Proteomic Analysis of the Functional Inward Rectifier Potassium Channel (Kir) 2.1 Reveals Several Novel Phosphorylation Sites

Abstract

Membrane proteins represent a large family of proteins that perform vital physiological roles and represent key drug targets. Despite their importance, bioanalytical methods aiming to comprehensively characterize the post-translational modification (PTM) of membrane proteins remain challenging compared to other classes of proteins in part because of their inherent low expression and hydrophobicity. The inward rectifier potassium channel (Kir) 2.1, an integral membrane protein, is critical for the maintenance of the resting membrane potential and phase-3 repolarization of the cardiac action potential in the heart. The importance of this channel to cardiac physiology is highlighted by the recognition of several sudden arrhythmic death syndromes, Andersen−Tawil and short QT syndromes, which are associated with loss or gain of function mutations in Kir2.1, often triggered by changes in the β-adrenergic tone. Therefore, understanding the PTMs of this channel (particularly β-adrenergic tone-driven phosphorylation) is important for arrhythmia prevention. Here, we developed a proteomic method, integrating both top-down (intact protein) and bottom-up (after enzymatic digestion) proteomic analyses, to characterize the PTMs of recombinant wild-type and mutant Kir2.1, successfully mapping five novel sites of phosphorylation and confirming a sixth site. Our study provides a framework for future work to assess the role of PTMs in regulating Kir2.1 functions.

Graphical Abstract

INTRODUCTION

Excitation−contraction coupling is dependent on the complex interaction of voltage and time-dependent ion channels in the cellular membrane, and perturbations of the action potential can result in arrhythmias and sudden death.^1,2 Kir2.1, the strong inward rectifier potassium channel, is the dominant molecular component for I_K1, which assists with terminal cardiac repolarization and maintains resting membrane stability.^3,4 Functional Kir2.1 channels are essential for cardiac excitability and mutations in KCNJ2, the gene encoding Kir2.1, causing genetic sudden cardiac death syndromes, and loss of Kir2 current is a major contributing factor to arrhythmogenesis in failing human hearts. Arrhythmogenesis related to Kir2.1 mutations or in acquired diseases such as heart failure can be influenced by the adrenergic tone and activation of protein kinases.^5–7 Despite the importance of understanding protein modifications and regulation, there have been no studies directed at specifically defining the influence of post-translational modifications (PTMs) on both normal functions and diseases.

Several studies have formed a basis for understanding Kir2.1’s structure, function, and some aspects of its regulation.^3–5,7–9 Kir2.1 assembles as heterotetramers to make a functional channel, and binding to phosphatidylinositol 4,5-bisphosphate is required for gating.³ Intracellular second messengers such as isoproterenol-mediated PKA activation regulate Kir2.1 by increasing the Kir2.1 outward current,⁶ yet there is only indirect evidence for a PKA phosphorylation site.¹⁰ Without direct evidence for the actual PKA activation site, unraveling complex PTM and channel regulation is confounded. This is highlighted by the recognition that some genetic sudden death syndromes related to KCNJ2 mutations are due to adrenergic tone-dependent Kir2.1 loss of function causing bidirectional and polymorphic ventricular tachycardia.⁵ Additionally, using mammalian overexpression models, inhibition of tyrosine phosphatase activity has also been shown to suppress Kir2.1 current.¹¹ Therefore, (1) phosphorylation appears to be a critical regulator of Kir2.1, (2) specific Kir2.1 phosphorylation sites are unknown, and (3) the site-specific effects on the channel function and phosphorylation state(s) are unknown.

