Axonal Gradient of Arachidonic Acid-containing Phosphatidylcholine and Its Dependence on Actin Dynamics*

Background: The neuronal distribution of arachidonic acid-containing phosphatidylcholine (AA-PC) remains unknown. Results: AA-PC axonal intensity showed a proximal-to-distal gradient, which was disrupted by actin inhibitors. Conclusion: AA-PC occupies a higher portion of PC at distal than at proximal axons and may be associated with actin dynamics. Significance: This research provides a better understanding of the neuronal spatial composition of PC. Phosphatidylcholine (PC) is the most abundant component of lipid bilayers and exists in various molecular forms, through combinations of two acylated fatty acids. Arachidonic acid (AA)-containing PC (AA-PC) can be a source of AA, which is a crucial mediator of synaptic transmission and intracellular signaling. However, the distribution of AA-PC within neurons has not been indicated. In the present study, we used imaging mass spectrometry to characterize the distribution of PC species in cultured neurons of superior cervical ganglia. Intriguingly, PC species exhibited a unique distribution that was dependent on the acyl chains at the sn-2 position. In particular, we found that AA-PC is enriched within the axon and is distributed across a proximal-to-distal gradient. Inhibitors of actin dynamics (cytochalasin D and phallacidin) disrupted this gradient. This is the first report of the gradual distribution of AA-PC along the axon and its association with actin dynamics.

Arachidonic acid (AA) 5 mediates several crucial functions such as neuronal firing (1,2), signaling (3), and long term potentiation (4). Phosphatidylcholine (PC) is a class of lipids and a major component of most intracellular membranes. It is composed of a choline head group, a phosphoglycerol backbone, and two acyl chains of various combinations, which generate various PC species. PC is synthesized by a de novo pathway (Kennedy pathway) and maintained by a remodeling pathway (Lands cycle) (5). Some intracellular bilayers include PC containing AA (AA-PC) (6), and actin dynamics have been known to mediate bilayer movement (7)(8)(9). However, the distribution of AA-PC within the neuron and its relationship with actin dynamics are not well known.
There have been attempts to visualize lipids. Recently, imaging mass spectrometry has been applied to lipid research, providing an opportunity to visualize endogenous lipids by their specific molecular weights. At a tissue level, we visualized intact PC distributions by matrix-assisted laser/desorption ionization (MALDI) with minimal fragmentation (10). However, the PC distribution has not been characterized by MALDI at an intracellular level. An elegant MALDI experiment measured signaling peptides in rat pituitary cells (11,12). In addition, time-offlight secondary-ion mass spectrometry (TOF-SIMS) has been used to characterize cell surface PC fragments at a single-cell level, revealing low distribution of the PC head group at the membrane fusion site of Tetrahymena (13) and high distribution of the PC head group at the nuclear membrane in blood cells (14). Although TOF-SIMS provides valuable information about the PC head group, it is difficult to distinguish intact PC species because of high fragmentation.
To visualize intrinsic neuronal PC at an intracellular level, we employed MALDI-imaging, wherein PC molecules can be detected with less fragmentation. Measurements were performed on explant cultures of mouse superior cervical ganglia (SCG). We found that PC showed a characteristic distribution depending on the acyl chains at the sn-2 position. Specifically, AA-PC exhibited a gradient of increasing intensity along the proximodistal axonal axis. Intriguingly, this characteristic distribution of AA-PC was disrupted by the inhibition of actin dynamics, indicating that the AA-PC distribution is dependent on actin dynamics.

EXPERIMENTAL PROCEDURES
Cultures-Procedures for the use and care of animals were compliant with the guidelines set forth by the Institutional Animal Care and Use Committee at Hamamatsu University School of Medicine. Cultures were performed and grown as reported previously (15), with modifications for performing MALDI-imaging (16). Briefly, SCG dissected from P0 ICR mice, following careful removal of other cell types, were cut into halves, each of which contained ϳ15,000 neuronal cell bodies (17). The SCG were placed on indium-tin-oxide glass slides (Bruker Daltonics, Leipzig, Germany), to which flexiperms were attached. The slides were coated with 1 mg/ml poly-L-lysine, followed by 10 g/ml laminin. The cells were grown in fresh feeding medium under the following conditions: 10% FBS /minimum essential medium supplemented with 50 ng/ml nerve growth factor; 15 M fluorodeoxyuridine, and 15 M uridine at 37°C in a humidified atmosphere of 5% CO 2 /95% air. On the following day, the anti-mitotic drug aphidicolin (6 M) was added to eliminate non-neuronal cells. Neurons were grown in the feeding medium for 4 -5 days prior to each experiment.
