Functional Redundancy of Cyclase-Associated Proteins CAP1 and CAP2 in Differentiating Neurons

Cyclase-associated proteins (CAPs) are evolutionary-conserved actin-binding proteins with crucial functions in regulating actin dynamics, the spatiotemporally controlled assembly and disassembly of actin filaments (F-actin). Mammals possess two family members (CAP1 and CAP2) with different expression patterns. Unlike most other tissues, both CAPs are expressed in the brain and present in hippocampal neurons. We recently reported crucial roles for CAP1 in growth cone function, neuron differentiation, and neuron connectivity in the mouse brain. Instead, CAP2 controls dendritic spine morphology and synaptic plasticity, and its dysregulation contributes to Alzheimer’s disease pathology. These findings are in line with a model in which CAP1 controls important aspects during neuron differentiation, while CAP2 is relevant in differentiated neurons. We here report CAP2 expression during neuron differentiation and its enrichment in growth cones. We therefore hypothesized that CAP2 is relevant not only in excitatory synapses, but also in differentiating neurons. However, CAP2 inactivation neither impaired growth cone morphology and motility nor neuron differentiation. Moreover, CAP2 mutant mice did not display any obvious changes in brain anatomy. Hence, differently from CAP1, CAP2 was dispensable for neuron differentiation and brain development. Interestingly, overexpression of CAP2 rescued not only growth cone size in CAP1-deficient neurons, but also their morphology and differentiation. Our data provide evidence for functional redundancy of CAP1 and CAP2 in differentiating neurons, and they suggest compensatory mechanisms in single mutant neurons.


Introduction
Mammalian cyclase-associated protein (CAP) and its yeast homolog suppressor of RAS2-V19 (SRV2, collectively referred to CAP in this manuscript) were both recognized, two decades ago, as actin-binding proteins (ABPs) [1,2], but their molecular functions remained largely unknown until recently. By exploiting recombinant proteins and mutant yeast strains, studies of the past few years implicated CAP in various steps of the actin treadmilling mechanism, thereby identifying it as a crucial regulator of actin dynamics [3][4][5][6][7][8]. Specifically, these studies revealed a cooperation with the actin-depolymerizing protein ADF/cofilin in the dissociation of actin subunits from filamentous actin (F-actin) and, hence, in F-actin disassembly [5,7,9]. Moreover, they revealed a role for CAP in nucleotide exchange on globular actin monomers (G-actin) relevant for G-actin recycling and F-actin assembly [4], and they reported an inhibitory function towards the F-actin assembly factor, inverted formin 2 (INF2) [6,8]. While these studies significantly advanced our knowledge of CAP's molecular activities, little is known about the cellular and physiological functions.
Unlike lower eukaryotes and most invertebrate species that possess a single CAP homolog, vertebrates express two closely related family members, CAP1 and CAP2, with different expression pattern [10,11]. In mice, CAP1 is present in most tissues except skeletal muscles, while CAP2 expression is restricted to a few tissues including skeletal muscle and the heart [10]. These findings led to the assumption that both proteins acquired cell-typespecific functions with CAP2 being the dominant family member in striated muscles [2]. Indeed, recent studies identified CAP2 as a crucial regulator of myofibril differentiation in both skeletal and cardiac muscles [12,13], and they reported a cardiomyopathy associated with dilated ventricles and impaired heart physiology as well as impaired skeletal muscle development for systemic CAP2 knockout (CAP2-KO) mice [12,[14][15][16][17]. Instead, systemic CAP1-KO mice died from unknown causes during embryonic development [18]. Differently from most other tissues, CAP1 and CAP2 are both expressed in the brain [10,11], and recent studies unraveled important functions for both in hippocampal neurons. Specifically, shRNA-mediated knockdown in isolated rat hippocampal neurons revealed a role for CAP2 in the morphology and function of dendritic spines [19], F-actin-enriched dendritic protrusions forming the postsynaptic compartment of most excitatory synapses in the brain [20,21]. Moreover, CAP2 dysregulation has been implicated in synaptic defects of Alzheimer's disease [19]. Instead, hippocampal neurons from brain-specific CAP1-KO mice displayed an altered morphology and function of growth cones [9], dynamic F-actinrich structures at the tip of neurites that sense environmental guidance cues and navigate axons through the developing brain to their target regions [22]. Consequently, neuron connectivity was compromised in CAP1-KO brains [9]. These findings are in line with a model in which CAP1 controls actin-dependent mechanisms during neuron differentiation, while CAP2 is relevant for actin regulation in differentiated neurons. However, the function of CAP2 during differentiation of hippocampal neurons has not been studied to date.
In the present study, we found CAP2 expressed during differentiation and abundant in growth cones from hippocampal neurons. We therefore hypothesized that CAP2 plays a crucial role in growth cones, similar to CAP1. Unlike CAP1-KO neurons, neuron differentiation and growth cone morphology were unchanged in neurons from CAP2-KO mice, and CAP2-KO brains did not display obvious brain developmental defects. However, overexpression of CAP2 rescued growth cone size and morphology in CAP1-KO neurons and normalized their differentiation. Our data revealed functional redundancy of CAP1 and CAP2 in neurons, and they suggest compensatory mechanisms in single KO mice.

