The human Na+/H+ exchanger 1 is a membrane scaffold protein for extracellular signal-regulated kinase 2
Extracellular signal-regulated kinase 2 (ERK2) is an S/T kinase with more than 200 known substrates, and with critical roles in regulation of cell growth and differentiation and currently no membrane proteins have been linked to ERK2 scaffolding.
Methods and results
Here, we identify the human Na+/H+ exchanger 1 (hNHE1) as a membrane scaffold protein for ERK2 and show direct hNHE1-ERK1/2 interaction in cellular contexts. Using nuclear magnetic resonance (NMR) spectroscopy and immunofluorescence analysis we demonstrate that ERK2 scaffolding by hNHE1 occurs by one of three D-domains and by two non-canonical F-sites located in the disordered intracellular tail of hNHE1, mutation of which reduced cellular hNHE1-ERK1/2 co-localization, as well as reduced cellular ERK1/2 activation. Time-resolved NMR spectroscopy revealed that ERK2 phosphorylated the disordered tail of hNHE1 at six sites in vitro, in a distinct temporal order, with the phosphorylation rates at the individual sites being modulated by the docking sites in a distant dependent manner.
This work characterizes a new type of scaffolding complex, which we term a “shuffle complex”, between the disordered hNHE1-tail and ERK2, and provides a molecular mechanism for the important ERK2 scaffolding function of the membrane protein hNHE1, which regulates the phosphorylation of both hNHE1 and ERK2.
KeywordsNHE1 Intrinsically disordered protein Phosphorylation MAPK Shuffle complex NMR Scaffold
Extracellular signal-regulated kinase 2 (ERK2) is a member of the mitogen-activated protein kinase (MAPK) family of kinases activated in response to numerous growth factors and cytokines, leading to phosphorylation and functional regulation of downstream targets. ERK2 has been linked to more than 200 different substrates whose phosphorylation by ERK2 is orchestrated by coordination of signaling networks through common binding to so-called scaffold proteins . The definition of a scaffold protein was recently refined and their identification as such suggested from qualities of multivalent binding, non-catalytic placeholders, and bidirectional process control . Several scaffold proteins have been described for the MAPKs such as kinase suppressor of Ras (KSR) , JNK-interacting protein (JIP) , IQ motif containing GTPase activating protein 1 (IQGAP1) , and β-arrestin , which interact with members of the MAPK cascade, providing multivalency, spatial concentration, and/or signaling fidelity. However, although MAPKs are known to regulate the action of several membrane proteins and receptors, none of these scaffold proteins are themselves membrane proteins, requiring additional mechanisms for colocalization of the scaffold protein, the membrane protein, as well as the kinases. Moreover, most of the available molecular insights are from structures of kinases in complex with folded domains or with small peptides of the scaffold proteins, and details regarding scaffolding by non-globular proteins are lacking.
MAPKs are S/T kinases that interact with targets and regulators via two types of domains, D-domains and F-sites [7, 8, 9, 10, 11]. D-domains, also known as docking sites for ERK and JNK, LXL (DEJL) domains, or kinase interaction motifs (KIMs) have the canonical sequence of 2–5 basic residues (R/K), spaced by 1–6 residues to a hydrophobic motif ΦXΦ, where Φ is generally V, L, or I [8, 9]. D-domains are found in MAPK substrates such as the transcription factor Elk-1 and p90 ribosomal S kinase (RSK1-3), as well as in other MAPK targets [8, 9, 11]. In ERK2, D-domains interact with the D-domain recognition site also known as the CD/ED (common docking domain/glutamate/aspartate docking) domain, located more than 10 Å from the active site [8, 9, 11]. The F-site recruitment site in ERK2 is much less studied and incompletely understood. It binds to F-sites, also called DEF (docking site for ERK, FXFP)-domains with the canonical FXFP sequence . F-sites allow for aromatic residues at the P1 (F, W) and P3 positions (F, Y, W) , and F-sites have been reported in substrates such as Elk-1 (FQFP)  and c-Fos (FTYP) , and within the nucleoporin FG-repeats (FXFG) [16, 17]. So far the only structure available of an F-site recruitment site-interacting protein is that of ERK2 in complex with the 15 kDa phosphoprotein enriched in astrocytes (PEA-15), which notably lacks any of the above-mentioned motifs .
