henhaobio
Sign in

Open BiologyNeurosciencePeer-reviewed journal article

NEK1 autophosphorylation is disrupted by amyotrophic lateral sclerosis-associated missense variants: activity biomarkers and structural insights

Shalini AgarwaliD, S.A. Abdul RehmaniD, Ivan M. MunoziD, Axel KnebeliD, Paul J. HopiD, Robert GourlayiD, Fiona BrowniD, Thomas MacartneyiD, Iolo Squires, Jan Herman VeldinkiD, Kevin Patrick KennaiD, John Rouse, Arpan R. MehtaiD

PDFDOIPreprintData
In-depth full-text digest available · Jump to digest ↓

Abstract

Rare variants in NEK1, encoding a serine/threonine kinase, are amongst the most consistently implicated genetic contributors to amyotrophic lateral sclerosis (ALS), reported in approximately 2-3% of cases. Yet, whilst recent studies have characterised the cell biological consequences of NEK1 loss-of-function, the biochemical effects of ALS-associated missense variants on kinase activity have not been directly investigated. This distinction is mechanistically important, because missense alleles encode mutant proteins rather than simply reducing protein dosage. Here, we provide the most comprehensive cell-based phosphoproteomic map of NEK1 phosphorylation to date, identifying ten recurrent phosphorylation sites across independent expression and acquisition conditions. We experimentally assign pSer14, pThr156 and pSer418 as NEK1 autophosphorylation sites using kinase-dead controls, targeted extracted ion chromatogram analysis, phosphosite mutagenesis and phosphospecific antibodies. Leveraging activation-loop pThr156 as a readout of NEK1 activity, we assessed nine ALS-associated missense variants spanning the major functional regions of the protein. Amongst catalytic-domain variants, R261C produced the most robust reduction in pThr156 autophosphorylation, R232C produced a smaller reduction, and R232H increased pThr156; the basic-region variant, A313T, also showed a smaller reduction. Structural modelling provides a mechanistic framework for understanding these variant-specific effects. Our study establishes the first activity-based framework for functional classification of NEK1 missense variants, and provides direct evidence that ALS-associated missense variants can alter NEK1 autophosphorylation through a mechanism distinct from simple haploinsufficiency. The phosphospecific antibodies, isogenic cell lines and curated phosphoproteomic datasets generated here provide a community resource for future studies of NEK1 regulation, variant interpretation and therapeutic target validation.

In-depth summary

Claude Sonnet 4.6 · xhigh effort · 8 Nov 2026 · from the full text of version 2 (3 Sept 2026) · CC BY 4.0

Kinase-dead controls, XIC quantification and phosphospecific antibodies establish pSer14, pThr156 and pSer418 as NEK1 autophosphorylation sites; a pThr156 assay shows ALS missense variants differ, with R261C reducing it most.

ALS-associated NEK1 missense variants encode full-length mutant proteins, yet their effects on kinase activity had not been measured directly, and no validated autophosphorylation sites or activity biomarkers existed for NEK1. The paper supplies validated sites, phosphospecific antibodies and an activity-based readout for sorting variants.

Key results

  1. ShownCell-based mass spectrometry of GFP-NEK1 re-expressed in NEK1-knockout U-2 OS cells found ten phosphorylation sites that recur across four acquisition conditions. One of them, pThr1090, is absent from PhosphoSitePlus.

    Conditions were transient vs stable expression, each with HCD or EThcD fragmentation. Nine of ten sites are in PhosphoSitePlus, and 25 of its 55 annotated isoform 3 sites were detected. pThr156 and pThr1090 were absent from kinase-dead cells; the other eight were also seen there. · Phosphoproteomic profiling identifies recurrent…

  2. ShownpSer14, pThr156 and pSer418 depend on NEK1 catalytic activity and were assigned as autophosphorylation sites, using kinase-dead controls, XIC quantification, phosphosite mutants and in-house phosphospecific antibodies.

    All eight XIC-tested candidates showed more than 5-fold WT/kinase-dead enrichment in at least one condition (maximum 80-fold for pThr1090 to 5,051-fold for pThr156). The three sites were detected in recombinant WT NEK1 (3/3 replicates) and absent from recombinant kinase-dead. Antibody signal was abolished or strongly reduced in D146A and in S14A, T156A and S418A, with some residual signal. · Integrated phosphosite mapping, mutagenesis and…; Figure legends

  3. ShownAmong nine ALS-associated missense variants, R261C gave the largest drop in activation-loop pThr156 autophosphorylation, and this effect survived multiple-comparison correction.

    pThr156 normalised to total NEK1 and relative to WT: R261C mean 0.369 (p=0.0027), kinase-dead D146A mean 0.046 (p=0.0002). Mean ± SEM from N=3 experiments, one-sample t-tests. Holm-adjusted p=0.024 for R261C. · ALS-associated NEK1 missense variants…; Figure legends

  4. SuggestedR232C and A313T showed smaller nominal reductions in pThr156, while R232H increased it. None of these three remained significant after correction, so they are suggestive only. AlphaFold3 models offer a structural rationale for the opposing arginine substitutions.

