By continuing to browse this site you agree to us using cookies as described in About Cookies
Notice: Wiley Online Library will be unavailable on Saturday 7th Oct from 03.00 EDT / 08:00 BST / 12:30 IST / 15.00 SGT to 08.00 EDT / 13.00 BST / 17:30 IST / 20.00 SGT and Sunday 8th Oct from 03.00 EDT / 08:00 BST / 12:30 IST / 15.00 SGT to 06.00 EDT / 11.00 BST / 15:30 IST / 18.00 SGT for essential maintenance. Apologies for the inconvenience.
H. S. Young, Department of Biochemistry, University of Alberta, Edmonton, AB T6G 2H7 Canada Fax: +1 780 492 0095 Tel: +1 780 492 3931 E-mail: email@example.com
K. S. Misono, Department of Biochemistry, University of Nevada School of Medicine, Reno, NV 89557, USA Fax: +1 775 784 1419 Tel: +1 775 784 4690 E-mail: firstname.lastname@example.org
Atrial natriuretic peptide (ANP) plays a major role in blood pressure and volume regulation. ANP activities are mediated by a cell surface, single-span transmembrane receptor linked to its intrinsic guanylate cyclase activity. The crystal structures of the dimerized ANP receptor extracellular domain (ECD) with and without ANP have revealed a novel hormone-induced rotation mechanism occurring in the juxtamembrane region that appears to mediate signal transduction [Ogawa H, Qiu Y, Ogata CM & Misono KS (2004) J Biol Chem279, 28625–28631]. However, the ECD crystal packing contains two major intermolecular contacts that suggest two possible dimer pairs: ‘head-to-head’ (hh) and ‘tail-to-tail’ (tt) dimers associated via the membrane-distal and membrane-proximal subdomains, respectively. The existence of these two potential dimer forms challenges the proposed signaling mechanism. In this study, we performed single-particle electron microscopy (EM) to determine the ECD dimer structures occurring in the absence of crystal contacts. EM reconstruction yielded the dimer structures with and without ANP in only the hh dimer forms. We further performed steady-state fluorescence spectroscopy of Trp residues, one of which (Trp74) occurs in the hh dimer interface and none of which occurs in the tt dimer interface. ANP binding caused a time-dependent decrease in Trp emission at 350 nm that was attributable to partially buried Trp74 in the unbound hh dimer interface becoming exposed to solvent water upon ANP binding. Thus, the results of single-particle EM and Trp fluorescence studies have provided direct evidence for hh dimer structures for unbound and ANP-bound receptor. The results also support the proposed rotation mechanism for transmembrane signaling by the ANP receptor.
Atrial natriuretic peptide (ANP) is a cardiac hormone that is secreted by the atrium of the heart in response to blood volume expansion. ANP stimulates renal salt excretion  and dilates blood vessels [2,3]. Through these activities, ANP participates in the regulation of blood pressure and salt–fluid volume homeostasis. ANP also has antigrowth activity on vascular cells, through which it regulates the maintenance and remodeling of the cardiovascular system [4–7]. These biological activities of ANP are mediated by the cell surface receptor for ANP, which possesses intrinsic guanylate cyclase (GCase) activity. The ANP receptor occurs as a homodimer of a single-transmembrane polypeptide, each containing an extracellular ANP-binding domain (ECD), a transmembrane domain, and an intracellular domain consisting of an ATP-binding regulatory domain and a GCase catalytic domain . ANP binding to the ECD stimulates the intracellular GCase domain, thereby generating the intracellular second messenger cGMP. The mechanism of this transmembrane signal transduction by the ANP receptor is only partially understood.
To understand the signaling mechanism, we earlier determined the crystal structures of the dimerized ECD with  and without  bound ANP. Comparison of the two structures has revealed that ANP binding causes a large change in the quaternary arrangement of the ECD dimer without significant intramolecular structure change. This change in the quaternary structure causes an alteration in the relative angular orientation of the two juxtamembrane domains in the dimer that is equivalent to rotating each by 24° . There is no appreciable change in the distance between the two juxtamembrane domains. On the basis of this finding, we have proposed that a novel hormone-induced rotation mechanism occurring in the juxtamembrane region may trigger transmembrane signal transduction [9,11]. However, this proposed signaling mechanism has been questioned because of uncertainty concerning the quaternary structure of the unbound ECD (apoECD) dimer.