Mass spectrometry (MS)-based proteomics has enabled the analysis of protein expression and PTMs on a large scale.^12,13 However, the study of membrane proteins, like Kir2.1, using the traditional bottom-up proteomic approach (wherein proteins are enzymatically digested into peptides before MS analysis) remains underdeveloped compared to targeted analysis of other protein classes.^14–16 The difficulty in studying membrane proteins arises because these biomolecules are generally hydrophobic (thus difficult to solubilize in aqueous solution) and expressed at low abundance in the proteome.¹⁷ Moreover, PTM information is often lost as a result of analyzing proteins after enzymatic digestion.¹⁸ Top-down proteomics (MS analysis of intact proteins) has emerged as the premier technique to comprehensively characterize PTMs and amino acid sequence variations providing important insights into protein regulation.¹⁹ However, challenges with intact protein solubility, complete fragmentation for full sequence coverage, and sensitivity have limited its global application, and further development is needed.²⁰ There are notable examples of studying membrane proteins using both the bottom-up^21–23 and top-down proteomics,^24–28 but characterizing the PTMs of specific, functional membrane proteins remains challenging. Thus, there is a great need for tools that enable the characterization of membrane proteins for a better understanding of their role and regulation in health and diseases.

Here, we report a targeted, MS-based proteomic method, integrating both top-down and bottom-up analyses, to identify the PTMs of recombinant, functional Kir2.1 (Figure 1). In this study, top-down proteomics enabled an overview of the present proteoforms^12,29 and their abundances in addition to the identification of a C-terminal phosphorylation site and secondary N-terminal sites that manifested after mutation of the C-terminal site. Next, bottom-up proteomic analysis uncovered additional C-terminal phosphorylation sites. Our work led to the confident identification of five novel phosphorylation sites highlighting the importance of developing novel methods for characterizing challenging membrane proteins like Kir2.1. Additionally, this study provides the basis for studying phosphorylation-mediated regulation of Kir2.1 across various mutations as well as other channels (and likely other membrane proteins) that may be regulated by a similar process given the structural and physicochemical similarities that they share.³⁰

Figure 1.

Workflow for characterization of recombinant Kir2.1^WT and Kir2.1^S425A using MS-based proteomics to identify PTMs and whole-cell patch-clamping to verify functions.

RESULTS

Construction of the Overexpressing Kir2.1 Cell Systems.

To effectively and reproducibly screen conditions for MS analysis and perform functional characterization in parallel, we developed a stable overexpressing MYC-tagged Kir2.1 cell system in HEK293 cells.^31,32 HEK293 cells have been previously used to study Kir2.1 functions^4,9 and contain the proper post-translational machinery; however, we acknowledge that Kir2.1’s environment could be different in cardiomyocytes. Cells were transfected with WT and S425A Kir2.1 constructs containing an MYC epitope at the N-terminus (Figure S1). They were then selected under geneticin, and resistant cells were diluted to isolate individual clones. Modifications identified and reported in this study correspond to the endogenous sequence, i.e., without the tag. Finally, the approach described here should also work for targets that might yield lower expression or lower solubility in MS-friendly detergents. Given the sensitivity of MS, our analysis was done with relatively small-scale preparations, and if needed, then these cells can be vastly scaled up in bioreactors.

Functional Assessment Using Whole-Cell Patch-Clamp Analysis.

To confirm that the function of the channel was unaltered by the MYC tag, we performed whole-cell patch-clamp analysis⁶ on cells expressing Kir2.1 with and without the tag. Average current−voltage (I−V) data was generated from pcDNA3.1-Kir2.1^WT and MYC-Kir2.1^WT. Both cell lines have similar levels of current, displaying typical N-shaped I−V relationships with maximal outward current at −50 mV. Overall, no statistical difference in I_K1 between the cells with or without the MYC tag was observed, and thus, the tag does not alter the channel function (Figure S2).

Development of Top-Down Proteomic Assay Using Recombinant, Functional Kir2.1.

Having confirmed that our cell system retained the Kir2.1 function, we next sought to develop an MS-based assay to precisely identify and monitor changes in channel PTMs. Like many membrane proteins, Kir2.1 represents an analytical challenge for proteomic analysis as it is relatively low in abundance (i.e., low copy number) and has hydrophobic properties, which can lead to poor solubility making purification challenging.^14,26 It is also relatively large (∼49.5 kDa) for top-down proteomics, which makes the precise characterization of its PTMs and their stoichiometry challenging as it can be difficult to obtain sufficient sequence coverage using online LC−MS/MS.^19,20,33 Thus, an enrichment strategy for Kir2.1 proteoforms is critical before top-down MS. Here, we utilized MYC-tag affinity purification³⁴ to enrich Kir2.1 from cell lysates generated using a nonionic surfactant, n-dodecyl β-d-maltoside (DDM). Successful enrichment by MYC-tag affinity purification was observed using Western blot analysis (Figure S3A).

