
==== Front
J Biol Chem
J Biol Chem
The Journal of Biological Chemistry
0021-9258
1083-351X
American Society for Biochemistry and Molecular Biology

S0021-9258(24)02089-1
10.1016/j.jbc.2024.107588
107588
Methods and Resources
Kinase-catalyzed biotinylation for discovery and validation of substrates to multispecificity kinases NME1 and NME2
Gary Chelsea R.
Acharige Nuwan P.N.
Oyewumi Tolulope O.
Pflum Mary Kay H. pflum@wayne.edu
∗
Department of Chemistry, Wayne State University, Detroit, Michigan, USA
∗ For correspondence: Mary Kay H. Pflum pflum@wayne.edu
18 7 2024
8 2024
18 7 2024
300 8 10758818 9 2023
28 6 2024
© 2024 The Authors
2024
https://creativecommons.org/licenses/by-nc-nd/4.0/ This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/).
Protein phosphorylation by kinases regulates mammalian cell functions, such as growth, division, and signal transduction. Among human kinases, NME1 and NME2 are associated with metastatic tumor suppression but remain understudied due to the lack of tools to monitor their cellular substrates. In particular, NME1 and NME2 are multispecificity kinases phosphorylating serine, threonine, histidine, and aspartic acid residues of substrate proteins, and the heat and acid sensitivity of phosphohistidine and phosphoaspartate complicates substrate discovery and validation. To provide new substrate monitoring tools, we established the γ-phosphate–modified ATP analog, ATP-biotin, as a cosubstrate for phosphorylbiotinylation of NME1 and NME2 cellular substrates. Building upon this ATP-biotin compatibility, the Kinase-catalyzed Biotinylation with Inactivated Lysates for Discovery of Substrates method enabled validation of a known substrate and the discovery of seven NME1 and three NME2 substrates. Given the paucity of methods to study kinase substrates, ATP-biotin and the Kinase-catalyzed Biotinylation with Inactivated Lysates for Discovery of Substrates method are valuable tools to characterize the roles of NME1 and NME2 in human cell biology.

Keywords

protein kinase
histidine kinase
phosphorylation
proteomics
substrate specificity
NME
ATP-biotin
Abbreviations

1-pHis N1-phosphohistidine

ACLY ATP-citrate lyase

CDP cytidine diphosphate

DMF dimethylformamide

FSBA 5’-(4-fluorosulfonylbenzoyl)adenosine hydrochloride

Gβ1 G protein subunit beta 1

K-BILDS kinase-catalyzed biotinylation with inactivated lysates for discovery of substrates

KSR1 kinase suppressor of Ras-1

LC-MS/MS liquid chromatography-tandem mass spectrometry

NME nonmetastatic enzyme

pAsp phosphoaspartic acid

PK pyruvate kinase

TRPV5 transient receptor potential cation channel subfamily V member 5

UMPS uridine 5′-monophosphate synthase

Reviewed by members of the JBC Editorial Board. Edited by Alex Toker
==== Body
pmcFor over 50 years, phosphorylation of serine (Ser), threonine (Thr), and tyrosine (Tyr) residues on proteins (Fig. 1A) has been widely recognized as a crucial posttranslational modification (1). Phosphorylation can alter the function, stability, and localization of proteins to regulate cellular events, including signal transduction and metabolism (2). With a critical role in cell biology, improper phosphorylation can lead to disease, including diabetes and cancers (3). Characterization of protein phosphorylation has been a rigorous area of study to understand the underlying mechanisms governing normal and disease states (1, 4).Figure 1 Kinase-catalyzed phosphorylation and K-BILDS.A, the structures of phosphoserine (pSer), phosphothreonine (pThr), phosphotyrosine (pTyr), the two forms of phosphohistidine (1-pHis and 3-pHis), and phosphoaspartic acid (pAsp). B, Ser, Thr, and Tyr kinase phosphorylation with ATP or ATP-biotin. C, NME1 and NME2 phosphorylation with ATP or ATP-biotin involves initial autophosphorylation to form a 1-pHis intermediate, with the phosphoryl group subsequently transferred to a substrate on His, Ser/Thr, or Asp. D, K-BILDS workflow begins with irreversible pan-kinase inhibitor FSBA (5’-(4-fluorosulfonlybenzoyl)adenosine) treatment to inactivate endogenous kinases in cell lysates, which is followed by FSBA removal by filtration. E, Inactivated lysates are incubated with active exogenous recombinant kinase (NME1 or NME2 in this case) and ATP-biotin to give biotinylated substrates. F, one negative control includes kinase-inactivated lysates and ATP-biotin alone. As a second negative control for bead binding, kinase-inactivated lysates are incubated without ATP-biotin or kinase. G, avidin resin enriches biotinylated proteins in the kinase reaction, which are observed by SDS-PAGE and Western blot or LC-MS/MS analysis in kinases reaction samples (E), but not controls (F). green: endogenous kinase, blue: substrates, black; non-substrate protein; orange and purple: cellular protein (CP), magenta: active exogenous kinase. LC-MS/MS, liquid chromatography-tandem mass spectrometry.

Kinase enzymes catalyze protein phosphorylation by transferring the γ-phosphoryl of ATP onto substrate hydroxyl groups (Fig. 1B) (1, 2, 3, 5). Kinases are generally classified by their amino acid specificity, with Ser/Thr kinases modifying Ser and Thr residues, Tyr kinases acting on Tyr, and dual-specificity kinases acting on all three residues of substrates. Beyond these common classes of kinases, a few human kinases catalyze the phosphorylation of additional residues, including histidine (His) and aspartic acid (Asp). Specifically, nucleoside diphosphate kinase A (or nonmetastatic enzyme 1 (NME1) or NM23-H1) and nucleoside diphosphate kinase B (NDPKB or NME2 or NM23-H2) (6) autophosphorylate to generate N1-phosphohistidine (1-pHis, Fig. 1A) intermediates in their active sites (Fig. 1C, first step) for subsequent transfer of the phosphoryl (Fig. 1C, second step) to itself on Ser, to another protein substrate, or to a nucleoside diphosphate substrate through a ping-pong mechanism (6, 7, 8, 9, 10, 11). With protein substrates, autophosphorylated NME1 and NME2 can transfer the phosphoryl to Asp, His, Ser, or Thr (Fig. 1C), making them some of the few multispecificity kinases and the only characterized human His kinases (6, 9, 10). Although understudied compared to Ser, Thr, and Tyr phosphorylation, the first antibodies recognizing both 1-pHis and N3-phosphohistidine (3-pHis, Fig. 1A) isomers identified over 700 pHis-containing proteins in mammalian cells (6, 12). Phosphoproteomics also showed wide presence of phosphoaspartic acid (pAsp, Fig. 1A) (13). The abundance of pAsp and pHis suggests roles in normal cell biology and diseases. In fact, NME1 is associated with metastatic tumor suppression and was the first metastatic suppressor gene identified (6, 14). While KO mice of either NME1 or NME2 were viable (15, 16), mice with double knockout died immediately after birth, suggesting redundant roles of NME1 and NME2 in development. Human NME1 and NME2 are 88% identical, active as homo- and hetero-hexamers, and abundant in many tissues and throughout the cell (17). Despite the link to cancer and likely role of cellular phosphorylation, NME1 and NME2 remain understudied (9).

To characterize the biological roles of NME1 and NME2, their protein substrates have been investigated. NME1 substrates include ATP-citrate lyase (ACLY), kinase suppressor of Ras-1 (KSR1), and aldolase C (9, 10, 18, 19, 20). Phosphorylation of ACLY on His760 by NME1 precedes the conversion of citrate and CoA to oxaloacetate and acetyl-CoA in the Krebs Cycle, which suggests a role for NME1 in metabolism (18, 21). Signaling protein KSR1 is phosphorylated on Ser392 by NME1 to inactivate KSR1 scaffold activity and subsequently suppress MAPK signaling, ultimately inhibiting tumor cell proliferation (19, 22). Phosphorylation of aldolase C on Asp319 by NME1 has been reported, though the biological relevance has not been widely studied (10). NME2 substrates include G protein subunit beta 1 (GNB1 or Gβ1), transient receptor potential cation channel subfamily V member 5 (TRPV5), and Ca2+-activated K+ channel KCa3.1 (KCNN4) (9). Phosphorylation on His266 leads to activation of Gβ1 and cAMP synthesis, which shows a role for NME2 in cell signaling (20). Ion channels TRPV5 and KCNN4 are activated upon phosphorylation by NME2 at His711 and His358, respectively (23, 24, 25). In one of the few mechanisms of pHis action characterized to date, His358 forms an inhibitory copper metal complex in KCNN4 that inhibits K+ efflux, which is disrupted upon phosphorylation by NME2 to activate the ion channel and thereby regulate membrane potential (6, 24, 25, 26). In these examples illustrate, the roles of NME1 and NME2 in metabolism, cell signaling, and membrane potential were revealed through substrate identification (9, 18, 21). The biological associations highlighted here demonstrate the need to characterize substrates to uncover the full role of NME1 and NME2 in biological events (6).

Substrate discovery and validation of putative NME1 and NME2 substrates has been a challenge. Common methods for kinase substrate validation involve incubation of recombinant kinase, recombinant substrate protein, and ATP, with subsequent monitoring of phosphorylation by gel methods (27, 28, 29). However, prior work documented that the enzymatic activity of NME2 biochemically purified from tissue was influenced by “scaffolding proteins,” which suggested that other cellular factors are needed for robust phosphorylation (20, 30). Additionally, the phosphoramidate (P-N) bond of pHis (Fig. 1A) is acid and heat labile (6, 8, 31), while the phosphoanhydride bond of pAsp (Fig. 1A) is acid and base sensitive (32). The pH and heat sensitivity of pHis and pAsp complicates the use of traditional methods to identify all NME1 and NME2 substrates (6, 9, 33). Specifically, traditional protein immunoprecipitation typically uses heat and/or acid for elution (6, 34), and liquid chromatography-tandem mass spectrometry (LC-MS/MS) analysis commonly uses acidic solvent in the LC step, which might have compromised prior NME1 and NME2 substrate identification (35). In contrast, phosphorylation of Ser, Thr, and Tyr forms phosphoester (P-O) bonds (Fig. 1A) that are stable to acid and heat, which has allowed thorough characterization of Ser/Thr, and Tyr kinase biology (6, 31). With the need for a cellular context and the intrinsic instability of some modified substrates, the few substrates identified are insufficient to fully characterize NME1 and NME2 biology, as well as possibly reveal common primary sequence substrate motifs for bioinformatic-based substrate prediction (36). Methods to both discover and validate NME1 and NME2 substrates in a cellular context with mild conditions compatible with phosphorylation of Ser, Thr, His, and Asp are needed.

