Restricted Morphological and Behavioral Abnormalities following Ablation of β-Actin in the Brain

The local translation of β-actin is one mechanism proposed to regulate spatially-restricted actin polymerization crucial for nearly all aspects of neuronal development and function. However, the physiological significance of localized β-actin translation in neurons has not yet been demonstrated in vivo. To investigate the role of β-actin in the mammalian central nervous system (CNS), we characterized brain structure and function in a CNS-specific β-actin knock-out mouse (CNS-ActbKO). β-actin was rapidly ablated in the embryonic mouse brain, but total actin levels were maintained through upregulation of other actin isoforms during development. CNS-ActbKO mice exhibited partial perinatal lethality while survivors presented with surprisingly restricted histological abnormalities localized to the hippocampus and cerebellum. These tissue morphology defects correlated with profound hyperactivity as well as cognitive and maternal behavior impairments. Finally, we also identified localized defects in axonal crossing of the corpus callosum in CNS-ActbKO mice. These restricted defects occurred despite the fact that primary neurons lacking β-actin in culture were morphologically normal. Altogether, we identified novel roles for β-actin in promoting complex CNS tissue architecture while also demonstrating that distinct functions for the ubiquitously expressed β-actin are surprisingly restricted in vivo.


Introduction
The localized polymerization of actin has critical functions in neurons where it contributes to almost every stage of neuronal development from migration [1] and growth cone guidance [2] to dendritic spine remodeling and learning [3]. Although the mechanisms regulating localized actin polymerization in neurons remain incompletely understood, current data support two nonmutually exclusive models. The first model relies on the regulated activity of a plethora of actin binding proteins, while the second model focuses on local synthesis of b-actin to mediate localized actin polymerization in neurons. b-actin is one of six actin isoforms expressed in mammalian cells and is the only isoform thought to be locally translated [4][5][6].
The molecular mechanism of the latter model has been extensively characterized in cell culture and is dependent on a 54 nucleotide sequence in the 39 untranslated region of b-actin mRNA called the ''zipcode'' [7]. The zipcode sequence of bactin mRNA is bound by zipcode binding protein-1 (ZBP1) which mediates the transport of b-actin mRNA from the cell body to the growth cone while also inhibiting its translation [8][9][10]. When attractive guidance cues are received at one side of the growth cone, a signaling cascade is initiated leading to the release and local translation of b-actin [9,[11][12][13]. However, growth cones are believed to constitutively express high concentrations of monomeric actin, and a general role for local translation in early growth cone guidance is controversial [14,15]. Furthermore, the significance of this mechanism in vivo has yet to be established. A knock-out of the mouse ZBP1ortholog Imp1 exhibited high perinatal lethality but no reported brain pathology [16], while motor neuron specific b-actin knock-out mice presented with no defects in motor neuron function or axonal regeneration [17].
In order to investigate the role of b-actin in the development and function of the mammalian CNS, we generated a central nervous system specific b-actin knock-out mouse (CNS-ActbKO). b-actin was rapidly and efficiently ablated from the brains of CNS-ActbKO mice while cand a sm -actin isoforms were upregulated to maintain total actin levels during development. However, nearly two thirds of these mice died postnatally. Examination of brains from viable CNS-ActbKO mice revealed specific defects in the morphology of the cerebellum and hippocampus that correlated with hyperactivity, poor cognitive performance in the Morris water maze, and maternal behavior defects. CNS-ActbKO mice also exhibited partial agenesis of the corpus callosum. Cultured primary neurons from CNS-ActbKO embryos exhibited only limited morphological defects however. Together, our study reveals a previously unreported role for b-actin in the morphogenesis of the hippocampus, corpus callosum, and cerebellum, while demonstrating that the local translation of b-actin may have a more restricted role in nervous system development and function than previously thought.