Immunostaining and Labeling with Calcein-Acetoxymethyl Ester (AM)-Cultured SCG neurons were fixed with 4% paraformaldehyde diluted in phosphate-buffered saline (PBS) for 20 min at room temperature, washed with 1ϫ PBS, and permeabilized with 0.1% Triton X-100/PBS for 10 min. The cells were then blocked with blocking reagent (5% goat/1% BSA/0.1% NaN 3 /PBS) for 1 h and incubated with the primary antibody diluted in blocking reagent overnight at 4°C. After washing, the cells were incubated with Alexa Fluor 488-or 568-conjugated anti-mouse IgG antibody (Molecular Probes, Eugene, OR) diluted with blocking reagent at room temperature for 1 h, then washed with 1ϫ PBS. For nuclear staining, TOTO-3 (Molecular Probes) was added at the washing step after the secondary antibody. For actin staining, the same procedures were used, but the cells were fixed at 37°C and incubated with 6.6 M Alexa Fluor 488 phalloidin (Molecular Probes) for 1 h at 37°C. The following mouse monoclonal antibodies were used: anti-Tau-1 (Chemicon, Temecula, CA), MAP2, and DM1A for ␣-tubulin staining (Sigma-Aldrich). Observations and image creations were carried out with a confocal microscope (FluoView1000; Olympus, Tokyo, Japan). For calcein-AM labeling (Dojindo, Kumamoto, Japan), SCG neurons were incubated with 500 nM calcein-AM diluted in medium for 1 h at 37°C in the dark. Observations and image creations were carried out with the Aqua Cosmos image acquisition and analysis system for microscopy (Hamamatsu Photonics, Hamamatsu, Japan).
Inhibitor Treatment-SCG neurons collected at 4 -5 days in vitro (DIV) were treated for 2 h with the following inhibitors or vehicles using the same amount of organic solvent that was used for inhibitor dissolution: 78.8 M cytochalasin D, 25 M phallacidin, 33.2 M nocodazole, 1 M cytosolic phospholipase (cPLA) 2␣ inhibitor, 10 M bromoenol lactone, or 10 M blebbistatin. Cytochalasin D, phallacidin, and nocodazole were purchased from Sigma-Aldrich; cPLA2␣ inhibitor and blebbistatin were purchased from Calbiochem. Bromoenol lactone was purchased from Cayman (Ann Arbor, MI).
Sample Preparation-For MALDI-imaging, sample preparation was performed as described previously (16). Briefly, cultures were washed rapidly and then immediately frozen on dry ice. After drying in a vacuum chamber for 3 h, 2,5-dihydroxybenzoic acid (Bruker Daltonics) matrix solution (50 mg/ml diluted in 70% methanol, 20 mM potassium acetate, and ultrapure water) was applied to the sample. For sample preparation of MALDI-mass spectrometry (MS) and liquid chromatography (LC)-electrospray ionization (ESI)-tandem MS (MS/MS), the freeze-dried cells were subjected to laser-capture microdissection (Leica Laser Microdissection LMD6000, Leica Microsystems, Inc., Germany) to divide the neurons into cell bodies and axonal parts. Samples were laser-detached from indiumtin-oxide slide glasses, harvested into tubes, and dissolved in methanol.
Measurement-For MALDI-imaging and MALDI-MS/MS, measurements were performed using a MALDI TOF/TOFtype instrument (Ultraflex 2 TOF/TOF; Bruker Daltonics) as described previously (16), with minor modifications. The instrument was equipped with a 355-nm Nd:YAG laser. Data were acquired in the positive reflectron mode under an accelerating potential of 25 kV by using an external calibration method. Signals between m/z 700 and m/z 1000 were collected. The raster size was 30 m. Image reconstruction was performed using FlexImaging 2.0 software (Bruker Daltonics). LC-ESI-MS/MS measurements were performed using a 4000Q-TRAP quadrupole linear ion trap hybrid mass spectrometer (Applied Biosystems/MDS Sciex, Concord, ON, Canada) with ACQUITY Ultra Performance Liquid Chromatography® (Waters, Milford, MA). The samples were resolved on an ACQUITY UPLCTM BEH C18 column (1.0 ϫ 150-mm inner diameter, 0.17-m particle) and subjected directly to ESI-MS/MS analysis (10).