The plasma membrane Na+/H+ exchanger 1 (NHE1, SLC9A1) is a major regulator of pH and volume in essentially all cells studied. Furthermore, NHE1 is involved in the regulation of cell proliferation, survival, motility, and other essential physiological processes, and its dysregulation contributes importantly to major human malignancies, including cancer and cardiovascular diseases [19, 20]. Numerous hormones and growth factors acting via receptor tyrosine kinases or GTP-binding protein-coupled receptors can elicit posttranslational regulation of NHE1 [21, 22, 23]. The MAPKs ERK1/2, p38 MAPK, and c-Jun N-terminal kinase (JNK) are widely implicated in NHE1 regulation [24, 25, 26, 27, 28], and direct phosphorylation of human (h) NHE1 by ERK1/2 was previously proposed based on 32P measurements  and mass spectrometry . Conversely, NHE1 has been reported to regulate signaling through regulation of ERK1/2 and p38 MAPK activity [26, 28, 30, 31, 32], and yeast two-hybrid screens have suggested the interaction of NHE1 with several members of the MAPK hierarchy . However, with the exception of the interaction with B-Raf , evidence from mammalian systems is lacking, and the possible sites of NHE1-MAPK interaction, its structural details, and possible functional consequences are unexplored. We recently showed by PONDR and DISOPRED predictions, as well as by nuclear magnetic resonance (NMR) spectroscopy and other biophysical techniques, that the distal ~ 130 residues of the hNHE1 C-terminal intracellular domain (hNHE1cdt), containing most of the known NHE1 phosphorylation sites, are intrinsically disordered (ID) [35, 36]. To our knowledge, no studies have yet addressed the mechanisms through which MAPKs interact with ID proteins (IDPs), although about one third of all proteins in higher eukaryotes contain significant ID regions (IDRs) , and ID is abundant in cellular signalling , scaffolding , as well as in MAPKs themselves .
Here, we demonstrate that hNHE1 acts as an ERK2 membrane protein scaffold in vivo that is necessary for ERK2 activation via direct interactions, and we show that loss of scaffolding by hNHE1 leads to decreased ERK2 activation. Using NMR spectroscopy we show that NHE1 scaffolds inactive (ia) ERK2 in a “shuffle complex” that involves a D-domain and two non-canonical F-sites. We characterize the order and kinetics of both previously reported and novel ERK2-mediated phosphorylations of hNHE1 in vitro. Our findings provide a molecular mechanism for the widely recognized and functionally important scaffolding function of hNHE1, and give mechanistic insight into the regulation of ERK2 activity by the intrinsically disordered hNHE1cdt.
The disordered tail of hNHE1 interacts with iaERK2
We next investigated whether interfering with any of these contact sites would affect the interactions and first exploited the knowledge that MAPK interaction is severely perturbed by mutations of ΦXΦ to AXA in D-domains . Hence, we constructed AXA variants of all three D-domains alone and in combination, both in the full-length hNHE1 (hNHE1-D1-AXA, hNHE1-D2-AXA, etc.) for cellular studies and in hNHE1cdt (D1-AXA, D2-AXA, D3-AXA) for in vitro studies. In D3-AXA, chemical shift perturbations in the two F-sites upon ERK2 addition were preserved in the interaction with iaERK2, whereas no perturbations were observed in the AXA-mutated D3-domain, implying that this site is important for the interaction (Fig. 2e and Additional file 2: Figure S2c). No notable effects of D1- and D2-AXA substitutions were observed (Additional file 2: Figure S2d–e), arguing against their involvement in the interactions. This data also indicated that the F-sites interacted with iaERK2 independently of the D-domain. To assess this further the F-sites were individually mutated by substituting FTP778–780 with ATP778–780 (denoting the F1-A variant) and FP811–812 with AA811–812 (denoting the F2-AA variant). Both F-site variants showed strongly decreased chemical shift perturbations at the mutation sites upon ERK2 addition, leaving the other F-site and the D3-domain unaffected (Fig. 2f–g and Additional file 2: Figure S2f–g). This conclusively identified all three sites as ERK2 interaction sites. Further, substitutions at each site left the other sites unaffected, indicating that these regions of hNHE1 interact independently with iaERK2.