The crystal packing of apoECD contains two major intermolecular contacts (Fig. 1A), which generate two possible dimer pairs: an hh dimer associated with the membrane-distal subdomain (Fig. 1B) and a tt dimer associated with the membrane-proximal subdomain (Fig. 1C). The buried surface areas in the hh and tt contacts in crystals are estimated to be 1100 Å2 and 1680 Å2, respectively . These values are both large and are within the range often found in physiological protein–protein interactions. Thus, it is not clear from the crystallographic data alone whether the hh or tt dimer represents the physiological structure. Similarly, the ANP–ECD complex (ANP–ECD) may also occur, at least theoretically, in an hh or a tt dimer form (Fig. 1E,F). We originally reported the structure of apoECD in the tt dimer configuration based on the fact that the tt contact was estimated to be larger than the hh contact . However, our subsequent site-directed mutagenesis studies of interface residues using the full-length ANP receptor expressed in COS cells showed that mutations in the hh interface, but not in the tt interface, affected signaling (stimulation of cGMP production by ANP) . These findings have suggested that the hh dimers, but not the tt dimers, represent the physiological structures.
On the other hand, it has been proposed that the hh dimer and tt dimer structures both occur, and represent the inactive and the hormone-activated states of the receptor, respectively [13,14]. It is hypothesized that a hormone-induced rearrangement of the ECD from the hh to the tt dimer structure brings the juxtamembrane domains into proximity, thereby mediating signal transduction . This proposed mechanism involving a ligand-induced domain approximation has been described in some reports as being well accepted for natriuretic peptide receptors [15,16], and been suggested to be similar to those of the G-protein-coupled metabotropic glutamamate receptor [15–17] and the erythropoietin receptor [18,19]. In contrast, our proposed rotation mechanism, which is based on the hh dimer structures for both apoECD and ANP–ECD, is mediated by a ligand-induced rotation of the juxtamembrane domains with essentially no change in the interdomain distance. To resolve this discrepancy over the ANP receptor signaling mechanism, it has become imperative to determine the ECD dimer structures in more physiological buffer solution conditions and in the absence of crystal contacts.
In this study, we have carried out single-particle image reconstruction of the ECD dimer with and without bound ANP using electron microscopy (EM). This method provides the ECD dimer structure as it occurs in solution free of crystal contacts. We reasoned that the crystal contacts, which occur under certain artificial and rather extreme sets of conditions used for protein crystallization, will not occur under solution conditions closer to the physiological state. Only the naturally occurring intermolecular contacts should remain. The results of our single-particle EM studies described in this article support the above reasoning, and have identified the hh dimer as the only form found in solution. The single-particle reconstructions for the apoECD dimer and ANP–ECD agree closely with the respective crystal structures, suggesting that crystal contacts have not appreciably altered the dimer structures. To further support our finding, we also present here steady-state fluorescence studies of Trp residues, taking advantage of the fact that Trp74 occurs at the hh interface and that its local environment changes upon ANP binding, whereas the environment of other Trp residues is largely unaltered. We observed quenching of Trp fluorescence concomitant with ANP binding, which is consistent with the apoECD being in the hh dimer structure. The implications of the results of single-particle EM and Trp fluorescence studies for the transmembrane signaling mechanism of the ANP receptor are discussed.
Results and Discussion
EM and single-particle reconstruction
From electron micrographs of negatively stained apoECD, more than 22 000 particles were selected (Fig. 2A). The particles were centered and grouped into self-similar groups by iterative multivariate statistical analysis-based classification. Class averages were then generated by iterative alignment and averaging. Among the 35 class averages generated, many showed clear two-fold symmetry, with several orientations consistent with the hh dimer (Fig. 2B). A set of Euler angles was then assigned to these class averages, using common lines in Fourier space (startAny command in eman), and an initial 3D model was built. The initial model was used for five iterations of refinement, or until convergence was achieved. The 3D reconstruction had the following approximate dimensions: width, 90 Å; height, 80 Å; and depth, 50 Å. This volume is consistent with an ECD dimer. The final reconstruction after a minimum of five rounds of refinement exhibited clear two-fold symmetry, which was enforced (Fig. 2C). The data were not corrected for the contrast transfer function (CTF), and only data within the first zero of the CTF were used. On the basis of the defocus series, this effectively limited the resolution of the reconstruction to 22 Å. Therefore, the reconstruction was low-pass-filtered at this resolution. The handedness of the reconstructions was determined by comparison with the known crystal structures of the dimers [9,10].
A similar approach was utilized for ANP–ECD, where the ECD was incubated with a 1.1-fold molar excess of ANP for 1 h before grid preparation. Visual inspection of electron micrographs of negatively stained ANP–ECD showed no apparent differences as compared to apoECD. More than 19 000 particles were selected, centered, and classified as described above. Reference-free 3D reconstruction and refinement resulted in a model that showed clear two-fold symmetry, consistent with the X-ray structure of ANP–ECD (Fig. 2D).
