Protein arginine methylation facilitates KCNQ channel-PIP2 interaction leading to seizure suppression

  1. Hyun-Ji Kim
  2. Myong-Ho Jeong
  3. Kyung-Ran Kim
  4. Chang-Yun Jung
  5. Seul-Yi Lee
  6. Hanna Kim
  7. Jewoo Koh
  8. Tuan Anh Vuong
  9. Seungmoon Jung
  10. Hyunwoo Yang
  11. Su-Kyung Park
  12. Dahee Choi
  13. Sung Hun Kim
  14. KyeongJin Kang
  15. Jong-Woo Sohn
  16. Joo Min Park
  17. Daejong Jeon
  18. Seung-Hoi Koo
  19. Won-Kyung Ho
  20. Jong-Sun Kang  Is a corresponding author
  21. Seong-Tae Kim  Is a corresponding author
  22. Hana Cho  Is a corresponding author
  1. Sungkyunkwan University School of Medicine, Korea
  2. Seoul National University College of Medicine, Korea
  3. Seoul National University Medical Research Center, Korea
  4. Korea Advanced Institute of Science and Technology, Korea
  5. Korea University, Korea
  6. College of Medicine, Kangwon National University, Korea
  7. Institute for Basic Science, Korea
  8. Seoul National University Hospital, Korea
  9. Advanced Neural Technologies, Republic of Korea
9 figures and 1 table

Figures

Spontaneous seizures and increased locomotor activities in Prmt1+/- mice.

(a) reduced expression of Prmt1 in Prmt1+/- mice, compared to WT control mice. (b) Representative traces of seizure activities: short (upper) or long (middle and lower) duration of epileptiform …

https://doi.org/10.7554/eLife.17159.003
Comparison of neuronal excitability in WT and Prmt1+/- mice.

(ab) a and b panels show representative trace in the whole-cell current-clamp recording from WT (a) and Prmt1+/- (b) mature dentate GCs in response to 1-s depolarizing current injection (200 pA), …

https://doi.org/10.7554/eLife.17159.005
Figure 3 with 1 supplement
KCNQ current deficiency contributed to the persistent hyperexcitability in Prmt1+/- mice.

(ab) Representative current traces of voltage clamp recordings from WT (a; n = 22, seven mice) and Prmt1+/- (b; n = 12, four mice) mature dentate GCs in response to the voltage protocol depicted …

https://doi.org/10.7554/eLife.17159.007
Figure 3—figure supplement 1
Changes in membrane potential in WT (+/+) and Prmt1+/- GCs (+/-) in response to 10 μM XE991.

Each connected line represents an individual neuron. Right panel gives the averaged changes. XE991 was less effective in mutants than in WT.

https://doi.org/10.7554/eLife.17159.009
Methylation suppression with MTA or furamidine, increases neuronal excitability via KCNQ channel.

(ab) Representative current traces recorded from MTA- (a) or furamidine- (b) pretreated WT GCs using the voltage protocol depicted below in the absence (black) or presence (blue) of 10 μM XE991. (c)…

https://doi.org/10.7554/eLife.17159.011
Figure 5 with 5 supplements
Prmt1 binds to and methylates KCNQ2 at R333, R345, R353, and R435.

(ac) Immunoblotting analysis showing the physical association of KCNQ2 and Prmt1. HEK293T cells were transfected with expression vectors, as indicated. Whole-cell lysates were immunoprecipitated …

https://doi.org/10.7554/eLife.17159.013
Figure 5—figure supplement 1
Expression of KCNQ2 and KCNQ3 in the hippocampus of WT and Prmt1+/- mice.
https://doi.org/10.7554/eLife.17159.014
Figure 5—figure supplement 2
Identification of in vivo arginine methylation sites of KCNQ2 MS/MS spectra of the methylated peptides.

(a) 332RRNPAAGLIQSAWR345, (b) 334NPAAGLIQSAWR345, (c) 346FYATNLSR353 and (d435RSPSADQSLEDSPSK449. R* represents monomethylated arginine and R** represents dimethylated arginine.

https://doi.org/10.7554/eLife.17159.015
Figure 5—figure supplement 3
The sequence alignment of 5 difference human KCNQ isoforms.

The arginine residues critical for methylation in KCNQ2 are marked in red. KCNQ3 and KCNQ4 have 3 conserved arginine residues while KCNQ5 and KCNQ1 have two or one conserved arginine residues, …

https://doi.org/10.7554/eLife.17159.016
Figure 5—figure supplement 4
In vitro methylation assays with GST-Prmt1 or myc-Prmt5 together with GST-KCNQ2 (320-449 aa).