NHE1 does not fold upon binding to ERK2 but may be a flexible wrapper
To further address how hNHE1cdt interacted with iaERK2 we analysed the complex by size exclusion chromatography (SEC) and compared elution profiles with those of the individual proteins (Additional file 3: Figure S3a–c). Since hNHE1cdt is an IDR, it has a larger hydrodynamic radius than iaERK, and hNHE1cdt thus eluted first from the column. Subtracting individual runs from that of the mixture revealed a broad peak with an elution volume smaller than that of iaERK2, yet larger than that of hNHE1cdt, suggesting that hNHE1cdt folds or wraps around iaERK2. Circular dichroism (CD) spectroscopic analyses (Additional file 3: Figure S3d), as well as NMR chemical shift analyses (Fig. 2b), did not indicate folding upon binding formation of significant secondary structure, suggesting that hNHE1cdt forms a relatively extended structure around iaERK2. To substantiate this conclusion, we recorded 15N transverse relaxation rates of the unbound (R 2 free) and the iaERK2-bound hNHE1cdt (R 2 bound), and analysed their differences (Fig. 2d). Since D3 residues broadened beyond detection in the complex, their R 2 values could not be extracted. However, for residues interacting with iaERK2, a significant increase in R 2 rates is expected compared to those of hNHE1cdt alone, due to the larger radius of gyration of the complex or due to chemical exchange between different states. Indeed, residues from both F-sites had substantially larger R 2 rates in the complex compared to hNHE1cdt alone, and many residues between these sites were also affected, although not to the same extent. This result supports the SEC results and suggests a substantial interaction area between the two proteins.
Finally, we used the NMR chemical shifts and relaxation data together with known structures of iaERK2 complexes to model the hNHE1cdt-iaERK2 interaction (Fig. 3b). When modelled at the D-domain recognition site of iaERK2, each individual D-domain (D1, D2, or D3) when bound to the D-domain recognition site allowed for either of the F-sites (F1 or F2) to reach the F-site recognition site. In each case the model predicted long ID linkers between the binding sites, which were long enough to allow NHE1 to wrap around iaERK2. This is in accordance with the NMR data and supports their uncoupled behaviour. Collectively, these data suggest that hNHE1cdt could interact with iaERK2 in a tripartite 1:1 interaction exploiting a D-domain (D3) and two F-sites.
ERK2 phosphorylates hNHE1cdt at six consensus sites in a distinct order and with different kinetics
D-domains play differential roles in scaffolding and activation
Apparent rate constants, k app for hNHE1cdt phosphorylation by aERK2, and effect of D-domain and F-site mutations
k app (h-1)
0.3489 ± 0.0101
0.3488 ± 0.0112
0.3193 ± 0.0109
0.3000 ± 0.009
0.4435 ± 0.0177
0.0821 ± 0.0018
0.0592 ± 0.0010
0.0610 ± 0.0011
0.0741 ± 0.0015
0.1274 ± 0.0018
0.0088 ± 0.0015
0.0070 ± 0.0011
0.0066 ± 0.0011
0.0047 ± 0.0008
0.0088 ± 0.0015
0.1648 ± 0.0022
0.1413 ± 0.0017
0.1572 ± 0.0021
0.1473 ± 0.0024
0.1441 ± 0.0018
0.0305 ± 0.0016
0.0325 ± 0.0016
0.0302 ± 0.0017
0.0088 ± 0.0019
0.0120 ± 0.0021
In detail, S693 phosphorylation was not affected by the D3-AXA mutation, but was slowed down by the D1D2-(AXA)2 and F1-A variants, and accelerated by the most distant F2-AA variant. Similarly, S723/S726 phosphorylations were slowed down by all mutations except the most distant one, F2-AA that again led to accelerated rates. T779 phosphorylation was slowed down by all mutations, and S785 was strongly decreased by both F-site mutants. Taken together, mutations of the D-domains and F-sites affected phosphorylation in a distance dependent manner. In three cases (S693 and S723/S726 phosphorylation of the F2-AA variant) the rate of phosphorylation went up, suggesting that the presence of this site was inhibitory. In all three cases this occurred for the site furthest away from the mutation, reflecting competition between sites. This suggests that the D-domains and F-sites are not mandatory for phosphorylation, but rather exert regulatory roles, and that each site uses the most optimal ERK2 interaction site to become phosphorylated.