Image Analyses-The imaging MS data were opened in Flex-Imaging 2.0 software. Each data point was normalized by total ion current to eliminate variation of ionization efficiency. To present data by PC component and to minimize noise, the data were loaded into Analysis Software for Shimadzu Imaging Mass Spectrometry (Shimadzu, Kyoto, Japan) and normalized by the sum of the ion currents of the major PC species. To analyze axonal parts, the data were further processed in Tissue view (Applied Biosystems). We divided the entire axonal region into five subregions and used four axonal regions for analysis. In each region, we averaged the MS signal intensities of each PC from every pixel and plotted the average signal values. Analytical details are described in Fig. 3.

AA-PC Levels Were Higher in Axon Than in Cell Body-
Throughout the study, we utilized explant cultures of SCG neurons to analyze and compare distinctive intracellular PC distributions within the cell body and axons. This was facilitated by the highly polarized and simple morphology of SCG explant cultures, consisting of a cell body mass (TOTO-3-positive) and axons (Tau1-positive and MAP2-negative) (Fig. 1A). There are many different PC molecular species with various combinations of two acyl chains (Fig. 1B). Initially, to investigate the overall PC molecular species composition in SCG neurons, we directly analyzed the neuronal culture by MALDI-MS and MS/MS. We detected ions of several PC molecular species and identified them by MS/MS (Table 1) by detecting the trimethylamine head group, i.e. choline (m/z 59) and/or phosphate (m/z 124) in the MS/MS product ion mass spectra (10).
Next, we investigated whether the composition of PC molecular species differed between the cell body and axon. Each com-position of PC molecular species was quantified by LC-ESI-MS/ MS, which provides precise quantitative information (18). The observed signal "spots" in the maps ( Fig. 2A) represent distinct molecular species of PC and sphingomyelin lipid classes; a precursor ion scan of m/z 184, used as an MS/MS method in the present study, specifically detected ions that contained a trimethylamine group in their structure, i.e. a head group of PC and sphingomyelin (19). Interestingly, we found that the inten- Interestingly, LCPUFA-PCs (dark gray field) accounted for a higher percentage at the axon than at the cell body/juxta, whereas LA-PCs (light gray field) accounted for a higher percentage at the cell body/juxta than at the axon. In particular, AA-PC accounted for the highest percentage among LCPUFA-PCs at the cell body/juxta as well as the axon. Moreover, AA-PC showed higher localization at the axon than at the cell body/juxta. Our data demonstrate that compared with SA/MUSA/LA-PCs, AA-PC is highly localized in the axon rather than in the cell body.
Increasing Intensity of AA-PC along Proximodistal Axis of Axon-Having demonstrated higher localization of AA-PC in the axon than in the cell body, we investigated whether AA-PC is distributed heterogeneously along the axon and, if so, where it localizes. MALDI-imaging clearly demonstrated the heterogeneous distribution of PC species within the axon (Fig. 3A). The most conspicuous feature of this PC distribution was an exponential-like increase of AA-PC along the proximodistal axis of the axon (Fig. 3A, left). Therefore, it could be said that AA-PC tends to localize in the distal part of the axon. In contrast, SA/MUSA/LA-PCs showed decreasing or almost flat distributions along the axonal length (Fig. 3A, middle and right). Because AA-PC showed an exponential-like increase, we performed a logarithmic transformation of the intensity and found a linear correlation between the log AA-PC signal intensity and distance from the cell body (Fig. 3B, left).
To confirm the reproducibility of this finding in several cells, we assessed the linear relationship between PC intensity and axonal distance by the image analyses described under "Experimental Procedures." We validated the linearity of the AA-PC plot for intensity (vertical axis) and axonal distance (horizontal axis) (Fig. 3C, left, R 2 ϭ 0.9). This linearity enabled us to calculate and utilize an axonal slope coefficient index, namely "k", as a simple representative for the axonal PC distribution pattern relative to the distance from the cell body; k is positive when PC intensity is increasing along the axon (Fig. 3C,  left), almost zero when the PC intensity does not exhibit a specific tendency along the axon (Fig. 3C, middle), and negative when the PC intensity is decreasing along the axon (Fig.  3C, right). The distributions of the 10 major PC molecular species are summarized in Fig. 3D. In particular, the intensity of AA-PC was higher at the distal axon than the proximal axon. Moreover, AA-PC showed the highest k value among the major PCs. In contrast, SA/MUSA/LA-PCs (of negative or near zero k values) showed a reduction or no specific changes in their intensity from the proximal to the distal axon (Fig. 3D).