Actin isoform expression in CNS-ActbKO brains
In order to examine the role of b-actin in the CNS, floxed Actb mice [18] were crossed to the transgenic Nestin-Cre line [19,20] to generate CNS-specific b-actin knock-out mice (CNS-ActbKO). Brains from embryonic day (E) 13.5, E18.5, and adult control and CNS-ActbKO mice were analyzed via quantitative Western blot analysis to confirm b-actin ablation ( Figure 1A-B). At E13.5, approximately three days after Cre expression and DNA recombination, b-actin levels in CNS-ActbKO brains were decreased 53.8% and further reduced to 30% of controls at E18.5. Adult CNS-ActbKO mice exhibited a 90.9% reduction of bactin relative to controls, with the remaining b-actin likely due to the presence of non-CNS derived cells ( Figure 1C). Using a panactin antibody that recognizes all actin isoforms, we measured similar levels of total actin in CNS-ActbKO and control embryos throughout development although we did observe a small but statistically significant decrease in total actin expression in adult brains ( Figure 1A-B). The most closely related actin isoform to bactin, c-actin, was upregulated on average 19.5% at all time points analyzed. Surprisingly, the largest increase in actin isoform expression was measured for vascular smooth muscle a-actin (a sm -actin), which was upregulated 35-fold at E18.5 and 3.4-fold in adults.
To examine the localization of actin in CNS-ActbKO brains, we performed immunofluorescence microscopy with actin isoform-specific antibodies on brain sections from E18.5 control and CNS-ActbKO embryos. While b-actin was expressed throughout the hippocampal region in control embryos, signal from CNS-ActbKO sections in the hippocampus was absent ( Figure 1C). The pattern of c-actin staining in CNS-ActbKO sections was indistinguishable from band c-actin staining in control sections ( Figure 1C). a sm -actin was only detectable around vasculature in controls ( Figure 1C arrows), but was surprisingly found in select populations of cells throughout the hippocampus in CNS-ActbKO embryos ( Figure 1C bottom right panel and inset).

Perinatal lethality and gross morphological phenotypes in CNS-ActbKO mice
We next determined whether the increased expression of cand a sm -actin was able to compensate for the ablation of b-actin and maintain the viability of CNS-ActbKO mice. The expected  (6-9 months). Data plotted as mean 6 standard error of the mean with an n$3 per genotype per age point. (C) Immunofluorescence with actin isoform-specific antibodies on saggital sections of the hippocampus at E18.5. b-actin staining was dramatically reduced in CNS-ActbKO brain sections. c-actin staining gave a similar localization pattern to b-actin on co-labeled control sections and was still present in CNS-ActbKO embryos. a sm -actin was restricted to blood vessels in control brains (arrows) but was prominent in select cells in CNS-ActbKO brain sections (see inset, inset scale bar 25 mm). Scale bar 100 mm. * indicates p,0.05, ** indicates p,0.01, *** indicates p,0.001. doi:10.1371/journal.pone.0032970.g001 number of CNS-ActbKO embryos were found immediately prior to birth at E18.5, but nearly two thirds were lost before weaning at postnatal day 21 (P21) due to unknown causes ( Figure 2A). E18.5 CNS-ActbKO embryos were morphologically indistinguishable from controls and of a similar mass (control: 1.09460.040 g, CNS-ActbKO: 1.18860.026 g). The small number of CNS-ActbKO mice that escaped perinatal lethality were on average 32.4% smaller than control littermates and maintained this decreased mass throughout adulthood despite having a caloric intake two times greater than controls ( Figure 2B-C). Surviving CNS-ActbKO mice also exhibited hind-limb clasping or contractures ( Figure 2D) which are generally indicative of neuromuscular dysfunction. Adult CNS-ActbKO mice were otherwise morphologically unremarkable however, and lived normal lifespans. Isolated brains of CNS-ActbKO mice revealed no obvious abnormalities other than a slightly decreased size which was proportional with their reduced body size ( Figure 2E).