NHE1 regulates ERK2 phosphorylation status in a cellular context
NHE1 activity after various stimuli is regulated by ERK1/2, and the NHE1 C-terminal tail is directly phosphorylated by ERK2 in vitro  and in vivo . Vice versa, NHE1 can regulate ERK1/2 activity [26, 28, 30, 31, 32], yet molecular details and mechanistic understanding of their interaction have been lacking. In conjunction with yeast two-hybrid studies suggesting interaction of NHE1 with MAPKs , these studies led us to hypothesize that NHE1 and ERK2 engage in direct physical interaction. Supporting this hypothesis, we report here that ERK2 and NHE1 interact directly in vivo as well as in vitro, and that ERK2 phosphorylates multiple sites in hNHE1cdt. Based on NMR analyses, in conjunction with various combinations of D-domain and F-site mutations, we suggest that hNHE1cdt scaffolds iaERK2 and that they interact in a non-cooperative modular manner that involves a D-domain (D3) and two F-sites. NMR titrations revealed the affinity of the D3-domain to be in the low micromolar range, in agreement with known D-domain affinities . D-domain and F-site mutations did not prevent ERK2-mediated hNHE1 phosphorylation in vitro, but altered its kinetics. In vivo, hNHE1 and ERK1/2 co-localized at the plasma membrane in a manner sensitive to ERK1/2 stimulation, and mutations in the hNHE1 D3-domain and F-site altered ERK1/2 activity. Thus, a central conclusion of this work is that NHE1 and ERK2 directly interact and engage in physical and functional reciprocal interactions. This provides a novel molecular framework for understanding previous reports of both NHE1-mediated scaffolding and regulation of ERK1/2  and ERK1/2-mediated phosphorylation of NHE1 .
NHE1 is, to the best of our knowledge, the first membrane protein described to scaffold members of the MAPK pathway, spanning all four levels of the MAPK hierarchy . Many soluble scaffold proteins acting together with, for example, G protein-coupled receptors (GPCRs) and growth factor receptors, have been described, but detailed interaction data have not provided insight into how scaffolding and regulation are coupled. In the cell, MEKs, phosphatases, and substrates all compete for the D-domain recognition site on ERK2 , and it is currently not known how NHE1 interacts with the other MAPK members, including MEK, and assembles a signalling complex. As NHE1 acts as a dimer in vivo [36, 63], we propose that upon release of the D3-domain from one NHE1 monomer due to competition with MEK, the shuffle complex organization keeps ERK2 in place via scaffolding by D-domains of the other NHE1 subunit in the NHE1 dimer, or by the F-sites (Fig. 7a, b). Furthermore, in a potential cellular complex, the remaining D-domains as well as F-sites will be available for further scaffolding of other members of the MAPK hierarchy (Fig. 7c), as suggested by yeast two-hybrid screens, potentially MEK, although this remains to be explored.
NHE1 is the first example of an ID substrate of ERK2 for which detailed interaction data now exist, and to the best of our knowledge, no other IDP or IDR has to date been experimentally linked to ERK2 phosphorylation or scaffolding. It has been noted that the D-domain of the tyrosine-phosphatase PTP-SL resides in a region with high disorder propensity . Disorder predictions of the nuclear pore protein Tpr (Additional file 7: Figure S7) show its ERK2-interacting F-site to reside in an IDR, similar to the nucleoporin FG-repeat regions and to hNHE1cdt. Thus, it appears that the F-site recruitment site interaction may be frequently exploited by IDPs. However, it remains to be seen whether the multi-site shuffle interaction is a novel canonical IDP/IDR-ERK2 interaction mode or if it is a unique scaffolding function specific to NHE1.
All six putative ERK2 phosphorylation sites of hNHE1cdt were phosphorylated by aERK2 in vitro (S693, S723, S726, S771, T779, and S785). Previous in vivo phosphoproteomics have mapped phosphorylation at five of these sites, yet without identification of the responsible kinases, and with no information on the sequence of individual phosphorylation events [48, 65]. Further, some of these sites, i.e. S693, T779, and S785, were previously identified upon in vitro phosphorylation of NHE1 by ERK2 . No non-canonical phosphorylation was detected in the present study, in contrast with previous reports identifying S766, S770, and S771 as ERK-dependent NHE1 phosphorylation sites [25, 29]. While it is possible that additional complexity may be introduced in the in vivo setting, our data underscore the major advantage of NMR for direct identification of phosphorylation sites.