Quantitative analyses of metabolites by MS can be performed by LC-ESI-MS/MS with the use of multiple reaction monitoring (18). To validate the PC distribution pattern, we introduced LC-ESI-MS/MS with multiple reaction monitoring and performed quantitative analyses of PC species between the proxi-  mal and distal axon. There was a slight reduction in the total amount of major PC species at the distal axon, which derived from the reduction of axonal numbers (data not shown). Thus, to compare PC amounts per axon, we normalized to the sum of major PC species; intriguingly, the relative amount was consistent with the MALDI-MS data. For example, we present the cases of AA-PC and LA-PC(diacyl-16:0/18:2), which showed the highest and lowest k values (Fig. 3D). This result validated the higher distribution of AA-PC and lower distribution of LA-PC at the distal axon compared with the proximal axon (Fig.  3E). The LC-ESI-MS/MS data suggested that the amount of AA-PC gradually increased as it progresses from the cell body at the single axon level. Taken together, our results indicate that AA-PC was heterogeneously distributed, with increasing intensity, along the axon.
Localization of AA-PC in Distal Axon in Early Stage Culture-To determine when the distribution of AA-PC along the axon is initiated, we prepared SCG cultures of DIV 1 to DIV 5 and used MALDI-imaging to assess the axonal PC distribution. We found that the gradient of AA-PC along the axon was present at DIV 1, although the axonal length was much shorter than it was at DIV 5. This pattern was observed from DIV 1 to DIV 5 (Fig.  4A, left). During the same period, PC(diacyl-16:0/18:0) consistently showed an almost flat distribution pattern (Fig. 4A, mid- A, representative ion distribution images and intensity graphs of major PC species. Briefly, SCG neurons were freeze-dried, sprayed with matrix, and immediately used for MALDI-imaging. In each image, the left side contains the cell body mass, and the right side corresponds to the distal axon. To obtain the graph, the averaged intensity of each row of pixels was numerically converted by Scion image (Scion, Frederick, MD). Each graph shows relative intensity (vertical axis) versus distance from the cell body (horizontal axis). The relative intensity is presented in color, with high and low intensity corresponding to white and black in the color bar. B, linear relationship between the log of the intensity ratio of each region to proximal (proxi) region (vertical axis) and distance (horizontal axis). C, image analyses of axonal intensity. Based on the linear relationship between AA-PC axonal intensity and distance, we derived the axonal slope, k, by dividing the indicated axonal region into four parts. D, graphs of axonal intensity and distance for major PCs. Three different explant cultures at DIV 5 were used for analyses. E, LC-ESI-MS/MS analysis of extracted PC from the proximal and distal parts of the axon. Axons were divided into two parts (inset) and were separately collected and used for LC-ESI-MS/MS. Six explant cultures were pooled for the measurement. dle). The decreasing distribution of PC(diacyl-16:0/18:2) was consistently observed in growing cultures (Fig. 4A, right). These consistent distribution patterns throughout the culture period (Fig. 4B) indicate that the major PC composition of proximal and distal axons is determined during early stage culture and maintained during axonal growth, at least in the culture system used here.
Analyses of the k value from DIV 1 to DIV 5 indicated a gradual reduction in AA-PC axonal slope (Fig. 4C, left). This may suggest that the high abundance of AA-PC at the distal part of the axon was relatively stable. Therefore, increasing axonal length resulted in a gradually decreasing slope. The k values of SA/MUSA/LA-PCs were negative or near zero (Ϫ0.3 to 0.05; Fig. 4C, right, six PCs), indicating that the intensities of SA/MUSA/LA-PCs showed consistently flattening or decreasing patterns along the axonal length. These results demonstrate that the increasing axonal intensity of AA-PC occurred from the early stage of culture (i.e. at DIV 1) and was maintained throughout the culture period.
