Communication Cite This: Biochemistry XXXX, XXX, XXX−XXX
pubs.acs.org/biochemistry
R‑Domain Phosphorylation by Protein Kinase A Stimulates Dissociation of Unhydrolyzed ATP from the First Nucleotide-Binding Site of the Cystic Fibrosis Transmembrane Conductance Regulator Luba A. Aleksandrov, Jonathan F. Fay,* and John R. Riordan* University of North Carolina, Chapel Hill, North Carolina 27599, United States
Downloaded via UNIV OF SOUTH DAKOTA on August 27, 2018 at 13:26:38 (UTC). See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.
S Supporting Information *
another part of CFTR is required for the trapping of ATP by NBD1. Biochemical studies that measured the long lifetime of ATP at NBD1, however, were performed without stimulation by PKA, which is necessary to activate the channel.2,3 Therefore, we hypothesized the R-domain plays a major role in ATP trapping at NBD1. Additionally, our recent single-molecule cryo-electron microscopy (cryo-EM) structural study of dephosphorylated and ATP present CFTR5 suggests that the dephosphorylated R-domain is located between the two halves of the CFTR. Importantly, we observed cryo-EM density for an R-domain region that appears to interact with intracellular loop 1 (ICL1) and around the NBD1 ATP-binding pocket (Figure 1A).
ABSTRACT: The cystic fibrosis transmembrane conductance regulator (CFTR) is an asymmetric ATPbinding cassette transporter in which ATP hydrolysis occurs only at the second of the two composite nucleotide-binding sites whereas there are noncanonical substitutions of key catalytic residues in the first site. Therefore, in widely accepted models of CFTR function, ATP is depicted as remaining bound at the first site while it is hydrolyzed at the second site. However, the long lifetime of ATP at nucleotide-binding domain 1 (NBD1) had been measured under conditions where the channel had not been activated by phosphorylation. Here we show that phosphorylation by protein kinase A (PKA), obligatory for channel activation, strongly accelerates dissociation of the unhydrolyzed ATP from NBD1 of both full-length and NBD2-deleted CFTR. This stimulation of nucleotide release results from phosphorylation of the CFTR regulatory domain (residues 634−835) (Rdomain). Mimicking phosphorylation by mutating the eight phosphorylation sites in the R-domain (8SE) has the same robust effect on accelerating the dissociation of ATP from NBD1. These findings provide new insight into relationships between R-domain phosphorylation and ATP binding and hydrolysis, the two main CFTR regulatory pathways.
Figure 1. R-Domain modulates ATP “trapping” at NBD1. (A) Closeup of our cryo-EM structure5 of dephosphorylated CFTR showing strong R-domain density (yellow) around the ATP-binding site of NBD1. (B) PKA phosphorylation greatly accelerates dissociation of ATP from NBD1 of WT CFTR with phosphorylation by PKA in the presence of the indicated ATP concentrations. Mean values ± the standard deviation are shown for at least three independent experiments.
