N-terminal and Central Segments of the Type 1 Ryanodine Receptor Mediate Its Interaction with FK506-binding Proteins

Background: The 12-kDa binding protein for the immunosuppressant, FK506 (FKBP), modulates type 1 ryanodine receptor (RyR1) activity via unknown RyR1 sequence determinants.

Results: Polyhistidine tags inserted in N-terminal and central RyR1 domains abolish FKBP binding, whereas high fluorescence resonance energy transfer is observed between FKBP and both domains.

Conclusion: FKBP binds to determinants within both domains.

Significance: Results support domain-switch malignant hyperthermia model.

Abstract

We used site-directed labeling of the type 1 ryanodine receptor (RyR1) and fluorescence resonance energy transfer (FRET) measurements to map RyR1 sequence elements forming the binding site of the 12-kDa binding protein for the immunosuppressant drug, FK506. This protein, FKBP12, promotes the RyR1 closed state, thereby inhibiting Ca²⁺ leakage in resting muscle. Although FKBP12 function is well established, its binding determinants within the RyR1 protein sequence remain unresolved. To identify these sequence determinants using FRET, we created five single-Cys FKBP variants labeled with Alexa Fluor 488 (denoted D-FKBP) and then targeted these D-FKBPs to full-length RyR1 constructs containing decahistidine (His₁₀) “tags” placed within N-terminal (amino acid residues 76–619) or central (residues 2157–2777) regions of RyR1. The FRET acceptor Cy3NTA bound specifically and saturably to these His tags, allowing distance analysis of FRET measured from each D-FKBP variant to Cy3NTA bound to each His tag. Results indicate that D-FKBP binds proximal to both N-terminal and central domains of RyR1, thus suggesting that the FKBP binding site is composed of determinants from both regions. These findings further imply that the RyR1 N-terminal and central domains are proximal to one another, a core premise of the domain-switch hypothesis of RyR function. We observed FRET from GFP fused at position 620 within the N-terminal domain to central domain His-tagged sites, thus further supporting this hypothesis. Taken together, these results support the conclusion that N-terminal and central domain elements are closely apposed near the FKBP binding site within the RyR1 three-dimensional structure.

Introduction

The ryanodine receptor (RyR)² is an intracellular Ca²⁺ release channel that plays a central role in muscle excitation-contraction (EC) coupling. Muscle-specific isoforms of the RyR, type 1 (RyR1) and type 2 (RyR2), control EC coupling in skeletal and cardiac muscle, respectively (1). Activation of these massive, homotetrameric proteins embedded within the sarcoplasmic reticulum releases Ca²⁺ from intracellular stores and thereby initiates muscle contraction. Although point mutations within RyR1 can result in human skeletal muscle disorders such as malignant hyperthermia (MH), central core disease, and multi-minicore disease, little is understood as to how these amino acid substitutions alter the structure of RyR1 and thereby result in these muscle disorders. One prominent structural hypothesis (2) suggests that N-terminal and central portions of RyR1 physically contact one another to form a “domain switch” that is disrupted by mutations, thereby leading to the uncontrolled Ca²⁺ release that is characteristic of these syndromes.

RyR1 serves as a scaffold for multiple accessory proteins, including the 12-kDa FK506-binding protein (FKBP12) (3). This small protein modulates RyR1 activity by promoting the closed state of the channel, thereby reducing the incidence of subconductance states (4, 5). FKBP12 is the only species expressed in skeletal muscle, but a small level of a closely related isoform, FKBP12.6, is expressed in heart where it tightly associates with RyR2 (6). Both FKBP isoforms possess cis-trans-prolyl isomerase activity in vitro (7, 8) and bind to RyRs with very high affinity (K_d < 1 nm) (9, 10). Although the FKBP12 isoform is predominantly expressed in mammalian skeletal muscle (6), FKBP12.6 can exchange with RyR1-bound FKBP12 (11) and thus behaves much like a constitutive subunit of RyR1 (12).

The location and orientation of bound FKBP on the RyR three-dimensional map have been well characterized by a combination of biophysical methods. Using cryo-electron microscopic (EM) single particle analysis, bound FKBP12 has been localized to a position between the “clamp” and “handle” domains of the large cytoplasmic “foot” structure of RyR1 (13, 14). In addition, FRET measurements indicate that the orientations of bound FKBP12 and FKBP12.6 on either RyR1 or RyR2 are similar (15), thus suggesting that the FKBP binding sites on each RyR isoform are in similar locations. However, despite this detailed understanding of the structural nature of the RyR1 FKBP binding site, the location of this site within the RyR1 protein sequence is controversial.

To date, three different FKBP binding locations within the RyR1/RyR2 sequence have been proposed. The first location, initially based on two-hybrid analysis of FKBP12 binding to the inositol trisphosphate receptor (16), places FKBP proximal to a valine-proline motif located at RyR1 position 2461 (17) as mutation of this valine disrupts FKBP binding. However, the equivalent mutation in RyR2 does not affect FKBP12.6 binding (18). Instead, an N-terminal FKBP binding site has been proposed based on sequence deletions that abolish FKBP12.6 binding to RyR2 (18). Finally, findings from a third group suggest the presence of an FKBP binding site in the C-terminal transmembrane assembly of RyR2, as fragments derived from this region can bind FKBP12.6 in vitro (19). However, a C-terminal FKBP binding site appears unlikely given the cytoplasmic localization of FKBP observed in cryo-EM reconstructions, which is >100 Å from the transmembrane assembly of RyR1 (13, 14).

