CDKL5 and Shootin1 Interact and Concur in Regulating Neuronal Polarization

Abstract

In the last years, the X-linked cyclin-dependent kinase-like 5 (CDKL5) gene has been associated with epileptic encephalopathies characterized by the early onset of intractable epilepsy, severe developmental delay, autistic features, and often the development of Rett syndrome-like features. Still, the role of CDKL5 in neuronal functions is not fully understood. By way of a yeast two hybrid screening we identified the interaction of CDKL5 with shootin1, a brain specific protein acting as a determinant of axon formation during neuronal polarization. We found evidence that CDKL5 is involved, at least in part, in regulating neuronal polarization through its interaction with shootin1. Indeed, the two proteins interact in vivo and both are localized in the distal tip of outgrowing axons. By using primary hippocampal neurons as model system we find that adequate CDKL5 levels are required for axon specification. In fact, a significant number of neurons overexpressing CDKL5 is characterized by supernumerary axons, while the silencing of CDKL5 disrupts neuronal polarization. Interestingly, shootin1 phosphorylation is reduced in neurons silenced for CDKL5 suggesting that the kinase affects, directly or indirectly, the post-translational modification of shootin1. Finally, we find that the capacity of CDKL5 to generate surplus axons is attenuated in neurons with reduced shootin1 levels, in agreement with the notion that two proteins act in a common pathway. Altogether, these results point to a role of CDKL5 in the early steps of neuronal differentiation that can be explained, at least in part, by its association with shootin1.

Fig 1. CDKL5 interacts with shootin1 in vivo.

(A) A yeast two-hybrid screening identified shootin1 as a CDKL5 interacting protein. The C-terminal region of hCDKL5, spanning amino acids 299–1030, was used as bait (upper, thick bar). The diagram below shows shootin1 with its coiled coil domains in black. The clones identified in the screen are indicated as black bars and the minimum CDKL5 interacting region as a black bar. (B) Coimmunoprecipitation of P5-7 brain lysates with anti-CDKL5 (upper, n = 3) or anti-shootin1 (lower, n = 3) antibodies (both rabbit). IgGs were used as negative control. The immunoprecipitates and inputs (5% of the brain lysates) were analyzed by immunoblotting for CDKL5 and shootin1 (using a goat anti-shootin1 antibody). Asterisks indicate the immunoglobulin heavy chains and the open circle an unspecific band detected with anti-CDKL5. (C) Coimmunoprecipitation of HeLa cells overexpressing either Flag-CDKL5 or shootin1 or both proteins together. Whole cell lysates were immunoprecipiated with an anti-Flag resin and inputs (5%) and immunocomplexes analyzed by western blotting as indicated. Asterisk shows an anti-shootin1 reactive protein that copurifies with CDKL5. (n = 3).

Fig 2. CDKL5 and shootin1 are coexpressed in brains and neurons.

(A) Western blot analysis showing CDKL5 and shootin1 levels in mouse brain at the indicated developmental stages using Tuj1 as loading control. (n = 2) (B) Shootin1 is expressed in the cortex, as early as E13, in the cortical plate (cp) and its levels increase ongoing with development (a,b,c,d); Cdkl5 (b’,c’,d’) follow the same pattern. Low but detectable levels of shootin1 and Cdkl5 mRNAs are present in cells migrating out of the ventricular and sub-ventricular zone (vz-svz) towards their final destination in the cortical plate (b,b’). At E18 shootin1 and Cdkl5 are strongly expressed throughout the whole thickness of the cortex (d,d’). Scale bars: 50 μm: b,b’,c,c’; 100 μm: d,d’; 200 μm: a. (C) Western blot showing CDKL5 and shootin1 levels in cultured primary hippocampal neurons at the indicated stages. A longer exposure of the 18 h time point is shown to the right. (n = 2). (D) Immunofluorescence analysis (left) of hippocampal neurons at stages 2–3 with antibodies against CDKL5 (green) and shootin1 (red). The small panels show the magnification of the minor processes/axons indicated with asterisks. Quantitative profiles showing the fluorescence intensities of shootin1 (red) and CDKL5 (green) from the soma to the distal tip of the neurites/axons indicated with asterisks are shown to the right. Scale bar: 10 μm.

Fig 3. CDKL5 knock-down causes aberrant neuronal polarization.