Restricted histopathology in CNS-ActbKO brains
Although surviving CNS-ActbKO mice presented with only limited gross phenotypes, examination of CNS-ActbKO brains histologically revealed several morphological abnormalities in select areas of the brain. Midsaggital sections through the adult cerebellum and stained with Cresyl violet revealed abnormal foliation patterns (denoted by asterisks in Figure 3A) that varied from extensive ( Figure 3A-left panel), to the aberrant positioning of only a single folia ( Figure 3A-middle panel). Interestingly, lamination of the cerebellum appeared unperturbed in the absence of b-actin, with no gross derangement of the molecular or granule cell layer.
Examination of the hippocampus in adult CNS-ActbKO mice also revealed prominent abnormalities. In saggital sections of the hippocampus, the dorsal arm of the dentate gyrus was abnormally positioned and approached the CA1 pyramidal neuron layer ( Figure 3B). In addition, the hippocampal fissure (denoted by arrows in Figure 3B) intersected with a portion of the CA1 pyramidal neuron layer correlating spatially with the invasion of the dentate gyrus (delineated by dashed lines in Figure 3B). A dramatic decrease in cell body staining in the CA1 pyramidal neuron layer was also apparent in the region invaded by the dentate gyrus and hippocampal fissure (area between dashed lines). These abnormalities were also apparent in coronal sections where the derangement of the dorsal arm of the dentate gyrus (black arrow in Figure 3C) and decreased staining in the CA1 neuronal layer (red arrow in Figure 3C) can be seen.
CNS-ActbKO mice also presented with partial agenesis of the corpus callosum. While the callosal axons from control and CNS-ActbKO mice completely crossed the midline in rostral sections ( Figure 3D), axons from CNS-ActbKO mice stopped short in more caudal sections (arrows in Figure 3E, right panel) and appeared to assemble into probst bundles indicative of growth cone aggregation due to an inability to traverse a substrate. Interestingly, the agenesis of the corpus callosum appeared to occur at levels that overlay the hippocampus, suggesting a possible interaction between the aberrant morphology of the hippocampus and corpus callosum. No other features of the brain appeared disturbed, including the cortex which was laminated properly into the six cortical layers ( Figure 3F). Thus, abnormal tissue architecture in CNS-ActbKO brains was apparent only in a subset of brain structures.
Behavioral abnormalities in CNS-ActbKO mice CNS-ActbKO mice exhibited no overtly abnormal cage behaviors with one striking exception. No pups born to a CNS-ActbKO female survived longer than one day regardless of genotype ( Figure 4A). This maternal behavior defect correlated with an apparent pup retrieval deficit ( Figure 4B) which may explain the pup mortality. Importantly, the maternal behavior defect observed was not due to the pups themselves, as pups born to CNS-ActbKO mothers could be fostered to control females where they were retrieved and survived ( Figure 4B, right).
In order to determine if the observed cerebellar histological anomalies resulted in behavioral phenotypes, cerebellar function was assessed using the Rotarod apparatus. CNS-ActbKO mice performed similarly to controls however, suggesting cerebellar functions in motor performance are largely intact ( Figure 5A). We also conducted an Open Field activity assay to examine motor function and basal activity levels of CNS-ActbKO mice. Surprisingly, CNS-ActbKO mice exhibited a profound hyperactive phenotype. Representative traces of mouse activity from the first five minutes of a fifteen minute trial are shown in Figure 5B. CNS-ActbKO mice traveled significantly farther and at a significantly greater average velocity than control mice ( Figure 5C-D). The profound hyperactivity may explain the significant increase in caloric demand observed in CNS-ActbKO mice ( Figure 2C) and Morphological abnormalities in the hippocampus as seen in saggital sections. Arrows indicate the position of the hippocampal fissure normally found between the dentate gyrus and CA1 neuronal layer. In CNS-ActbKO mice, the hippocampal fissure was displaced (arrows) which correlated with an abnormal invagination of the dentate gyrus. The region corresponding to this displacement is indicated by dashed lines which also reveals a dramatic decrease in cell body staining in the CA1 neuronal layer. Scale bar 0.5 mm. (C) The abnormal morphology of the hippocampus in CNS-ActbKO mice is further illustrated in coronal sections where the dentate gyrus displacement (black arrows) and decrease in CA1 region staining (red arrows) can also be seen. Scale bar 0.5 mm. (D) Axons of the corpus callosum crossed the midline normally in rostral sections from CNS-ActbKO brains (arrows) but failed to cross the midline in more caudal sections (arrows in E). Scale bars in D-E, 0.5 mm. (F) The lamination of cortical layers 1-6 was preserved in 6 month old control and CNS-ActbKO mice. Scale bar 0.5 mm. doi:10.1371/journal.pone.0032970.g003 ActbKO mothers exhibited a pup retrieval deficit. While control mothers retrieved pups into a nest, pups born to CNS-ActbKO females were scattered throughout the cage. The pup retrieval deficit was independent of the pups however, as pups born to CNS-ActbKO females were retrieved normally by foster control females. doi:10.1371/journal.pone.0032970.g004 the reduced size of these animals throughout their lifetime ( Figure 2B). Interestingly, the hippocampus has also been implicated in the regulation of mouse basal activity, suggesting that the morphological abnormalities in the hippocampus could contribute to the hyperactive phenotype observed [21,22].
To more directly examine hippocampal and cognitive function, CNS-ActbKO and control mice were subjected to a Morris water maze. CNS-ActbKO mice performed similarly to controls with a cued platform, suggesting that vision is not likely impaired in these animals ( Figure 5F). During a nine day hidden platform training series, CNS-ActbKO mice appeared to perform comparably to controls until the last day, at which time CNS-ActbKO mice took significantly longer than controls to find the platform ( Figure 5F). Traces from the probe trial revealed that CNS-ActbKO mice employed an unusual search pattern ( Figure 5E) and spent a significantly smaller percentage of time in the platform quadrant compared to controls ( Figure 5G). This was also corroborated by quantification of distance traveled in the platform quadrant ( Figure 5H) and platform area crosses ( Figure 5I), both of which were significantly decreased in CNS-ActbKO mice. The abnormal performance of CNS-ActbKO mice suggests these mice suffer from cognitive dysfunction with a possible basis in aberrant hippocampal organization.