Next, it was important to screen elution conditions that were compatible with downstream MS analysis as common elution conditions for gel-based techniques, e.g., Laemmli buffer, are incompatible with MS.³⁵ We tested an anionic photocleavable surfactant,²⁸ low pH,²⁴ and mixtures of organic solvents³⁶ as MS-compatible elution conditions. Based on results from SDS-PAGE analysis using Coomassie Brilliant Blue stain, we found that the low pH condition (0.2 M glycine HCl, pH 2) led to the highest recovery (Figure S3B,C). This condition yields relatively good MS results but was further improved by replacing glycine HCl with 1% formic acid + 5 mM tris(2-carboxyethyl)-phosphine hydrochloride (TCEP), which acts as an antioxidant. The formic acid elution condition minimized sample handling as it did not require offiine desalting, which is generally required when utilizing glycine HCl for elution.³⁷

Reversed-phase liquid chromatography (RPLC) was then employed to separate Kir2.1^WT from a few copurified proteins as well as the large residue DDM that was observed despite washing the affinity beads with ammonium bicarbonate (Figure S4). Using online RPLC−MS, a mass of 49,515.7 Da was observed, which is 0.1 Da (2 ppm) lower than the mass corresponding to the amino acid sequence with an acetylation (42.0 Da) modification (Figure 2A,B). Collisionally activated dissociation (CAD) enabled the localization of the acetylation to the N-terminus of the protein (Figure 2C,D). The 17 b- and 19 y-type ions were confined to the N- and C-termini, which is common for online fragmentation of large protein species.^38,39 We found that our top-down method was generally reproducible with only a small degree of prep-to-prep variability observed (Figure S5).

Figure 2.

Top-down proteomics of Kir2.1^WT using online RPLC−MS/MS. (A) Representative charge state distribution of a mass spectrum with a zoom-in image of the +56 ion. (B) Deconvoluted mass spectra demonstrating high mass accuracy. (C,D) Representative tandem MS spectra and fragmentation map generated from CAD.

Top-Down Proteomic Analysis for Kir2.1 for PTM Characterization.

Having established a robust top-down proteomic method using online RPLC−MS/MS, we next sought to map the phosphorylation sites of Kir2.1^WT, characterized by a +80 Da mass shift. The cells were incubated in a PKA stimulating “cocktail” (IBMX and forskolin) to increase the phosphorylation of the channel followed by the enrichment and MS analysis.⁵ Using top-down proteomics, we observed both mono- and bis-phosphorylated Kir2.1^WT and successfully localized a phosphorylation site to S425 (Figure 3).

Figure 3.

Identification of S425 phosphorylation using top-down MS. (A) Intact mass spectra of Kir2.1^WT with and without PKA stimulation. Mass spectra were deconvoluted using the maximum entropy algorithm.⁶⁹ The intensity of each peak was quantified using the sum peak algorithm and is represented as a percentage $\left(\frac{{\text{intensity}}_{\text{proteoform}}}{{\text{intensity}}_{\text{total}}}\times 100\right)$. (B,C) Fragmentation map and representative fragment ions demonstrating successful localization of phosphorylation to S425 using online CAD. Circles represent theoretical isotopic distribution. Phosphorylation sites are reported based on a native sequence (i.e., without the MYC tag).

Based on the identification of S425 phosphorylation, we mutated the site to A to prevent phosphorylation. MS analysis of intact Kir2.1^S425A yielded a measured molecular weight of 49,500.3 Da, which is 0.5 Da (10 ppm) higher than the theoretical value of 49,499.8 Da (with N-terminal acetylation). Interestingly, mono- and bis-phosphorylated species were still observed with PKA stimulation with little change in the overall levels of phosphorylation after mutating S425 (Figure 4). We employed both CAD and electron transfer dissociation (ETD)⁴⁰ enabling the identification of phosphorylation site(s) at S13 and/or S14 (Figure 4B,C), which were not detected in the Kir2.1^WT samples (Figure S6), as well as verifying the success of the mutation (Figure S7).

