Fast axonal transport of protein was examined in regenerating goldfish optic axons after a lesion of either the optic tract or optic nerve, which revealed changes in the original intact optic axon segments or in the newly regenerated axon segments, respectively. In animals killed either 6 or 24 hr after injection of 3H-proline into the eye, labeling of total fast-transported protein in the original axon segments was increased by 2 d after the lesion, reached a peak of nearly 20 X normal at 2 weeks, and then declined to a level somewhat above normal at 12 weeks. When the labeling of individual transported proteins was examined by 2-dimensional gel electrophoresis, it was found that no new labeled proteins appeared during regeneration, but all proteins examined showed an increase in labeling. Among the various proteins, there was great variation in the magnitude and time course of the labeling increase. The largest increase, to nearly 200 X normal with 6 hr labeling, was seen in a protein with a molecular weight of 45 kDa and a pl of about 4.5, resembling a protein that has previously been designated a "growth-associated protein" (GAP-43; Skene and Willard, 1981a). The proteins showing increased labeling included a small fraction of cytoskeletal proteins (alpha-tubulin, beta-tubulin, and actin) that was apparently transported at a much faster rate than is usually expected of these constituents. In the new axon segments, the total protein labeling was increased by 1 week after the lesion, remained elevated at a nearly constant level of about 7 X normal from about 2 to 5 weeks, and then declined to levels somewhat above normal by 12 weeks. The 45 kDa protein again showed the largest increase, and became the single most prominently labeled constituent in the new axons. On the basis of the time course of labeling in both original and new axon segments during regeneration, the fast-transported proteins were tentatively separated into 5 classes that may represent groups of proteins that are coregulated during regeneration. They may conceivably correspond to different functional or structural entities within the neuron.

Growth cones of Aplysia californica neurons were observed with video-enhanced contrast-differential interference contrast (VEC-DIC) microscopy as they turned at a border between poly-L-lysine-treated and untreated glass. Growth cones that turned generally developed 2 distinct active areas of filopodial and veil formation, much in the way of growth cones undergoing branching. Both active areas advanced, but turning of the neurite occurred through the selective resorption of the incipient branches developing on the untreated substrate. Thus, micropruning of developing regions of the growth cone, rather than the asymmetric extension of filopodia or veils, was primarily responsible for directing neurite growth. We present the hypothesis that abrupt turns by growing neurites are mediated by 2 sets of signals, one causing growth cone splitting, and a second set regulating the survival of the separate branches.

The growth cone at the front of a growing neurite often has F-actin- rich structures--digitate filopodia and sheet-like veils and lamellipodia--whose protrusion advances the leading edge. Microtubules and other cytoplasmic constituents later fill the protruded area, transforming it into new neuritic length. Growth can be initiated from an axon by transecting it. We have used video-enhanced contrast- differential interference contrast microscopy to observe the early events following transection of Aplysia axons in culture. Many filopodium-like protrusions (FLPs) grew rapidly (average instantaneous velocity of 1.6 microns/sec) from the sides and end of the axon stump within minutes of transection. Some of these displayed bidirectional transport of swellings, at a rate similar to fast axonal transport. Dihydrocytochalasin B, which blocks actin polymerization, only halved the number of FLPs that formed within 10 min of transection, and actually increased the number of transporting FLPs. Nocodazole, a microtubule-specific drug, also halved the number of FLPs, but none of them displayed transport of swellings. No FLPs formed in the presence of both drugs. In transected axons that had not been exposed to either drug, removal of the plasma membrane revealed fibers in many of the FLPs; immunofluorescence showed these fibers to be microtubules. Thus, a substantial number of the FLPs that form soon after axotomy are microtubule based, rather than actin based, underscoring the potential of microtubules to drive the rapid extension of neuritic precursors.

How is axonal transport in regenerating neurons affected by contact with their synaptic target? We investigated whether removing the target (homotopic) lobe of the goldfish optic tectum altered the incorporation of 3H-proline into fast axonally transported proteins in the regenerating optic nerve. Regeneration was induced either by an optic tract lesion (to reveal the changes in the original axon segment that remained connected to the cell body) or by an optic nerve lesion (to reveal the changes in the newly formed axon segment). Of 26 proteins analyzed by 2-dimensional gel electrophoresis and fluorography, all but one showed increased labeling as a result of tectal lobe ablation. By 2 d after the lesion, significantly increased labeling of some proteins was seen with a 6-hr labeling interval, but not with a 24-hr labeling interval. This is probably indicative of an increased velocity of transport, which may have been a nonspecific consequence of the surgery. Otherwise, tectal lobe removal had relatively little effect until 3 weeks, when there was a transitory increase in labeling of transported proteins in the new axon segments of the tectum-ablated animals. Beginning at 5 weeks, tectal lobe ablation caused considerably higher labeling of many of the proteins in the original axon segments. Because this was seen with both 6-hr and 24-hr labeling intervals, it is probably indicative of increased protein synthesis. The increased synthesis lasted until at least 12 weeks, though some proteins were beginning to show a diminished effect at this time. In the late stages of regeneration (8-12 weeks), there was also increased labeling of proteins in the new axon segments as a result of the absence of the target tectal lobe. This included a disproportionately large increase in the relative contribution of cytoskeletal proteins and of protein 4, which is the goldfish equivalent of the growth-associated protein GAP- 43 (neuromodulin). We conclude that, after the regenerating axons begin to innervate the tectum, the expression of most of the proteins in fast axonal transport is down-regulated by interaction between the axons and their target. However, the changes in expression may be preceded by a modulation of the turnover and/or deposition of proteins in the newly formed axon segment.