CNS-ActbKO hippocampal neurons are comparable to controls in culture
While the restricted nature of histological abnormalities in CNS-ActbKO brains was unexpected, agenesis of the corpus callosum is often linked to axon guidance defects [23] where b-actin is believed to have an essential role. In order to investigate whether the corpus callosum agenesis was due to defective neuronal development, we cultured primary hippocampal neurons from control and CNS-ActbKO embryos to examine neuronal morphology. Hippocampal neurons derived from E15.5 CNS-ActbKO embryos and cultured three days in vitro were null for b-actin by immunofluorescence ( Figure 6A). Immunolocalization of band c-actin in control neurons revealed that both cytoplasmic actin isoforms colocalized throughout the neuronal cell body ( Figure 6A) and growth cone ( Figure 6B).
We next analyzed morphological parameters in hippocampal neurons to determine if the absence of b-actin adversely affected neuritogenesis or differentiation in vitro. Quantification of neurons labeled with a bIII-tubulin antibody revealed no significant differences in the length of axon or axon branches, number of axonal branches, number of dendrites or dendritic branches, or length of dendritic branches ( Figure 6C-D). However, we did observe a small but significant decrease in the length of primary dendrites ( Figure 6D). The number of dendritic branches was also reduced in CNS-ActbKO neurons, but did not reach statistical significance ( Figure 6D). Overall, the lack of dramatic effects on neuronal differentiation in vitro suggests b-actin is not essential in neurons for normal morphogenesis.