Figure 4.

Identification of S13 and S14 phosphorylation using top-down MS. (A) Top-down proteomics comparing Kir2.1^WT and Kir2.1^S425A both with PKA stimulation. Mass spectra were deconvoluted using the maximum entropy algorithm.⁶⁹ The intensity of each peak was quantified using the sum peak algorithm and is represented as a percentage $\left(\frac{{\text{intensity}}_{\text{proteoform}}}{{\text{intensity}}_{\text{total}}}\times 100\right)$. (B,C) Fragmentation map and representative fragment ions demonstrating successful localization of N-terminal phosphorylation using ETD. Circles represent theoretical isotopic distribution. The gray region in the protein sequence indicates the MYC tag.

Bottom-Up Proteomic Analysis of Kir2.1 for PTM Characterization.

To further investigate the post-translational modifications of Kir2.1^WT and Kir2.1^S425A, we performed bottom-up proteomics, which involves proteolytic digestion of the channel before RPLC−MS/MS. The channel was reduced, alkylated, and digested with trypsin in the presence of 4-hexylazophenylsuflonate (0.05%), which has previously been shown to aid in membrane protein digestion.⁴¹ RPLC−MS/MS using a timsTOF Pro (Bruker Daltonics) mass spectrometer and MaxQuant data analysis⁴² yielded 84.1 and 75.4% sequence coverage for Kir2.1^WT and Kir2.1^S425A, respectively (Figures S8 and S9 and Tables S1 and S2). Additionally, novel phosphorylation sites were identified (1% false discovery rate, MaxQuant scores of >100) at S313, Y326, and T347 (Figure S10). Bottom-up proteomic analysis of Kir2.1^S425A enabled the identification of a mono-phosphorylated S13 and S14 peptide as well as a bis-phosphorylated peptide with both sites occupied (Figure 5). Interestingly, S13 and S14 phosphorylation was also observed in the Kir2.1^WT samples using the bottom-up approach; however, these sites had the lowest relative intensity, which explains why they were not detected using the less sensitive top-down approach. Overall, these results are significant because our data demonstrate multiple independent phosphorylation sites on the Kir2.1 channel on both the N- and C-termini.

Figure 5.

Bottom-up mass spectrometry characterization of Kir2.1^S425A for localization of phosphorylation at S13, S14, and Y326. Representative MS/MS spectra annotated using Skyline software.⁶⁷ Purple, blue, and gray peaks denote b, y, and unassigned ions, respectively. Brackets indicate the identified mass shifts in Da for the residues.

Unfortunately, the small tryptic peptide 424−427, corresponding to the previously localized phosphorylation at S425, was not identified using bottom-up analysis. Therefore, we employed a middle-down approach using an Asp-N digestion strategy. Although this approach yielded significantly lower sequence coverage (16%) than with trypsin, the C-terminal phosphorylation at S425 was successfully identified providing further validation of this phosphorylation site (Figure S11).

DISCUSSION

Phosphoproteomic Analysis of Kir2.1.

To date, there are no reported directly identified phosphorylation sites for human Kir2.1 (https://www.uniprot.org/uniprot/P63252) according to the UniProt database⁴³ with the only reported amino acid modification being N-myristoyl glycine⁴⁴ and S-nitrosocysteine.⁴⁵ However, studies have indirectly probed the potential role of phosphorylation on the channel function.^46,47 In this study, we performed MS-based proteomics to characterize the PTMs of functional Kir2.1 for the first time. A robust top-down proteomic assay was developed for normally functioning Kir2.1 in a stable cell line enriched by MYC-tag affinity purification with an MS-compatible elution of 1% formic acid and purification and analysis using online RPLC−MS/MS. This approach enabled us to gain insight into the relative abundance of novel phosphorylation and localized sites at the N- and C-termini (S13, S14, and S425).