As an avenue for NME1 and NME2 substrate monitoring, we previously employed γ-phosphate–modified ATP analogs as cosubstrates of Ser/Thr and Tyr kinases to label and study phosphoproteins (37, 38, 39, 40). ATP-biotin (Fig. 1B) was particularly useful by transferring biotin to a protein substrate, which allowed subsequent phosphoprotein isolation and/or imaging using avidin-based commercial reagents (41). As an application of ATP-biotin, we established the kinase-catalyzed biotinylation with inactivated lysates for discovery of substrates (K-BILDS) method to identify the substrates of a target kinase. The K-BILDS method involves inactivation of endogenous kinases in lysates using the covalent pan kinase inhibitor 5’-(4-fluorosulfonylbenzoyl)adenosine hydrochloride (FSBA), which reacts with an active site catalytic lysine in the kinase (Fig. 1D) (42, 43). After removal of excess FSBA, an exogenous active kinase and ATP-biotin are added to the inactivated lysates to biotinylate cellular substrates of the active kinase (Fig. 1E, top). A control reaction without kinase is also prepared to eliminate nonspecifically or endogenously biotinylated proteins (Fig. 1F, bottom). Enrichment of biotinylated substrates with avidin resin is followed by the identification of substrate proteins either by gel or LC-MS/MS analysis (Fig. 1G). Initial studies with K-BILDS focused on substrates of the well-studied Ser/Thr kinases, cAMP-dependent PKA, and 56 out of roughly 250 known substrates, as well as many novel substrates, were identified (42). Related to NME1 and NME2 substrates, K-BILDS offers the advantages of avoiding extreme pH conditions and heat prior to substrate capture, as well as bypassing the need for direct pHis or pAsp identification, which can overcome limitations of prior methods (42).

To develop new methods for studying NME1 and NME2 substrates, ATP-biotin and the K-BILDS method were explored here. Initial studies established the compatibility of ATP-biotin as a cosubstrate for NME1 and NME2. Next, K-BILDS with gel analysis validated a previously characterized NME2 substrate, KCNN4 (24). K-BILDS was then coupled with LC/MS-MS analysis to discover seven candidate NME1 and three candidate NME2 substrates. Finally, confirmation using K-BILDS analysis documented that PSMA4 and Rho-GDI are NME1 substrates, while uridine 5′-monophosphate synthase (UMPS) is an NME2 substrate. In total, kinase-catalyzed biotinylation and the K-BILDS method represent powerful validation and discovery tools to study NME1 and NME2 substrates and functions in human cell biology.

Results

ATP-biotin is a cosubstrate of NME1 and NME2

Given the different mechanisms of Ser/Thr and Tyr kinases compared to NME1 and NME2 (Fig. 1, B and C), the compatibility of γ-phosphate–modified ATP analogs was first determined by incubating recombinant NME1 and NME2 with ATP-biotin, followed by separation by SDS-PAGE and biotin visualization. Biotinylation of both NME1 (Fig. 2A, lane 2) and NME2 (Fig. 2A, lane 6) was observed in the presence of ATP-biotin, but not in the absence of ATP-biotin (Fig. 2A, lanes 1 and 5), which is consistent with autophosphobiotinylation (Fig. 1C). As a control, biotinylation was reduced when excess ATP was preincubated with NME1 and NME2 prior to introduction of ATP-biotin (Fig. 2A, lanes 3 and 7), which demonstrated that ATP-biotin accesses the ATP binding active site of the kinases. As a second control, biotinylation was also reduced when NME1 and NME2 were denatured with SDS prior to ATP-biotin addition (Fig. 2A, lanes 4 and 8), which showed kinase activity dependence. As another control, heating the reaction product, which is known to degrade pHis, reduced biotinylation (Fig. S2, lane 5), which suggested phosphorylbiotinylation of His. Also consistent with the generation of biotinylated 1-pHis, biotinylated NME1 was recognized by the 1-pHis antibody (Fig. S2, lane 2). Three independent trials were quantified, and NME1 and NME2 were only biotinylated reproducibly in the presence of ATP-biotin (Fig. 2B, lanes 2 and 6). These results documented that ATP-biotin is compatible with NME1 and NME2 autophosphorylation.Figure 2 NME1 and NME2 biotinylation with ATP-biotin.A, recombinant NME1 (lanes 1–4) or NME2 (lanes 5–6) were incubated with ATP (2 mM, lanes 1 and 5) or ATP-biotin (2 mM, lanes 2–4 and 6–8) for 2 h at 31 °C before SDS-PAGE separation (without heat denaturation). Controls included pre-incubation of NME1 and NME2 with excess ATP (10 mM, lanes 3 and 7) prior to the addition of ATP-biotin and the presence of SDS denaturant (2%, lanes 4 and 8) during the reaction to assure kinase activity dependence. Biotinylated proteins were visualized with streptavidin-Cy5 (SA-Cy5), and total proteins were observed using SYPRO Ruby stain. Molecular weight markers (kDa) are indicated on the left side of each gel. Three independent trials and full gel images are shown in Fig. S1. B, band intensities from the SA-Cy5 gel images from part A (Fig. S1A) were quantified from three independent trials to generate the mean and SD, with statistical power determined using a Student t test (Fig. S1B). The sample numbers in the histogram correspond to the gel lanes in part A. ∗p ≤ 0.05, ∗∗p ≤ 0.01, ns = not significant.

The efficiency of ATP-biotin as a cosubstrate of NME1 and NME2 was further assessed using an NADH-linked enzyme-coupled kinetics assay (7, 39, 41, 44). As previously described, cytidine diphosphate (CDP) was used as the phosphate acceptor, which takes advantage of the NDPK activity of NME1 and NME2 (7). CDP and various concentrations of ATP or ATP-biotin were incubated with recombinant NME1 or NME2, and initial rate data were obtained by monitoring the loss of NADH absorbance over time. After fitting the data to the Michaelis–Menten equation, the Michaelis–Menten constant (KMapp, Table 1) was elevated 7.1-fold (NME1) or 11-fold (NME2) with ATP-biotin compared to ATP, whereas the catalytic rate (kcatapp, Table 1) was 2.0-fold (NME1) or 1.7-fold (NME2) higher for ATP than ATP-biotin. Taken together, the efficiency of the reaction (kcatapp/KMapp) decreased by 14-fold (NME1) or 18-fold (NME2) with ATP-biotin in comparison to ATP (Table 1). While the reaction efficiency was reduced, ATP-biotin was compatible not only with NME1 and NME2 autophosphorylation but also subsequent phosphorylation of the substrate.Table 1 Kinetic constants of NME1 and NME2 with ATP and ATP-biotina

Kinase	KMapp (μM)b	kcatapp (s−1)b	kcatapp/KMapp (mM s)−1	
ATP	ATP-biotin	Ratioc	ATP	ATP-biotin	Ratiod	ATP	ATP-biotin	Ratioe	
NME1	9.8 ± 2.4	70 ± 18	7.1	7.3 ± 0.7	3.7 ± 0.3	2.0	740	53	14	
NME2	11 ± 3	120 ± 40	11	9.3 ± 0.7	5.5 ± 0.5	1.7	850	46	18	
a Curve fits and rate plots are shown in Fig. S3, and the mean and SD are shown.

b Due to unnatural and limiting substrate, apparent Michaelis–Menten constants (KMapp) and apparent turnover (kcatapp) were calculated.

c KMapp values for ATP-biotin were divided by ATP to determine the ratio.

d kcatapp values for ATP were divided by ATP-biotin to determine the ratio.

e kcatapp/KMapp values for ATP were divided by ATP-biotin to determine the ratio.

To gain better insight into the influence of the modified γ-phosphoryl of ATP-biotin on kinase activity, docking was performed. ATP and ATP-biotin were docked in the crystal structures of NME1 (pdb 3L7U) (45) and NME2 (pdb 1NUE) (46). While NME1 and NME2 assemble as hexamers with dihedral (D3) symmetry and an active site present in each of the six subunits (36), only a single active site was used for docking. Among the 100 docking poses generated, the poses selected were the highest energy showing ATP or ATP-biotin bound similarly to ADP cocrystallized with NME1 (pdb 2HVD, Fig. S4, A and B) (47) or GDP cocrystallized with NME2 (pdb 1NUE, Fig. S4, C and D) (46). As expected based on prior work with Ser/Thr/Tyr kinases (48), the PEG linker in ATP-biotin positions the bulky biotin tag outside of the active site to avoid interference with substrate binding with both NME1 (Fig. 3C) and NME2 (Fig. S4G). To assess interaction, distances were measured between the γ-phosphorus of ATP or ATP-biotin and nitrogen (N1) of the active site histidine of NME1 and NME2. Whereas ATP displayed distances of 3.6 Å (NME1, Figs. 3B) and 5.5 Å (NME2, Fig. S4F), ATP-biotin bound with a longer distance of 7.7 Å (NME1, Fig. 3C) or 5.6 Å (NME2, Fig. S4G). The altered binding of the modified γ-phosphoryl of ATP-biotin within the kinase active sites is consistent with the reduced kinetic efficiency (Table 1).Figure 3 Docking of NME1 with ATP and ATP-biotin. The crystal structure of NME1 (A, pdb 3L7U) was docked with ATP and ATP-biotin using Autodock Vina 1.2.0 (90, 91). The distance between the γ-phosphate of ATP or ATP-biotin and N1 of the catalytic histidine (H118) in NME1 was measured as 3.6 Å (B) and 7.7 Å (C), respectively, using PyMOL. H118 (C-yellow, N-blue, H-gray) and ATP-biotin (C- pink, H-gray, O-red, N-blue, P-orange, S-yellow) are color coded and labeled. Docking data with NME2 are shown in Fig. S4, E–G.

KCNN4 peptide substrates were insufficient for NME2 phosphorylation

The next step was to validate the compatibility of ATP-biotin and NME with a protein substrate, such as the well-characterized NME2 substrate KCNN4. Phosphorylation of KCNN4 by NME2 on His358 is needed for activation of the ion channel by relieving copper inhibition (24, 25). Similar to the NME2 biotinylation study, NME2 was incubated with ATP-biotin and a KCNN4 peptide containing Phe301-Lys360, which includes the His358 phosphorylation site at the C terminus. Biotinylation of KCNN4(301–360) by NME2 was absent (Fig. 4A, lane 3), although NME2 biotinylation via autophosphorylation was observed (Fig. 4A, lanes 3 and 4). As a positive control, ATP was incubated with NME2 and KCNN4(301–360), and phosphorylation was assessed using two independent methods previously used to visualize pHis: Phos-tag phosphoprotein separation gel and ProQ Diamond phosphoprotein stain (33, 49, 50, 51). Similar to the biotinylation studies, phosphorylation of KCNN4(301–360) by NME2 was absent in the presence of ATP (Fig. 4B, lane 3 and Fig. S6B, lane 2), although NME2 phosphorylation was observed (Fig. 4B, lanes 1 and 3 and Fig. S6B, lanes 2 and 3). A potential explanation for the absence of KCNN4(301–360) phosphorylation was that the pHis358 site was only two amino acids away from the C terminus of this KCNN4(301–360) fragment, which was insufficient for robust phosphorylation (25). To increase the number of amino acids surrounding the His358 phosphorylation site and possibly encourage phosphorylation by NME2, the larger KCNN4(328–427) fragment was used in kinase reactions. Monitoring phosphorylation using pHis antibodies again failed to observe pHis-modified KCNN4(328–427) (Fig. S7), despite robust pHis modification with NME2 (Fig. S7).Figure 4 Attempted phosphorylation of KCNN4 fragments.A and B, dialyzed KCNN4 (301–360) was incubated for 2 h at 31 °C with or without NME2 in the presence of ATP-biotin or ATP and separated by SDS-PAGE (without heat denaturation). Proteins in the gel were visualized using SYPRO Ruby total protein stain or transferred to a membrane and visualized using streptavidin-Cy5 (SA-Cy5) for biotinylation or Phos-tag SDS-PAGE with Coomassie for phosphorylation. Molecular weight markers (kDa) are indicated on the left side of each gel. Full gel images and independent trials are shown in Figs. S5 and S6. C, KCNN4 peptide (3 mM) was incubated with NME2 (2 μg) and ATP (2 mM) for 2 h at 31 °C, followed by analysis by HPLC with detection at 214 nm using a gradient beginning with 95% buffer A (0.1% TFA in water) and 5% buffer B (0.1% TFA in acetonitrile), moving to 80% A and 20% B over 5 min, followed by 80% to 60% A over 10 min and then 60% to 5% A for the next 5 min. Control samples for HPLC are shown in Fig. S11.