Discussion
Here we report the role of b-actin in the mammalian brain and identify distinct contributions of b-actin to normal brain structure CNS-ActbKO mice also crossed an area 50% larger than that of the actual platform significantly less times than control mice. n = 4 male mice for each genotype. * indicates p,0.05, ** p,0.01, *** p,0.001. doi:10.1371/journal.pone.0032970.g005 and function. Although b-actin is believed to comprise one-half to two-thirds of the cytoplasmic actin in the mammalian brain [24,25], we showed that total actin levels are maintained during development in CNS-ActbKO embryos due to an upregulation of cand a sm -actin. However, CNS-ActbKO mice presented with significant perinatal lethality, with only 32% surviving to adulthood. Characterization of brain histology in CNS-ActbKO mice revealed dramatic perturbations in the structure of the cerebellum and hippocampus that correlated with behavioral and cognitive abnormalities. Finally, CNS-ActbKO mice exhibited partial agenesis of the corpus callosum while cultured neurons from CNS-ActbKO embryos had largely normal morphology.
The most prominent phenotype we identified in CNS-ActbKO mice was the disrupted architecture of the hippocampus and cerebellum. Interestingly, gross neuronal lamination in the hippocampus, cerebellum, and cortex ( Figure 3) was normal, indicating neuronal migration along radial tracts is not likely bactin dependent. Instead, the morphological abnormalities observed appear to stem primarily from defects in tissue architecture, suggesting impaired regulation of localized cell-cell or cell- extracellular matrix interactions. Alternatively, the outgrowth of cerebellar folia [26] and folding of the developing hippocampus [27,28] are reminiscent of the complex tissue remodeling that occurs during neural tube closure in embryonic development. Hingepoints at developing tissue folds form when select cells undergo apical constriction mediated by localized actin-myosin contraction [29,30]. A longstanding question in the field has been how actin and myosin are localized to the apical domain of hingepoint cells [30,31]. One intriguing hypothesis is that perhaps the local translation of b-actin may facilitate the localized accumulation of actin. Interestingly, Dugina et al. previously reported b-actin was strongly enriched at the cleavage furrow in dividing cells in culture while c-actin was localized ubiquitously along the cortical cytoskeleton, providing precedent for such a role of b-actin in a similar cellular constriction process [32]. The fact that CNS-ActbKO embryos do not present with neural tube closure defects is most likely because formation of hingepoints during neurulation occurs between E8.5 and E11, while Nestin-Cre expression commences at approximately E10.5 [20,29,33]. Pre-existing b-actin mRNA and protein present before Cre expression are thus likely sufficient to facilitate proper neural tube development.
The actin filaments involved in mediating tissue morphogenic movements must be dynamic to facilitate cellular shape changes, but at other times more stable to endure mechanical strain encountered during these movements [34]. Although actin binding proteins are hypothesized to play the primary role in regulating the dynamics and stability of actin filaments, the composition of actin itself within filaments may also contribute. A recent study by Bergeron et al. characterized the biochemical properties of individual actin isoforms and found that b-actin exhibits more rapid polymerization, phosphate release, and depolymerization than c-actin, suggesting bactin may be the more dynamic cytoplasmic actin isoform [35]. Thus, it is plausible that cells may also spatiotemporally regulate the properties of actin filaments by controlling their relative composition of band c-actin, with filaments more rich in b-actin being relatively more dynamic. It also warrants mention however that aactin filaments are reported to be more stable than bor c-actin filaments [36,37]. Given the significant upregulation of a sm -actin in CNS-ActbKO brains ( Figure 1C), the incorporation of the more stable a sm -actin into otherwise more dynamic filaments could also perturb tissue morphogenesis and lead to the abnormal brain architecture seen in CNS-ActbKO mice.
Intriguingly, the altered morphology of specifically the hippocampus may also inform on the mechanism of the partial corpus callosum agenesis in CNS-ActbKO mice. Corpus callosum agenesis is a relatively complex phenomenon with significant contributions derived from neuronal axon guidance as well as glial support populations and axonal growth substrates [23,38]. Pioneer callosal axons begin crossing the midline in most regions of the mouse brain at E15.5 [39]. In regions where the corpus callosum lies above the hippocampus, callosal axons are hypothesized to follow the pre-existing hippocampal commissures derived from hippocampal neuron axons to facilitate midline crossing [23,40]. The dorsal hippocampal commissure, which lies immediately ventral to the corpus callosum in the region depicted in Figure 3E, also does not cross the midline in CNS-ActbKO brains, and thus the later arriving callosal axons may fail to cross because their normal growth substrate is absent. While defects in axon guidance due to the inability to locally translate b-actin may contribute to the axonal defects observed, disruptions in local glial populations or growth substrates caused by ablation of b-actin may also contribute substantially to the phenotypes observed, and perhaps explain the highly restricted nature of the defect.
The data presented here along with our previous report characterizing the role of b-actin in motor neuron function and axonal regeneration [17] further suggests that distinct roles for bactin may be unexpectedly confined to select nervous system cell populations. By using a CNS-wide conditional knock-out approach as described here with a Nestin-Cre line, we were able to identify cell populations where b-actin appears to be particularly important. Although CNS-ActbKO mice exhibit significant perinatal lethality hindering studies in adult mice, future studies could take advantage of more selective Cre lines to help further define the molecular mechanisms of the phenotypes described here and the role of b-actin and other actin isoforms in cytoskeletal regulation in vivo.