Following the successful identification of these phosphorylation sites in Kir2.1WT, we ablated the S425 to A425 to enable further investigation of the changes in PTMs and/or the channel function. Notably, in previous works, mutating S425 to N425 in rat and human Kir2.1 channels (expressed in Cos-7 cells) blocks PKA-mediated regulation of the channel function.^4,6,48 In our study, we observed that after mutating S425 to A425, the overall phosphorylation levels appear to be minimally affected by the mutation (Figure 5). There could be a difference between the S425N mutation previously studied and the S425A mutation studied here. The larger polar side chain for asparagine may have different conformational influences on the channel in tetrameric conformation from that of alanine. The smaller nonpolar side chain of alanine may permit the “unblocking” of other PKA phosphorylation sites.

Previous studies have also demonstrated that inhibiting tyrosine phosphates can alter the channel function with Y242 being reported;^11,47 however, our work represents the first successful direct identification of tyrosine phosphorylation on Kir2.1 at a new site (Y326). An important caveat, however, is that distinguishing phosphotyrosine-containing peptides from their sulfotyrosine counterparts, which constitutes a delta mass of 9.6 mDa, using mass spectrometry is challenging and requires resolution beyond what was used in this work.⁴⁹ Given previous evidence of tyrosine phosphorylation¹¹ for this channel, we believe that this modification is more likely. While only the mono- and bis- phosphorylated species were commonly observed using the top-down approach, there are likely multiple different mono- and bis-phosphorylated species due to partial site occupancy,^50–53 which can account for the additional sites of phosphorylation. Moving forward, precise characterization of Kir2.1 proteoforms will likely require further development of the intact membrane protein sequence including employing alternative MS2 techniques to achieve higher sequence coverage and the ability to robustly quantify the level of phosphorylation at each site.^{20,38,54–56} In particular, techniques such as electron-transfer/higher-energy collision dissociation and activated ion electron transfer dissociation would likely improve sequence coverage in the transmembrane domain as they generally require thermal denaturation in the gas phase to achieve fragmentation.^55,57

Overall, proteomic analysis of Kir2.1 led to the identification of five novel sites of phosphorylation (S13, S14, S313, Y326, and T347) and confirmed S425 as a site of phosphorylation. Figure 6 shows the locations of the sites directly identified along with two sites previously reported (S10 and Y242) on mouse Kir2.1 and chicken Kir2.2 structures.^58,59 Interestingly, Y326 is not solvent-accessible, suggesting either that large-scale movement in the secondary structure can occur in the intracellular domains or the intracellular part of this structure is not in its native state since it is lacking much of the N- and C-termini. Further, sequence alignment with Kir2.2 and Kir2.3 in Figure S12 shows that except for Y326, phosphorylation sites are conserved between Kir2.1, 2.2, and 2.3. This suggests that multiple monomers within the heterotetrameric channel are phosphorylated to modulate I_K1.

Figure 6.

Summary of phosphorylation identifications. (A) Tetrameric structure (amino acids 43−372) of chicken Kir2.2 showing the transmembrane domain (TMD) and N- and C-terminal intracellular domains (PDB 3JYC).⁵⁸ Phosphorylation sites were identified in blue (S313, Y326, and T347) and those previously reported but not identified (Y242 and S10) in green.^47,70,71 S10, S13, S14, and S425 are not present in the truncated structures. The missing C-terminus is shown with a dashed line. (B) Intracellular domains (N-terminal amino acids 41−63 in orange and C-terminal amino acids 190−368 in gray) of mouse Kir2.1 (PDB 1U4F)⁵⁹ in a similar orientation to (A) with only two of the four subunits shown for clarity. Missing N- and C-termini are shown by dashed lines with phosphorylation sites labeled. (C) Accessible surface area (ASA) for each residue calculated by ASAView⁷² for each structure. Sites listed are for Kir2.1 and are conserved between species except for Y326, which is a Phe in the Kir2.2 structure (see Figure S13). Distal N- and C-terminal residues are not in the structures, and so, the ASA is not available (n/a). Shaded areas are buried.