Transgenic Mice
CAP2 −/− mice were generated by breeding heterozygous CAP2 mice (CAP2 +/− ) obtained from the European Conditional Mouse Mutagenesis Program (EUCOMM). Generation of conditional CAP1 mice has been reported before [9]. CAP1-deficient hippocampal neurons were obtained from conditional CAP1 mice (CAP1 flx/flx ) additionally expressing a Cre transgene under control of the nestin promoter [23]. Mice were housed in the animal facility of Marburg University on 12 h dark-light cycles with food and water available ad libitum. Treatment of mice was in accordance with the German law for conducting animal experiments and followed the guidelines for the care and use of laboratory animals of the U.S. National Institutes of Health. Sacrificing of mice was approved by internal animal welfare authorities at Marburg University (AK-5-2014-Rust, AK-6-2014-Rust, AK-12-2020-Rust), breeding of brain-specific CAP1 mutant mice was approved by the RP Giessen (G22-2016).

Growth Cone Morphology
Growth cone morphology was assessed by determining growth cone circularity (growth cone area divided by growth cone perimeter) and solidity (growth cone area divided by hull area), similar to previous studies [9].

Live Cell Imaging
For life cell imaging, neurons were seeded in a 22 mm glass-bottom dish coated with PLL as described above and cultured for 1 d. DIC imaging was done in a chamber maintained at 37 • C and the medium was exchanged with CO 2 -saturated HBS solution. Images were acquired every 5 s for 10 min at a Leica DMi8 setup.

Histology and Immunohistochemistry
Nissl staining and immunohistochemical staining was performed as described previously [28,29]. Briefly, E18.5 mice were killed by decapitation and brains were fixed for 2 h in PBS containing 4% PFA. Thereafter, 25 µm transversal brain sections were generated by a Leica CM3050 S cryostat. For immunohistochemistry, brain sections were incubated for 1 h with 2% BSA, 3% goat serum, 10% donkey serum, and 0.5% NP40 in PBS and stained overnight at 4 • C with following primary antibody in 2% BSA and 0.5% NP40 in PBS: rabbit anti-neurofilament 200 (1:80, Sigma-Aldrich, St. Louis, MO, USA). As secondary antibody, AlexaFluor488 anti-rabbit IgG was used. For Nissl staining, brain sections were incubated in staining solution according to the manufacturer's instructions. Image acquisition was done with a Leica TCS SP5 II confocal microscope setup and Leica M80 equipped with a Leica DFC295 camera (Leica Microsystems, Wetzlar, Germany).

Statistical Analysis
Statistical tests were done in R or SigmaPlot. For comparing mean values between groups, Student's t-test was performed. Analyzing the rescue conditions and protein expression over time and in different brain areas, ANOVA with post-hoc test (pairwise t-test with correction for multiple testing) was used. Stage distribution was tested for differences with χ 2 -test.