In this study we utilized a cell-based FRET approach (20) to define binding determinants of FKBP in the RyR1 sequence. Using FKBP12.6 labeled with FRET donor, energy transfer was measured to a FRET acceptor targeted to 10-residue histidine (His₁₀) tags engineered into either N-terminal or central positions of RyR1. We observed strong FRET from labeled FKBP to His₁₀ tags at both the N-terminal position 519 and central position 2341 insertion sites within RyR1, suggesting that the bound FKBP12.6 is within 50–60 Å from these positions. Moreover, insertion of His₁₀ tags at N-terminal position 619 or central domain positions 2157 and 2502 completely disrupted FKBP binding to the RyR, suggesting that these positions are even closer to, or may be part of, the FKBP binding site. Taken together, these results suggest that the FKBP binding site on RyR1 consists of both N-terminal and central sequence elements. Moreover, the physical proximity of these elements to the labeled FKBP and, therefore, to each other lends support for the domain switch hypothesis of MH pathogenesis.

EXPERIMENTAL PROCEDURES

Synthesis and Purification of FRET Donors

Fluorescent derivatives of FKBP12.6 were prepared as previously described (15). A maleimide derivative of Alexa Fluor 488 (AF488; Invitrogen) was used to label single cysteines substituted at positions 14, 32, 44, 49, and 85 into a null-cysteine variant of human FKBP12.6. These AF488-labeled FKBPs were used as FRET donors and were denoted D-FKBP, or D14-, D32-, D44-, D49-, and D85-FKBP to indicate the labeling site.

Synthesis and Purification of FRET Acceptors

Cy3NTA was synthesized and purified using thin-layer chromatography as described previously (20). Yields quantified spectrophotometrically (Cy3 ϵ₅₅₀ = 150,000 m⁻¹cm⁻¹) were normally ∼40% of starting material. Before use in FRET experiments, a dried, 10-nmol aliquot of Cy3NTA was charged with 20 nmol of NiCl₂ as described (20, 21).

cDNA Cloning

Full-length cDNA constructs containing His₁₀ tags in the N-terminal domain of RyR1 were created via excision of green fluorescent protein (GFP) cDNA from previously constructed GFP-RyR1 fusion cDNAs containing His₁₀ tags at positions 76, 181, 290, 519, or 619 (21). His tags were placed into central domain positions 2157, 2341, 2502, and 2777 using DNA oligos substituted into unique restriction sites in a RyR1 subclone ranging from position 4935 to 8331 of the rabbit RyR1 cDNA. All oligos had a core sequence consisting of 5′-CACCATCACCATCACCATCACCATCACCATGGATATC-3′, which encoded the protein sequence His₁₀GlyTyr and contained an EcoRV restriction sequence (in bold) to track successful clones. Flanking sequences were then added to the core sequence above to ensure formation of sticky ends necessary to ligate the oligos into unique restriction sites at cDNA positions 6471 (XhoI), 7023 (HpaI), 7506 (NcoI), and 8331 (BsiW1) corresponding to protein sequence positions 2157, 2341, 2502, and 2777, respectively. The proper insertion and reading frame of these His tag-encoding cDNAs were confirmed via DNA sequencing and restriction digest analysis.

Cell Culture

HEK-293T cells were grown and transiently transfected with His-tagged RyR1 cDNAs using polyethyleneimine as described (20). Cells were used in functional or FRET-based assays 2 days after transfection.

FRET Imaging

HEK-293T cells expressing His-tagged RyR1 constructs were permeabilized in FRET buffer consisting of 125 mm NaCl, 5 mm KCl, 6 mm glucose, 25 mm HEPES, pH 7.6, supplemented with 0.1% saponin, 3 μm Cy3NTA, and 10 nm D-FKBP. The cells were then incubated at 37 °C for 30 min to allow entry of the fluorophores. Next, cells were imaged using a Nikon 60× 1.20 NA water immersion objective housed within a Nikon Eclipse TE2000-U inverted microscope. Donor fluorescence was recorded via illumination with a YFP cubeset consisting of a 480/30-nm bandpass excitation filter, a 505-nm dichroic mirror, and a 535/40-nm bandpass emission filter (Chroma Technology, Bellows Falls, VT) using a 300-watt xenon lamp housed within a Lambda DG-4 lightsource (Sutter Instruments, Novato, CA). Initial donor fluorescence was recorded using a Stanford Photonics XR-Mega 10 CCD camera (Stanford Photonics, Palo Alto, CA) as a series of 60 16-bit 672 × 516-pixel images across a Z-stack 60 μm in thickness. FRET was quantified after selective photobleaching of Cy3NTA for 4 min at maximum light output using a ReAsH filter cubeset consisting of a 570/20-nm bandpass emission filter, 585-nm dichroic mirror, and a 620/60-nm emission filter (Chroma). Donor fluorescence of the cells was then re-measured as above, and the FRET efficiency (E) was calculated from the resulting increase in fluorescence after the bleaching of Cy3NTA according to Equation 1,

where F_prebleach and F_postbleach are donor fluorescence intensities before and after photobleaching of Cy3NTA. Donor fluorescence was quantified from background-corrected images derived from these experiments using the ImageJ image processing and analysis software version 1.45m (National Institutes of Health). FRET efficiencies were converted to intramolecular distances using Equation 2,

where R represents the distance between donor and acceptor, R₀ represents the Förster distance for the AF488/Cy3 FRET pair (59 Å), and E represents the measured FRET efficiency. Distance calculations assumed random orientation (κ² = 2/3) between donor and acceptor, an assumption supported by the agreement of distance triangulations from a redundant number of donor positions (see “Results”).