Approximately 150 disease-associated missense variants (ATS, SQTS, and CPVT) have been reported in ClinVar, a database of genetic variant interpretation.⁶⁰ Many of these variants may disrupt Kir2.1 phosphorylation either directly by disrupting kinase sequence recognition sites (e.g., Y326N would abolish tyrosine kinase activity) or indirectly through conformational changes. For example, S314F associated with ATS is adjacent to the phosphorylation site S313, which may affect PKA activity. Furthermore, the phosphorylation site S313 is near the Golgi export determinant Y315 that is important for Kir2.1 trafficking, and phosphorylation may modulate this process.⁶¹ Understanding how these variants can affect the various phosphorylation sites described above can be addressed using the proteomic techniques described herein. Finally, this proteomic approach should apply to studying PTMs of other inward rectifying potassium channels³ as well as any relatively small integral membrane protein.

CONCLUSIONS

We investigated the PTMs of functional Kir2.1 and have directly identified five novel sites of phosphorylation (S13, S14, S313, Y326, and T347) and confirmed phosphorylation at S425 using MS-based proteomics. Looking forward, we plan to investigate the structural−functional influence of phosphorylation,⁶² lipid-binding,^63,64 and/or interacting partners⁹ using MS and evaluate their effect on the channel function to gain a holistic understanding of the channel’s regulator mechanisms. Additionally, improving the sensitivity of our method could enable the analysis of Kir2.1 in disease-relevant models such as induced pluripotent stem cell-derived cardiomyocytes or in vivo. We envision that the MS-based methods that we have developed here can help unravel the functional consequence of clinically observed mutations in Kir2.1, linking PTM with altered channel functions. Finally, we believe that our method can be broadly applied to other ion channels as an effective tool for understanding channelopathies.

MATERIALS AND METHODS

Materials.

Chemicals and reagents were purchased from MilliporeSigma, Inc. (St. Louis, MO, USA) without further purification unless otherwise noted. Tris(2-carboxyethyl)-phosphine (TCEP), 2-chloroacetamide, the BCA reagent, fetal bovine serum (FBS), Dulbecco’s modified eagle medium (DMEM), phosphate-buffered saline (PBS), and penicillin/streptomycin were obtained from Thermo Fisher Scientific (Waltham, MA, USA). Water (HPLC grade), isopropanol (HPLC grade), and acetonitrile (HPLC grade) were purchased from Fisher Scientific (Fair Lawn, NJ, USA). Trypsin gold and Asp-N were purchased from Promega (Madison, WI, USA). The polymeric reverse phase (PLRP) material was purchased from Agilent (Santa Clara, CA, USA) and was packed into capillaries of PEEK tubing from VICI (Houston, TX). A C18 column (AUR2–2575C18A-CSI) was purchased from IonOpticks (Parkville, Victoria, Australia). 4-Hexylphenylazosulfonate was synthesized in-house as described previously.²⁸

Expression Constructs.

WT Kir2.1 in pcDNA3 previously reported⁵ was modified to include an MYC tag by PCR as shown in Figure S1. The S425A mutation was made using the QuikChange II Xl kit from Agilent (Santa Clara, CA) using the MYC-tagged Kir2.1 as a template. Mutagenic primers were designed with the Agilent Primer Design Program and ordered from Integrated DNA Technologies. The mutation was verified by sequencing at the University of Wisconsin Biotechnology Center.

Expression and Purification of Kir2.1.