CAP2 Is Not Relevant for Early Neuron Differentiation
To test whether CAP2 was relevant for neuron differentiation, we analyzed hippocampal neurons isolated from CAP2-KO mice [12,30]. Immunoblots confirmed efficient CAP2 inactivation and revealed unchanged CAP1 expression levels in brain lysates from CAP2-KO mice (Figure 2A; CAP1: CTR: 1.00 ± 0.40, CAP2-KO: 0.86 ± 0.29, n = 3, p = 0.79; CAP2: CTR: 1.00 ± 0.20, CAP2-KO: 0.05 ± 0.01, n = 3, p < 0.05). To test whether CAP2 was relevant for neuron differentiation, we isolated hippocampal neurons from E18.5 CAP2-KO and compared them to neurons isolated from CAP2 +/+ littermates that served as controls (CTR). First, we stained neurons at various time points after plating with an antibody against the neurite marker doublecortin (Dcx, Figure 2B). This allowed us to categorize neurons according to their differentiation stage, similar to previous studies [31]. After five hours in vitro (HIV5), we found the majority of CTR and CAP2-KO neurons in stage 1, i.e., they formed lamellipodia, but not yet neurites ( Figure 2C; (%) CTR: 70.74 ± 2.93, CAP2-KO: 72.61 ± 2.46, n > 300 neurons from three mice). All other neurons possessed minor neurites, but not yet an axon and, hence, were assigned to stage 2 ((%) CTR: 29 We further exploited Dcx-stained neurons to determine their morphology by counting the numbers of primary neurites and neurite endpoints and by calculating the ratio of neurite endpoints and primary neurites (endpoint/neurite ratio) that we used as a readout for neurite branching. None of these parameters were changed in stage 2 CAP2-KO neurons at DIV1 or DIV2 ( Figure 2D  The abundance of GFP-CAP2 in growth cones forced us to test whether CAP2 was relevant for growth cone size or morphology. To do so, we determined growth cone size in phalloidin-stained stage 2 neurons at HIV5 to DIV2. None of the investigated time

CAP2 Is Dispensable for Brain Development
The data we presented so far revealed CAP2 expression in differentiating neurons and abundance in growth cones, similar to CAP1. However, while we found CAP1 to be relevant for neuron differentiation, growth cone morphology, and growth cone motility [9], none of these processes was impaired in CAP2-KO neurons. Impaired differentiation of CAP1-KO neurons was associated with hypomorphic fiber tracks in brains from CAP1-KO mice and a somewhat altered hippocampus morphology [9]. Such changes were not present in brains from CAP2-KO mice. Specifically, Nissl-stained transversal brains sections revealed no obvious differences in cerebral cortex or hippocampus anatomy between CTR and CAP2-KO mice at E18.5 or in adult mice ( Figure 4A,B). Moreover, antibody staining against the axon marker neurofilament suggested a normal appearance of fiber tracks in CAP2-KO brains ( Figure 4C), different from CAP1-KO mice [9]. Together, CAP2-KO mice did not display any gross defects in brain development.

CAP2 Is Dispensable for Brain Development
The data we presented so far revealed CAP2 expression in differentiating neurons and abundance in growth cones, similar to CAP1. However, while we found CAP1 to be relevant for neuron differentiation, growth cone morphology, and growth cone motility [9], none of these processes was impaired in CAP2-KO neurons. Impaired differentiation of CAP1-KO neurons was associated with hypomorphic fiber tracks in brains from CAP1-KO mice and a somewhat altered hippocampus morphology [9]. Such changes were not present in brains from CAP2-KO mice. Specifically, Nissl-stained transversal brains sections revealed no obvious differences in cerebral cortex or hippocampus anatomy between CTR and CAP2-KO mice at E18.5 or in adult mice ( Figure 4A,B). Moreover, antibody staining against the axon marker neurofilament suggested a normal appearance of fiber tracks in CAP2-KO brains ( Figure 4C), different from CAP1-KO mice [9]. Together, CAP2-KO mice did not display any gross defects in brain development.