Ethics Statement
The experimental protocols in this study were reviewed and approved by the University of Minnesota Institutional Animal Care and Use Committee (IACUC) and approved on August 13 th , 2010 (IACUC Protocol #0907A69551).

Mouse Lines, breeding, and ethics statement
CNS-ActbKO mice were generated by crossing a floxed Actb mouse line [18] to the Nestin-Cre transgenic line (Stock# 003771, Jackson Labs, Bar Harbor, ME) as described previously [17]. Homozygous floxed animals lacking the Cre transgene were used as controls. Because of the scarcity of CNS-ActbKO animals and a maternal behavior defect in CNS-ActbKO females, breeding of Actb flox/+; Nestin-cre animals to Actb flox/flox animals was required. For timed matings, the morning a vaginal plug was found was designated E0.5.

Brain lysate preparation and quantitative Western blot analysis
Control and CNS-ActbKO embryo and adult (6-9 month old) mouse brains were snap frozen and pulverized in liquid nitrogen. Pulverized tissue was solubilized in a 1:10 (m:V) dilution of SDS lysis buffer (1% SDS, 5 mM EGTA, 1 mM Benzamadine, 10 mM Leupeptin, 0.2 mM PMSF) at 100uC for 5 minutes. Lysates were then cleared by centrifugation at 25,0006g for 2 minutes. The resulting supernatant was collected and analyzed with the Bio-Rad D C protein assay kit to determine protein concentration. 25 mg of control and CNS-ActbKO brain lysates were then run side by side and in duplicate on 8-16% SDS-polyacrylamide gels, transferred to PVDF membrane, and blocked for 1 hour in 5% milk. The following primary antibodies were used: C4, a pan-actin antibody that recognizes all actin isoforms (LMAB-C4, Seven Hills Bioreagents, Cincinatti, OH, 1:1000), anti-b-actin antibody AC15 (A1978, Sigma Aldrich, St. Louis, MO, 1:1000), anti-cactin antibody 2-4 ([41], 1:1000) and anti-a sm -actin antibody 1A4 (A5228, Sigma Aldrich, 1:500). All membranes were also incubated with an anti-GAPDH antibody (G9545, Sigma Aldrich, 1:2000) as a loading control. Infrared dye conjugated secondary antibodies were purchased from Li-Cor Biosciences. The infrared signals from duplicate blots were quantitatively measured with the Li-Cor Odyssey Imager system (Li-Cor Biosciences, Lincoln, NE, USA), normalized to the loading control signal and averaged.

Brain isolation, histology and immunofluorescence
Brains for histology were isolated from age matched adult (4-18 month old) control and CNS-ActbKO mice and perfused, postfixed and cryoprotected as described previously [17]. Brains were then blocked in the desired orientation and frozen in OCT (TissueTek, Torrance, CA) on dry ice. Sixteen micron cryosections were cut and stored at 220uC until processing for histology or immunofluorescence. For histology, sections were equilibrated to room temperature and stained with 0.1% Cresyl Violet acetate (C1791, Sigma Aldrich, St. Louis, MO) following standard protocols. Stained sections were mounted with Eukitt (03989, Sigma Aldrich).