(A) Western blot showing CDKL5 and shootin1 levels in primary hippocampal neurons infected with shRNAs against CDKL5, shootin1 or, as control, LacZ. Neurons were infected at DIV0 and cell lysates prepared after 96 h. Tuj1 was used as loading control. The graph shows the quantified protein levels as means of n≥3 experiments, *p<0,05, **p<0,01, ***p<0,001 (Student’s t test). (B) Representative images of neurons 96 h after plating and infection of viral particles expressing GFP together with shRNAs against LacZ, CDKL5 or shootin1. The image is a composed panel of GFP-positive neurons. Scale bar: 50 μm. (C) Graph showing polarization of neurons expressing shRNAs against LacZ, CDKL5 or shootin1. Polarized neurons are indicated with black bars, those with numerous axons with dark grey bars and neurons with no axon with light grey bars. Data present percentage of neurons as means ±SEM (n≥250 neurons/condition from n≥3 experiments), *p<0,05, **p<0,01, ***p<0,001 (Student’s t test). (D) Quantification of the length of the longest axon and dendrite of shRNA-expressing neurons. Data present neurite length as means ±SEM (a total of 110 neurons were analyzed in 4 independent experiments). (E) Neurite length of GFP-positive neurons bearing one axon (a total of 74 neurons were analyzed in 3 independent experiments). Data in D and E present neurite length as means ±SEM, *, p<0,05, ***p<0.001. (Student’s t test).

Fig 4. CDKL5 promotes axon formation.

(A) Western blot showing CDKL5 levels in primary hippocampal neurons nucleofected before plating with bicistronic vectors expressing GFP alone or together with CDKL5 or CDKL5-K42R. Cell lysates were prepared at DIV5 and analyzed for CDKL5 levels using Tuj1 as loading control. (B) Representative images showing the localization of endogenous and exogenous CDKL5 in the soma of nucleofected GFP-positive neurons at DIV5. The exposure time of the GFP-expressing neuron (left) was increased to reveal the staining of endogenous CDKL5. (C) Representative images showing hippocampal neurons at DIV5 transfected with vectors expressing GFP together with CDKL5 or the K42R derivative. GFP and CDKL5 signals are in green and red, respectively. The arrows indicate neurons with increased CDKL5 levels. Scale bar: 20 μm. (D) Quantitative analysis of neuronal polarization. Axon specification was analyzed at DIV5 by determining the number of neurons with a single axon (polarized, black bars), multiple axons (dark grey bars) and neurons with no axon (light grey bars). Data are expressed as mean of 4 independent experiments ±SEM; ***p<0,001, **p<0,01, *p<0,05 (n≥100 neurons/condition, Student’s t test). (E) Graph showing the length of the longest axon and dendrite of transfected neurons. Data present neurite length as means ±SEM (n>28 neurons/condition, 4 independent experiments); **p<0,01. (Student’s t test).

Fig 5. CDKL5 regulates axon outgrowth through shootin1.

(A) Immunofluorescence of neurons nucleofected before plating with a bicistronic vector expressing GFP alone or together with CDKL5 and subsequently infected with lentiviral particles expressing shRNAs against shootin1 or LacZ. At DIV5 neurons were stained for GFP, Tau1, and CDKL5 (green, blue, and red, respectively). (B) Quantification of neuronal polarization of GFP-positive neurons with increased CDKL5 expression. Data are expressed as means ±SEM. **p<0,01, *p<0,05. (n≥24 neurons/condition in 3 independent experiments; ANOVA two-way). Scale bar: 20 μm.

Fig 6. CDKL5 influences shootin1 phosphorylation in primary cortical neurons.

(A) Total cell extracts of DIV7 cortical neurons were treated with or without lambda phosphatase (λ-PPase) and analyzed by two-dimensional gel electrophoresis and immunoblotting with antibodies against shootin1, β-actin and, as control for the λ-PPase treatment, phopho-ERK1/2. (B) Primary cortical neurons were infected with shLacZ- or shCDKL5#1-expressing viral particles at DIV0 and total cell lysates were prepared at DIV7 and subjected to two-dimensional gel electrophoresis. Shootin1 and NFL were detected by immunoblotting; the single NFL-spot was used as internal control for alignment. Silencing of CDKL5 was confirmed by western blot (right panel). (n = 3).