Prediction of FRET Efficiencies to Known RyR1 Positions

Measured FRET efficiencies were compared with predicted FRET values for known donor/acceptor positions (when possible) as follows. The atomic structures of the RyR1 N-terminal ABC domain (PDB ID 2XOA (22)) and FKBP12 (PDB ID 1D6O (23)) were docked to the cryo-EM reconstruction (24) of the closed state of RyR1 (EM DataBank ID 1606 (25)). This was achieved by manually positioning the structures in the RyR1 map as previously published, then fitting in the density using the Fit in Map function of UCSF Chimera (24). The fit was optimized for correlation with a map simulated from the FKBP coordinates using the same resolution as the RyR map, i.e. 10.2 Å. All other settings were as default. Using Chimera, distances were measured from the Cβ of residues substituted by cysteines in FKBP12.6 (see above) to β carbons of insertion sites of His tags within ABC domains docked to the two subunits closest to the docked FKBP. Then the predicted FRET efficiency to each subunit was calculated according to

where R₀ represents the Förster distance for AF488/Cy3NTA, and R represents the measured distance between the two positions. Predicted FRET values were then compared with the measured FRET efficiency values in Table 1, and R₀ = 59 Å was found to minimize the aggregate difference between prediction and measurement.

TABLE 1.

Summary of FRET values and calculated distances between D-FKBP and His tags in the RyR1 N-terminal domain

N/A, not applicable.

Donor Positiona	FRET, S1 (predicted)b	FRET, S2 (predicted)b	FRET (observed)c	Nd	Apparent D-A distancee
wtRyR1
14	N/A	N/A	0.04 ± 0.09	55	N/A
32			0.08 ± 0.07	21
44			0.02 ± 0.06	43
49			0.00 ± 0.08	48
85			0.05 ± 0.09	24
His76
14	0.02	0.03	0.09 ± 0.06	45	N/Sf
32	0.10	0.06	0.33 ± 0.10	21	66 (61–71)
44	0.06	0.07	0.14 ± 0.07	36	78 (73–87)
49	0.03	0.07	0.13 ± 0.07	46	80 (74–91)
85	0.07	0.13	0.25 ± 0.11	15	70 (64–78)
His181
14	0.02	0.03	0.06 ± 0.09	52	N/Sf
32	0.14	0.08	0.31 ± 0.10	10	68 (62–73)
44	0.06	0.06	0.09 ± 0.06	35	86 (79–101)
49	0.04	0.07	0.07 ± 0.08	32	90 (79–101)
85	0.12	0.19	0.23 ± 0.08	14	72 (67–79)
His290
14	0.04	0.01	0.16 ± 0.11	38	N/Dg
32	0.21	0.02	0.29 ± 0.12	18
44	0.10	0.01	0.11 ± 0.13	59
49	0.07	0.01	0.16 ± 0.09	45
85	0.28	0.03	0.18 ± 0.08	34
His519
14	0.20	0.01	0.18 ± 0.07	41	75 (70–83)
32	0.86	0.01	0.58 ± 0.08	30	56 (53–59)
44	0.50	0.01	0.25 ± 0.04	39	70 (68–73)
49	0.33	0.01	0.22 ± 0.04	46	73 (70–76)
85	0.82	0.03	0.39 ± 0.09	43	63 (60–67)

^a Amino acid residue attachment site of AF488 on FKBP.

^b Predicted FRET efficiency to His tag insertion site on ABC domain docked to two nearest subunits (S1 and S2, in Fig. 5A).

^c Observed FRET efficiency (mean ± S.D.).

^d Number of measurements.

^e Average and range (in parentheses) of donor-acceptor distances (in Å) calculated from the mean and S.D. of the measured FRET efficiencies.

^f FRET value non-significant (p < 0.01) compared to wtRyR1 tested with equivalent D-FKBP.

^g FRET values not used for distance calculations due to fractional Cy3NTA binding to His²⁹⁰ tag.

Spatial Triangulation of His₁₀-tagged Positions

For each acceptor-labeled His tag, we acquired FRET measurements from five FKBP residues (14, 32, 44, 49, and 85). Based on the assumption that probes are randomly distributed around their point of attachment, we set the effective donor position at the Cβ atom of the labeled residue on FKBP. These donor coordinates and FRET-derived distances served as initial constraints to calculate a locus in space corresponding to the acceptor position. We developed software that performs this analysis in several steps. First, uniformly spaced coordinate points were placed on the surfaces of a set of concentric spheres of radii ranging from the minimum to the maximum FRET-derived distances for a particular D-A pair. The density of points on each sphere was 1 or 0.4 Å⁻². The spacing between concentric spheres was set to 1 or 2 Å. When no FRET was detected for a particular D-A pair, it was assumed that the distance was just outside the limit of detectability and the minimum radius was set to 105 Å. Then, spheres corresponding to all five FRET measurements were intersected by retaining coordinate points that were within a threshold distance (which we set at 2 Å for these calculations).

We further constrained the resultant loci by excluding coordinate points that were located >20 Å from the RyR surface. EM map conversion and manipulations to allow these calculations were done with the SITUS software package.

Intracellular Ca²⁺ Imaging

To test His-tagged RyR1 constructs for function, caffeine-induced changes in intracellular Ca²⁺ were measured using Fluo-4 based Ca²⁺ imaging as described previously (20). Normalized Ca²⁺ responses at each caffeine concentration were averaged and plotted against the individual caffeine concentrations, and the EC₅₀ values for each RyR1 construct were then calculated from these curves using Prism 5.0 software (GraphPad Inc., San Diego, CA), as described previously (21). A one-way analysis of variance was used to compare the EC₅₀ values followed by a Dunnett's post-test. A p < 0.05 was considered to be a significant difference in EC₅₀ values derived from each His-tagged construct relative to wtRyR1.