One to three 150 mm plates of HEK cells stably expressing MYC-WT Kir2.1 or MYC-S425A Kir2.1 as described in the results were grown to ∼80% confluence (∼15−45 million cells) followed by a 2 h incubation in a PKA-stimulating cocktail of 100 µM forskolin and 10 µM 3-isobutyl-1-methylxanthine (IBMX). Cells were then washed once with Tris-buffered saline solution, pelleted at 1000g for 5 min, and lysed in 500−750 µL of DDM buffer (25 mM Tris-HCl pH 7.5, 150 mM NaCl, 1% DDM, the phosphatase inhibitors sodium fluoride (50 mM) and sodium orthovanadate (5 mM), and an EDTA-free protease inhibitor cocktail (Bimake.com)), for at least 20 min on ice with intermittent vortexing. Insoluble debris was pelleted at 16,000g for 10 min. The soluble fraction was then bound in ∼50 µL of MYC resin, washed, and pre-equilibrated in DDM buffer and rotated overnight. Bound beads were washed in 250 µL of DDM buffer for 20 min (3×) followed by a 250 µL wash with 25 mM ammonium bicarbonate to remove residual DDM buffer, before elution in 30 µL of 1% formic acid containing 5 mM TCEP for 10 min with intermittent vortexing. This was done 3× to obtain 90 µL of the total eluent for MS. The sample was stored at 4 °C for less than a day or snap-frozen and stored at −80 °C (the best results were achieved with no freeze−thaw, but the successful analysis was observed after storage for a few weeks). For immunoblotting, washed beads were eluted in ∼50 µL of DDM buffer and an equal volume of sample buffer (125 mM Tris, pH 6.8, 4% SDS, 8 M urea, 20% glycerol, 200 mM DTT, and 0.02% pyronin Y) and boiled for 5 min to elute. Samples were run on 7% SDS-PAGE and either stained with Coomassie Brilliant Blue (BioRad) or transferred to PVDF membranes and probed with 1:2000 dilution of a rabbit polyclonal Kir2.1 primary antibody (Santa Cruz Biotech) followed by a donkey anti-rabbit-HRP secondary antibody (1:10,000). Signals were detected with a ProSignal ECL detection kit (Prometheus Protein Bio Prod). The UniProt⁴³ accession ID for KCNJ2 is P63252.

Electrophysiology.

Electrophysiology experiments were performed using an Axopatch 200B amplifier and pCLAMP 10 (Molecular Devices, Sunnyvale, CA). Voltage clamp data were recorded at room temperature from transiently transfected cells expressing pcDNA3.1-Kir2.1^WT or MYC-Kir2.1^WT and stably expressing Kir2.1^WT plated on poly-l-lysine-coated coverslips using the whole-cell technique. Transiently transfected cells were used 24−48 h post-transfection. Borosilicate glass pipettes were pulled to resistances of 2−3 MΩ when filled (Model P-97, Sutter Instruments, Novato, CA). Bath solution for I_K1 measurements contained (mM) NaCl 148, KCl 5.4, CaCl₂ 1.8, MgCl₂ 1, HEPES 15, NaHPO₄ 0.4, and d-glucose 5.5, and the pH was adjusted to 7.4 (NaOH). Pipette filling solution contained (mM) K-gluconate 150, EGTA 5, MgATP 5, and HEPES 10, and the pH was adjusted to 7.2 (KOH). I_K1 was recorded from a holding potential of −50 mV with sequential 10 mV steps from −120 to +50 mV in 100 ms steps. Cells were perfused with bath solution containing 0.5 mM barium chloride, and the I−V protocol was repeated. This was then subtracted from previously recorded I−V protocols in addition to measuring I−V relationships at 500 ms of the protocol to eliminate possible contaminating currents. Capacitance measurements were taken before barium perfusion, and currents were normalized to cell capacitance. Data were analyzed using pCLAMP 10 (Axon Instruments) and Origin 8 software (OriginLab Corporation).

Top-Down Mass Spectrometry Analysis.

Samples were concentrated to 50 µL using a vacuum centrifugal concentrator. Proteins (10 µL of aliquot) were separated with a home-packed analytical column (250 × 0.250 mm, PLRP-S, 5 µm, 1000 Å) heated to 50 °C using a flow rate of 6 µL/min on a Waters nanoAcquity HPLC (M-Class).^24,65 The following conditions provided effective separation: 0−5 min, 20% B; 5−45 min, 20−60% B; 45−50 min, 95% B; 50−60 min, 95% B; 60−61 min, 20% B; 61−65 min, 20% B. The mobile phase solvent A consisted of 99.8% water and 0.2% formic acid, while the mobile phase solvent B consisted of 49.9% acetonitrile, 49.9% isopropanol, and 0.2% formic acid. The eluted proteins were directly infused into a maXis II ETD Q-TOF (Bruker Daltonics) via electrospray ionization with a capillary voltage of 4500 V and an endplate offset of 500 V. MS1 scans were collected from 200 to 3000 m/z at 1 Hz. Tandem MS (MS2) was performed via collisionally activated dissociation (CAD) using different current biases to optimize fragmentation (14, 16, and 18 eV). Precursor selection was performed in the quadrupole with a narrow window of 2 m/z. Electron transfer dissociation was also performed using a precursor ion accumulation of 1000 ms with a reagent (3,4-hexanedione) injection duration of 7 ms with a 4 m/z (isolating both the mono- and bis-phosphorylated species) as well as a wide isolation window of 20 m/z (for increased signal-to-noise of fragment ions). The MS spectra were deconvoluted by the maximum entropy algorithm with a low resolving power of 1000 (as isotopic resolution was not achieved) using Bruker DataAnalysis 4.3, and the average mass was reported. All fragment ions were processed using MASH Explorer. Deconvolution and peak picking were performed using the eTHRASH algorithm with the following settings: signal/noise, 3; delete intensity threshold, 10; min. charge, 1; max. charge, 50; num. of peaks for shoulder, 1; min. m/z, 1; max. m/z, 10,000; min. fit (%), 60. The assignment of fragment ions was accepted with a mass accuracy cutoff of ±25 ppm and verified manually. The monoisotopic mass was reported for fragment ions. Top-down experiments were performed at least three times with reproducible phosphorylation levels and sites observed across replicates.