Behavioral Assays
Rotarod. An accelerating Rotarod apparatus (Accelerating Model, Ugo Basile Biological Research Apparatus, Varese, Italy) was used to assess motor coordination in gender matched 6 month old control and CNS-ActbKO mice (n = 8 controls and 6 CNS-ActbKO mice). The Rotarod was configured to accelerate from 4-40 RPM in five minutes, and continue at the maximal speed for another five minutes. Mice were tested for four consecutive days with four trials per day and 10 minute rest intervals between each trial.
Open Field Activity Assay. Activity of age matched adult (6-9 month old) male control (n = 4) and CNS-ActbKO (n = 4) mice in a 50650640 cm test cage was videotaped and analyzed using the Topscan System (Clever System). Mice were allowed to explore the cage for 15 minutes while their activity was tracked.
Water Maze. A water maze analysis was conducted in a 1.2 m diameter opaque pool with water maintained at 2460.5uC. The testing room was decorated with posters and other fixtures to serve as spatial cues. Age matched adult (6-9 month old) male control (n = 4) and CNS-ActbKO (n = 4) mice were trained to locate a 10 cm platform submerged 0.5 cm under the water in an arbitrary quadrant during 90 second training trials. On the first day, the platform was raised to the surface of the water and fitted with a 15 cm flag to serve as a cued control for possible visual differences between control and CNS-ActbKO animals. Hidden platform learning sessions were conducted on the following nine days, where mice performed two 90 second trials per day separated by 30 minutes. Mice were started from randomized positions approximately equidistant from the hidden platform. For the probe trial, the platform was removed and mice were started from a novel starting position for a 30 second probe trial. Video data was analyzed with the Topscan Video Analysis System (Clever System).
Food Consumption. Food consumption from adult (4-6 month old) male control and CNS-ActbKO mice (n = 4 for each genotype) was analyzed by weighing the amount of food initially placed in the cage and the amount consumed per 24 hours for three consecutive days. The data from the three days was then averaged.
For all experiments, neurons were cultured for three days in vitro and then fixed in 4% PFA in PBS warmed to 37uC for 15 minutes. The coverslips were then rinsed in PBS, permeabilized for 10 minutes in 0.2% Triton X-100 in PBS, rinsed again in PBS, and blocked with 3% BSA in PBS for 30 minutes. Coverslips to be used for morphological analysis were then stained with a mouse monoclonal anti-bIII tubulin (G712A; Promega, Madison, WI, 1:2000) followed by an anti-mouse Alexa 488 fluor conjugated secondary antibody (Invitrogen). Coverslips were finally washed again in PBS and mounted in SlowFade Gold antifade reagent with DAPI (Invitrogen, S36938). For cultures stained with actin isoform specific antibodies, coverslips were post-fixed in ice cold 100% Methanol at 220uC for 10 minutes, rinsed for 5 minutes in PBS, and then processed as described above. The directly labeled anti-b-actin antibody described above was used along with a rabbit polyclonal anti-c-actin 7577 [41] and an anti-rabbit Alexa 568 conjugated secondary antibody. Images were acquired on an Olympus FluoView FV1000 laser scanning confocal microscope or a Zeiss Axiovert 25 micrscope under identical conditions for control and CNS-ActbKO coverslips.
Stage 3 neurons were characterized for morphological analysis where the axon was specified as the primary neurite at least twice as long as the next longest primary neurite. All measurements were made with the NeuronJ ImageJ plugin [43]. Data for analysis of morphological parameters in control (n = 127 neurons) and CNS-ActbKO neurons (n = 149) was collected from at least two independent cultures. Axonal branches were defined as processes extending at orthogonal angles to the axon, while distal axon segments were identified as the process that remained parallel to the axonal segment proximal to the branch point as described previously [44].

Statistics
All data are presented as mean 6 standard error of the mean and calculated with GraphPad Prism 5 software (GraphPad Software, Inc.). T-tests were conducted to determine statistical significance when only two groups were compared while one-way ANOVA was used for groups of three or more accompanied by a Tukey post hoc test. A p-vale of ,0.05 was considered significant.