FKBP12.6/RyR1 Co-localization

Relative D-FKBP binding to the various His-tagged RyR1 constructs was determined by comparing D-FKBP fluorescence intensity to the expression level of each construct determined using anti-RyR immunocytochemistry as follows. HEK-293T cells expressing His-tagged constructs were loaded with 10 nm D14-FKBP for 30 min as described above. After washing cells 3 times with phosphate-buffered saline (PBS; 137 mm NaCl, 2.7 mm KCl, 10 mm Na₂HPO₄, 1.8 mm KH₂PO₄, pH 7.4), cells were fixed in 0.5% formalin for 30 min at room temperature. Cells were then washed 3 times in PBS followed by a 30-min incubation in blocking solution consisting of 2.5% normal goat serum and 0.1% saponin. After washing, the cells were treated for 30 min with 34C anti-RyR monoclonal antibody (Developmental Studies Hybridoma Bank, Iowa City, IA) (26) diluted 1:200. The cells were then washed 3 times with PBS and then exposed for 30 min to a rhodamine-conjugated goat anti-mouse secondary antibody (Sigma) diluted 1:2000. After washing, cellular fluorescence was visualized at 60× magnification using either YFP (for D-FKBP) or ReAsH (for RyR1 immunocytochemistry) cubesets (see FRET imaging). Fluorescence was quantified from 60-image stacks acquired under identical conditions using the ImageJ software as described for FRET imaging above. The ratio of D-FKBP fluorescence to RyR1 immunofluorescence was calculated and normalized to the corresponding ratio derived from wtRyR1-expressing cells. Pseudo-colored images were presented as Z-projections that depict the Z-stack maximum fluorescence values at each image pixel, thus enabling cells at different Z-planes to be displayed in the same image.

RESULTS

Experimental Approach

To identify RyR1 sequence determinants responsible for FKBP binding, we measured FRET from RyR1-bound AF488-FKBP12.6 variants to Cy3NTA (Fig. 1A) bound to His tags inserted at discrete positions into each of two RyR1 regions, an N-terminal domain spanning RyR1 residues 76–620 and a central domain spanning residues 2157–2777. Within the RyR1 cryo-EM map, FKBP12.6 binds to a well defined site between the clamp and handle structural subdomains (Fig. 1B) in the same location and orientation as FKBP12, the naturally occurring FKBP isoform in skeletal muscle (15). However, because FKBP12.6 dissociates from RyR1 extremely slowly, and has an almost identical backbone structure as FKBP12, it has been used for FRET-based analysis of RyR structure in previous studies (12, 15, 27, 28) as well as in the current study.

FIGURE 1.

A, shown is a schematic of the FRET-based assay used in this study. AF488-FKBP acting as a FRET donor can transfer energy to Cy3NTA acting as FRET acceptor when bound to a His₁₀ tag inserted within the RyR1 primary structure (black bar). The degree of energy transfer is indicative of the relative spatial proximity of these fluorophores targeted to RyR1. B, shown are positions of surface representations of the FKBP atomic structure (yellow; Ref. 23) and ribbon diagrams of the N-terminal ABC domains (22) within each of the four subunits of the cryo-EM reconstruction of wtRyR1 (25) as viewed from the top (i.e. the portion that would face the t-tubule in situ). Scale bar, 50 Å. C, shown is a ribbon diagram of the atomic structure of FKBP in same orientation as surface representation of the FKBP atomic structure #1 in B. Positions substituted by cysteine and then labeled with AF488 donor are indicated in red. Scale bar, 10 Å. D, shown are HEK-293T cells expressing wtRyR1 incubated with 10 nm D14-FKBP (left panel) and then analyzed for RyR1 immunoreactivity (middle panel). The right panel is an overlay of the two channels.

The FRET donor (AF488) was covalently attached to unique cysteine residues substituted via site-directed mutagenesis at FKBP12.6 positions 14, 32, 44, 49, or 85, which are located on the surface of the protein and are widely separated in three dimensions (Fig. 1C). D-FKBP bound in situ to wtRyR1 expressed in HEK-293T cells (Fig. 1D) at concentrations as low as 1 nm, whereas no labeling of control HEK-293T cells was observed, even using D-FKBP concentrations as high as 33 nm (data not shown).

Analysis of D-FKBP Binding to RyR1 Constructs with His Tags in the N-terminal Domain

To evaluate interactions between D-FKBP and sequence elements within the first 620 amino acids of RyR1, we assessed the binding of 10 nm D-FKBP to full-length RyR1 constructs containing single His₁₀ insertions at either positions 76, 181, 290, 519, or 619 (Fig. 2A). For wtRyR1 and constructs His⁷⁶, His¹⁸¹, His²⁹⁰, and His⁵¹⁹, D-FKBP fluorescence and RyR1 immunofluorescence co-localized and had similar intensity ratios (Fig. 2, B and C). However, despite abundant expression of His⁶¹⁹, no D-FKBP fluorescence was detected for this construct (Fig. 2, B and C) even when D-FKBP levels as high as 100 nm were tested (data not shown), thus indicating that the His tag at position 619 disrupts D-FKBP binding.

FIGURE 2.

A, shown are the positions of His tags inserted into the RyR1 N-terminal domain (gray bar). Black dots represent positions and numbers of MH mutations within the RyR1 protein sequence. AA, amino acids. B and C, shown is D-FKBP binding analysis via fluorescence microscopy. B, shown is the relative fluorescence ratio of D-FKBP:RyR1 immunoreactivity for the indicated His-tagged RyR1 constructs. Values represent the mean ± S.E. for the number of cells indicated in each bar. C, representative images of HEK-293T cells expressing either His⁷⁶, His¹⁸¹, His²⁹⁰, His⁵¹⁹, or His⁶¹⁹ RyR1 constructs labeled with 10 nm D14-FKBP (left panels, green) and then fixed and stained for RyR1 immunoreactivity (middle panels, red). The overlays (right panels) indicate overlap between the two fluorophores.

Functional Analysis of RyR1 Constructs Containing His₁₀ Tags in the N-terminal Domain

All N-terminal domain His-tagged RyR1 constructs released Ca²⁺ in response to the RyR agonist, caffeine (Fig. 3 A and B). The EC₅₀ values for caffeine activation were statistically unchanged relative to wtRyR1 (EC₅₀ = 1.06 mm) except for His⁶¹⁹, which had a significantly enhanced EC₅₀ = 0.35 mm (Fig. 3D). Therefore, these His-tagged RyR1 constructs function as intracellular Ca²⁺ release channels, as we have shown previously (21). The enhanced caffeine sensitivity observed for the His⁶¹⁹ construct may be related to the proximity of its His tag to the well characterized human R614C MH mutation (29), which also sensitizes RyR1 to caffeine activation.