In vitro methylation assays with myc-Prmt5 purified from HEK293T cells and GST-Prmt1 by using GST-KCNQ2 (aa 320–449) as a substrate and histone as a positive control. Coomassie blue stainings show …

https://doi.org/10.7554/eLife.17159.017
Figure 5—figure supplement 5
Expression of Prmt8 in Prmt1+/+ and Prmt1+/- brain lysates. 
https://doi.org/10.7554/eLife.17159.018
Methylation of KCNQ2 regulates its channel activity.

(a) Representative current recordings from HEK293T cells expressing KCNQ2 with control shRNA vector (left) or a Prmt1 shRNA (Prmt1 sh) (right). Currents were elicited by voltage steps from -70 mV to …

https://doi.org/10.7554/eLife.17159.020
Figure 7 with 5 supplements
Methylation of KCNQ2 determines its PIP2 affinity.

(a) Pooled data show 10 μM neomycin-induced rundown of WT (●) and R333K (○) KCNQ2 currents. KCNQ2 currents were normalized to KCNQ2 current at t = 0. (b) Dose-response curves for neomycin measured …

https://doi.org/10.7554/eLife.17159.022
Figure 7—figure supplement 1
Surface expression of KCNQ channels was not affected by Prmt1 knockdown or 4RK mutation.

Membrane surface biotinylation assays show that the ratio of channels on the membrane to those in the total lysates was not significantly different in control cells and Prmt1 knockdown cells (left). …

https://doi.org/10.7554/eLife.17159.024
Figure 7—figure supplement 2
Quantitative determination of the sensitivity of KCNQ2 channels to activation of Dr-VSP in HEK293T cells.

Tail current amplitudes were used to measure current inhibition by Dr-VSP activation and its recovery. (ae) Time course of tail current amplitude in a cell transfected with Dr-VSP and WT KCNQ2 (a), …

https://doi.org/10.7554/eLife.17159.025
Figure 7—figure supplement 3
Comparison of XE991 sensitivity in GCs.

Changes in firing rate (a) and threshold current (b) in WT (+/+) neurons, mutants (+/-), and mutants loaded with 20μM diC8-PIP2 (+/-; PIP2) in response to 10 μM XE991. Each connected line represents …

https://doi.org/10.7554/eLife.17159.027
Figure 7—figure supplement 4
Application of exogenous PIP2 had little effect on the excitability in WT neurons.

(a) representative trace in the whole-cell current-clamp recording from WT GCs with no PIP2 (left) or 20 μM diC8-PIP2 in the patch pipette (right) in response to 1-s depolarizing current injection …

https://doi.org/10.7554/eLife.17159.029
Figure 7—figure supplement 5
The effects of SK channel block and BK channel block on neuronal excitability of control or furamidine-pretreated GCs.

Spike trains were evoked by injecting 1-s depolarizing current pulses of different intensities (100–250 pA) into the cell before and after application of apamin (af) or paxilline (gl) in control …

https://doi.org/10.7554/eLife.17159.031
Prmt1 is involved in Oxidative stress-induced silencing of neurons.

Spike trains were evoked by injecting 1-s depolarizing current pulses of different intensities (100–250 pA) into the cell before (black) and after (red) application of H2O2. Left panels show …

https://doi.org/10.7554/eLife.17159.033
Figure 9 with 1 supplement
Activation of KCNQ2 channels by H2O2 is sensitive to methylation.

(ag) Time course for the effect of 500 μM H2O2 on KCNQ2 currents. WT KCNQ2 (a-c), R333K (d), R345K (e), R353K (f) or R435K (g) were expressed in HEK293T cells. For experiment (b) & (c), cells were …

https://doi.org/10.7554/eLife.17159.035
Figure 9—figure supplement 1
Time course for the effect of 500 μM H2O2 on Triple-Cys mutant of KCNQ2 currents (left).

Summarized data for Triple-Cys mutant of KCNQ2 currents activation induced by H2O2 or retigabine (right). (b) Treatment of MTA induces a reduction of KCNQ2(WT) and Triple-Cys mutant of KCNQ2 …

https://doi.org/10.7554/eLife.17159.037

Tables

Table 1

Mono- and dimethylated peptides identified from KCNQ2.

https://doi.org/10.7554/eLife.17159.019
Peptide*ModificationCharge state
332RRNPAAGLIQSAWR345Mono-methylated R333+3
334NPAAGLIQSAWR345Mono-methylated R345+2
346FYATNLSR353Di-methylated R353+2
435RSPSADQSLEDSPSK449Di-methylated R435+3
  1. *Methylated amino acid residues are highlighted in bold.

Download links