As an alternative to fragments of KNCC4, synthetic peptides derived from KCNN4 were also tested. A peptide containing His358 and five surrounding amino acids from KCNN4 (N-QVRLKHRKLRE-C, Fig. S8A) was generated, along with a negative control alanine mutant peptide (N-QVRLKARKLRE-C, Fig. S8B). ATP was incubated with NME2 and the peptides, and phosphorylation was monitored by HPLC and kinetics analysis. With HPLC, no new phosphopeptide peaks were observed (Fig. 4C). One possibility is the cleavage of pHis due to the slightly acidic solvent system in the HPLC. To avoid acid and heat conditions, kinetics assays were used similarly to the prior studies with CDP as the substrate (7, 52). When the kinetics analysis was performed with the KCNN4 peptide substrate in the place of CDP in the enzyme-coupled assay, no initial rate was observed (Fig. S3C). Previous work showed that the C-terminus of KCNN4, particularly Arg355-Met368, was crucial for ATP-dependent activity (53). Our data suggests that peptides were insufficient for phosphorylation by NME2; a method using full-length KCNN4 is likely needed for efficient phosphorylation.

NME substrate validation using K-BILDS

To overcome challenges using peptide substrates to monitor NME1- and NME2-mediated phosphorylation, we sought a method that uses full-length substrate proteins. Additionally, NME2 phosphorylation of Gβ1 has only been successfully observed in the presence of an unknown scaffold protein (30, 54, 55), making lysates necessary for some substrates. Finally, a method that avoids heat and extreme pH conditions prior to substrate capture was needed. The previously developed K-BILDS method (Fig. 1, D–F) is ideal to monitor NME phosphorylation by involving full length proteins in a lysate and avoiding heat and extreme pH conditions prior to substrate capture. K-BILDS also has the advantage of distinguishing NME1 and NME2 substrates.

K-BILDS was performed to monitor KCNN4 phosphorylation by NME2, but not NME1, as documented in prior work (56). FSBA-inactivated HeLa lysates were incubated without or with ATP-biotin, along with NME1 or NME2, to biotinylate substrate proteins. Biotinylated proteins captured by avidin enrichment were separated using SDS-PAGE and analyzed by Western blotting for KCNN4. KCNN4 was biotinylated and enriched in the presence of NME2 (Fig. 5A, lane 4), but not NME1 (Fig. 5A, lane 3), which is consistent with prior data that showed KCNN4 is only activated by NME2 (56). Three independent trials were quantified, and KCNN4 was biotinylated and enriched reproducibly with NME2 (Fig. 5B). K-BILDS represents a helpful assay option for NME substrate validation, with the advantage of discriminating NME1 from NME2 substrates.Figure 5 K-BILDS substrate analysis of KCNN4 with NME1 and NME2.A, FSBA-inactivated HeLa lysates were incubated with or without ATP-biotin in the presence or absence of NME1 or NME2. Following the kinase reaction, biotinylated proteins were enriched using NeutrAvidin beads, separated by SDS-PAGE, and western blotted with a KCNN4 antibody. KCNN4 levels were measured before (input as a load control) and after (eluate) enrichment. Molecular weight markers (kDa) are indicated on the left side of each gel. Full gel images and independent trials are shown in Fig. S12A. B, three trials were quantified to generate the mean and SD shown, with statistical power determined using a Student t test (Fig. S12B). The sample numbers in the histogram correspond to the gel lanes. ∗p ≤ 0.05, ns = not significant.

NME substrate discovery using K-BILDS

In addition to NME1 and NME2 substrate validation, K-BILDS also has application to substrate discovery by combining with LC-MS/MS analysis, similar to previous work (42). A strength of K-BILDS as a discovery tool is the ability to separately identify NME1 and NME2 substrates. To identify substrates using K-BILDS, FSBA-inactivated HeLa lysates were incubated with ATP-biotin in the absence or presence of NME1 or NME2. HeLa lysates were chosen due to prior studies of NME1 and NME2 phosphorylation (6, 12, 57). Biotinylated proteins in each sample were enriched using streptavidin beads. To control for nonspecific bead binding, untreated FSBA-inactivated lysates were also enriched with avidin beads. Following desalting by SDS-PAGE and in-gel trypsin digestion, LC-MS/MS and label-free quantitation (58, 59, 60) were performed to identify 1887 total proteins in two independent trials (Tables S1 and S2). To eliminate proteins nonspecifically bound to avidin beads, proteins with higher intensities in samples with untreated FSBA-inactivated lysates than samples treated with ATP-biotin and NME1 or NME2 were removed. Potential substrates were distinguished by dividing protein intensities in the samples containing NME1 or NME2 by the intensities of that same protein in samples without NME1 and NME2. Seven candidate NME1 and three candidate NME2 substrates were identified using enrichment values of at least 1.5-fold or 2.0-fold in samples containing NME1 or NME2, respectively, compared to samples without kinase in two independent trials (Tables S3 and S4). The 1.5-fold or 2.0-fold fold enrichment values were chosen based on the enrichment of NME1 and NME2, which are built-in positive substrate controls due to autophosphorylation, as well as previously identified pHis-containing proteins, including regulator of nonsense transcripts 1 (UPF1) in the NME1 list and peroxiredoxin-6 (PRDX6) in the NME2 list (Table 2) (6). To assess if K-BILDS could have identified substrates containing an unstable modification, the hit were compared to known pHis-containing proteins from previous studies (6). Six of the twelve K-BILDS hits (50%) contain pHis according to prior work (Table 1), which is remarkable given the low overlap among reports of pHis-containing proteins (6, 9, 12, 51).Table 2 Candidate NME1 and NME2 substrates from K-BILDS

Protein	Gene name	Contain pHisa	Enrichment Valueb	
Trial 1	Trial 2	
NME1 Hits					
 Erythrocyte band 7 integral membrane protein	STOM	No	∞	∞	
 Proteasome subunit alpha type-4	PSMA4	Yes	∞	∞	
 Profilin-1	PFN1	No	∞	∞	
 Lanosterol synthase	LSS	No	∞	∞	
 Rho GDP-dissociation inhibitor 1	ARHGDIA	No	∞	∞	
 DnaJ homolog subfamily B member 1	DNAJB1	Yes	∞	∞	
 Nucleoside diphosphate kinase A	NME1	Yes	6.4	2.6	
 Regulator of nonsense transcripts 1	UPF1	Yes	1.5	2.7	
NME2 Hits					
 Uridine 5′-monophosphate synthase	UMPS	No	∞	∞	
 Protein-glutamine gamma-glutamyltransferase 2	TGM2	No	∞	3.4	
 Peroxiredoxin-6	PRDX6	Yes	∞	2.4	
 Nucleoside diphosphate kinase B	NME2	Yes	9.8	3.5	
a Identified as a pHis-containing protein in a prior study (6).

b Enrichment values were calculated by dividing protein intensities in the samples containing NME1 or NME2 by the intensities of that same protein in samples without NME1 and NME2. Proteins that were infinitely enriched (no presence in the control sample without NME1 or NME2) are denoted as “∞“. Full data are shown in Tables S3 and S4.

The biological functions of the 10 candidate protein hits were investigated to possibly reveal the roles of NME1 and NME2 in cellular events. Hits from the NME1 study were associated with diverse functions, including cell cycle growth and apoptosis (PSMA4) (61), cellular migration and proliferation (ARHGDIA or Rho-GDI GTPase) (62), cytoskeleton dynamics (PFN1) (63), proteostasis (DNAJB1 chaperone) (64), transcription (UPF1 helicase) (65), cholesterol domains (STOM) (66), and biosynthesis (LSS) (67). Among the NME2 hit list, proteins associated with biosynthesis and cell signaling were identified, including nucleic acid biosynthesis (UMPS) (68) and protein signaling (TGM2 and PRDX6) (69, 70). The candidate substrates participate in a variety of biological pathways, which suggest unanticipated functions of NME1 and NME2 in human biology.

Abundance analysis of the hits was performed to determine if protein quantity in the cell affected the K-BILDS enrichment. Given that the phosphoryl biotin tag is not sensitive to removal by phosphatases (71), substrates should be identified regardless of abundance. The abundance distribution of all proteins in HeLa cells spans from 0.01 to 10,000 ppm, and PKA substrates from the prior K-BILDS study were identified across that range (0.7–9000 ppm) (42). HeLa cell data for the 10 K-BILDS hits from the Pax-db database (http://pax-db.org) (72, 73) indicated abundances from 40 to 1229 ppm (Table S5). The abundance range of the candidate NME substrates was more narrow than with PKA, which might be due to the instability of pHis and pAsp, as well as the low activity of NME1 and NME2 compared to PKA (6, 9, 32, 74). Given the challenges associated with NME substrate discovery, K-BILDS represents a powerful method to identify multiple candidate substrates.