FIGURE 3.

The concentration dependence of caffeine-induced Ca²⁺ release is shown for HEK-293T cells expressing either wtRyR1, constructs containing His tags in the N-terminal domain (A and B), or constructs containing His tags in the central domain (C). All His-tagged constructs released Ca²⁺ in response to caffeine except His²⁷⁷⁷. D, EC₅₀ values for caffeine activation of each RyR construct, 95% confidence interval of the measured EC₅₀ values, and number of measurements (N) are shown. Asterisks indicate EC₅₀ values significantly different relative to wtRyR1 (p < 0.05).

FRET Measurements to the RyR1 N-terminal Domain

In control experiments, we first measured FRET from each of the five D-FKBPs to Cy3NTA when these fluorophores were both targeted to wtRyR1. No saturable dependence of the measured FRET efficiency relative to the added Cy3NTA concentration was observed (Fig. 4A, inset), whereas low FRET efficiencies ranging from 0.00 (for D49-FKBP) to 0.08 (for D32-FKBP) were measured using 3 μm Cy3NTA as FRET acceptor. These low background FRET efficiencies suggest that nonspecific binding sites for Cy3NTA on RyR1 (if they exist) are far from the D-FKBP donor and thus should not contribute to FRET measurements made using Cy3NTA targeted to His tags.

FIGURE 4.

FRET efficiency values from each of five donor-labeled positions on FKBP (indicated on x axes) to the Cy3NTA acceptor after both probes were targeted to wtRyR1 or constructs containing His tags in the N-terminal domain. A, FRET efficiency values from each FKBP to Cy3NTA incubated with wtRyR1. B, Cy3NTA titration curves for FRET from D85-FKBP to Cy3NTA targeted to each of the His-tagged constructs. EC₅₀ values and Hill coefficients (K^H) are indicated for each construct. C–F, FRET efficiency values from each FKBP to Cy3NTA targeted to either His⁷⁶ (C), His¹⁸¹ (D), His²⁹⁰ (E), or His⁵¹⁹ (F). Values represent the mean and S.E. for the number of cells indicated in Table 1.

Next, we measured FRET from each of the five RyR1-bound D-FKBPs to Cy3NTA targeted to His⁷⁶, His¹⁸¹, His²⁹⁰, and His⁵¹⁹. We observed saturable FRET efficiencies relative to the added Cy3NTA concentration for each His-tagged construct except His²⁹⁰ (Fig. 4B). Thus, 3 μm Cy3NTA produced >90% binding saturation to His⁷⁶, His¹⁸¹, and His⁵¹⁹, which enabled us to determine maximum FRET efficiencies to these positions for triangulation experiments (see below).

For His⁷⁶ and His¹⁸¹ labeled with 3 μm Cy3NTA, similar FRET profiles were observed from each D-FKBP (Fig. 4, C and D). For example, D32- and D85-FKBP produced the highest FRET efficiencies, whereas FRET efficiencies from D14-, D44-, or D49-FKBP to His⁷⁶ and His¹⁸¹ were significantly lower (Fig. 4, C and D, and Table 1). For His²⁹⁰, FRET efficiencies from D32- and D44-FKBP were comparable to those seen with His⁷⁶ and His¹⁸¹ (Fig. 4E), whereas FRET efficiencies from D14- and D49-FKBP were slightly higher (Table 1). Finally, FRET from each of the D-FKBPs to Cy3NTA targeted to His⁵¹⁹ were the highest compared with the other N-terminal domain His-tagged constructs (Fig. 4F) with values ranging as high as 0.58 for D32-FKBP and 0.39 for D85-FKBP.

Triangulation of Cy3NTA Targeted to His Tags within the N-terminal Domain

The atomic structure of the first 559 amino acids of RyR1 has been solved and docked to the RyR1 cryo-EM map (22), which enabled comparison between our FRET results and predicted distances between the AF488 donor attached to FKBP and the Cy3NTA acceptor attached at His-tagged positions within this domain. We converted the measured FRET values to donor-acceptor distances and then centered spheres on each donor position of FKBP docked to RyR1. The radii and thickness of these spheres corresponded to distances calculated from the mean and S.D. of the FRET measurements, respectively (Table 1). The position of Cy3NTA bound to each His tag (except His²⁹⁰, which did not achieve binding saturation with Cy3NTA) was then triangulated at the intersection of these spheres (see “Experimental Procedures” for detailed description of the triangulation procedure).

Using this approach, FRET-derived loci of Cy3NTA targeted to each of the three saturable His-tagged positions in the N-terminal domain were found within the docked ABC domain (Fig. 5, A and B). Loci corresponding to Cy3NTA bound to either His⁷⁶ (purple spheres, Fig. 5) or His¹⁸¹ (yellow spheres) were relatively broad, overlapping both the A and B domains of the docked ABC atomic structure. The locus of Cy3NTA bound to His⁵¹⁹ (green spheres) localized to a much more discrete zone within the C domain (Fig. 5, C and D). Thus, these results are consistent with the docking of the ABC domain to its central location within the RyR1 cryo-EM map (22) as positions 76 and 181 are found in the A domain, and position 519 is in the C domain (white spheres, Fig. 5).

FIGURE 5.