Physiological roles have been proposed for all six phosphorylations, although their interplay and dynamics have never previously been assessed. S723 and S726 (corresponding to S722 and S725 in rabbit NHE1) were reported to be phosphorylated by p38 MAPK in murine pro-B-cells , and phosphorylation of S726 was suggested to mediate apoptosis-induced alkalinization by NHE1 . Based on studies of NHE1 mutants expressed in NHE1-deficient CHO cells, S771 was reported to mediate ERK-dependent NHE1 activation , and later, S771, T779, and S785, but not S693, S723, or S726, which were assigned roles in ERK-dependent NHE1 phosphorylation after sustained acidosis in renal cells . In Amphiuma erythrocytes, phosphorylation of residues corresponding to S693 and S785 (S701 and S783) were detected by MS, where S785 (S783) was constitutively phosphorylated . The precise downstream effects of these phosphorylations are not currently known, but they are likely to both impact NHE1 structural dynamics and hence activity, and to contribute to the scaffolding role of NHE1 in regulation of ERK, hence fine-tuning cellular ERK signaling. Timing of signalling events is crucial for many cellular functions, and phosphorylation events that are interdependent or distributive with very different rate constants are possible ways of controlling signal duration and strength. Indeed, it has been suggested that such temporally ordered phosphorylations serve as platforms for signal integration . Our findings provide evidence for a distinct temporal order of ERK2 phosphorylation of hNHE1cdt with the occurrence of specific intermediates. These intermediates could function as tightly regulated docking sites or thresholds that convert graded signals to switch-like responses . Such dynamics in ERK2 signalling have been observed to affect the half-life of an ERK2 downstream effector, the transcription factor c-Fos . Timed phosphorylation events in hNHE1cdt may therefore similarly partake in control of the ERK2 signal duration. The hNHE1cdt has other confirmed phosphorylation sites than those demonstrated in the present study, and many more putative ones , several of which are close to the ERK2 interaction sites, for example S703, phosphorylated by RSK . These sites may mediate interactions with other binding partners, introducing additional layers of complexities, for example of pathway crosstalk.
The close proximity of the primary S693 and T779 phosphorylation sites to the D3-domain and F1-site, respectively, suggests regulatory role(s) for the interaction of hNHE1cdt with, and phosphorylation by, ERK2. Analogously, the first phosphorylation events in hNHE1cdt (S693 and T779) may change the binding mode and/or dynamics with ERK2 in a way that regulates the phosphorylation of succeeding sites, although we showed that they do not act as priming sites. The effects of hNHE1 variants on the phosphorylation kinetics support the hypothesis that all sites are at play within the shuffle complex, where the sites closest to the phosphorylation site are exploited for interaction with ERK2. The current data does not allow us to evaluate whether all six sites are phosphorylated during NHE1 activation. Some sites, and most likely the slowest ones observed here, may only react under certain physiological conditions. We hypothesize that such conditional phosphorylations could be important for a rheostatic regulation of both ERK2 and NHE1. Consequently, the results of this work open a series of new questions, both regarding the generality of shuffle complexes in scaffolding by IDPs, but also concerning the functional roles and spatial and temporal interconnectivity of the six identified phosphorylation sites in hNHE1.
In this work we have demonstrated that the intrinsically disordered region of hNHE1 acts as a membrane scaffold engaging ERK2 in a multi-site shuffle complex. We show that the interaction is recapitulated in vivo, and that co-regulation of hNHE1 and ERK2 manifests in distinct in vivo effects on ERK1/2 activity and in vitro effects on hNHE1 phosphorylation. Our work provides a molecular mechanism for the important scaffolding function of NHE1, and characterizes a direct interaction between the intrinsically disordered hNHE1cdt and ERK2, leading to hNHE1cdt phosphorylation and regulation of ERK1/2 activity.