he cystic fibrosis transmembrane conductance regulator (CFTR) is a unique ATP-binding cassette (ABC) ion channel mutated in patients with cystic fibrosis. CFTR is composed of two transmembrane domains (TMDs) and two nucleotide-binding domains (NBDs), like many other ABC transporters, and a CFTR regulatory domain (residues 634− 835) (R-domain), which is unique to CFTR and plays an important role in the regulation of channel activity. ATP binding and hydrolysis at the NBDs and phosphorylation of Rdomain orchestrate channel gating.1 Because of noncanonical substitutions of key catalytic residues, ATP is not hydrolyzed in NBD1.2,3 Previously, we had shown that the release of ATP from NBD1 essentially did not occur at low temperatures and ATP release was still very slow (T1/2 ∼ 30 min) at physiological temperatures.4 Moreover, deletion of NBD2 did not appear to alter the release of ATP from NBD1.4 Interestingly, purified NBD1 alone, without the context of the full protein, did not “trap” ATP.4 These findings suggest NBD2 is not required for ATP trapping at NBD1, and thus,
T
© XXXX American Chemical Society
Consistent with NMR data,6 the cryo-EM R-domain density is no longer observed after PKA treatment in our threedimensional reconstructions of phosphorylated CFTR.5 Similar, although not identical, observations have been made in zebrafish CFTR cryo-EM maps.7,8 As CFTR is constantly present in an environment with millimolar concentrations of ATP in cells and ATP turnover at NBD1 is slow, it is reasonable to postulate that R-domain phosphorylation, one of the two main regulatory pathways of CFTR channel gating, may play a role in modulating the interaction of ATP with NBD1. Thus, in this work, we sought to investigate if Received: June 12, 2018 Revised: August 8, 2018
A
DOI: 10.1021/acs.biochem.8b00646 Biochemistry XXXX, XXX, XXX−XXX
Communication
Biochemistry
Using the basic ATP transporter architecture as an anion channel, CFTR requires an additional level of control other than ATP binding and hydrolysis; otherwise, it would be continuously active as its ATP ligand concentration in cells does not change appreciably and the permeating anions, although they may have modulating effects,15 do not strongly influence activity. Instead, the overriding control is provided mainly by changes in the phosphorylation state of the Rdomain. The simplest model of primary control of channel function would be for the unphosphorylated state of the Rdomain to plug the channel. This model is consistent with the functional ΔNBD2 CFTR mutant in which PKA phosphorylation is still a prerequisite for channel activation, despite the lack of NBD2.13,16 In our observed cryo-EM conformation of dephosphorylated ATP-bound CFTR, the R-domain appears to plug the intracellular vestibule and interact with NBD1 and thus could block NBD1 ATP release.5 The interaction between the R-domain and NBD1 is weakened upon phosphorylation.6,17 This is consistent with the lack of the R-domain observed in our cryo-EM PKA-treated, phosphorylated CFTR.5 Again, PKA phosphorylation activates ΔNBD2 CFTR channels,13 an effect that is very pronounced when single gain of function mutations are introduced into membrane-spanning domains.16 We have shown here that one effect of phosphorylation of ΔNBD2 CFTR is the accelerated release of occluded ATP from NBD1. It is unknown if this acceleration of nucleotide dissociation by phosphorylation, which reflects an increased rate of ATP turnover at NBD1, is involved in channel activation as there may also be more direct impacts of phosphorylation on the TMDs. Kirk and colleagues have argued that PKA activation of ΔNBD2 CFTR containing a gain of function mutation is independent of ATP as it still occurs when bulk ATP has been removed from the solution.18 How does phosphorylation of the R-domain regulate ATP turnover at NBD1? Our recent cryo-EM data suggest it involves interaction with ICL1 and NBD1 itself (Figure 1A). This interaction is consistent with altered interactions of the Rdomain with NBD1 demonstrated by Baker et al.6 or as yet undefined allosteric couplings proposed by Wang et al.19 An interpretation of these results, which are consistent with our cryo-EM structural observations, is depicted in Figure 2. As illustrated, the R-domain when unphosphorylated (yellow)