Candidate substrate confirmation using K-BILDS

Several candidate substrates were chosen for confirmation experiments based on their biological function. For NME1, the protein degradation protein proteosome subunit alpha type-4 (PSMA4) and the GDP-GTP exchange factor Rho GDP-dissociation inhibitor 1 (Rho-GDI, ARHGDIA) were selected. For NME2, the pyrimidine synthase uridine-5'-monophophate synthase (UMPS) was chosen. To confirm these three candidate substrates, K-BILDS with gel analysis was conducted in the same manner used to validate KCNN4 as an NME2 substrate (Fig. 5). Following biotinylation using K-BILDS with NME1 or NME2 and avidin enrichment, SDS-PAGE and Western blot analysis was performed to monitor enrichment via biotinylation of each candidate substrate. Consistent with the proteomics study, the candidate NME1 substrates PSMA4 and Rho-GDI were enriched more effectively by K-BILDS in the presence of NME1 (Fig. 6, A and C, lane 3) than NME2 (Fig. 6, A and C, lane 4). In contrast, the candidate NME2 substrate UMPS was enriched more effectively in the presence of NME2 (Fig. 6E, lane 4) than NME1 (Fig. 6E, lane 3). Quantification of at least three independent trials showed that NME1 candidate substrate PSMA4 had the highest level of enrichment with NME1 (Fig. 6B), whereas NME2 candidate substrate UMPS was enriched only in the presence of NME2 (Fig. 6F). Quantification of Rho-GDI after enrichment showed higher levels with both NME1 and NME2 (Fig. 6D, lanes 3 and 4), although only samples containing NME1 were statistically significant (Fig. 6D, lane 2 compared to lane 3). Therefore, K-BILDS confirmed that PSMA4 and Rho-GDI are putative NME1 substrates, whereas UMPS is a putative NME2 substrate.Figure 6 K-BILDS analysis of candidate NME1 and NME2 substrates. FSBA-inactivated HeLa lysates were incubated without (lane 1) or with ATP-biotin (lanes 2–4) in the absence (lanes 1–2) or presence of NME1 (lane 3) or NME2 (lane 4). Following the kinase reaction, biotinylated proteins were enriched using NeutrAvidin beads, separated by SDS-PAGE, and western blotted with PSMA4 (A), Rho-GDI (C), and UMPS (E) antibodies. PSMA4, Rho-GDI, and UMPS levels were measured before (input as a load control) and after (eluate) enrichment. Molecular weight markers (kDa) are indicated on the left side of each gel. At least three independent trials were quantified, with the mean and SD shown and statistical power determined using a Student t test for PSMA4 (B), Rho-GDI (D), and UMPS (F). The sample numbers in the histogram correspond to the gel lanes. Full gel images, repetitive trials, and quantified data are shown in Figs. S13–S15. ∗p ≤ 0.05, ns = not significant.

Discussion

Characterizing the roles played by NME1 and NME2 in human biology remains a challenge due to the instability of two of its possible products, pHis and pAsp, which complicates the use of common methods such as mass spectrometry and traditional immunoprecipitation (6, 9, 32, 75, 76). Additionally, given the difficulties associated with the phosphorylation of protein fragments in vitro by NME1 and NME2 (Fig. 4) (54), methods are needed that rely on full-length protein substrates. While previous work has produced lists of phosphorylated proteins, including pHis and pAsp sites (6, 12, 77, 78), the kinase linked to each phosphoprotein substrate is lacking. New methods are needed that promote the unbiased discovery and validation of all NME1 and NME2 substrates.

The goal of this study was to explore ATP-biotin and K-BILDS as tools to study NME1 and NME2 substrates. ATP-biotin was compatible as a cosubstrate with NME1 and NME2 (Fig. 2), with kinetics studies showing efficient phosphorylation with ATP-biotin (Table 1). While the studies herein focused on NME1 and NME2, prior work documented the compatibility of γ-phosphate–modified ATP analogs with bacterial His kinases, which autophosphorylate to generate pHis intermediates, similar to NME1 and NME2 (75, 79). The kinetics of an alkyne-tagged ATP analog with the bacterial His kinase HK853 from Thermotoga maritima revealed a KMapp of 6.9 ± 3.3 μM, kcatapp of 0.15 ± 0.03 s−1, and kcatapp/KMapp of 22 mM−1s−1 (75). Compared to the kinetics of human NME1 and NME2 with ATP-biotin (Table 1), the human kinases display a roughly 2-fold higher kcatapp/KMapp than the bacterial His kinases. We note, however, that conditions varied in the two kinetics experiments, with different enzyme concentrations (25 nM with NME1 and NME2 compared to 1–3 μM with bacterial kinases) and assay platforms (enzyme-coupled assay compared to gel-based analysis). Bacterial and human His kinases are also quite different structurally, with bacterial His kinases existing as dimers that each contain a catalytic domain along with a dimerization and histidine phosphorylation domain (80), whereas human His kinases are hexamers with a dihedral (D3) symmetry and an active site present in each of the six subunits (81). Despite the differences, NME1 and NME2 were similarly compatible with γ-phosphate–modified ATP analogs compared to HK853, which shows the general utility of ATP analogs as tools to probe kinases forming pHis intermediates for phosphotransfer.

An unexpected observation in this work was that peptides of KCNN4 (301–360, 328–427, or 353–363) failed to show phosphorylation by NME2 with ATP or ATP-biotin (Fig. 4), despite success in the past with Ser/Thr and Tyr kinase peptide substrate phosphorylation (41). KCNN4(301–360) also contained multiple Ser, Thr, His, and Asp residues, including a hexa-His tag at the N terminus, which suggests selective recognition of substrates by NME2. Prior work documented that a large portion of the C terminus of KCNN4 (Arg355-Ala413) was required for ATP-dependent channel activation, with Arg355-Met368 especially critical (53). Subsequent studies demonstrated that Arg355-Met368 was sufficient for NME2-dependent KCNN4 channel activation (24). In another example, NME1 required a large portion of the N-terminus of immunoprecipitated KSR1 (Met1-Ala539) substrate for efficient Ser392 phosphorylation (19). The crystal structure of KCNN4 in complex with calmodulin showed that His358 is positioned within a long α-helical structure (Fig. S16) (82) that might be needed for efficient phosphorylation by NME2. Taken together, the present and prior work document that efficient NME1 and NME2 phosphorylation likely requires large protein surfaces and perhaps distal binding domains. Beyond the full-length sequence, factors present in lysates were needed for phosphorylation of Gβ1 by NME2 (30, 54, 55), suggesting that recombinant proteins are insufficient for some substrates. The combined data suggest that future substrate studies with NME1 and NME2 would benefit from the use of full-length proteins in cellular mixtures.

Given the likely importance of full-length proteins in a cellular context for NME1 and NME2 activity, K-BILDS was explored as a tool to study NME1 and NME2 substrates. In fact, several key features make K-BILDS a potentially ideal NME1 and NME2 substrate monitoring tool. First, lysates are used as the source of substrates, rather than recombinant proteins or peptides (42). Incorporating cell lysates in K-BILDS ensures that full-length proteins and any necessary associated proteins are available for robust phosphorylation (54). Second, full-length biotinylated proteins were enriched with avidin resin prior to the use of acid or heat, which potentially preserves the labile pHis and pAsp modifications during the critical capture step (42). Finally, both gel and LC-MS/MS analyses after K-BILDS monitor the full-length protein hit, not the modification directly. By monitoring enriched full-length proteins, the acid or heat needed for SDS-PAGE, trypsin digestion, and LC-MS/MS that will degrade the labile pHis and pAsp modifications are inconsequential. In total, K-BILDS offers multiple features that have the potential to overcome the intrinsic challenges of studying all NME1 and NME2 substrates.

With multiple advantages, K-BILDS successfully validated KCNN4 as an NME2, but not an NME1, substrate (Fig. 5), consistent with prior work (56). In addition to KCNN4, PSMA4 and Rho-GDI were confirmed as NME1, but not NME2, substrates (Fig. 6, B and D), and UMPS was confirmed as an NME2, but not NME1, substrate (Fig. 6F). As these confirmation studies show, a key strategic advantage of K-BILDS is the ability to distinguish substrates of NME1 and NME2 in a cellular context. In fact, a paucity of methods is available to validate substrates of NME1 and NME2, most involving radiolabeling. For example, active and inactive mutants of NME1 or NME2 immunoprecipitated from lysates were incubated with recombinant KCNN4 and [γ-32P]GTP to monitor radiolabeled protein products (24). In a similar approach, immunoprecipitated KSR was incubated with 32P-autophosphorylated NME1 or inactive mutant to observe KSR radiolabeling (19). Complementing radiolabeling, functional assays monitored changes in KCNN4 channel activity in the presence of recombinant NME1 and NME2, which documented the NME2 specificity of activation (56). In a similar study, channel activity of TRPV5 increased following cotransfection of NME2 with TRPV5 (23). As an alternative validation method, we used commercially available pSer (Sigma, SAB5200086) and pHis (Sigma, ABS1670) antibodies to monitor NME1- and NME2-mediated phosphorylation of immunoprecipitated endogenous PSMA4 and UMPS without success (data not shown), which we speculate was due to the small quantities of immunoprecipitated proteins that limited detection, as well as heating to elute the immunoprecipitated proteins that could have degraded pHis. While radiolabeling and functional assays are effective options, K-BILDS provides a useful complementary method for studying candidate NME1 and NME2 substrates.

Another valuable application of K-BILDS was in substrate discovery. Using HeLa cell lysates as the source of substrates, K-BILDS identified seven candidate NME1 and three candidate NME2 substrates (Table 2), with three validated in follow-up K-BILDS experiments (Fig. 6). Prior work used radiolabeled [γ-32P]ATP to identify ACLY and aldolase as an NME1 substrate by initially forming the pHis-containing NME1 and then incubating with lysates to identify radiolabeled proteins (10, 18). More recently, the KALIP method identified NME1 and NME2 substrates from a mammalian cell–derived peptide library by monitoring directly the pHis group using LC-MS/MS (35). With a combination of advantages, including use of cell lysates, early enrichment, and analysis of full-length unlabeled proteins, K-BILDS represents a complementary tool to study NME1 and NME2 substrates.

The 10 candidate substrates identified by K-BILDS (Table 2) are associated with diverse functions not previously linked to NME1 and NME2, including cell growth and maintenance, actin-binding, and protein chaperoning (6, 61, 62, 63, 64, 65). Among the three hits that were confirmed using K-BILDS, PSMA4 was the only one previously shown to contain pHis (6). In terms of biological function, PSMA4 is an α subunit of the 20S proteasome and participates in protein degradation, which suggests a novel role of NME1 in protein homeostasis (6, 83). Rho-GDIs regulate Rho GTPases to play a role in a variety of cell processes that include migration and proliferation (62). Mechanistically, phosphorylation of Rho-GDI disrupts the interaction with Rho GTPase to promote subsequently activation by RhoGEF (62), which suggests that NME1-mediated phosphorylation of Rho-GDI might play a role in the regulation of Rho GTPase activity. Given that Rho-GDI is linked to human cancer progression by regulation Rho GTPases, NME1 might play a role in cancer. Finally, UMPS and NME2 work together in nucleic acid biosynthesis to maintain UTP levels in the cell (68, 84). Specifically, UMPS is a bifunctional enzyme that initially converts orotate to orotidine monophosphate, which is followed by the synthesis of UMP (85). After generation of UDP from UMP by cytidine monophosphate kinase, the nucleoside-diphosphate kinase activity of NME2 biosynthesizes UTP from UDP. Based on our data, NME2 might regulate UTP biosynthesis in the cells through phosphorylation-dependent regulation of UMPS, which influences UMP levels, as well as direct generation of UTP. Future studies are needed to further characterize the candidate NME1 and NME2 substrate hits. For example, mutagenesis can be used to determine the phosphorylated amino acid on the substrates (86). Subsequently, the known cellular function of the substrates can be monitored after overexpression of the mutant, as well as NME1 or NME2 knockdown, which can tie NME-mediated phosphorylation to novel molecular mechanisms cell biology (24).