Comparison of the 3D reconstructions by EM and the crystal structures
In the crystal packing of apoECD, the buried surface areas in the hh and tt dimers are within the range typically found for physiological protein–protein interactions. Thus, it is not possible from the crystallographic data alone to determine which dimer structure represents the physiological state. To identify the correct apoECD dimer, the crystal structures for apoECD in the hh dimer (Fig. 1B) and tt dimer (Fig. 1C) forms were superimposed on the 3D reconstruction of apoECD obtained by single-particle EM (Fig. 3A,C). The molecular envelope of the hh dimer crystal structure agreed closely with the EM density map, whereas that in the tt dimer form clearly showed a large structural discrepancy. These results demonstrate that apoECD, in the absence of crystal contacts, assumes the hh dimer structure.
In the crystal packing of ANP–ECD, two ECD monomers form an hh dimer, with one molecule of ANP captured in between these monomers . In this structure, ANP binding involves a very large buried surface area (1450 Å2 with one ECD monomer and 1320 Å2 with the other monomer, for a total buried surface area of 2770 Å2), which strongly supports the notion that the hh dimer structure represents the physiological ANP–ECD structure. The crystal structure of ANP–ECD in the hh dimer form (Fig. 1E), when superimposed on the 3D reconstruction obtained by single-particle EM, agreed closely (Fig. 3B). On the other hand, the tt dimer model (Fig. 1F) showed a large discrepancy with the EM reconstruction (Fig. 3D).
We also performed reference-based single-particle reconstruction using the hh and tt dimer crystal structures as initial models (Fig. S1). The reconstruction of apoECD and ANP–ECD using the hh dimers as the initial models quickly converged within five refinement cycles on a reconstruction that was similar to the hh dimer described above. In contrast, the refinements using the tt dimer as the initial model quickly diverged from the initial models within five cycles of refinement. By 20 cycles, the solution converged on a reconstruction similar to the hh dimer. These results suggest that both apoECD and ANP–ECD occur entirely in the hh dimer form in solution. Hence, the tt contacts in crystals are artificial interactions that only occur under the conditions used for crystallization and do not occur in more physiological solution conditions. Additionally, the close agreement of the EM reconstructions with their respective crystal structures indicates that the crystal contacts did not appreciably alter the quaternary structures of the dimers.
Steady-state fluorescence studies of ANP-induced structural change
Each ECD monomer contains 10 Trp residues. Of these, one, Trp74, occurs in the hh interface (Fig. 4A,B). No Trp residue is present in the tt interface. In the apoECD hh dimer model (Fig. 4A), Trp74 of one monomer interacts with Trp74 of the other monomer and contributes to the hh dimer contact . In ANP–ECD (Fig. 4B), these two Trp74 residues are pulled apart and are exposed to the solvent. We have shown previously that ANP binding causes no appreciable intramolecular structural change in the ECD monomers (rmsd of Cα atoms between the apo and the complex structures, 0.64 Å) . Furthermore, no Trp residues make contact with ANP in the bound complex. Therefore, if the ECD assumes the hh dimer structures, only the Trp74 residue should undergo a significant change in its environment. On the other hand, if the ECD assumes the tt dimer structures, no change is expected in the Trp environment in response to ANP binding. On the basis of the above structure analyses, we utilized Trp fluorescence to examine the solution structures of apoECD and ANP–ECD.
The fluorescence emission spectra of apoECD and ANP–ECD are shown in Fig. 4C. Comparison of the spectra shows that addition of ANP causes an approximately 7% decrease in the fluorescence emission intensity at the lambda maximum 350 nm. This drop in the fluorescence intensity was time-dependent and was largely complete in about 30 min (not shown). The course of this intensity drop matches closely the course of ANP binding measured using [125I]ANP . These findings are consistent with the hh dimer structures for both apoECD and ANP–ECD, where the two partially buried Trp74 residues at the apoECD hh dimer interface become exposed upon ANP binding [9,12] and quenched by water. The difference spectrum obtained by subtracting the ANP–ECD emission from the apoECD emission revealed a shift to a longer wavelength (Fig. 4C). This red shift in the emission difference spectrum is consistent with the two Trp74 residues that are localized at the edge of the apo dimer interface in a partially exposed, polar environment . The decrease in Trp emission intensity from the total emission intensity from 10 Trp residues in each ECD monomer was relatively small (7%). The quantum yield of Trp residues is known to vary widely, depending on the environment. The relatively small decrease may be due to quenching of the two Trp74 residues at the apoECD hh dimer by a staggered face-to-face interaction between the two indole rings (Fig. 4A).
