All experimental protocols were performed in accordance with the guidelines for animal research of the Physiological Society of Japan and this study was approved by the Institutional Animal Care and Use Committee (Permission number; 26-10-02) and carried out according to the Kobe-University Animal Experimentation Regulations. All biochemical procedures were performed at 0~4℃ unless otherwise described.

In a previous study, we found that NAP-22 fraction, prepared from the detergent-resistant low-density fraction of rat brain, contains several lipid components including gangliosides and the interaction of NAP-22 and calcineurin was regulated by these lipids 3. NAP-22 was also shown to bind and regulate the GTPase activity of dynamin 4. Since the GTPase activity of dynamin was known to be regulated with several acidic membrane lipids such as phosphatidylserine and phosphatidylinositol, the presence of lipid components in a purified dynamin fraction prepared from brain was studied. Figure 1A shows the elution pattern of B-dynamin from the affinity column of amphiphysin II SH3-domain. The obtained dynamin fraction showed fairly high purity as described 7. Extraction of lipid components using chloroform/methanol was attempted and the TLC analysis of the fraction showed the presence of several lipid components (Figure 1B, lane 1). Six lipid bands were observed (band 1~6 from bottom to top) and their mobility was compared with standard lipids. Band 1 and band 2 showed similar mobility to PC and PE, respectively. Band 3 and 4 were similar to cerebroside (cerebroside separates into two bands usually). Although the mobility of band 5 did not correspond to applied standard lipids, the mobility of band 6 (top) was similar to CHL. After TLC development, lipid containing silica gel was scraped and lipids were extracted as above. These lipid fractions were further analyzed with LC-MS/MS and obtained data were processed by LipidSerach®. The presence of PC in band 1, PE in band 2, and cerebroside in band 3/4 was confirmed (Table 1). Table 1 also showed the presence of TG in band 6.

The developing buffer of TLC was changed slightly, for the reduction of methanol content was found to reduce the mobility of lipids and hence was useful to separate high hydrophobic lipids. Under this condition, the band 6 in Figure 1B separated into two band (band 6 and band 7 in Figure 1C) and the mobility of band 7 was the same to that of TG. Since the mobility of band 6 in Figure 1C was identical with that of CHL (Figure 1C, lane 2), its enrichment in band 6 of Figure 1B (lane 1) was confirmed after calculation of the amount of CHL in each band (Figure 1D). Under this condition, band 5 split into two bands (5a and 5b) and their mobility was near that of standard OHCs (Figure 1C). Since the standard annotation list of LipidSerach® did not contain CHL and hydroxycholesterols (OHCs), these lipids were not detected in Table 1.

In order to confirm the presence of OHCs in the lipid fraction, LC-MS analysis was attempted using commercially available three OHCs (A), the lipid fraction (B), and a mixture of the lipid fraction and OHCs (C) (Figure 2). The molecular mass of OHCs is 402.65. In MS analysis, OHCs were detected as m/z = 367.3371 (after the deletion of two H₂O and addition of one H⁺) or m/z = 385.3465 (after the deletion of one H₂O and addition of one H⁺). The elution pattern of these regions was shown (upper lane, m/z = 367.3338-367.3374; lower lane, m/z= 385.3443-385.3481). In Figure 2A, peaks of 7OHC (a), 27OHC (b), and 24OHC (c) were shown with arrows. In Figure 2B, five peaks were detected in the lipid fraction (#1~#5). Since the lipid amount of these peaks were low compared to that of Figure 2A (NL value of Figure 2A; 8.04 x 10⁸ and 9.56 x 10⁷, Figure 2B; 1.54 x 10⁶ and 2.67 x 10⁶), the peaks of OHCs of the lipid fraction were not evident in the mixture (Figure 2C). Figure 3 shows the mass-spectrum at the retention time of each peak. Peaks of #1, #3, and #4 were weak and their retention time and mass-spectrum did not coincide with that of three standard OHCs. On the other hand, the signal of #2 peak was strong and the mass-spectrum and the retention time were coincided with that of 7OHC. Another strong peak, #5, showed similar mass-pattern to 27OHC, but the retention time was different from that of 27OHC. In summary, the presence of 7OHC in the dynamin lipid fraction was confirmed and the presence of other hydroxycholesterol species different from 24OHC and 27OHC was suggested.

