fig3

( A–C) Dual-labelling of epithelial sheaths and all somatosensory neurons. Maximum projections of confocal stacks show ( A) mRFP-labeled sensory neurons (magenta) together with epidermal PLC^δ-PH-GFP signal (green) or ( B) epidermal PLC^δ-PH-GFP signal on its own. ( C) PLC^δ-PH-GFP-positive sheaths are pseudocolored with dendrite-associated sheaths shaded in magenta and sheaths without apposed dendrites shaded in green. ( D, E) Dual-labeling of epithelial sheaths and c3da/c4da sensory neurons. ( D) tdTomato-labeled c3da and c4da neurons (magenta) are shown together with epidermal PLC^δ-PH-GFP signal (green). ( E) Image showing epidermal PLC^δ-PH-GFP signal with c3da-containing sheaths shaded green and c4da-containing sheaths shaded magenta. ( F) Histogram depicting mean and standard deviation values for the portion of the dendrite arbor of different classes of da neurons ensheathed by the epidermis using PLC^δ-PH-GFP or cora immunostaining as a marker for ensheathment. ( G–K) The extent of ensheathment was inversely correlated with trigeminal (TG) axon complexity in zebrafish. Examples of single TG neurons labeled by transient injection of  Tg(isl1[ss]:LEXA-VP16,LEXAop:tdTomato) with low ( G, H) or high ( I, J) branch density. Epidermal sheaths are shaded in magenta in ( H, J). ( K) Scatterplot of axon branches versus percentage of axon length ensheathed from tracings of 12 individual TG neurons. Note the inverse linear regression (blue line)