    Mean relative to WT: R232C 0.474 (p=0.0395), A313T 0.501 (p=0.0364), R232H 1.331 (p=0.0218). Holm-adjusted p ≥ 0.17 for R232H, A313T and R232C. Models predict R232C and R261C abolish charge-based contacts that R232H and R261H partly keep; A313T was not modelled. · ALS-associated NEK1 missense variants…; Figure legends

  5. SuggestedThe variant with the strongest genetic support, R261H, did not significantly change pThr156. Neither did the nominally protective K648E, and V223M, R440Q and R721Q showed only non-significant lower means at N=3.

    R261H and K648E did not differ significantly from WT; V223M, R440Q and R721Q reductions did not reach significance (N=3, unadjusted tests). The authors conclude missense variants do not act as a single functional class. · ALS-associated NEK1 missense variants…; Discussion; Figure legends

Methods, data and reproducibility

GFP-NEK1 (wild-type or D146A kinase-dead) was re-expressed in a CRISPR NEK1-knockout U-2 OS line and phosphosites were mapped by LC-MS/MS with HCD and EThcD, then confirmed by recombinant WT/kinase-dead mapping, Skyline XIC quantification, mutagenesis and antibodies. Nine ALS variants were stably expressed and pThr156 was quantified by immunoblot.

  • GFP-NEK1 WT and D146A kinase-dead phosphosite mapping in NEK1-knockout Flp-In T-REx U-2 OS cells, transient and stable expression, HCD and EThcD · human (U-2 OS cell line, human NEK1 isoform 3) · n = four wild-type cell-based acquisition conditions · PXD079815
  • Recombinant full-length NEK1 WT and D146A co-expressed with C21ORF2 in Sf9 insect cells, phosphosite mapping · human NEK1 expressed in Spodoptera frugiperda Sf9 cells · n = three independent biological replicates
  • Stable lines expressing nine ALS-associated NEK1 missense variants (V223M, R232C, R232H, R261C, R261H, A313T, R440Q, K648E, R721Q), pThr156 immunoblot · human (U-2 OS cell line) · n = N=3 independent experiments
  • Published exome-wide rare variant association summary statistics used to select NEK1 missense variants (P < 0.05) · human · n = 13,138 ALS cases and 69,775 controls
  • NEK1 orthologue sequences for conservation and ancestral-state analysis · 26 vertebrate species · n = 26 species
  • Source data, plasmid and knockout sequencing data · 10.5281/zenodo.22118136
Methods & tools
Orbitrap Lumos (HCD, EThcD), Proteome Discoverer 2.4, Mascot 2.8.3, Skyline v25.1, AlphaFold3, ConSurf, MAFFT, IQ-TREE v3.0.1, PastML, PhosphoSitePlus, RVAT, GraphPad Prism 11
Reagents & environment
Proteome Discoverer 2.4, Mascot v2.8.3, Skyline v25.1, IQ-TREE v3.0.1, Fiji v2.16.0/1.54p, Jalview 2.11.5.1, GraphPad Prism 11.
Reproducibility gaps
Cell lines, reagents and antibodies are available only on reasonable request. Custom code for the UpSet plot and ConSurf-to-Jalview mapping is mentioned but no repository is given. Annotated spectra are Mascot reports rather than raw spectra.
Limitations (5)
  • Work was confined to a single osteosarcoma-derived line (U-2 OS); the autophosphorylation order has not been checked in neurons or motor neurons. · stated by the authors
  • Variants were expressed as a single allele in a homozygous knockout background, which does not model the heterozygous state of most patients or dominant interactions with wild-type NEK1. · stated by the authors
  • In vitro kinase assays and mixing experiments were not done, so cis versus trans autophosphorylation is unresolved; site occupancy and co-occurrence on one molecule were not measured. · stated by the authors
  • The pThr156 antibody retains residual signal in T156A lanes, possibly from cross-reactivity with pThr141; this was not tested. The Ser155/Thr156 coupling data are exploratory and need replicated quantification. · stated by the authors
  • Variant screen used N=3 and unadjusted tests; only R261C survives correction, so claims for R232C, A313T and R232H are tentative. · noted in this summary

Read this if you work on NEK1 or ALS variant interpretation and want validated phosphosites, antibodies and a pThr156 activity readout for classifying missense alleles.

Citations
1
FWCI
4.20
Field percentile
top 6%

Reproducibility ledger

Nobody has reported trying to reproduce this paper yet. Did you replicate an experiment or run the code? Your report saves the next lab a week.

Sign in to file a reproducibility report.

Related papers

Similarity Map
SelectedSame fieldCross-field
THISBloom '22Marriott '22Γεωργιάδου '26Vrijsen '23Liu '22Huang '24
64% semantic matchNeuroscience
A. Joseph Bloom, Xianrong R. Mao, Amy M. Strickland+3 more·Molecular Neurodegeneration·Jan 6, 2022· 117 citations