Triangulation of Cy3NTA bound to His tags in the RyR1 N-terminal domain. A, the position of Cy3NTA bound to either His⁷⁶ (purple spheres), His¹⁸¹ (yellow spheres), or His⁵¹⁹ (green spheres) is depicted relative to the RyR1 cryo-EM map (gray) and the ribbon diagrams of the N-terminal ABC domain (A, B, and C) docked to the two subunits (S1 and S2) nearest to the FKBP ribbon diagram (FKBP). Cryo-EM reconstruction is viewed from the top (i.e. the portion that would face the t-tubule in situ). B, shown is the same structure in A viewed from the side along sectioning plane indicated by a dashed line in A. C, shown is a magnified view of Cy3NTA triangulations outlined by the dashed box in A. White spheres depict the position of β carbon atoms for amino acid residues 76, 181, and 519, used for His tag insertions. D, shown is a magnified view of Cy3NTA triangulations outlined by the dashed box in B. Scale bars, 50 Å (A and B), 20 Å (C and D).

Analysis of D-FKBP Binding to RyR1 Constructs with His₁₀ Tags in the Central Domain

We conducted a similar analysis for four His-tagged positions in the RyR1 central domain, which spans amino acid residues 2157–2777 and contains at position 2461 the critical valine-proline motif that has been implicated in FKBP12 binding to RyR1 (17) (Fig. 6A). D-FKBP fluorescence and RyR1 immunofluorescence co-localized only for the His²³⁴¹ and His²⁷⁷⁷ RyR1 constructs, whereas no specific binding was observed for D-FKBP to the His²¹⁵⁷ and His²⁵⁰² constructs (Fig. 6, B and C) even when D-FKBP concentrations as high as 100 nm were used (data not shown). Thus, insertion of His tags at sequence positions 2157 and 2502 disrupts the binding of D-FKBP. The relatively high D-FKBP:RyR1 ratio observed for His²⁷⁷⁷ may result from partial disruption of the epitope of the 34C anti-RyR primary antibody (26, 30), which is proximal to this His-tag insertion.

FIGURE 6.

A, the positions of His tags inserted into the central portion of the wtRyR1 sequence (black bar) are shown. Black dots represent positions and numbers of MH mutations within the RyR1 sequence. VP indicates the location of the FKBP12 dipeptide binding site identified previously (17). AA, amino acids. B and C, D-FKBP binding analysis via fluorescence microscopy is shown. B, shown is the relative fluorescence ratio of D-FKBP:RyR1 immunoreactivity for the indicated wild type and His-tagged RyR constructs. Values represent the mean ± S.E. for the number of cells indicated in each bar. C, shown are representative images of HEK-293T cells expressing either His²¹⁵⁷, His²³⁴¹, His²⁵⁰², or His²⁷⁷⁷ RyR1 constructs labeled with D14-FKBP (left panels, green) and then fixed and stained for RyR1 immunoreactivity (middle panels, red). Overlays (right panels) indicate overlap between the two fluorophores.

Functional Analysis of RyR1 Constructs Containing His₁₀ Tags within the Central Domain

All central domain His-tagged RyR1 constructs released Ca²⁺ in response to caffeine, except His²⁷⁷⁷ (Fig. 3C). EC₅₀ values for caffeine activation of His²¹⁵⁷ and His²³⁴¹ were unchanged relative to wtRyR1 (Fig. 3D), whereas the EC₅₀ for activation of His²⁵⁰² was statistically higher.

FRET Measurements to the RyR1 Central Domain

FRET efficiencies measured between D-FKBP and Cy3NTA bound to the His²³⁴¹ and His²⁷⁷⁷ constructs were higher than those seen for the N-terminal domain His-tagged positions (except His⁵¹⁹). Relative to His²⁷⁷⁷, His²³⁴¹ showed significantly higher FRET to each labeled position of D-FKBP (Fig. 7, A and B; Table 2). However, FRET efficiencies to either His²³⁴¹ or His²⁷⁷⁷ showed little variation across the different donor-labeled positions.

FIGURE 7.

Shown are FRET efficiency values from each of the five labeled positions on the FKBP FRET donor (indicated on x axes) to Cy3NTA after both molecules were targeted to either His²³⁴¹ (A) or His²⁷⁷⁷ (B) in permeabilized HEK-293T cells. Values represent the mean and S.E. for the number of cells indicated in Table 2. Insets indicate Cy3NTA titration curves for FRET from D85-FKBP to Cy3NTA targeted to either of the His-tagged constructs. C–E, triangulation of Cy3NTA bound to the His tag at position 2341 is indicated (purple spheres) within the RyR1 cryo-EM map (gray) viewed either from the side (C) or the top (i.e. the portion that would face the t-tubule in situ) (D). E, shown is a magnified view of Cy3NTA triangulation outlined by the dashed box in D. The FKBP ribbon diagram (FKBP) and ABC domain (A, B, and C) docked to the RyR1 cryo-EM map are indicated. Scale bars, 50 Å (C and D), 20 Å (E).

TABLE 2.

Summary of FRET values and calculated distances between D-FKBP and His tags in the RyR1 central domain

Donor Positiona	FRET (observed)b	Nc	Apparent D/A distanced
His2341
14	0.47 ± 0.06	30	61 (58–63)
32	0.41 ± 0.12	9	63 (58–69)
44	0.52 ± 0.09	46	58 (55–62)
49	0.46 ± 0.13	30	61 (56–67)
85	0.45 ± 0.12	16	61 (56–67)
His2777
14	0.23 ± 0.06	53	73 (69–77)
32	0.30 ± 0.08	28	68 (64–73)
44	0.25 ± 0.07	62	71 (67–77)
49	0.30 ± 0.07	51	69 (65–72)
85	0.27 ± 0.10	37	70 (64–76)

^a Amino acid residue attachment site of AF488 on FKBP.

^b Observed FRET efficiency (mean ± S.D.).

^c Number of measurements.

^d Average and range (in parentheses) of donor-acceptor distances (in Å) calculated from the mean and S.D. of the measured FRET efficiencies.