Cloning and mutagenesis
The human NHE1 C-terminal distal tail was extended by six residues to M + I680-Q815 (hNHE1cdt; primer sequences presented in Additional file 8). The variants D1D2-(AXA)2, D3-AXA, D1D2D3-(AXA)3, F1-A, and F2-AA; S693A and T779A were prepared from the WT hNHE1cdt plasmid and the WT full-length hNHE1 in a pcDNA3.1 plasmid using a QuikChange II Kit (Stratagene). Final constructs were confirmed by sequencing (Eurofins MWG Operon).
Protein expression and purification
The expression and purification of unlabeled human ERK2, and of hNHE1cdt and 15N- and 13C,15N- labeled hNHE1cdt were performed essentially as in . All protein preparations were > 95 % pure judged from SDS-PAGE (Additional file 9: Figure S8). Details are presented in Additional file 8: Supplementary materials and methods.
Putative D-domains and consensus ERK2 phosphorylation sites in the hNHE1cdt were predicted by Scansite 3  and aligned with Clustal Omega . Intrinsic disorder was analysed using PONDR-FIT  and DISOPRED 3.1 .
All NMR spectra were recorded on Varian INOVA 750 MHz or 800 MHz 1H NMR spectrometers with a 5 mm triple resonance probe and a Z field gradient at 5 °C. Chemical shift referencing was relative to 4,4-dimethyl-4-silapentane-1-sulfonic acid (DSS), and spectra zero-filled, apodized, Fourier transformed, and baseline-corrected in NMRDraw , and analysed manually in CCPN Analysis . Backbone resonance assignments of hNHE1cdt and variants were done at 5 °C using 1.0–1.5 mM samples of 15N,13C-hNHE1cdt in PBS pH 7.2, 0.5 mM DSS, 10 mM dithiothreitol (DTT), and 10 % (v/v) 99.96 % D2O by standard 3D triple resonance experiments as described . Intrinsic random coil referencing was done from assignments of hNHE1cdt in 8 M urea from similar experiments. Chemical shift perturbations of hNHE1cdt WT and variants from interaction with unlabeled iaERK2 were determined using 0.1 mM 15N-labeled hNHE1cdt WT or variants in the presence/absence of equimolar iaERK2 and 15N,1H-HSQC spectral analyses, dialyzed against PBS pH 7.4, 5 mM EDTA, and added 0.5 mM DSS, 2 mM DTT, 10 % (v/v) 99.96 % D2O prior to recording. Chemical shift perturbations of hNHE1cdt WT by variation of pH were determined using 0.1 mM 15N-labeled hNHE1cdt WT in PBS, 2 mM DTT, 0.5 mM DSS, 10 % (v/v) 99.96 % D2O at pH 7.2 and 7.4 and 15N,1H-HSQC spectral analyses. 15N transverse relaxation times (R 2 ) were determined using standard 15N,1H-HSQCs at a 750 MHz proton frequency field at 5 °C with relaxation decays extracted from a nine-step relaxation delay (0.01, 0.05, 0.09, 0.13, 0.17, 0.19, 0.21, 0.23, and 0.25 s). R 2 values were calculated by fitting the height of each peak to a single exponential decay function, and each fit was manually reviewed.
Native mass spectrometry
Protein samples were dialyzed against 200 mM ammonium acetate supplemented with 0.5 mM DTT. To detect the complex 12 μM ERK2 (MW 42343.7 Da) was mixed with a 4 × molar excess of NHE1cdt (48 μM, 14755.3 Da). Intact mass spectrometry experiments were performed on a Waters SYNAPT HDMS modified for high mass transmission as previously described . Gold-coated capillaries prepared in-house  were filled with sample and held at 1.6 kV against a sample cone of 200 V. Trap and transfer collision cells were maintained at 10 V, with an argon collision gas at 6.6 × 10-2 mbar. Backing pressure in the early ion optics was increased to 5.5 mbar to improve transmission of high m/z ions. Spectra were assigned using the UniDec software as previously described .
Model of hNHE1-ERK2 complex
The structure of iaERK2 in complex with PEA [PDB:4IZ5]  was used for F-site recruitment site modelling and the structure of iaERK2 in complex with a MAP kinase interacting kinase peptide [PDB:4H3Q]  for the D-recruitment site modelling. From the templates, the ERK2-NHE1 complex was modelled using Modeler version 9.11 , generating an ensemble of 1,000 models. These were clustered using the Linkage algorithm and the average structure from the most populated cluster selected as a final model. In the final model, the remaining linker regions were analysed to assess whether they were able to provide a structure of the complex compatible with the hydrodynamic radius observed experimentally. Since linker locations in the complex were ambiguous, they are not shown in the figure.
