, 2004 and Riley, 1981). Thus, synapse elimination seems to be underway just as animals are being born. Because of the large volume being reconstructed, it was possible in some cases to trace the axons back far enough to assess whether the same axon Palbociclib in vitro was innervating

more than one of the three adjacent neuromuscular junctions. Of the 26 terminal axon branches innervating these three junctions, seven were traceable back to branch points where they bifurcated to give rise to innervation to two of the three junctions (Figure 5). In six of the seven cases, the axons innervated comparably sized percentages of each of the junctions (6% versus 10%; 16% versus 10%; 8% versus Selleck Afatinib 17%; 4% versus 10%; 17% versus 14%; 21% versus 16%). In one case, however, we saw that one of the axon branches did not establish a synaptic contact with the neuromuscular junction site but rather terminated in a bulb

just proximal to one of the junctions. The ultrastructural appearance of this axonal bulb suggested, as described above, that it was a retracting axon, i.e., there were nearby local shed axosomes, and it had a smooth shape (Bishop et al., 2004), rather than a growth cone (i.e., it showed no filopodia or lamellopodia). This result suggested that an axon branch was already in the process of retracting in the first postnatal day. Another branch of the same axon innervated 15% of the neuromuscular junction area on an adjacent neuromuscular junction

(axon 7 in Figure 5). Thus, this early stage of branch loss is occurring asynchronously among the branches of one axon. This result lends further support to the conclusion that the initial axon pruning decisions are being made at the level of terminal branches and not more proximally in the axon arbor. Moreover, the fact that most axons are being maintained at a neuromuscular junction while one is being removed supports the idea that beginning at birth, during the earliest stages of synapse elimination, different axons are being sequentially removed from junctions rather than synchronously. The serial reconstructions also provided information about the way multiple until axons coinnervated neuromuscular junctions at birth. Many of these features were different from both adult singly innervated neuromuscular junctions and later-stage multiply innervated junctions. The synaptic contacts of the axons were highly intermixed, showing no evidence of the interaxonal segregation found at later stages of the elimination process (Gan and Lichtman, 1998) (Figure 4A). The branches of the different axons were not only intermixed but also were closely juxtaposed to each other, with their membranes abutting without intervening Schwann cell processes (Figure S1A, boxed region).