CAP2 Can Rescue Neuron Morphology and Differentiation in CAP1-KO Neurons
From our data and our previous analysis in CAP1-KO neurons we concluded that CAP2 is dispensable for neuron differentiation and brain development and that CAP1 is the limiting factor in differentiating neurons [9]. Absence of any defects in CAP2-KO neurons could be explained by functional redundancy of CAP1 and CAP2 during neuron differentiation. To determine whether both CAPs share redundant functions, we tested whether overexpression of CAP2 can rescue morphological changes in CAP1-deficient neurons. Hippocampal neurons deficient for CAP1 were isolated from brain-specific CAP1-KO mice that we achieved by crossing a conditional CAP1 strain and Nestin-Cre transgenic mice (CAP1 flx/flx,Nestin-Cre ) [9,23]. Immunoblots confirmed absence of CAP1 from hippocampal lysates obtained from E18.5 CAP1 flx/flx,Nestin-Cre mice (termed CAP1-KO),

Discussion
The present study aimed at deciphering the role for CAP2 in early neuron differentiation, which we found expressed in the embryonic and perinatal brain and abundant in growth cones from hippocampal neurons, similar to its close homolog CAP1 [9,11,32,33]. While we recently reported important CAP1 functions in neuron differentiation, growth cone morphology, and neuron connectivity in the mouse brain [9], neuron differentiation and growth cone morphology was normal in CAP2-KO neurons, and CAP2-KO mice did not display obvious defects in brain development or neuron connectivity. Interestingly, CAP2 overexpression rescued morphological changes in isolated hippocampal neurons from CAP1-KO mice, suggesting that CAP1 and CAP2 share redundant functions in differentiating neurons.
Normal stage distribution and morphology of isolated hippocampal CAP2-KO neurons together with normal growth cone size and morphology and absence of any obvious histological changes in CAP2-KO brains revealed that CAP2 was dispensable for neuron differentiation and brain development. Instead, a previous study unraveled important functions for CAP2 in differentiated neurons [19]. Specifically, this study implicated CAP2 in regulating the morphology of dendritic spines, the F-actin-enriched postsynaptic compartments of most excitatory synapses in the brain. Moreover, it revealed an interaction of CAP2 with the actin-depolymerizing protein cofilin1, a key regulator of synaptic actin dynamics, spine morphology, synaptic plasticity, brain function, and behavior [21,[34][35][36][37][38][39][40]. Further, this study implicated the interaction of CAP2 with cofilin1 in structural plasticity of excitatory synapses, and it provided evidence that CAP2-cofilin1 interaction is compromised in Alzheimer's disease patients, which may contribute to the underlying disease mechanism [19]. Spine morphological changes upon CAP2 inactivation were also reported in a second study, which unfortunately did not provide detailed mechanistic insights [41]. Together, these studies and our findings in CAP2-KO neurons and brains suggest that CAP2 acquires important functions in differentiated neurons, but is dispensable during neuron differentiation.
Unlike most other cell types, hippocampal neurons express both CAP family members, and our group recently demonstrated important functions for CAP1 in regulating the actin cytoskeleton during neuron differentiation [9]. Specifically, we showed that CAP1 controls organization and dynamics of F-actin in growth cones. F-actin defects impaired growth cone function in CAP1-KO neurons and retarded their differentiation, which likely caused compromised neuron connectivity in CAP1-KO mouse brains [9]. Interestingly, we found a cooperation of CAP1 and cofilin1 during neuron differentiation, and our data suggested functional interdependency for both actin regulators in growth cones. Hence, CAP1 was relevant for cofilin1-dependent actin dynamics in growth cones of differentiating neurons, while CAP2 controls cofilin1 in dendritic spines from differentiated neurons.
Absence of any defects in differentiating CAP2-KO neurons led us suggest that CAP1 is the dominant family member during neuron differentiation and brain development. However, CAP2 overexpression rescued neuron differentiation as well as neurite width and total neurite length in CAP1-KO neurons, and it partially rescued growth cone size. These data revealed that CAP2 can compensate CAP1 inactivation in isolated neurons during differentiation, suggesting that CAP1 and CAP2 share overlapping and redundant