Triangulation of Cy3NTA Bound to His-tagged Positions in the Central Domain of RyR

We used calculated distances based on these FRET measurements to triangulate the position of Cy3NTA bound to His²³⁴¹ and His²⁷⁷⁷. For His²³⁴¹ we identified a rather broad locus near the clamp domain of the RyR1 cryo-EM map using distances from each of the D-FKBPs except D32-FKBP (Fig. 7, C–E, purple spheres). This region was 50–60 Å from the various donor-labeled positions on D-FKBP bound to RyR1. In contrast, no locus could be triangulated for Cy3NTA bound to His²⁷⁷⁷. The reason for this is not clear, although it may be related to structural disruption of the channel by this particular His tag insertion, which also abolished RyR channel function (Fig. 3).

FRET Measurements between MH Zones 1 and 2

These FRET measurements suggest that FKBP binds proximal to sites within both the N-terminal and central domains of RyR1. Thus, positions within these two domains may be very close to each other, a notion previously formalized as the domain switch hypothesis used to explain the molecular pathogenesis of malignant hyperthermia (31). To test this hypothesis more directly, we fused GFP either to the N terminus (GFP(1)) or position 620 (GFP(620)) of RyR1 and then measured FRET to Cy3NTA bound to His tags inserted into the central domain.

In positive control experiments, high FRET efficiencies were observed from each GFP donor position to Cy3NTA targeted to adjacent His tags (Fig. 8, A and B). However, no significant energy transfer was detected from GFP(1) to Cy3NTA targeted to any His tags in the central domain (Fig. 8A). In contrast, robust FRET was detected from GFP(620) in domain C to the His tags inserted within clusters of MH mutations (i.e. His²¹⁵⁷, His²³⁴¹, and His²⁵⁰²), whereas no significant energy transfer was observed from GFP(620) to Cy3NTA targeted to His¹⁹¹⁵ or His²⁷⁷⁷ (Fig. 8B), which are outside the disease hot-spot.

FIGURE 8.

FRET efficiency measurements for GFP-RyR1 fusion constructs containing His tags in the central domain and GFP fused at either the N terminus (A) or position 620 (B). Positions of the GFP donors and the central domain containing the His tags (gray bar) within the wtRyR1 sequence (black bar) are indicated for each panel. FRET efficiency levels for each GFP-RyR1 fusion construct are indicated. Asterisks indicate significant differences in FRET efficiencies (p < 0.01) relative to non His-tagged constructs as determined using one-way ANOVA followed by a Dunnett's post test.

DISCUSSION

Labeling System

This FRET-based method has several advantages compared with previous techniques used to investigate RyR structure via FRET (20, 32). Our approach requires only the insertion of short His tags to make measurements, which should result in minimal disruption of the native RyR structure when compared with the use of fluorescent protein fusions. In addition, the relationship between the position of the FRET probes and their insertion sites within RyR1 is better defined than for fused fluorescent proteins, which introduce significant distances from their insertion sites that complicate FRET analysis, as we have shown (21). Finally, the locations of both FKBP (13, 15) and some of the His-tagged sites (22) within the RyR1 cryo-EM map (25) are known, so we could validate our approach via triangulation of Cy3NTA bound to these His tags.

For positions 76, 181, and 519, good general agreement was seen between our triangulations of Cy3NTA bound to these sites and the location of the insertion sites themselves (Fig. 5). Thus, Cy3NTA targeted to either His⁷⁶ or His¹⁸¹ within the A domain, localized to a broad area encompassing parts of both the A and B domains, whereas Cy3NTA bound to His⁵¹⁹ was localized to a much smaller area near its insertion site in the C domain. It is intriguing that our triangulations of positions 76 and 181 did not localize to a more centralized area between ABC domains docked at two adjacent RyR subunits, as comparable FRET to both subunits is expected based on predicted distances from FKBP donors to these positions (see Table 1). The contribution of FRET from a second subunit may be one reason why these triangulated areas are much broader compared with His⁵¹⁹ (where no FRET to a second subunit is expected). Alternatively, the orientation of the A domain in the full-length protein may be slightly shifted compared with its orientation in the ABC crystal structure such that FRET predominantly results from a single acceptor site.

We did not attempt to triangulate position 290, as Cy3NTA binding to this position did not saturate (Fig. 4C). The decreased Cy3NTA binding affinity to position 290 suggests that this site is more occluded than the other His-tagged sites used in this study. Indeed, surface representations of the ABC domain suggest that this position is partially buried. Thus, these saturation binding measurements using Cy3NTA may prove useful in testing accessibility of unknown sites in future FRET-based experiments.

FKBP Localization

The location of the binding site for FKBP12 within the RyR1 protein sequence remains controversial. Although initial reports indicated a binding site proximal to the VP dipeptide at position 2461 of RyR1 (17, 33), subsequent reports examining in vitro binding of FKBP12.6 to RyR2 fragments (18, 19, 34) failed to support this finding. However, FKBP binding data to RyR fragments or deletion mutants is difficult to interpret because this technique is predicated on the unlikely assumption that all RyR fragments fold and behave exactly as they would within the full-length RyR. What this in vitro work does suggest is that the FKBP binding site on the RyR is most likely not comprised exclusively of a contiguous stretch of the RyR sequence, as none of 11 overlapping fragments encompassing the entire primary structure of RyR2 bound FKBP12.6 (19). Rather, this previous work and our current study suggest that the binding site for FKBP is composed of elements from non-contiguous sequence regions, as outlined below.