A possible limitation of K-BILDS is a bias towards the identification of relatively high abundance substrates (Table S5). In fact, abundance might explain why some known NME1 and NME2 substrates were not observed. For example, the known substrates KSR1, TRPV5, and KCNN4 are present in HeLa cells but have lower abundances than the NME candidate substrates found here, according to the Pax-db database (72, 73). In fact, KSR1, TRPV5, and KCNN4 were not observed in the LC-MS/MS raw data (Tables S1 and S2). Future work would likely benefit from the use of higher lysate or kinase amounts for LC-MS/MS, which could overcome limited protein quantities. Additionally, given that many of the known substrates are membrane bound, optimizing the detergent amounts in the cell lysis buffers could increase membrane protein concentrations for subsequent enrichment (87, 88). Finally, given that NME1 and NME2 modify multiple amino acids, substrate identification (with K-BILDS or KALIP) (35) suffers from the lability of pHis and pAsp in the mass spectrometer. A current need in the field is the development of methods to not only discover NME1 and NME2 substrates but also identify the residue being modified.

Following kinase-catalyzed biotinylation and kinetics studies with ATP-biotin, K-BILDS was established here for NME1 and NME2 substrate validation and discovery. K-BILDS successfully confirmed both known and new substrates and was able to distinguish NME1 and NME2 substrates. In addition, K-BILDS could discover new candidate full-length protein substrates from a lysate mixture. In total, K-BILDS and ATP-biotin represent enabling tools to characterize kinase-mediated phosphorylation and contribute to the growing evidence that NME1 and NME2 play a critical role in cellular events.

Experimental procedures

Synthesis of ATP-biotin

The synthesis and characterization of ATP-biotin was previously described (71).

Phosphorylation or biotinylation with recombinant NME1 and NME2

Protein kinase buffer (PK; 50 mM Tris–HCl, 10 mM MgCl2, 0.1 mM EDTA, 2 mM DTT, 0.01% Brij 35, pH 7.5) and either NME1 (1 μg, at least 0.2 units) or NME2 (1 μg, at least 0.3 units) were combined. ATP-biotin (2 mM) was added to initiate the reaction. As controls, ATP (2 mM) was added instead of ATP-biotin, or competitive ATP (10 mM) was added prior to ATP-biotin. As another control, denaturant SDS (2%) was added prior to ATP-biotin. All reactions were incubated for 2 h at 31 °C with no shaking in a final volume of 30 μl (Fig. 2). As a final control, the incubated reaction was heated for 10 min at 95 °C prior to gel analysis (Fig. S2). For attempted biotinylation of KCNN4(301–360), PK buffer, NME2 (6 μg), KCNN4(301–360) (2 μg), and either ATP-biotin (4 mM) or ATP (4 mM) were used, as described above, with incubation for three or 16 h at 31 °C (Fig. S5). Because the commercial KCNN4(301–360) stock contained urea (8 M), reactions were also performed with KCNN4(301–360) samples where the urea was removed by dialysis by using a 3.5 kDa Slide-A-LyzerTm and dialysis buffer (0.5 ml of KCNN4 solution in 100 ml of 20 mM tris, 1 mM DTT, pH = 8.0) for 1.5 h at 4 °C three times, with the final dialysis performed for 16 h. Following dialysis, glycerol was added to a final concentration of 10% for storage at −80 °C. Reactions contained PK buffer, NME2 (3 μg), dialyzed KCNN4(301–360) (1.5 μg), and either ATP-biotin (4 mM) or ATP (4 mM) and incubated for 2 h at 31 °C (Fig. S6). For attempted biotinylation of KCNN4(338–427), PK buffer, NME2 (0.6 or 2 μg), KCNN4(328–427) (0.3 or 0.6 μg), and ATP (4 or 5 mM) were used, as described above, with incubation for 2, 4, or 6 h at 31 °C (Fig. S7). Laemmli sample buffer was added to each reaction without heat denaturation to maintain the stability of pHis. SDS-PAGE (16%) was used to separate the protein products, and SYPRO Ruby total protein stain was used to confirm equal protein loading. After transfer onto a PVDF membrane, streptavidin-Cy5 stain to visualize biotinylation and western blotting with the 1-pHis, 3-pHis, or global pHis antibodies, followed by rabbit secondary antibody, were used. Phos-tag gel experiments were conducted based on the protocol with zinc from Wako Pure Chemical Industries, Ltd, with the exception that Phos-tag (16%) was used in the SDS-PAGE gel (50). Reactions contained NME2 (2 μg), KCNN4 (301–360) (1 μg), and a high concentration of ATP (15 mM) to encourage an elevated level of phosphorylation and were incubated for 2 h at 31 °C. Proteins were separated by SDS-PAGE (16%) and gels were stained with Coomassie (Fig. 4B). Pro-Q Diamond phosphoprotein stain was conducted based on previous work (51). Reactions consisted of NME2 (2 μg), KCNN4(301–360) (1 μg), and ATP (4 mM) and were incubated for 2 h at 31 °C. After SDS-PAGE (16%) separation, the gel was fixed for 30 min in 50% MeOH at 4 °C and then washed three times for 10 min each with purified water. The gel was then stained with Pro-Q (60 ml) for exactly 1 h, which was followed by two washes for 5 min each with purified water. The gel was immediately imaged using the Typhoon (Fig. S7B).

NME1 and NME2 kinetics

The NADH-linked enzyme-coupled assay was conducted as previously described (7, 44). Briefly, initial rates of NME1 or NME2 kinetic activity with ATP and ATP-biotin were obtained in a Corning clear polystyrene 96-well microplate. The reactions consisted of Hepes buffer (50 mM Hepes, 150 mM NaCl, 1 mM MgCl2, 10 mM KCl, pH 7.5), BSA (0.5 mg/ml), NADH (250 μM), pyruvate kinase (PK, 3 units/ml), lactate dehydrogenase (5.25 units/ml), phosphoenolpyruvate (300 μM), and CDP (250 μM) in a 100 μl reaction volume. The final concentrations of ATP were 1, 3, 10, 30, and 100 μM and ATP-biotin were 10, 30, 90, 270, and 810 μM. The reaction was initiated by the addition of NME1 or NME2 (0.04 μg, 25 nM), and absorbance of NADH was monitored at 340 nm every 30 s for 30 min. A standard curve with NADH (0, 50, 100, 200, and 400 μM) was generated with each trial. The NADH concentration consumed was plotted versus time. The slope was determined for the initial rate (Fig. S3, right). The Michaelis–Menten equation (v = Vmax[S]/(KM + [S]), v = reaction rate and [S] = substrate concentration) was used to fit the rate versus substrate to obtain KappM and Vappmax (Fig. S3, left) using nonlinear regression with KaleidaGraph (Synergy software). To obtain kappcat values, Vmax was divided by enzyme concentration (25 nM). Mean and standard error for three independent trials are shown in Table 1. The enzyme-coupled assay was also used to assess phosphorylation of the KCNN4 peptide by NME2. The same assay conditions described above were used, except the CDP substrate was replaced with varying concentrations of WT or mutant KCNN4 peptide (50, 100, 200, and 400 μM) and a constant amount of ATP (2 mM) was used. No change in absorbance was observed (Fig. S3C).

Docking of ATP and ATP-biotin with NME1

The structures of ATP and ATP-biotin were drawn in ChemDraw 22.0. Energy-minimized structures were generated using Chem3D 22.0 and saved as sdf files, which was converted to a pdb file using PyMOL 2.5.4 (Schrodinger LLC). The crystal structure of NME1 was downloaded from RCSB Protein Data Bank (3L7U) and saved in different folders for each ATP and ATP-biotin pdb file. AutoDock Tools 1.5.7 (89) and Windows Notepad (Microsoft) were used to remove water, magnesium, and cocrystalized ligands. AutoDock Tools 1.5.7 was used to add hydrogen atoms and Kollman charges to prepare the separate NME1 pdbqt output files. The ATP and ATP-biotin pdb files were saved as ligand output pdbqt files using AutoDock Tools 1.5.7 in the same folder as the NME1 pdbqt. The grid boxes were created using AutoDock Tools 1.5.7 and for NME1, the size was x = 70, y = 70, and z = 70 and the coordinates used were x = −37.896, y = 12.333, and z = −24.749. A configuration file was generated using Notepad that included the file name of NME1, ligand, grid box size and coordinates, energy range, exhaustiveness of 20, and mode/pose number of 100. Autodock Vina 1.2.0 (90, 91) was saved in a separate folder. Calculations were carried out using Windows Command Prompt with Python 2.7.11 to generate the output pdbqt. The best ATP-biotin pose was picked based on the best comparison to ATP (Fig. S4). PyMOL 2.5.4 was used to measure the distance between the γ-phosphate group of ATP and N1 of the active site histidine in NME1.

Attempted labeling of KCNN4 peptide by NME2

Synthesis of KCNN4 peptides

Solid phase peptide synthesis was conducted to obtain KCNN4 WT (N-QVRLKHRKLRE-C, where His358 is underlined) and mutant (N-QVRLKARKLRE-C, Ala358 mutation is bold and underlined) peptide (24, 92). Fmoc-Glu(OtBu)-Wang resin (100–200 mesh, 0.65 mmol/g, 0.15 mmol) was swollen in dimethylformamide (DMF, 5 ml) for 1 hour. The resin was then rinsed once with DMF (5 ml), followed by incubation with 20% piperidine in DMF (5 ml) for 15 min at room temperature twice to deprotect the amino acid. The beads were washed 10 times with DMF (5 ml) and the Kaiser test (93) was performed to ensure deprotection before the addition of the next amino acid. Each amino acid was coupled in sequential order, with Fmoc-Arg(Pbf)-OH, Lys(Boc)-OH, Fmoc-Leu-OH, Fmoc-Val-OH, and Fmoc-Gln(Trt)-OH (0.45 mmol for each amino acid) and HOBt (0.72 mmol) dissolved in 5 ml DMF added to the resin, followed by DIC (0.58 mmol) addition and shaking overnight at room temperature. For the first Fmoc-Leu-OH, the second Arg(Pbf)-OH, Fmoc-His(Trt)-OH, and Fmoc-Ala-OH, each amino acid (0.73 mmol) dissolved in DMF (5 ml) was added to the resin and incubated with HATU (0.71 mmol) and DIPEA (1.5 mmol) using Biotage SP Wave Initiator + at 75 °C for 10 or 15 min. Coupling was monitored using the Kaiser test (93), and deprotections between couplings were performed as described earlier. After all couplings and final deprotection, the resin was washed with DMF (5 ml) 10 times, and N-terminal acetylation took place by adding acetic acid and DIPEA (1:1, 20 equiv) in DMF (5 ml) with shaking for 1 h at room temperature. The resin was washed with DMF (5 ml) 10 times, THF (5 ml) four times, a DCM and ethyl ether mixture (1:1, 5 ml total) three times, and ethyl ether (5 ml) 10 times, before drying with forced air. The cleavage mixture (2 ml; 82.5% TFA, 5% thioanisole, 5% phenol, 2.5% EDT, and 5% deionized, purified water) was added and shaken for 2 hours at room temperature. The cleaved peptides in solution were precipitated by the addition of cold ethyl ether (30 ml; precooled in −80 °C) and incubation at −20 °C for 10 min. The precipitated peptides were centrifuged at −9 °C for 5 min at 4000 rpm, the ethyl ether was decanted off, and the peptides were washed with cold ethyl ether (30 ml; precooled in −80 °C). After the ethyl ether was decanted for the second time, the peptides were air dried and then stored at 4 °C until HPLC purification. Peptides were dissolved in a solution of 95% water and 5% acetonitrile (20 ml) and were filtered using a Millex-GP filter unit (PES 0.22 μM). Semi-preparative HPLC was performed to obtain pure peptide using 0.1% TFA in water (buffer A) and 0.1% TFA in acetonitrile (buffer B). The gradient began with 95% A and 5% B to 80% A and 20% B over 5 min, followed by 80% to 60% A over 10 min and then 60% to 5% A for the next 5 min. Peptides were monitored at 214 nm. Pure peptides collected were lyophilized to afford a white solid (WT peptide, 43.5 mg, 20% and mutant peptide, 13.6 mg, 7%). Peptides were dissolved in water and neutralized with Tris base (10 mM, pH 7) prior to use. WT peptide ESI MS: [M + H]- expected 1504.8, observed: 1504.0. Mutant peptide ESI MS [M + H]- expected: 1438.7, observed: 1438.5. The WT and mutant peptide purities were 96% and 98%, respectively (Figs. S9 and S10).