To confirm that the decrease in the fluorescence intensity is due to the change in Trp74 environment, we measured the fluorescence emission of an ECD mutant, W74R. We have shown previously that the W74R mutant binds ANP with an affinity similar to that of the wild-type . The fluorescence emission spectrum of the W74R mutant was similar to that of the wild-type, with a peak at around 350 nm, but with a slightly reduced intensity because of the Trp to Arg mutation. As shown in Fig. 4D, addition of ANP to the W74R mutant caused no appreciable change in the emission intensity. This finding confirms that the decrease in Trp fluorescence observed upon ANP binding to the wild-type ECD was caused by solvent exposure and the resulting quenching of Trp74 emission in ANP–ECD.
Comparison of the apoECD and ANP–ECD EM reconstructions
To evaluate the structural change induced by ANP binding, the 3D reconstructions of apoECD and ANP–ECD were aligned with each other for comparison, using the align3d command in eman (Fig. 5). For clarity, the reconstructions are contoured at 70% of the expected molecular volume for an ECD dimer. Despite the low resolution of the reconstructions, the ANP–ECD structure is more detailed, with a shape characteristic of the crystal structure. Nonetheless, both EM reconstructions exhibit dimeric shape and monomer orientations that closely agree with those observed by X-ray crystallography. In the front view, there is no appreciable change in the distance between the two monomers (Fig. 5). Viewed from the side, each monomer in the ANP–ECD reconstruction is displaced in a clockwise direction, reminiscent of the twist motion observed by X-ray crystallography . Viewed from the bottom (i.e. in the direction from the presumed transmembrane regions; Fig. 5, bottom view), the two juxtamembrane domains are displaced in opposite directions upon binding of ANP, without an appreciable change in the distance between the two.
Proposed mechanism for transmembrane signal transduction
On the basis of the hh dimer pairs demonstrated above, the X-ray structures of ECD with  and without  bound ANP show that ANP binding causes a large change in the quaternary structure of the ECD dimer without appreciable intramolecular structural change. ANP binding causes each of the two ECD monomers to undergo a twisting motion while retaining the two-fold symmetry in the dimeric complex . This twisting motion causes the two juxtamembrane domains in the dimer to undergo parallel translocation in the opposite direction, with essentially no change in the distance between the two (Fig. 6A). This movement causes an alteration in the relative angular orientation of the two juxtamembrane domains that is equivalent to rotating each domain by 24° (Fig. 6B). We have proposed that this hormone-induced rotation mechanism occurring in the juxtamembrane region may trigger ANP receptor signaling [9,11]. The ANP-induced structural change observed here by single-particle EM closely resembles that recognized by X-ray crystallography, thus supporting the proposed signaling mechanism.
In summary, the 3D reconstructions by single-particle EM, which were obtained in the absence of crystal contacts, yielded the hh dimer structures for both apoECD and ANP–ECD. Comparison of the 3D reconstructions with and without ANP showed the ANP-induced structural change in the dimer that was surprisingly close to that observed by X-ray crystallography. The quenching of Trp74 fluorescence emission concomitant with ANP binding is also in agreement with apoECD and ANP–ECD in hh dimer structures. Thus, the results of our complementary approaches, single-particle EM, fluorescence spectroscopy and X-ray crystallography, together demonstrate a novel hormone-induced structural change in the ECD dimer that generates a rotation mechanism in the juxtamembrane regions and possibly mediates transmembrane signal transduction.
Preparation of ECD and ANP–ECD
ECD consisting of residues 1–435 of the rat ANP receptor was expressed by slight modification of the method described previously , as follows. CHO cells were transfected with pcDNA3–NPRA, and stably transfected, high-producer cells were cloned by selection with G-418. The cloned cells were cultured in roller bottles, and the conditioned medium containing the expressed ECD was collected every 2 days. The ECD was purified by ANP affinity chromatography as previously described . ANP–ECD was prepared by incubating ECD (1 mg·mL−1) with a 1.1-fold molar excess of a truncated ANP peptide, Cys-Phe-Gly-Gly-Arg-Ile-Asp-Arg-Ile-Gly-Ala-Gln-Ser-Gly-Leu-Gly-Cys-Asn-Ser-Phe-Arg, representing residues 7–27, in 5 mm Hepes buffer (pH 7.0) containing 20 mm NaCl at room temperature for 60 min.