Our results suggest that regions within the N-terminal portion of RyR1 are proximal to FKBP (Fig. 10A). For example, we observed a high degree of energy transfer (E = 0.58) from D32-FKBP to Cy3NTA at His⁵¹⁹ (Fig. 4F), and either a GFP- or a His-tag insertion at position 619 eliminated FKBP binding (Fig. 2 and Ref. 35). Disruption of FKBP binding by insertions at position 619 suggests that this position is within the FKBP binding site. Supporting evidence for this finding comes from secondary structure analysis, which suggests that residue 619 lies at the C-terminal end of 10 predicted α helices between amino acid residues 390 and 630. These helices most likely comprise the entire C domain, determined from crystallographic analysis of the N-terminal domains of RyR1 and the IP₃ receptor (22, 36, 37) (Fig. 9A). Within the C domain, the first 5 (though 6 in the inositol trisphosphate receptor; Refs. 36 and 37)) of these helices pack in antiparallel fashion into a tight armadillo-fold (38). If the final five RyR1 helices also fold into this structural domain, then their location within the RyR1 cryo-EM map can be predicted by identifying areas connecting the ABC domain to the rest of the protein. We found only two such locations (Fig. 9, B–D), a small connecting “pier” and a region of high electron density directly adjacent to the C domain (Fig. 9D). The small pier contains a hairpin loop connecting two antiparallel α helices and thus may connect the C domain with a non-consecutive RyR domain. That leaves the large density (red in Fig. 9D) as a probable location of the remainder of the C domain, including position 619, which itself could then be part of the FKBP binding site.

FIGURE 10.

Shown is a model of the FKBP binding site (green circle) based on FRET from D-FKBP to His-tagged positions within the RyR1 N-terminal domain (A) or central domain (B). Positions 619, 2157, and 2502 are most likely within the FKBP binding site because insertion of His tags at these positions disrupts D-FKBP binding. Positions 519 and 2341 are also close to the FKBP binding site, as suggested by high FRET efficiencies observed from D-FKBP to these positions. C, shown is an integrated model that takes into account our structural measurements to both the N-terminal and central domains. The FKBP binding site on RyR1 is most likely composed of determinants from both regions.

FIGURE 9.

Localization of the N-terminal C domain within the cryo-EM reconstruction of RyR1. A, protein sequence alignment of RyR1 and RyR2 between amino acids 383 and 681 of rabbit RyR1 is shown. The top scale corresponds to the rabbit RyR1 sequence. Highlighted residues correspond to positions with documented MH (red) or CPVT/ARVD2 mutations (green). Secondary structure elements are indicated either as blue (E) to denote β strands or as red (for RyR1) or pink (H) (for inositol trisphosphate receptor) to denote α helices. Predicted secondary structure elements are enclosed in the black box. B, shown is a top view (as would be viewed from the t-tubule membrane in situ) of the cryo-EM reconstruction of RyR1 indicating the position of the ABC domain docked to each of the four subunits, as well as FKBP (yellow) docked to one subunit. The red zone indicates the proposed position of the remaining 5 helices of the C domain between residues 559 and 623 of rabbit RyR1. The dotted line indicates the sectioning plane used to visualize the structure shown in C. C, a side view of the structure shown in B is indicated. The dotted box indicates the structure shown in higher magnification in D. D, a magnified view of the proposed location of the C domain extension is indicated. Location of ABC domain and FKBP are shown. The pink helix in the C domain indicates the position of the sixth α helix in the inositol trisphosphate receptor C domain after aligning the position of these domains and docking to the RyR1 cryo-EM structure. Pier refers to potential connecting structure between the ABC domain and the rest of the cryo-EM structure (see under “Discussion”). Scale bars, 50 Å in B and C, 25 Å in D.

Our results also suggest that sites within the central domain of RyR1 contribute to FKBP binding (Fig. 10B). FRET efficiencies from D-FKBP to His²³⁴¹ in the central domain ranged from 0.41 to 0.52, and we triangulated Cy3NTA bound to this site to a region around 60 Å from the bound FKBP (Fig. 7). In addition, His tag insertions at positions 2157 and 2502 completely disrupted FKBP binding (Fig. 6), which suggests that these sites may compose part of the FKBP binding site. Because position 2502 is near the VP dipeptide motif at position 2461 initially identified as the FKBP12 binding site on RyR1 (17), it is not surprising that a His tag insertion at this position might disrupt FKBP binding, particularly if the VP dipeptide plays an important role in this process. Thus, our data suggest that the FKBP binding site on RyR1 is composed of elements both from the N-terminal and central portions of the RyR, thus reconciling the seemingly contradictory findings from previous reports.

Domain Switch Hypothesis

Our structural findings also have implications for the domain switch hypothesis (2, 31), originally proposed to explain the molecular pathogenesis of malignant hyperthermia. In this hypothesis, zones within the N-terminal and central portions of the RyR physically contact each other in the closed channel (zipping), and MH mutations in either of these zones destabilize (unzips) this interaction, resulting ultimately in the MH phenotype. Our data, indicating that the N-terminal and central regions of RyR1 comprise the FKBP binding site, support this hypothesis, as the data suggest that these sites most likely are in close proximity to each other (Fig. 10C). This is further supported by our FRET results indicating that GFP fused at position 620 is proximal to His tags placed at positions 2157, 2341, and 2502 (Fig. 8B), the same positions that either result in high FRET efficiencies to D-FKBP or disruption of D-FKBP binding to the RyR. Finally, the first discovered MH mutation lies in the N-terminal domain at position 615 (29), whereas there are >45 mutations sites in the area between amino acids 2157 and 2502 in the central domain (39, 40). Thus, these two MH zones may contact each other in the vicinity of the FKBP binding site, and MH mutations could disrupt this interaction. With our FRET-based system, we intend to test this hypothesis directly in future studies.

Summary

These results provide new insights into the location of the FKBP binding site within the RyR1 sequence. Given that some of the His₁₀ insertions disrupt FKBP binding, further resolution of the binding site may be achieved using FRET measurements to smaller tags placed in the RyR1 sequence near positions 619, 2157, and 2502. Furthermore, structural information gleaned here will inform future studies aiming to resolve conformational changes associated with RyR1 gating.