Analysis of KCNN4 peptide phosphorylation by NME2

NME2 (2 μg), WT or mutant KCNN4 peptide (3 mM), and ATP (2 mM) were incubated in PK buffer at 31 °C for 2 h in a 15 μl final volume. The samples were diluted to 120 μl with water and analyzed by HPLC by injecting the entire sample with the same gradient as the peptide purification (Figs. 4C and S11).

ADP-Glo assay analysis of KCNN4 peptide phosphorylation by NME2

ADP-Glo assay samples consisted of NME2 (2 μg), WT KCNN4 peptide (400 μM), and ATP (100 μM) in PK buffer. Samples were incubated for 2 hours at 31 °C and the assay was performed according to the Promega Technical Manual for ADP-Glo Kinase Assay. No ADP production was observed (data not shown).

K-BILDS method

The K-BILDS method was performed as previously described, with exceptions noted (42). HeLa cells (1 L growth) were lysed in lysis buffer (4 ml; 50 mM Tris, pH 8.0, 150 mM NaCl, 10% glycerol, 0.5% Triton X-100) and 1x protease inhibitor cocktail at 4 °C for 30 min with end over end rotation, followed by centrifugation at 4 °C for 20 min to remove cell debris. Bradford assay was used to determine protein concentration of the soluble lysates. Single-use lysate aliquots were stored at −80 °C until use. FSBA inactivation included HeLa lysates (1 mg), FSBA (5 mM), and DMSO (20%) in lysis buffer (50 mM Tris, pH 8.0, 150 mM NaCl, 10% glycerol, 0.5% Triton X-100) at a 500 μl total volume, which was incubated for 2 h at 31 °C with 800 rpm rocking. To remove precipitate, samples were centrifuged for 2 min at 1000 rcf at room temperature. Soluble lysates were collected, diluted in buffer (2 ml; 50 mM Tris and 2 mM DTT), and centrifuged in a 3 kDa centriprep spin column at 14,200 rcf for 45 min twice to remove excess FSBA. FSBA-treated lysates were used immediately or aliquoted and stored at −80 °C until use. For the kinase reaction, FSBA-inactivated HeLa lysates (125 μg) with or without ATP-biotin (4 mM) were incubated with NME1 or NME2 (2 μg) in lysis buffer in a total volume of 19 μl for 2 h at 31 °C. ATP-biotin was removed by diluting to 400 μl with phosphate binding buffer (PBB: 380 μl; 28 mM NaH2PO4 · H2O, 72 mM Na2HPO4 · 7H2O, 150 mM NaCl, pH = 7.2) and centrifuging in an Amicon Ultra-0.5 Centrifugal Filter Units at 14,200 rcf for 30 min twice. Biotinylated proteins were enriched using prewashed (once with 400 μl PBB buffer) Pierce NeutrAvidin agarose beads (400 μl slurry) and incubation at room temperature for 1 h (KCNN4 samples) or 30 min (PSMA4, Rho-GDI, and UMPS samples). After enrichment, beads were washed 10 times with PBB (400 μl) for KCNN4 samples or high-stringency PBB (400 μl; 28 mM NaH2PO4 · H2O, 72 mM Na2HPO4 · 7H2O, 500 mM NaCl, pH = 7.2) for PSMA4, Rho-GDI, and UMPS samples, followed by five water (400 μl) washes. Proteins were eluted by mixing with 2% SDS in water (200 μl), heating for 7 minutes at 95 °C, drying using a SpeedVac Concentrator, and storing at −20 °C until use. Eluted proteins were denatured with Laemmli sample buffer without heat, separated by SDS-PAGE (16%), and visualized by Western blot analysis using KCNN4, PSMA4, Rho-GDI, or UMPS primary antibodies and rabbit secondary antibody (Figs. 5 and 6).

NME substrate discovery using K-BILDS

Sample prep and in-gel trypsin digestion

K-BILDS for proteomics analysis was conducted in the same manner as the gel analysis experiment except with FSBA-inactivated HeLa lysates (1 mg) in the presence or absence of NME (8 μg) and with or without ATP-biotin (4 mM). Following reaction incubation as described earlier, enrichment was performed using pre-washed (once with 400 μl PBB buffer) streptavidin resin (400 μl slurry). The enriched proteins were desalted using SDS-PAGE (10%) with no stacking layer in the gel and only running for 1 cm at 100 V. Gels were fixed and stained with SYPRO Ruby total protein stain as described by Thermo Fisher Scientific. A previously published protocol was followed for in-gel trypsin digestion with modifications (94). Gel bands were excised, cut into cubes (1 mm3), and placed into LoBind Eppendorf tubes. Gel pieces were destained by the addition of ammonium bicarbonate (NH4HCO3 50 mM, 250 μl) and incubation for 10 min at room temperature, followed by removal of the solution. Acetonitrile (MeCN, 250 μl) was added to the gel pieces and incubated for 15 min. The MeCN solutions were removed, fresh MeCN (250 μl) was added, samples incubated for 5 min or until gel pieces were dehydrated, and the MeCN removed again. The sequential process of NH4HCO3 destaining and MeCN dehydration was repeated but only incubating gel pieces with NH4HCO3 for 5 min. The gel pieces were dried with a SpeedVac Concentrator. Dried gel pieces were incubated for 20 min at 37 °C in a freshly prepared reducing buffer (250 μl per sample; 50 mM tris(2-carboxyethyl)phosphine, and 25 mM NH4HCO3) and the solution was removed. A solution of iodoacetamide (55 mM) in NH4HCO3 (50 mM in water) was freshly prepared, added to the gel pieces (250 μl), and incubated for 1 h at room temperature in the dark with gentle rocking. The iodoacetamide solution was removed, and the gel pieces were incubated for 15 min at room temperature with 1:1 MeCN:NH4HCO3 (250 μl; 50 mM in water) twice. After removal of the solution, gel pieces were incubated with MeCN (250 μl) for 5 min or until dehydrated. The MeCN was removed from the gel pieces and samples were dried with the SpeedVac Concentrator. Trypsin (20 μg) was dissolved in aqueous HCl (100 μl; 1 mM) and the resulting solution was added to digestion buffer (900 μl; 40 mM NH4HCO3 and 9% MeCN in water). Gel pieces were covered in the trypsin solution (250 μl each sample) and incubated for 16 h at 37 °C. After the incubation period, gel pieces were placed on ice and the remaining steps were conducted at 0 °C, unless otherwise indicated. The solutions in each sample were transferred to LoBind tubes, extraction buffer (250 μl; 1:1 MeCN:water with 0.002% formic acid) was added to the gel pieces, and the samples sonicated for 15 min to release digested peptides. The extraction buffer from each sample was combined with the trypsin solution from that same sample in the LoBind tubes. Samples were dried in a SpeedVac Concentrator at −50 °C and then stored at −20 °C until use. The dried peptide samples were resuspended in aqueous TFA (10 μl; 1%) and sonicated for 1 min. Pierce C18 tips were washed twice with MeCN (10 μl; 50% in water) and then equilibrated twice using aqueous TFA (10 μl; 0.1%). Samples were aspirated 10 times with the tip and the solutions were discarded. Peptides in the tip were washed twice with wash buffer (10 μl; 0.1% formic acid and 5% MeCN in water). Peptides were eluted in elution buffer (10 μl; 0.1% formic acid and 70% MeCN) by aspirating five times and dispensing to new LoBind tubes. Samples were dried using a SpeedVac concentrator and then stored at −20 °C until use.

Mass spectrometry analysis

Samples were analyzed in a similar manner as the previous K-BILDS study (42). An EASY nLC-1000 UHPLC (Thermo Fisher Scientific) was used to separate peptides in an acidic solvent system (0.1% formic acid in water). Analysis was performed using an Orbitrap Fusion Tribrid mass spectrometer. MS1 spectra were scanned from 350 to 1600 m/z at a resolution of 120,000. Higher energy collision-induced dissociation was used to fragment peptides with +2 to +7 charges. MS2 fragmentation was conducted using a 1.6 m/z window with a collision energy of 30% and a dynamic exclusion of 15 s. A ≤1% false discovery rate was set for proteins to be considered a positive identification. For label-free quantitation, NME1 candidate substrates were analyzed using the Proteome Discoverer database (58, 59), whereas MaxQuant (60) was used for NME2 candidate substrates. Proteins with higher intensities in samples with untreated FSBA-inactivated lysates than samples treated with ATP-biotin and NME1 or NME2 were removed. To identify candidate NME substrates, enrichment values were calculated by dividing the protein intensities in the NME1- or NME2-containing reactions by the intensities of the same protein in the samples without added NME1 or NME2. For NME1 candidate substrates, a ≥1.5-fold enrichment value was required. For NME2 candidate substrates, a ≥2.0-fold enrichment value was required (Table 2).

Data analysis

Gel bands were quantified using ImageJ software (National Institutes of Health, imagej.net) and normalized to the experimental sample (set to 1.0). Statistical analyses were performed using Excel (Microsoft 365) to calculate the SD of the mean and unpaired Student t test with unequal variance. Biological functions of the K-BILDS hit proteins were assessed through literature searches. Abundance analysis was performed by using Pax-db database (72, 73).

Data Availability

The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium with the dataset identifier PXD051411 and 10.6019/PXD051411.

Supporting information

Additional materials, methods, kinetics data, independent trials, and quantification of data (PDF). Raw data from LC-MS/MS study (Tables S1 and S2) (XLSX). Enrichment analysis of LC-MS/MS raw data, which includes protein hits (Tables S3 and S4) (XLSX). This article contains supporting information (72, 73, 90, 91, 95, 96).

Conflict of interest

The authors declare that they have no conflicts of interest with the contents of this article.

Supporting information

Supporting information

Table S1

Table S2

Table S3

Table S4

Acknowledgments

We thank the Stockdill Lab and L. Mendoza, along with T. Hendrickson, for technical support and E. Davis, H. Bremer, and A. Herppich for comments on the manuscript.

Author contributions

C. R. G. and M. K. H. P. writing–review and editing; C. R. G. writing–original draft; C. R. G., N. P. N. A., and T. O. O. validation; C. R. G., N. P. N. A., and T. O. O. investigation; C. R. G., N. P. N. A., and T. O. O. formal analysis; C. R. G. and M. K. H. P. conceptualization; M. K. H. P. visualization; M. K. H. P. supervision; M. K. H. P. resources; M. K. H. P. project administration; M. K. H. P. methodology; M. K. H. P. funding acquisition.