Inhibition of Actin Dynamics Reduces Axonal Gradient of AA-PC-It is well known that cytoskeletal structures modulate the movement of intracellular membranes. The dynamics of the actin cytoskeleton can become a driving force that mobi-lizes intracellular components (7), influencing the redistribution of membrane compartments. We investigated whether actin dynamics play a role in the distribution of AA-PC. We treated neurons with cytochalasin D, which inhibits the association and dissociation of globular (G)-actin by binding to the barbed end of filamentous (F)-actin (20). Cytochalasin D yielded normal staining of ␣-tubulin but weak staining of F-actin (Fig. 5A, left), suggesting an actin-specific disruption. We confirmed that cytochalasin D did not affect cell viability by observing that cytochalasin D-treated neurons were calcein AM-positive (Fig. 5A, right). We treated SCG neurons with vehicle or cytochalasin D for 2 h and analyzed the results using MALDI-imaging (Fig. 5B). Intriguingly, we found that cytochalasin D reduced the signal intensity of AA-PC, particularly in the distal part of the axon, compared with the vehicle control (Fig. 5B, left). Moreover, the k for AA-PC was significantly reduced following cytochalasin D treatment (k ϭ 0.05 Ϯ 0.002, average Ϯ S.E.) compared with the vehicle control (k ϭ 0.1 Ϯ 0.001, Fig. 5C).
To investigate further the effect of actin dynamics on AA-PC axonal distribution, we used phallacidin, a cell-permeable analog of phalloidin that inhibits F-actin transition to G-actin by preventing depolymerization (21), with negligible cytotoxicity (22). We first confirmed that phallacidin treatment did not affect cell viability (Fig. 5D) and then performed MALDI-imaging. Intriguingly, the distribution of AA-PC in the phallacidin-treated neurons showed a dramatic reduction in axonal intensity (Fig. 5E), consistent with the results of cytochalasin D treatment. Moreover, the k value for AA-PC was significantly   0.002, Fig. 5F). The consistency between the results of phallacidin and cytochalasin D treatments indicates that the observed axonal slope reduction of AA-PC is caused by inhibition of actin dynamics.
Effects of Other Inhibitors on Axonal Distribution of AA-PC-Having demonstrated that inhibition of actin dynamics significantly influences axonal localization of AA-PC, we investigated whether the disruption of another major cytoskeleton, i.e. microtubule, has a similar effect (Fig. 6A, B). We used nocodazole, which disrupts microtubules, inducing axonal swelling (Fig. 6A). However, it did not produce significant changes in the axonal slope for AA-PC (Fig. 6B), indicating that the axonal slope of AA-PC is not significantly contributed by microtubule. The significant changes of AA-PC axonal slope in neurons treated with actin inhibitors raise a possibility that a myosin motor walking on actin contributes to the distribution. However, this possibility was also dismissed by myosin motor blockage using blebbistatin; AA-PC axonal slope was not changed after blebbistatin treatment (Fig. 6C).
To understand the formation of AA-PC axonal slope, we treated neurons with several inhibitors that block specific pathways of AA-PC metabolism. Once PC is synthesized, it enters into a remodeling pathway in which phospholipase plays roles to make LCPUFA-PCs. Phospholipase might affect on axonal distribution of AA-PC. To check this hypothesis, we treated cells with cPLA2␣ inhibitor or calcium-independent phospholipase (iPLA) 2 inhibitor (bromoenol lactone) (23). However, FIGURE 6. Inhibitor treatment for other possible factors that may affect the axonal slope did not change the axonal slope of AA-PC. A, immunostaining of ␣-tubulin in neurons treated with medium, vehicle, and nocodazole (10 g/ml). Scale bar, 10 m. B, blue and red colors representing the coefficient k from vehicle-and nocodazole-treated neurons, respectively. Three different explant cultures were used for each treatment. C, summary of inhibitor treatments. Upper, changes in relative axonal intensity of AA-PC by inhibitor treatment. Treatments for actin dynamic inhibitors, i.e. cytochalasin D (Student's t test, ** p Ͻ 0.01, n ϭ 3) and phallacidin (Student's t test, * p Ͻ 0.05, ** p Ͻ 0.01, n ϭ 3) decreased the slope, as did sodium azide. Lower, changes in relative axonal intensity of SA-PC. Each experiment was performed at least twice. they did not induce significant changes in the distribution of AA-PC on axon (Fig. 6C), suggesting that AA-PC axonal distribution is not significantly correlated with the phospholipase activity. The axonal gradient of AA-PC might be also derived from different intake ratio of the medium depending on the axon region. To investigate this, we replaced the culture medium with PBS and measured AA-PC axonal intensity. This hypothesis was also discarded because of no significant changes on AA-PC axonal intensity by PBS replacement.