In order to know the biological role of these lipids, the effect on the dynamin GTPase was studied and the lipid fraction was found to activate of the GTPase activity of dynamin (Figure 4A). Since PC and PE were found in band 1 and band 2, their effect on the GTPase was studied and they showed no effect (Figure 4B). The results of TLC and LC-MS analysis suggested that bands in Figure 1B (lane1) were cerebroside (band 3 and 4), OHCs (band 5), CHL and TG (band 6), respectively. The effect of these lipids on the GTPase activity of dynamin was studied. Interestingly, these lipids showed the activation of the GTPase.

In this study, we showed the presence of several lipid species in the dynamin fraction prepared from rat brain. Since no lipid was recovered in the binding fraction of amphiphysin II-SH3 domain after the heat treatment of the dynamin fraction in boiling water for 5 min and centrifugation at 20,000 xg, for 30min, at 4℃ to remove the aggregated protein (data not shown), these lipids were considered to be associated with dynamin. Since the dynamin fraction prepared using a Ca²⁺-dependent binding to the membrane fraction and through an affinity purification is of fairy high purity (Figure 1A), some of these lipids were considered to have dynamin-binding property. Activation of the GTPase activity of dynamin by CHL, TG, OHCs, and cerebroside confirmed their binding to dynamin.

Identification of lipid species was done using TLC and LC-MS/MS, and the presence of PC, PE, cerebroside, and TG was shown. In order to screen the presence of some OHCs, three OHCs were selected for the internal standard of LC-MS. 24OHC is the major brain CHL metabolites produced exclusively in neurons and eliminates CHL from neurons 910. Formation of lipid droplet-like structures and induction of neuronal cell death through esterification of 24OHC is reported 11. 27OHC is a circulatory breakdown product of CHL that is transported into the brain 12. 7OHC is a possible biomarker of cellular lipid peroxidation 13. In this study, we identified the presence of 7OHC in the dynamin lipid fraction. Although the presence of another hydroxycholesterol was detected, its identity is unknown at present. Activation of the GTPase activity of dynamin by these OHCs showed their direct interaction to dynamin.

Although the molecular components of this lipid fraction have not been fully identified, absence of sphingomyelin and phosphatidylserine in the fraction (Figure 1B) suggests that the lipid fraction does not derived from fragmented cell membranes. Presence of TG and CHL in the fraction suggests that the lipid droplet could be one of the source 1415. In adipocytes, dynamin- and protein kinase C dependent cholesterol-induced trafficking of caveolins to lipid droplets has been reported 16. Participation of dynamin II for the lipid metabolism in the neuromuscular junctions on muscle cells was reported 17. Screening of marker proteins of lipid droplet could be a good clue to identify the nature of this lipid fraction 18. Presence of OHCs in the fraction is another interesting point because the importance of CHL metabolism in brain is well recognized 19 and various functions of oxysterols have been elucidated 20.

Recent studies disclosed another role of dynamin in the actin bundling through co-working with cortactin 2122. The binding of dynamin I to septin filament and dynamin II to girdin are also reported 2324. The presence of multiple dynamin related proteins and their functions are also paid much attention 25. The interaction of dynamin with various lipids could be important in these processes.

The purity of a protein fraction is usually analyzed with SDS-PAGE. In this case, we showed the presence of several lipid components in a purified dynamin fraction from rat brain. Multiple regulations of protein functions through post translational modifications are well recognized. Studies on the participation of lipid components in protein functions will be interesting issues.