Funding and additional information

We thank the 10.13039/100000001 National Science Foundation (1904670 ) and 10.13039/100006710 Wayne State University for funding. We also thank the 10.13039/100000002 National Institutes of Health (P30 ES020957 , P30 CA022453 , and S10 OD030484 ) that support the 10.13039/100006710 Wayne State University Proteomics Core. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Science Foundation.
==== Refs
References

1 Manning G. Whyte D.B. Martinez R. Hunter T. Sudarsanam S. The protein kinase complement of the human genome Science 298 2002 1912 12471243
2 Hunter T. Protein kinases and phosphatases: the Yin and Yang of protein phosphorylation and signaling Cell 80 1995 225 236 7834742
3 Cohen P. The role of protein phosphorylation in human health and disease Eur. J. Biochem. 268 2001 5001 5010 11589691
4 Kussmann M. Hauser K. Kissmehl R. Breed J. Plattner H. Roepstorff P. Comparison of in vivo and in vitro phosphorylation of the exocytosis-sensitive protein PP63/parafusin by differential MALDI mass spectrometric peptide mapping Biochemistry 38 1999 7780 7790 10387018
5 Lawrence D.S. Chemical probes of signal-transducing proteins Acc. Chem. Res. 36 2003 401 409 12809526
6 Fuhs S.R. Meisenhelder J. Aslanian A. Ma L. Zagorska A. Stankova M. Monoclonal 1- and 3-phosphohistidine antibodies: new tools to study histidine phosphorylation Cell 162 2015 198 210 26140597
7 Yu B.Y.K. Tossounian M.-A. Hristov S.D. Lawrence R. Arora P. Tsuchiya Y. Regulation of metastasis suppressor NME1 by a key metabolic cofactor coenzyme A Redox Biol. 44 2021 101978
8 Mourad N. Parks R.E. Jr. Erythrocytic nucleoside diphosphokinase. II. Isolation and kinetics J. Biol. Chem. 241 1966 271 278 5903722
9 Kalagiri R. Hunter T. The many ways that nature has exploited the unusual structural and chemical properties of phosphohistidine for use in proteins Biochem. J. 478 2021 3575 3596 34624072
10 Wagner P.D. Vu N.D. Histidine to aspartate phosphotransferase activity of nm23 proteins: phosphorylation of aldolase C on Asp-319 Biochem. J. 346 2000 623 630 10698688
11 Freije J.M. Blay P. MacDonald N.J. Manrow R.E. Steeg P.S. Site-directed mutation of Nm23-H1. Mutations lacking motility suppressive capacity upon transfection are deficient in histidine-dependent protein phosphotransferase pathways in vitro J. Biol. Chem. 272 1997 5525 5532 9038158
12 Hardman G. Perkins S. Brownridge P.J. Clarke C.J. Byrne D.P. Campbell A.E. Strong anion exchange-mediated phosphoproteomics reveals extensive human non-canonical phosphorylation EMBO J. 38 2019 e100847
13 KOSHLAND D. Effect of catalysts on the hydrolysis of acetyl phosphate - nucleophilic displacement mechanisms in enzymatic reactions J. Am. Chem. Soc. 74 1952 2286 2292
14 Steeg P.S. Bevilacqua G. Kopper L. Thorgeirsson U.P. Talmadge J.E. Liotta L.A. Evidence for a novel gene associated with low tumor metastatic Potential2 J. Natl. Cancer Inst. 80 1988 200 204 3346912
15 Arnaud-Dabernat S. Bourbon P.M. Dierich A. Le Meur M. Daniel J.Y. Knockout mice as model systems for studying nm23/NDP kinase gene functions. Application to the nm23-M1 gene J. Bioenerg. Biomembr. 35 2003 19 30 12848338
16 Di L. Srivastava S. Zhdanova O. Sun Y. Li Z. Skolnik E.Y. Nucleoside diphosphate kinase B knock-out mice have impaired activation of the K+ channel KCa3.1, resulting in defective T cell activation J. Biol. Chem. 285 2010 38765 38771 20884616
17 Gilles A.M. Presecan E. Vonica A. Lascu I. Nucleoside diphosphate kinase from human erythrocytes. Structural characterization of the two polypeptide chains responsible for heterogeneity of the hexameric enzyme J. Biol. Chem. 266 1991 8784 8789 1851158
18 Wagner P.D. Vu N.-D. Phosphorylation of ATP-citrate lyase by nucleoside diphosphate kinase J. Biol. Chem. 270 1995 21758 21764 7665595
19 Hartsough M.T. Morrison D.K. Salerno M. Palmieri D. Ouatas T. Mair M. Nm23-H1 metastasis suppressor phosphorylation of kinase suppressor of Ras via a histidine protein kinase pathway J. Biol. Chem. 277 2002 32389 32399 12105213
20 Cuello F. Schulze R.A. Heemeyer F. Meyer H.E. Lutz S. Jakobs K.H. Activation of heterotrimeric G proteins by a high energy phosphate transfer via nucleoside diphosphate kinase (NDPK) B and Gbeta subunits. Complex formation of NDPK B with Gbeta gamma dimers and phosphorylation of His-266 IN Gbeta J. Biol. Chem. 278 2003 7220 7226 12486123
21 Verschueren K.H.G. Blanchet C. Felix J. Dansercoer A. De Vos D. Bloch Y. Structure of ATP citrate lyase and the origin of citrate synthase in the Krebs cycle Nature 568 2019 571 575 30944476
22 Mátyási B. Farkas Z. Kopper L. Sebestyén A. Boissan M. Mehta A. The function of NM23-H1/NME1 and its homologs in major processes linked to metastasis Pathol. Oncol. Res. 26 2020 49 61 31993913
23 Cai X. Srivastava S. Surindran S. Li Z. Skolnik E.Y. Regulation of the epithelial Ca2+ channel TRPV5 by reversible histidine phosphorylation mediated by NDPK-B and PHPT1 Mol. Biol. Cell 25 2014 1244 1250 24523290
24 Srivastava S. Li Z. Ko K. Choudhury P. Albaqumi M. Johnson A.K. Histidine phosphorylation of the potassium channel KCa3.1 by nucleoside diphosphate kinase B is required for activation of KCa3.1 and CD4 T cells Mol. Cell 24 2006 665 675 17157250
25 Srivastava S. Panda S. Li Z. Fuhs S.R. Hunter T. Thiele D.J. Histidine phosphorylation relieves copper inhibition in the mammalian potassium channel KCa3.1 eLife 5 2016 e16093 27542194
26 Brown B. Pressley B. Wulff H. KCa3.1 channel modulators as potential therapeutic compounds for glioblastoma Curr. Neuropharmacol. 15 2017 618 626
27 Xue L. Tao W.A. Current technologies to identify protein kinase substrates in high throughput Front. Biol. (Beijing) 8 2013 216 227 25110472
28 Xue L. Wang W.H. Iliuk A. Hu L. Galan J.A. Yu S. Sensitive kinase assay linked with phosphoproteomics for identifying direct kinase substrates Proc. Natl. Acad. Sci. U. S. A. 109 2012 5615 5620 22451900
29 Sardon T. Pache R.A. Stein A. Molina H. Vernos I. Aloy P. Uncovering new substrates for Aurora A kinase EMBO Rep. 11 2010 977 984 21072059
30 Klinker J.F. Seifert R. Nucleoside diphosphate kinase activity in soluble transducin preparations biochemical properties and possible role of transducin-beta as phosphorylated enzyme intermediate Eur. J. Biochem. 261 1999 72 80 10103035
31 Garrison A.W. Boozer C.E. The acid-catalyzed hydrolysis of a series of phosphoramidates J. Am. Chem. Soc. 90 1968 3486 3494
32 Anthony R.S. Spector L.B. Phosphorylated acetate kinase: its isolation and reactivity J. Biol. Chem. 247 1972 2120 2125 4552689
33 Wilke K.E. Francis S. Carlson E.E. Activity-based probe for histidine kinase signaling J. Am. Chem. Soc. 134 2012 9150 9153 22606938
34 Sousa M.M.L. Steen K.W. Hagen L. Slupphaug G. Antibody cross-linking and target elution protocols used for immunoprecipitation significantly modulate signal-to noise ratio in downstream 2D-PAGE analysis Proteome Sci. 9 2011 45 21816076
35 Leijten N.M. Heck A.J.R. Lemeer S. Histidine phosphorylation in human cells; a needle or phantom in the haystack? Nat. Methods 19 2022 827 828 35726056
36 Adam K. Ning J. Reina J. Hunter T. NME/NM23/NDPK and histidine phosphorylation Int. J. Mol. Sci. 21 2020 5848 32823988
37 Green K.D. Pflum M.K.H. Kinase-catalyzed biotinylation for phosphoprotein detection J. Am. Chem. Soc. 129 2007 10 11 17199263
38 Senevirathne C. Green K.D. Pflum M.K.H. Kinase-catalyzed biotinylation of peptides, proteins, and lysates Curr. Protoc. Chem. Biol. 4 2012 83 100 25177527
39 Green K.D. Pflum M.K.H. Exploring kinase cosubstrate promiscuity: monitoring kinase activity through dansylation ChemBioChem 10 2009 234 237 19107758
40 Suwal S. Pflum M.K. Phosphorylation-dependent kinase-substrate cross-linking Angew. Chem. Int. Ed. Engl. 49 2010 1627 1630 20108289
41 Senevirathne C. Embogama D.M. Anthony T.A. Fouda A.E. Pflum M.K.H. The generality of kinase-catalyzed biotinylation Bioorg. Med. Chem. 24 2016 12 19 26672511
42 Embogama D.M. Pflum M.K.H. K-BILDS: a kinase substrate discovery tool Chembiochem 18 2017 136 141 27860220
43 Knight J.D.R. Tian R. Lee R.E.C. Wang F. Beauvais A. Zou H. A novel whole-cell lysate kinase assay identifies substrates of the p38 MAPK in differentiating myoblasts Skeletal Muscle 2 2012 5 22394512
44 Cook P.F. Neville M.E. Jr. Vrana K.E. Hartl F.T. Roskoski R. Jr. Adenosine cyclic 3',5'-monophosphate dependent protein kinase: kinetic mechanism for the bovine skeletal muscle catalytic subunit Biochemistry 21 1982 5794 5799 6295440
45 Han B.G. Min K. Lee B.I. Lee S. Refined structure of human NM23-H1 from a hexagonal crystal Bull. Korean Chem. Soc. 31 2010 1397 1399
46 Morera S. Lacombe M.L. Xu Y. LeBras G. Janin J. X-ray structure of human nucleoside diphosphate kinase B complexed with GDP at 2 A resolution Structure 3 1995 1307 1314 8747457
47 Giraud M.-F. Georgescauld F. Lascu I. Dautant A. Crystal structures of S120G mutant and wild type of human nucleoside diphosphate kinase A in complex with ADP J. Bioenerg. Biomembr. 38 2006 261 264 16944299
