 (transverse tubule marker) and (3)H-ryanodine (terminal cisternae marker) binding sites. By pelleting crude microsomes between 2,000 an 12,000 x g without any rehomogenizations, we improved both the yield and specific activity of transverse tubule and terminal cisternae markers in crude microsomes by approximately 4-fold to 1000-3000 pmoles binding sites (starting material: approximately 400 grams wet weight fast twitch skeletal muscle), with 10-15 pmoles/mg. Rehomogenization of the 1,000 x g pellet, which is typically discarded, allowed recovery of an additional 5000 pmoles PN200-110 binding sites and an additional 8000 pmoles ryanodine binding sites. Crude microsomes from the rehomogenized 1,000 x g pellets typically displayed specific activities of 20-25 pmoles binding/mg for both (3)H-PN200-110 and (3)H-ryanodine. Separation of crude microsomes on a sucrose gradient increased specific activity up to a maximum of 50 pmoles/mg in a specific fraction, a five- to ten-fold increase over standard triadic or terminal cisternae preparations. The mean specific activity for enriched triads was 30-40 pmoles/mg for both PN200-110 and ryanodine in pooled fractions, while pooled fractions of enriched terminal cisternae displayed low (3)H-PN200-110 binding (3-5 pmoles/mg) and high (3)H-ryanodine-specific activity (30-40 pmoles/mg).

Doxorubicin directly binds to the cardiac-type ryanodine receptor.

The clinical use of doxorubicin, an antineoplasmic agent, is limited by its extensive cardiotoxicity which is mediated by the mobilization of intracellular Ca2+ from SR. In order to elucidate the mechanism of Ca2+ release, we analyzed the binding sites of doxorubicin on rabbit cardiac SR (sarcoplasmic reticulum). One of the binding sites was identified as cardiac-type ryanodine receptor (RyR2) which was purified by immunoprecipitation from solubilized cardiac SR in the presence of DTT. Ligand blot analysis revealed the direct binding of doxorubicin to RyR2. The binding of doxorubicin to RyR2 was specific and displaced by caffeine. Both doxorubicin and caffeine enhanced [3H]-ryanodine binding to RyR2 in a Ca2+ dependent manner. These results suggest that there is a doxorubicin binding site on RyR2.

A novel and rapid approach to isolating functional ryanodine receptors.

Conventional methods of isolating and reconstituting ryanodine receptors (RyRs) from native membranes into proteoliposomes take a minimum of 2 days to complete. We have developed an alternative strategy that can be used to isolate and reconstitute functional RyRs in just 3 h with a similar degree of purification. RyRs isolated by this method display characteristic functional behaviour as assessed by radioligand binding and single channel analyses.

Identification of functional type 1 ryanodine receptors in mouse dendritic cells.

Ca(2+) signaling plays an important role in the function of dendritic cells (DC), the specialized antigen-presenting cells of the immune system. Here we describe functional ryanodine receptor (RyR) Ca(2+) release channels in murine, bone marrow-derived DC. RT-PCR analysis identified selective expression of the type 1 RyR, with higher levels detected in immature rather than mature DC. The RyR activators caffeine, FK506, ryanodine and 4-chloro-m-cresol mobilized Ca(2+) in DC, and responses to 4-chloro-m-cresol were inhibited by dantrolene. Furthermore, activation of RyRs both inhibited subsequent inositol trisphosphate-mediated Ca(2+) release and provoked store-operated Ca(2+) entry, suggesting a functional interaction between these intracellular Ca(2+) channels. Thus, the RyR1 channel may play an intrinsic role in Ca(2+) signaling in DC.

RyR3 amplifies RyR1-mediated Ca(2+)-induced Ca(2+) release in neonatal mammalian skeletal muscle.

The neonatal mammalian skeletal muscle contains both type 1 and type 3 ryanodine receptors (RyR1 and RyR3) located in the sarcoplasmic reticulum membrane. An allosteric interaction between RyR1 and dihydropyridine receptors located in the plasma membrane mediates voltage-induced Ca(2+) release (VICR) from the sarcoplasmic reticulum. RyR3, which disappears in adult muscle, is not involved in VICR, and the role of the transiently expressed RyR3 remains elusive. Here we demonstrate that RyR1 participates in both VICR and Ca(2+)-induced Ca(2+) release (CICR) and that RyR3 amplifies RyR1-mediated CICR in neonatal skeletal muscle. Confocal measurements of intracellular Ca(2+) in primary cultured mouse skeletal myotubes reveal active sites of Ca(2+) release caused by peripheral coupling between dihydropyridine receptors and RyR1. In myotubes lacking RyR3, the peripheral VICR component is unaffected, and RyR1s alone are able to support inward CICR propagation in most cells at an average speed of approximately 190 microm/s. With the co-presence of RyR1 and RyR3 in wild-type cells, unmitigated radial CICR propagates at 2,440 microm/s. Because neonatal skeletal muscle lacks a well developed transverse tubule system, the RyR3 reinforcement of CICR seems to ensure a robust, uniform, and synchronous activation of Ca(2+) release throughout the cell body. Such functional interplay between RyR1 and RyR3 can serve important roles in Ca(2+) signaling of cell differentiation and muscle contraction.