To ask whether Nak participates in the localization of other endocytic organelles, we examined the colocalization between YFP-Nak and various fluorescently tagged Rab proteins, including Rab5-GFP (early endosome), Rab4-mRFP (fast recycling endosome), Rab11-GFP (recycling endosome Epacadostat nmr to plasma membrane and Golgi apparatus), and ManII-GFP (Golgi outposts)(de Renzis et al., 2002 and Ye et al., 2007). While these Rab and ManII fusion proteins were widely distributed in dendrites as puncta, they only colocalized partially with YFP-Nak (Figures

S5G–S5J). Furthermore, while GFP-Clc puncta were significantly reduced in nak-RNAi da dendrites, the numbers and distributions of these puncta Ceritinib remained unchanged, as shown by the quantification of puncta distribution along the proximodistal axis ( Figure 5I). Thus, Nak specifically regulates the localization of clathrin puncta in dendrites of da neurons. To understand whether CME has a role in Nak dendrite localization, we examined the distribution of YFP-Nak puncta in da neurons deficient in α-adaptin or dynamin. Although YFP-Nak was present at similar levels in the soma, the YFP-Nak-positive puncta in dendrites were reduced in α-Adaptin-RNAi neurons ( Figures 5F and S5F). Similarly, in shits1-expressing da neurons

shifted to 30°C for 17 hr, YFP-Nak puncta were specifically absent in dendrites ( Figures 5G and S5F). The loss of Nak-positive puncta in shits1-expressing da dendrites was reversible, as they reappeared after being shifted back to 22°C for 100 min (arrows in Figure 5H). These results suggest that the dendritic localization of YFP-Nak puncta requires components of CME. To test whether the presence of clathrin puncta in higher-order dendrites regulates local dendrite dynamics, we imaged higher-order dendrites in live early third-instar wild-type and nak mutant larvae. In wild-type class IV da

dendrites labeled by mCD8-GFP, SB-3CT terminal arbors were highly mobile, undergoing extension (red arrows), retraction (yellow arrows), stalling, and turning during the 40 min recording period ( Figure 6A). By tracking the endpoints of terminal branches in successive frames, we found that distances were larger and frequencies were higher when comparing extension to retraction, resulting in a net growth ( Figures 6C–6E). By contrast, terminal branches extended and retracted with similar distance and frequency in nak2 mutants ( Figures 6B–6D). In addition, more branches were stalled in nak2 mutants than in wild-type larvae ( Figure 6D). Thus, only a small net movement of nak2 terminal branches was detected during the 40 min period ( Figure 6E). These results suggest that Nak modulates dynamic behaviors of terminal branches, contributing to dendrite growth.