48 Suwal S. Senevirathne C. Garre S. Pflum M.K. Structural analysis of ATP analogues compatible with kinase-catalyzed labeling Bioconjug. Chem. 23 2012 2386 2391 23116557
49 Kinoshita E. Kinoshita-Kikuta E. Takiyama K. Koike T. Phosphate-binding tag, a new tool to visualize phosphorylated proteins Mol. Cell Proteomics 5 2006 749 757 16340016
50 Yamada S. Nakamura H. Kinoshita E. Kinoshita-Kikuta E. Koike T. Shiro Y. Separation of a phosphorylated histidine protein using phosphate affinity polyacrylamide gel electrophoresis Anal. Biochem. 360 2007 160 162 17092477
51 [preprint] Adam K. Fuhs S. Meisenhelder J. Aslanian A. Diedrich J. Moresco J. A non-acidic method using hydroxyapatite and phosphohistidine monoclonal antibodies allows enrichment of phosphopeptides containing non-conventional phosphorylations for mass spectrometry analysis bioRxiv 2019 10.1101/691352
52 Zegzouti H. Zdanovskaia M. Hsiao K. Goueli S.A. ADP-glo: a bioluminescent and homogeneous ADP monitoring assay for kinases ASSAY Drug Dev. Tech. 7 2009 560 572
53 Gerlach A.C. Syme C.A. Giltinan L. Adelman J.P. Devor D.C. ATP-Dependent activation of the intermediate conductance, Ca2+-activated K+ channel, hIK1, is conferred by a C-terminal domain J. Biol. Chem. 276 2001 10963 10970
54 Wieland T. Hippe H.-J. Ludwig K. Zhou X. Korth M. Klumpp S. Reversible histidine phosphorylation in mammalian cells: a teeter-totter formed by nucleoside diphosphate kinase and protein histidine phosphatase 1 Methods Enzymol. 471 2010 379 402 20946858
55 Orlov N.Y. Orlova T.G. Nomura K. Hanai N. Kimura N. Transducin-mediated, isoform-specific interaction of recombinant rat nucleoside diphosphate kinases with bleached bovine retinal rod outer segment membranes FEBS Lett. 389 1996 186 190 8766826
56 Zhou X.-B. Feng Y.-X. Sun Q. Lukowski R. Qiu Y. Spiger K. Nucleoside diphosphate kinase B–activated intermediate conductance potassium channels are critical for neointima formation in mouse carotid arteries Arteriosclerosis, Thromb. Vasc. Biol. 35 2015 1852 1861
57 Biondi R.M. Engel M. Sauane M. Welter C. Issinger O.-G. Jiménez de Asúa L. Inhibition of nucleoside diphosphate kinase activity by in vitro phosphorylation by protein kinase CK2 Differential phosphorylation of NDP kinases in HeLa cells in culture FEBS Lett. 399 1996 183 187 8980148
58 Eng J.K. McCormack A.L. Yates J.R. An approach to correlate tandem mass spectral data of peptides with amino acid sequences in a protein database J. Am. Soc. Mass Spectrom. 5 1994 976 989 24226387
59 Orsburn B.C. Proteome discoverer-A community enhanced data processing suite for protein informatics Proteomes 9 2021 15 33806881
60 Cox J. Mann M. MaxQuant enables high peptide identification rates, individualized p.p.b.-range mass accuracies and proteome-wide protein quantification Nat. Biotechnol. 26 2008 1367 1372 19029910
61 Almond J.B. Cohen G.M. The proteasome: a novel target for cancer chemotherapy Leukemia 16 2002 433 443 11960320
62 Cho H.J. Kim J.-T. Baek K.E. Kim B.-Y. Lee H.G. Regulation of Rho GTPases by RhoGDIs in human cancers Cells 8 2019 1037 31492019
63 Alkam D. Feldman E.Z. Singh A. Kiaei M. Profilin1 biology and its mutation, actin(g) in disease Cell Mol. Life Sci. 74 2017 967 981 27669692
64 Zarouchlioti C. Parfitt D.A. Li W. Gittings L.M. Cheetham M.E. DNAJ Proteins in neurodegeneration: essential and protective factors Philos. Trans. R. Soc. Lond. B Biol. Sci. 373 2018 20160534 29203718
65 Frischmeyer P.A. Dietz H.C. Nonsense-mediated mRNA decay in health and disease Hum. Mol. Genet. 8 1999 1893 1900 10469842
66 Rungaldier S. Umlauf E. Mairhofer M. Salzer U. Thiele C. Prohaska R. Structure-function analysis of human stomatin: a mutation study PLoS One 12 2017 e0178646
67 Phillips R.E. Yang Y. Smith R.C. Thompson B.M. Yamasaki T. Soto-Feliciano Y.M. Target identification reveals lanosterol synthase as a vulnerability in glioma Proc. Natl. Acad. Sci. U. S. A. 116 2019 7957 7962 30923116
68 Okesli A. Khosla C. Bassik M.C. Human pyrimidine nucleotide biosynthesis as a target for antiviral chemotherapy Curr. Opin. Biotechnol. 48 2017 127 134 28458037
69 Lai T.S. Greenberg C.S. TGM2 and implications for human disease: role of alternative splicing Front. Biosci. (Landmark Ed.) 18 2013 504 519 23276939
70 Schattauer S.S. Land B.B. Reichard K.L. Abraham A.D. Burgeno L.M. Kuhar J.R. Peroxiredoxin 6 mediates Gαi protein-coupled receptor inactivation by cJun kinase Nat. Commun. 8 2017 743 28963507
71 Senevirathne C. Pflum M.K.H. Biotinylated phosphoproteins from kinase-catalyzed biotinylation are stable to phosphatases: implications for phosphoproteomics ChemBioChem 14 2013 381 387 23335220
72 Wang M. Weiss M. Simonovic M. Haertinger G. Schrimpf S.P. Hengartner M.O. PaxDb, a database of protein abundance averages across all three domains of life Mol. Cell Proteomics 11 2012 492 500 22535208
73 Nagaraj N. Wisniewski J.R. Geiger T. Cox J. Kircher M. Kelso J. Deep proteome and transcriptome mapping of a human cancer cell line Mol. Syst. Biol. 7 2011 548 22068331
74 Taylor S.S. Yang J. Wu J. Haste N.M. Radzio-Andzelm E. Anand G. PKA: a portrait of protein kinase dynamics Biochim. Biophys. Acta 1697 2004 259 269 15023366
75 Espinasse A. Wen X. Goodpaster J.D. Carlson E.E. Mechanistic studies of bioorthogonal ATP analogues for assessment of histidine kinase autophosphorylation ACS Chem. Biol. 15 2020 1252 1260 32043868
76 Hunter T. A journey from phosphotyrosine to phosphohistidine and beyond Mol. Cell 82 2022 2190 2200 35654043
77 Potel C.M. Lin M.-H. Heck A.J.R. Lemeer S. Widespread bacterial protein histidine phosphorylation revealed by mass spectrometry-based proteomics Nat. Methods 15 2018 187 190 29377012
78 Gao Y. Lee H. Kwon O.K. Cheng Z. Tan M. Kim K.-T. Profiling of histidine phosphoproteome in Danio rerio by TiO2 enrichment Proteomics 19 2019 1800471
79 Chase O.M. Espinasse A. Wilke K.E. Carlson E.E. Exploration of the effects of γ-phosphate-modified ATP analogues on histidine kinase autophosphorylation Biochemistry 57 2018 4368 4373 29944360
80 Wang C. Sang J. Wang J. Su M. Downey J.S. Wu Q. Mechanistic insights revealed by the crystal structure of a histidine kinase with signal transducer and sensor domains PLoS Biol. 11 2013 e1001493
81 Webb P.A. Perisic O. Mendola C.E. Backer J.M. Williams R.L. The crystal structure of a human nucleoside diphosphate kinase, NM23-H2 J. Mol. Biol. 251 1995 574 587 7658474
82 Lee C.H. MacKinnon R. Activation mechanism of a human SK-calmodulin channel complex elucidated by cryo-EM structures Science 360 2018 508 513 29724949
83 Bi M. Du X. Jiao Q. Chen X. Jiang H. Expanding the role of proteasome homeostasis in Parkinson’s disease: beyond protein breakdown Cell Death Dis. 12 2021 154 33542205
84 Qiu Y. Zhao D. Butenschön V-m Bauer A.T. Schneider S.W. Skolnik E.Y. Nucleoside diphosphate kinase B deficiency causes a diabetes-like vascular pathology via up-regulation of endothelial angiopoietin-2 in the retina Acta Diabetol. 53 2016 81 89 25900369
85 Wortmann S.B. Chen M.A. Colombo R. Pontoglio A. Alhaddad B. Botto L.D. Mild orotic aciduria in UMPS heterozygotes: a metabolic finding without clinical consequences J. Inherit. Metab. Dis. 40 2017 423 431 28205048
86 Matz A. Halamoda-Kenzaoui B. Hamelin R. Mosser S. Alattia J.-R. Dimitrov M. Identification of new Presenilin-1 phosphosites: implication for γ-secretase activity and Aβ production J. Neurochem. 133 2015 409 421 25458374
87 Koley D. Bard A.J. Triton X-100 concentration effects on membrane permeability of a single HeLa cell by scanning electrochemical microscopy (SECM) Proc. Natl. Acad. Sci. U. S. A. 107 2010 16783 16787 20837548
88 Landa S. Verlov N. Fedorova N. Filatov M. Pantina R. Burdakov V. Extracellular particles as carriers of cholesterol not associated with lipoproteins Membranes (Basel) 12 2022 618 35736324
89 Goodsell D.S. Olson A.J. Automated docking of substrates to proteins by simulated annealing Proteins 8 1990 195 202 2281083
90 Trott O. Olson A.J. AutoDock Vina: improving the speed and accuracy of docking with a new scoring function, efficient optimization, and multithreading J. Comput. Chem. 31 2010 455 461 19499576
91 Eberhardt J. Santos-Martins D. Tillack A.F. Forli S. AutoDock Vina 1.2.0: new docking methods, expanded force field, and Python bindings J. Chem. Inf. Model. 61 2021 3891 3898 34278794
92 Ji T. Corbalán-García S. Hubbard S.R. Crystal structure of the C-terminal four-helix bundle of the potassium channel KCa3.1 PLoS One 13 2018 e0199942
93 Kaiser E. Colescott R.L. Bossinger C.D. Cook P.I. Color test for detection of free terminal amino groups in the solid-phase synthesis of peptides Anal. Biochem. 34 1970 595 598 5443684
94 Shevchenko A. Tomas H. Havlis J. Olsen J.V. Mann M. In-gel digestion for mass spectrometric characterization of proteins and proteomes Nat. Protoc. 1 2006 2856 2860 17406544
95 Joiner W.J. Wang L.J.Y. Tong M.D. Kaczmarek L.K. hSK4, a member of a novel subfamly of calcium-activated potassium channels Proc. Natl. Acad. Sci. U. S. A. 94 1997 11013 11018 9380751
96 Herráez A. Biomolecules in the computer: Jmol to the rescue Biochem. Mol. Biol. Educ. 34 2006 255 261 21638687