phosphorylation of the R-domain can modulate the release of ATP from NBD1. Previously, we have developed a well-validated ATP dissociation assay2,4that provides a means of monitoring the rate of release of ATP from NBD1 under various conditions at physiological temperatures. Using this approach, we found that phosphorylation of membranes containing wild-type (WT) CFTR with PKA in the presence of 50 μM ATP caused a large increase in the apparent dissociation rate of [γ-32P]-8-N3ATP prebound at NBD1 (Figure 1B). PKA treatment at 2 mM ATP, a concentration well above the Km of the enzyme,9 caused a further increase in the rate of [γ-32P]-8-N3ATP release (Figure 1B). Similar results were found with [α-32P]-8-N3ATP, demonstrating that release but not hydrolysis is responsible for this effect (data not shown). To further test the role of R-domain phosphorylation, we utilized the gain of function substitution of the serines at eight of the sites with glutamates (8SE, 660, 686, 700, 712, 737, 768, 785, and 813) as phosphorylation mimics, a maneuver that produces essentially complete channel activation.10,11 There was a rapid dissociation of [γ-32P]-8-N3ATP from 8SE-CFTR without phosphorylation (Figure S1A), similar to that with PKA treatment of WT CFTR (Figure 1B). Moreover, no additional phosphorylation or the presence of cold ATP is required for the observed accelerated release of ATP from NBD1. This result provides evidence that mimicking R-domain phosphorylation sites alone is sufficient to cause accelerated release of unhydrolyzed ATP from NBD1. We also determined if the PKA-stimulated release from the full-length protein involved NBD2. Comparison of the release of [γ-32P]-8-N3ATP from NBD1 in full-length form and a mutant lacking NBD2, ΔNBD2 CFTR, showed that PKA phosphorylation (with 2 mM ATP) elicited strong responses in both cases (Figure S1B). These observations further confirm that R-domain modulation of the release of ATP from NBD1 can occur in the absence of NBD2. This study has addressed a novel aspect of the relationship between the two regulatory mechanisms that control the activity of the CFTR anion channel, i.e., ATP binding and turnover at the degenerate nucleotide-binding site and phosphorylation by protein kinase A. PKA phosphorylation remains essential for nonhydrolytic channel activity supported by nonhydrolyzable ATP analogues12 and ΔNBD2 constructs in which interactions between the two NBDs play no role.13 Therefore, phosphorylation must have another mode of action, different from preventing NBD dimerization. Indeed, our recent cryo-EM structure of dephosphorylated CFTR indicates direct interaction of the R-domain with NBD1, around its ATP-binding site.5 Thus, we tested the possibility that it might influence the trapping of the unhydrolyzed ATP at NBD1. Repeating the experimental protocols that we had used previously to demonstrate ATP trapping at this site4 with the addition of PKA, exactly as in channel activation experiments,14 revealed that the bound intact nucleotide was released at a greatly accelerated rate (Figure 1B). In addition, 8SE CFTR, which mimics the phosphorylated state of the Rdomain, enhanced the dissociation of ATP from NBD1 to an extent similar to that of PKA-treated WT CFTR (Figure S1A). A similar effect of PKA phosphorylation-stimulated release of ATP from NBD1 was found in ΔNBD2 CFTR (Figure S1B), indicating dissociation of ATP from NBD1 can occur in the absence of NBD2.
Figure 2. Cartoon depicting the involvement of the R-domain (yellow) in modulating nonhydrolytic ATP exchange from NBD1 (blue) before (left) and after PKA phosphorylation (right; orange circles indicate R-domain phosphorylation). Note the R-domain interacts either directly or indirectly with the ATP-binding site under dephosphorylated conditions (left), and after phosphorylation, the Rdomain no longer interacts with NBD1 (right), thereby allowing faster nonhydrolytic ATP turnover (indicated by the larger arrows) at the degenerate NBD1 site at physiologically relevant temperatures. B
DOI: 10.1021/acs.biochem.8b00646 Biochemistry XXXX, XXX, XXX−XXX
Communication
Biochemistry
(9) Li, C., Ramjeesingh, M., Wang, W., Garami, E., Hewryk, M., Lee, D., Rommens, J. M., Galley, K., and Bear, C. E. (1996) J. Biol. Chem. 271, 28463−28468. (10) Rich, D. P., Berger, H. A., Cheng, S. H., Travis, S. M., Saxena, M., Smith, A. E., and Welsh, M. J. (1993) J. Biol. Chem. 268, 20259− 20267. (11) Seibert, F. S., Chang, X. B., Aleksandrov, A. A., Clarke, D. M., Hanrahan, J. W., and Riordan, J. R. (1999) Biochim. Biophys. Acta, Biomembr. 1461, 275−283. (12) Aleksandrov, A. A., Chang, X., Aleksandrov, L., and Riordan, J. R. (2000) J. Physiol. 528 (Part 2), 259−265. (13) Cui, L., Aleksandrov, L., Chang, X. B., Hou, Y. X., He, L., Hegedus, T., Gentzsch, M., Aleksandrov, A., Balch, W. E., and Riordan, J. R. (2007) J. Mol. Biol. 365, 981−994. (14) Anderson, M. P., Berger, H. A., Rich, D. P., Gregory, R. J., Smith, A. E., and Welsh, M. J. (1991) Cell 67, 775−784. (15) Broadbent, S. D., Ramjeesingh, M., Bear, C. E., Argent, B. E., Linsdell, P., and Gray, M. A. (2015) Pfluegers Arch. 467, 1783−1794. (16) Kirk, K. L., and Wang, W. (2011) J. Biol. Chem. 286, 12813− 12819. (17) Bozoky, Z., Krzeminski, M., Muhandiram, R., Birtley, J. R., AlZahrani, A., Thomas, P. J., Frizzell, R. A., Ford, R. C., and FormanKay, J. D. (2013) Proc. Natl. Acad. Sci. U. S. A. 110, E4427−4436. (18) Wang, W., Wu, J., Bernard, K., Li, G., Wang, G., Bevensee, M. O., and Kirk, K. L. (2010) Proc. Natl. Acad. Sci. U. S. A. 107, 3888− 3893. (19) Wang, W., Roessler, B. C., and Kirk, K. L. (2014) J. Biol. Chem. 289, 30364−30378.