Bottom-Up Mass Spectrometry Analysis.

The sample was diluted with 100 mM ammonium bicarbonate containing 0.05% 4-hexylphenylazosulfonate (photocleavable surfactant), and the pH was adjusted to ∼8 with ammonium hydroxide. After reduction with 5 mM TCEP and alkylation with 15 mM 2-chloroacetamide, samples were digested with trypsin (1:50 enzyme:protein) or Asp-N (1:50 enzyme:protein) overnight at 37 °C. The surfactant was degraded with 5 min of UV irradiation using a 100 W mercury lamp (Nikon housing with a Nikon HB-10101AF power supply; handle with caution), and the peptides were desalted using C18 Tips (100 µL bed) according to the manufacturer’s protocol and dried using a vacuum centrifugal concentrator. The sample was reconstituted in 30 µL of the mobile phase solvent A (99.8% water and 0.2% formic acid), and 200 ng was loaded onto an IonOpticks column (25 cm × 75 µm, C18 1.6 µm) heated to 55 °C. The separation was performed using the following gradient: 0−60 min, 2−17% B; 60−90 min, 17−25% B; 90−100 min, 37% B; 100−110 min, 37−85% B; and 110−120 min, 85% B using a flow rate of 400 nL/min and the mobile phase B consisting of 99.8% ACN and 0.2% formic acid. Eluting peptides were directly ionized via electrospray ionization (CaptiveSpray) using a capillary voltage of 1500 V, a dry gas of 3.0L/min, and a dry temperature of 180 °C. Ions were measured from 100 to 1700 m/z using a timsTOF Pro Q-TOF (Bruker Daltonics) operating in PASEF mode⁶⁶ with an ion mobility range (1/k0) of 0.60 to 1.60 V s/cm². For M2, the following parameters were used: the number of PASEF MS/MS scans (10); total cycle time (1.16 s); target intensity (20,000); the intensity threshold (2500); the charge range (0−5); active exclusion on, release after 0.4 min and reconsider if current intensity was 4× of the previous; isolation width (2 m/z for m/z< 700 and 3 m/z for m/z > 700); collisional energy (20−59 eV). Data were processed using MaxQuant V1.6.17.0 software.⁴² For searches, we used the reviewed human UniProt sequences (https://www.uniprot.org, December 4, 2020; 26,575 entries) using a 1% false discovery rate. All searches performed with carbamidomethyl (C) had a fixed modification and oxidation (M), protein N-terminal acetylation, and phosphorylation (STY) set as variable modifications. Two missed cleavages were allowed for both trypsin and Asp-N digestions. Match between runs was enabled. The data was further filtered using a cutoff MaxQuant score of 100 for phosphorylation site identifications. Otherwise, the MaxQuant parameters were not changed from their default values for the timsTOF Pro instrument. The tandem MS (MS/MS) spectra were inspected manually and annotated using Skyline 20.2.⁶⁷ Bottom-up experiments were performed with trypsin at least three times and Asp-N a single time.