To confirm whether the AA-PC axonal slope is derived from healthy neurons and not from dead ones, we treated neurons with sodium azide and compared them with normal neurons. The results showed that the AA-PC axonal slope was not present in neurons treated with sodium azide, indicating that AA-PC localization is ATP-dependent and not present in dead cells. Throughout the experiments, PC(diacyl-16:0/18:0) consistently showed k values that were near zero, indicating no significant changes between the vehicle and the indicated reagent (Fig. 6C, lower). In summary, the increasing axonal intensity of AA-PC along the axon was specifically influenced by inhibition of actin dynamics, but not by disruption of myosin, microtubules, PC metabolism, or medium depletion. deacylation is mediated by more than 19 different isoforms of PLA2, which cleaves fatty acids from the sn-2 position of phospholipids. They can be divided into three major groups: cPLA2, iPLA2, and secretory PLA2 (36). cPLA2 exhibits high selectivity for AA at the sn-2 position of the glycerol backbone (37), and iPLA2 is considered a major source for the release of DHA (23) but is also reported as a source for AA (38). PLA2 activities were observed in SCG, and some of them depend on Ca 2ϩ , but others remain active in the absence of Ca 2ϩ , indicating that both cPLA2 and iPLA2 exist in SCG (39). Although PLA2 activities exist in SCG, the possibility of their contribution to the axonal gradient of AA-PC was dismissed by the unaltered axonal gradient of AA-PC after treatment of cells with inhibitors of cPLA2 or iPLA2 (Fig. 6). We did not test soluble PLA2 because its mRNA is not constitutively expressed; it only is induced by certain conditions in the peripheral nervous system and has no fatty acid preference (40), so it does not specifically affect the metabolism of AA-PC. Thus, it seems that reacylation/deacylation is not the main contributor to the AA-PC axonal gradient in our culture system.
Microtubule Transport-Another mechanism that may drive the axonal gradient of AA-PC is microtubule or actin transport. Microtubules are the major longitudinal cytoskeletal filament in axons. Kinesins and dyneins move along microtubules in vesicle transport (41). If the enzymes that mediate AA-PC or its metabolism are transported on microtubules, the axonal gradient of AA-PC would be altered by inhibition of microtubule transport. However, the gradient was unchanged by nocodazole treatment (Fig. 6), indicating that the axonal gradient of AA-PC is not derived from microtubule-dependent transport.
Transport by Actin via Myosin-Actin transport can be mediated by myosin walking along an actin filament or by actin dynamics. The former hypothesis was dismissed by the unaltered axonal gradient of AA-PC on neurons treated with blebbistatin, a cell-permeable inhibitor of ATPase activity of the A and B isoforms of non-muscle myosin II (42) (Fig. 6C). Although there are other types of myosin such as class I, V, and VI, myosin II is suggested to be the most crucial at the growth cone (43).
Transport by Actin Dynamics-Finally, we found out that the axonal gradient of AA-PC was decreased significantly by inhibition of actin filament polarization and depolarization (Fig. 5). Thus, it seems that actin dynamics maintain the axonal gradient of AA-PC. Although it is remained to be investigated how the inhibition of actin dynamics changed the axonal gradient of AA-PC, we can speculate on its mechanism; actin dynamics may restrict the distribution of AA-PC because actin polymerization (7,44) and depolymerization (45,46) could affect the intracellular motility of vesicles containing AA-PC. The turnover of actin filaments is an energy-consuming process that requires ATP. Thus, we also tested sodium azide, which inhibits ATP synthesis (47). Sodium azide treatment reduced the axonal gradient of AA-PC (Fig. 6), indicating that ATP is required for maintaining the gradient.
In our previous study, AA-PC was localized more to the hippocampal area than to other parts of the rodent brain (10). This study revealed that the polarity of AA-PC distribution exists within the cells. Our elucidation of the actin-related distribution of AA-PC inside neurons reveals a new aspect of PC localization and regulation in neurons.