clearly interacts with NBD1 [as evidenced by cryo-EM (Figure 1A)], and this interaction can block ATP release. Upon phosphorylation, necessary for channel activation, the phosphorylated R-domain (yellow and orange) is no longer observed interacting with NBD1, in line with our cryo-EM observations.5 In summary, the nonhydrolytic release of ATP from NBD1 at physiological temperatures can occur more rapidly (Figure 1B) without the inhibitory modulation of the R-domain, in the activated channel. To reiterate, under physiological conditions, ATP is bound more tightly to NBD1 when the R-domain is unphosphorylated.4 After PKAmediated phosphorylation of the R-domain, the release of ATP from NBD1 is accelerated, allowing for nucleotide turnover at this site. These new findings provide a potential mechanism for how phosphorylation of the R-domain can facilitate the release of ATP from the NBD1-binding site in CFTR.
■
ASSOCIATED CONTENT
* Supporting Information S
The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.biochem.8b00646.
■
Materials and Methods, release rates for 8SE and ΔNBD2 (Figure S1) and representative autoradiograms for Figures 1B and S1 (Figure S2) (PDF)
AUTHOR INFORMATION
Corresponding Authors
*E-mail:
[email protected]. *E-mail:
[email protected]. ORCID
Jonathan F. Fay: 0000-0003-1822-2384 Funding
This work was supported by National Institutes of Health Grants R01-DK051870 and P01-HL10873-01 and by grants from the Cystic Fibrosis Foundation FAY16F0 and RIORDA07XX0. Notes
The authors declare no competing financial interest.
■
ACKNOWLEDGMENTS The authors thank Tim Jensen and Adrei Aleksandrov for insightful discussions and critical reading of the manuscript.
■
REFERENCES
(1) Hwang, T. C., Yeh, J. T., Zhang, J., Yu, Y. C., Yeh, H. I., and Destefano, S. (2018) J. Gen. Physiol. 150, 539−570. (2) Aleksandrov, L., Aleksandrov, A. A., Chang, X. B., and Riordan, J. R. (2002) J. Biol. Chem. 277, 15419−15425. (3) Basso, C., Vergani, P., Nairn, A. C., and Gadsby, D. C. (2003) J. Gen. Physiol. 122, 333−348. (4) Aleksandrov, L., Aleksandrov, A., and Riordan, J. R. (2008) Biochem. J. 416, 129−136. (5) Fay, J. F., Aleksandrov, L. A., Jensen, T. J., Cui, L. L., Kousouros, J. N., He, L., Aleksandrov, A. A., Gingerich, D. S., Riordan, J., and Chen, J. Z. (2018) bioRxiv, n/a. (6) Baker, J. M., Hudson, R. P., Kanelis, V., Choy, W. Y., Thibodeau, P. H., Thomas, P. J., and Forman-Kay, J. D. (2007) Nat. Struct. Mol. Biol. 14, 738−745. (7) Zhang, Z., and Chen, J. (2016) Cell 167, 1586−1597. (8) Zhang, Z., Liu, F., and Chen, J. (2017) Cell 170, 483−491. C
DOI: 10.1021/acs.biochem.8b00646 Biochemistry XXXX, XXX, XXX−XXX