Aliquots (3 μL) of ECD at 0.03 mg·mL−1 in the absence (apoECD) and presence (ANP–ECD) of ANP were applied to glow-discharged, carbon-coated grids. The grid was washed with two drops of 2% uranyl acetate, and then a third drop of 2% uranyl acetate was allowed to sit on the grid for 1 min (4 °C). The excess stain was removed by blotting with filter paper, and the sample was allowed to air dry. Data were collected on a Tecnai F20 (FEI Company) located in the Microscopy and Imaging Facility at the University of Calgary (Calgary, Canada). The microscope was operated at 200 keV, and images were recorded on Kodak SO-163 film under low-dose conditions at a magnification of ×50 000, with a defocus ranging from −1.5 to −2.5 μm. Micrographs were digitized with a Nikon Super Coolscan 9000 with a scanning resolution of 6.35 μm·pixel−1, and this was followed by pixel averaging to achieve a final resolution of 3.81 Å·pixel−1.
Image processing and reconstruction were performed with the eman program package . Seventeen micrographs with minimal drift and astigmatism were selected for reconstruction of apoECD. Similarly, 20 micrographs were used for ANP–ECD. Particles were selected semiautomatically and extracted as 40 × 40 pixel images (boxer). In total, 22 778 and 19 600 particle images were selected for apoECD and ANP–ECD, respectively. No correction for the CTF was applied. Reference-free classification was performed to generate 35 class averages (refine2d.py), and an initial set of Euler angles was then assigned to these class averages (startAny). The initial three-dimensional models built using common lines in Fourier space were then refined in eman for up to 20 cycles of refinement (refine). The assignment of Eulerian angles from class averages by common lines results in two possible enantiomeric solutions. The X-ray crystallographic structures were used to determine the handedness of the reconstructions. Because the expected two-fold symmetry for the two ECD monomers in apoECD and ANP–ECD was observed, C2 symmetry was applied throughout the refinement procedure. The first zero of the CTF for the lowest defocus images effectively limited the resolution of the final reconstruction to ∼ 22 Å. This resolution limit was confirmed by calculating the Fourier shell correlation between two independent half datasets (eotest command in eman; 0.5 FSC criterion). Therefore, the final density maps were low-pass-filtered to 22 Å resolution. The final 3D maps were visualized and analyzed, and figures were created using the UCSF chimera package . A protein partial specific volume of 0.73 cm3·g−1 was used to set the isosurface threshold that corresponded to the correct molecular volume.
Because of the availability of apoECD and ANP–ECD crystal structures, we also performed reference-based refinement (eman) as a means of evaluating agreement of the single-particle data with the X-ray crystallographic data. The crystal structures of apoECD (Protein Data Bank code: 1DP4) and ANP–ECD (Protein Data Bank code: 1T34) each contain tt dimer and hh dimer pairs. Density maps were created from the hh and tt dimer pairs at a resolution comparable to the EM data (pdb2mrc; 22 Å resolution) for each of apoECD and ANP–ECD. These density maps were then used as starting models for the refine command in eman. Up to 20 cycles of refinement were performed. Depending on whether the hh or tt dimer map was used as the starting model, the refinement quickly diverged from an incorrect solution, and it converged on the correct solution within 20 cycles of refinement. Finally, fitting of the atomic coordinates of the hh or tt dimer pairs to the EM reconstructions was performed with eman (foldhunterp). Calculated density maps from each atomic model were used as reference structures for the calculation.
Steady-state fluorescence spectroscopic studies of Trp residues
Fluorescence emission spectra were acquired in a Fluorolog-222 fluorescence spectrometer using fluorescence software over the wavelength range from 305 to 500 nm with excitation at 291 nm and an emission slit width of 2 nm. All experiments were carried out at 22 °C.
ECD or mutated ECD W74R , in which Trp74 was replaced by Arg, at 1 mg·mL−1 concentration in 5 mm Hepes buffer (pH 7.0), containing 20 mm NaCl was used in the experiments. Fluorescence emission spectra of ECD or ECD W74R were acquired before and after the addition of a 1.1-fold molar excess of the truncated ANP peptide. The change in the emission spectrum was followed at 2 min intervals over a period of 60 min.
The work was supported by HL54329 to K. S. Misono and by grants to H. S. Young from the Canadian Institutes for Health Research, the Canada Foundation for Innovation, and the Alberta Science and Research Investments Program. H. S. Young is a Senior Scholar of the Alberta Heritage Foundation for Medical Research.