In vitro ERK2 phosphorylation assays by NMR
Time course experiments were run at 25 °C. Assignments were transferred to 25 °C by recording 15N,1H-HSQCs at 5, 10, 15, 20, and 25 °C. NMR samples of 400 μL of 100 μM or 200 μM 15N-labeled hNHE1cdt or variants were prepared in PBS buffer, 5 mM EDTA, 5 mM ATP, 15 mM MgCl2, 0.01 % (w/v) NaN3, 1 mM PMSF, 0.5 mM DSS, 2 mM DTT, 10 % (v/v) 99.96 % D2O, pH 7.0. A reference 15N,1H-HSQC spectrum was recorded before addition of kinase. Phosphorylation was started by addition of 10 μL of 0.1 mg/mL (55 nM, 401,000 units/mg) unlabeled active ERK2 (proteinkinase.de), resulting in a molar excess of hNHE1cdt:ERK2 of 1,800:1. Phosphorylation was monitored from extraction of peak position and volumes from a series of 15N,1H-HSQCs. Peak intensities were normalized to unperturbed residues (Q815). Kinetics of S693, S771, T779, and S785 phosphorylation were extracted from non-linear least squares fittings of disappearing peaks of the unphosphorylated state and/or appearing peaks of the phosphorylated state, as well as reporting neighbours to single exponentials (S693 (D692, S693, and R700), S771 (S770, S771, and G773), T779 (V777 and T779), and S785 (S785, S787, and S788)). Kinetics of S723 and S726 require fitting to bi-exponentials due to their crosstalk. For this purpose peak intensities reporting on the disappearing unphosphorylated state, the appearing phosphorylated state, and both intermediates I1 and I2 of both S723 and S726 were fitted. For the comparison of the NHE1cdt variants, one peak was chosen for each site and each variant, i.e. S693, S723, S771P, V777 (reporting on T779), and S785. The disappearing peak of S723 reports on both, S723 and S726 phosphorylation. Fully phosphorylated 15N,13C-hNHE1cdt was assigned from standard 3D triple resonance NMR experiments as above.
Mammalian cell culture and transfections
AP-1 cells (a kind gift from Dr. S. Grinstein, University of Toronto, ON, Canada), which are CHO-derived cells with no endogenous NHE activity  and no recovery from an acid load in the nominal absence of HCO3 - [41, 82], were used for all experiments in mammalian cells. AP-1 cells were grown at 37 °C, 5 % CO2, 95 % humidity in α-Minimum Essential Medium Eagle (Sigma) with 10 % fetal bovine serum, 1 % L-glutamine, 1 % penicillin/streptomycin (Gibco). Every 3–4 days, cells were passaged by gentle trypsination, and only passages 5–35 were used for experiments. WT and variant hNHE1 were expressed in AP-1 cells as in . Transfectants were selected for resistance to 600 μg/ml G418 (Calbiochem), individual clones picked, and hNHE1 expression verified by immunoblotting and immunofluorescence analysis.
EGF-mediated stimulation of ERK1/2 activity in AP-1 cells
Untransfected AP-1 cells or AP-1 cells expressing WT or variant hNHE1 were grown to ~ 80 % confluence in 10 cm Petri dishes, and incubated for 15 min in absence or presence of 100 ng/ml recombinant human EGF (Sigma). Cells were subsequently lysed and processed for immunoblotting as described below.
Immunoblotting was carried out essentially as in . Antibody descriptions and experimental details are provided in Additional file 8. For quantifications, blots were scanned, and band intensities quantified using Un-Scan-IT Graph Digitizer software (Silk Scientific). The pERK1/2 and ERK1/2 bands were normalized to those of the loading control (tubulin) from the same gel to eliminate gel-to-gel differences, and subsequently, pERK1/2 was taken relative to total ERK1/2 from the same experiment.
Immunofluorescence analysis was carried out essentially as in . Antibody descriptions and experimental details are provided in Additional file 8. Line scan quantification of immunofluorescence was performed using Olympus image analysis software, as the average pixel intensity at each wavelength across the line indicated. Co-localization was quantified as the percentage of cells with NHE1-ERK1/2 co-localization in both membranes, based on representative immunofluorescence images. Data are shown as mean percentage with SEM error bars, based on analysis of at least 60 cells in three to five independent replicates per condition.
Proximity ligation assay
Proximity ligation assay was carried out with the Duolink II Detection Reagents Red kit from Sigma Aldrich. AP-1 WT cells were seeded on coverslips the day before assaying. Cells were washed in ice-cold PBS, fixed in 4 % PFA for 20 min on ice, and washed in Duolink II Buffer A. Quenching was carried out in 0.1 M glycine for 15 min followed by permeabilization in 0.5 % Triton X-100. After permeabilization, cells were washed in Duolink II Buffer A and added O-link blocking solution for 30 min. Incubation with primary antibodies for 60 min in a humidity chamber at 37 °C was followed by 60 min incubation with PLA probes diluted 1:5 in Duolink II Antibody Diluent buffer at 37 °C. Coverslips were washed in Duolink II Buffer A and added ligation solution diluted 1:5 for 30 min at 37 °C, followed by wash in Duolink II Buffer A. Amplification solution diluted 1:5 was carried out for 100 min at 37 °C. After amplification, coverslips were washed in Duolink Buffer A, incubated with phalloidin488 for 1 h, and treated with DAPI to stain nuclei. Finally, coverslips were washed in Duolink Buffer A, mounted on object glass with mounting buffer, and sealed with nail polish. Imaging was carried out with an Olympus BX-61 epifluorescence microscope using cellSens Dimensions V1.6 software. Images were taken as z-stacks and z-projection images were created. Further image processing and quantification were carried out in ImageJ.
Data analysis and statistics
Data from mammalian cell culture on NHE1 function, immunofluorescence, and immunoblotting are shown as individual experiments representative of at least n = 3, or as means ± standard error of the mean (SEM) as indicated. ANOVA with Tukey post-test, or Student’s t-test, as appropriate, were used to test for statistically significant differences, with p < 0.05 as the significance level.
Availability of data and materials
Data supporting the results of this article are available in the Additional files 1, 2, 3, 4, 5, 6, 7, and 9, and further details on the materials and methods can be found in Additional file 8. Backbone assignments of the NHE1cdt have been deposited in the BioMagResBank [BMRB:26755].
We are grateful to J. B. Johansen, V. Nylander, K. Mark, G. W. Haxholm, and S. A. Sjørup for excellent technical assistance; S. Grinstein (University of Toronto, ON, Canada) for the AP-1 cells; M. Musch (University of Chicago, IL, USA) for the polyclonal NHE1 antibody; and G. W. Haxholm and L. Ellgaard for critical reading of the manuscript.
This study was supported by the Novo Nordisk Foundation (SFP), the Danish research council (SFP: FNU: 12-126942). BBK is supported by the VELUX FOUNDATION and grants from the Danish national research councils (12-128803; 4181-00344).
- 30.Javadov S, Baetz D, Rajapurohitam V, Zeidan A, Kirshenbaum LA, Karmazyn M. Antihypertrophic effect of Na+/H+ exchanger isoform 1 inhibition is mediated by reduced mitogen-activated protein kinase activation secondary to improved mitochondrial integrity and decreased generation of mitochondrial-derived reactive oxygen species. J Pharmacol Exp Ther. 2006;317:1036–43.CrossRefPubMedGoogle Scholar
- 54.Pedraz-Cuesta E, Fredsted J, Jensen HH, Bornebusch A, Nejsum LN, Kragelund BB, et al. Prolactin signaling stimulates invasion via the Na+/H+ exchanger NHE1 in T47D human breast cancer cells. Mol Endocrinol. 2015; in revisionGoogle Scholar
- 67.Grenier AL, Abu-Ihweij K, Zhang G, Ruppert SM, Boohaker R, Slepkov ER, et al. Apoptosis-induced alkalinization by the Na+/H+ exchanger isoform 1 is mediated through phosphorylation of amino acids Ser726 and Ser729. Am J Physiol Cell Physiol. 2008;295:C883–96.CrossRefPubMedPubMedCentralGoogle Scholar
Open AccessThis article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated.