Changes in rapidly transported proteins associated with development of abnormal projections in the diencephalon

265

Brain Research, 586 (1992) 265-272 © 1992 Elsevier Science Publishers B.V. All rights reserved 0006-8993/92/$05.00 BRES 17931

Changes in rapidly transported proteins associated with development of abnormal projections in the diencephalon K e n n e t h L. M o y a ~, Larry I. B e n o w i t z b, B e r n h a r d A. Sabel a,! a n d G e r a l d E. S c h n e i d e r

a

" Department of Brain and Cognitive Sciences, Massachusetts Institute of Technology, Cambridge, MA (USA) and ~'Depanmolt of &4rgery and Program in Neuroscience, Han,ard Medical School and Department of Surgery, Children's Hospital, Boston, MA (USA) (Accepted 3 March 1992)

Key words: Growth-associated protein; Visual system; Axonal transport; Abnormal connection; Developmental neurobiology

The development of the hamster visual system is accompanied by striking changes in the pattern of proteins that are synthesized in retinal ganglion cells and rapidly transported to their nerve terminals. To determine whether any of these protein changes are regulated by interactions between the developing nerve endings and the cells with which they form synapses, we induced retinofugai axons to form abnormal pr.ojections in the lateral posterior (LP) nucleus of the thalamus and dense patches of hyperinnervation in the lateral geniculate nucleus (LGN) by removing their principal target, the superior colliculus (SC), the day after birth. Under these experimental conditions, two rapidly transported proteins, including the neural cell adhesion molecule, NCAM, showed significant changes in their time course of expression. NCAM, identified here using a monospecific antibody, is normally synthesized and transported at high levels at early stages of development and then declines during the second and third postnatal weeks. However, this decline was delayed when optic fibers were re-routed. A second rapidly transported protein, M r - 67 kDa, pl ,= 4,7, normally shows a rise in its synthesis and transport during terminal arbor formation and a subsequent decline, but it also remained elevated for a prolonged period when the SC was absent. These findings cannot be accounted for by a simple delay in the retinal ganglion cells' program of axonal L~rowtll, since other rapidly traqsported proteins, including the growth-associated protein GAP-43, showed a norm'.l developmental timecourse when the SC was removed. Target interactions therefore appear to influence the retinal ganglion cells' expression of different proteins in a specific fashion,

INTRODUCTION One approach to understanding the development of neuronal connections is to relate the events that take place during the course of ontogeny to the pattern of molecules expressed in specific neuronal populations. The hamster retinofugal projection is particularly well-suited for examining such relationships in vivo in view of the extensive information available on the developmental sequence of anatomical events and the accessibility of this system for molecular analysis. The morphogenesis of retinal ganglion cell fibers is marked by an initial rapid elongation of axons from the eye to the brain, where they invade target structures, elaborate terminal arbors, and finally establish and refine their synaptic relationships with postsynaptic neurons ,~,2t,23,24,40,43.This developmental progression is accompanied by striking changes in the pattern of pro-

teins synthesized in the retinal ganglion cells and transported to neuronal endings 2~. Our previous results suggested that some of these molecular species may be particularly involved in early axon outgrowth, while others appear to be related to terminal arborization and the formation or maintenance of mature connections. Some of the most marked changes in neuronal membrane proteins occur when retinal axons contact their targets, an observation that raises the question of whether interactions with target cells influence these changes. This question can be addressed in the hamster retinofugal pathway in rive by manipulating the growth and target selection of retinal ganglion cell axons. T h e n o r m a l d e v e l o p m e n t a l s e q u e n c e can be a l t e r e d by lesions early in the p o s t n a t a l period that d e s t r o y the u p p e r layers of the s u p e r i o r colliculus (SC) a n d t h e r e b y e l i m i n a t e a significant p o r t i o n of the n o r m a l retinal

Correspondence: K.L. Moya. Present address: INSERM U334, Service Hospitalier Fr~d6ric Joliot--C.E.A., 4, Place du G~n~ral Leclerc, 91401 Orsay cedex, France. Fax: (33) (1) 69 86 77 68. ! Present address: Institute of Medical Psychology, University of Munich Medical School, Goethestrasse 31, Miinchen, Germany.

265 axon terminal space, Following such lesions, retinal axons form anomalous connections in the diencephalon, and also regenerate over the damaged area and cross into the medial portions o f the remaining intact SC 42,43,45. W h e n the e x t e n t o f t h e recrossing p a t h w a y is m i n i m i z e d , retinal axons f o r m d e n s e p a t c h e s of h y p e r i n n e r v a t i o n in the l a t e r a l g e n i c u l a t e n u c l e u s ( L G N ) a n d a large a b n o r m a l p r o j e c t i o n to the lateral p o s t e r i o r n u c l e u s (LP), a t h a l a m i c n u c l e u s which normally receives little or no direct visual input in t h e adult r o d e n t =5,3o..~,~(for studies in the rat see ref. 35). In the p r e s e n t studies we e x a m i n e d the c h a n g e s in nerve t e r m i n a l p r o t e i n s w h e n d e v e l o p i n g retinal axons fail to e n c o u n t e r t h e i r n o r m a l t a r g e t c o m p l e m e n t . O u r results show t h a t the d e v e l o p m e n t a l time course o f certain p r o t e i n s is a l t e r e d significantly w h e n the optic fibers are i n d u c e d to form a b n o r m a l c o n n e c t i o n s in the LGN a n d LP. MATERIALS AND METHODS

Earl)' lesions of the SC Syrian hamsters, bred in the laboratory, were :mesthetized on postnatal day I (Pl; day of birth = P0) by hypothermia and subjected to surgery as described previously .m,.w. The scalp was opened and the SC visualized through the as yet uncalcified skull. The lesion was made by briefly placing the head of a heated pin over the right SC, while pin temperature was controlled by the amount of electrical current flowing through an attached resistor device. The scalp wound was sutured closed and pups were allowed to recover under a heat lamp before being returned to the nest, In order to verify the extent of the lesions and the resulting pattern of retinal axon growth, 2-3 hamsters from several different operated litters were anesthetit.ed and injected in the left eye with horseradish peroxidase (HRP, I.I) mg in 2/.tl of 2% DMSO/saline) at Pl5 or in adulthood, Animals were allowed to survive 24 h, then were deeply anesthetized with a mixture of chloral hydrate, magnesium sulfate and pentobarbital (Chloropent, 0,35 ml/100 g body weight, Fort Dodge Laboratories) and perfused with phosphatebuffered 4% paraformaldehyde, or 2% paralormaldehyde plus 0.5% glutaraldehyde, Brains were removed, postfixed 2-3 hours, and cryoprotected in 30% sucrose, Transverse sections were cut frozen at 40 ~.m and reacted to visualize anterogradely transported HRP using the TMB-AHM method ,z, then counterstained with neutral red,

Labeling of proteins transported to the diencephalon The methods for labeling rapidly transported pp.~teins, 2-dimensional gel electrophoresis, and quantitation of protein labeling have been described in detail :'~, Neonate animals (P2 and P5) were anesthetized by hypothermia and injected intraocularly with 100/,tCi of [~~S]-methionine (New England Nuclear, Boston, MA) or Tran ('~bS]-Iabel (['~S]t.-Methionine Labeling Reagent, ICN, ]rvine, CA) dissolved in I pl of phosphate-buffered saline (PBS), PI2, PiT, and adult animals were anesthetized with Chloropent (0,35 ml/100 g body weight) and injected intraocularly with 100 ~.Ci of radiolabeled amino acid dissolved in 2 p,I PBS. Injections were made into the eye contralateral to the lesion (i.e. left eye), in neonatal hamsters, radioactive amino acids that are injected into the eye diffuse into the circulation, readily cross the immature blared-brain barrier, and are incorporated into locally synthesized proteins, resulting in significant background labeling :". We have shown that this problem can be minimized by giving large systemic injections of non-radioactive amino acids which compete for access to the brain with the label that has diffused from the eye za. Therefore, P2 and P5 animals were also

injected i. p. with non=radioactive leucine (1 mg/g body weight; leucine and methionine utilize the same amino acid transport mechanism to cross the blood-brain barrier) to reduce the background. After allowing 4 h for newly synthesized proteins to ,he rapidly transported from the retinal ganglion cells to optic nerve terminals in the diencephalon, animals were given an overdose of anesthetic and an area including the LGN and LP contralateral to the injected eye was dissected out rapidly under visual guidance. One hemisphere from the neonatal animals was also removed for analysis of systemic labeling. Samples were frozen immediately on dry ice and stored individually at - 70°C. Tissue from 3 to 6 animals at each age was pooled and homogenized in ice-cold sucrose buffer (0.32 M sucrose (Schwarz-Mann, Cambridge, MA), 50 mM Trig, pH 7.4 (Sigma)) with l0 strokes in a teflon-glass homogenizer at high speed. Homogenized samples were centrifuged at 100,000x g for 60 min and the pellets were resuspended in 50 mM Tris for protein determination, After lyophilization, 400/~g of protein was solubilized for 30 rain at 30°C in 9,5 M urea (Bio-Rad) containing 2% NP-40 (Nonidet, Particle Data Laboratories, Elmhurst, IL), 6% ampholines in a ratio of 2:2:! (pH 3.5-5.0, pH 5.0-8.0, pH 3.5-10.0: LKB, Gaithersburg, MD), and 5% /3.mercaptoethanol. lsoelectric focusing was carried out for 16 h at 300 V, then 4 h at 400 V. Gels were equilibrated and proteins were then separated in the second dimension on linear gradient (5-15%) polyacrylamide gels .~i fixed in acetic acid/methanol and stained with Coomassie brilliant blue. Gels were prepared for fluorography by impregnation with Au:ofluor (National Diagnostics, Somerville, NJ), dried onto filter paper (Whatman No. I) and exposed to preflashed X-ray film (Kodak X-Omat, AR) for 30-50 days, Three to 4 gels from each time point were quantified as described below,

Quantitation of rapidly transported proteins Templates were prepared from each fluorogram and used as a guide to excise labeled proteins from the gels. Gel pieces were solubili',ed in 6% Protosol in Econofluor (NEN, Boston, MA) and the radioactivity was determined by liquid scintillation counting, Some background labeling was observed in neonatal animals, as evidenced by the presence of labeled actin and #.tubulin in the LGN and LP (see Fig, 2A) and the presence of labeled proteins in fltlorograms of cerebral cortex (not shown), The cortex, which does not receive a direct retinal illput, showed a pattern of systemically labeled proteins similar to the pattern of background labelinB of control tissues in our previous studies "", Thus, in the present experiments, the cortex provided a good index of the systemic labeling in the neonatal brains and tht; measured background radioactivity was quantitatively eliminated as in our previous studies, As before, we observed virtually no systemic labeling at PI2 and later ages, The background-corrected values for the neonatal ages and the directly measured values for the older time points were used in subsequent calculations, Since the time course of virtually all transported proteins changed during development, it was not possible to compare the labeling of a given protein to a particular reference protein, Therefore, the radioactivity in each p:otein was normalized to the total radioactivity transported to the LGN and LP (i,e, the sum of radioactivity in all rnpidly trnnsported proteins o,z a given gel) and the data were expressed as a percentage of the total transported radioacitivity, A more detailed discussion of this procedure has appeared previously z,~,

Identification of the 230 kDa acidic protein as NCAM Whole brains from P3 hamster were homogenized and 400/~g of the particulate fraction were subjected to 2-dimensional gel electrophoresis as above, At the completion of the run, proteins were electrophoretically transferred to nitrocellulose at 250 mA overnight at 4°C using a buffer containing 0.025 M "Iris (Sigma), pH 8.8, 0.192 M 81ycine,0.1% (w/v) SDS and 20% (v/v) methanol. After transfer, Western blots were fixed in methanol/acetic acid for 10 min and allowed to dry. The blots were rinsed in 2 changes of distilled water for I rain, then washed in 50 mM Tris, PH 7.4 (Sigma), 0.15 M NaCI, 0.5% Tween 20 (Sigma) (TBST) for 30 min. Blots were then blocked in 3% BSA (Fraction V, Sigma) in TBST for 60 rain, washed 3 times for l0

267 min and incubated overnight at 4°C with an affinity-purified antibody raised against the neural cell adhesion molecule, NCAM (gift of Dr, U. Rutishauser) at a concentration of 50/zg/ml in TBST containing 1% BSA. After rinsing in 3 changes of TBST, the blots were incubated for 60 min with biotinylated goat anti-rabbit igG (Vector Labs, Burlingame, CA; 1:250 in TBST plus 1% BSA), rinsed again, and placed in avidin-biotin conjugated HRP complex in TBST prepared as directed (Vector Labs). The blots were rinsed once in TBgT, followed by 2 changes in buffer without the detergent. HRP was visualized using 4-chloronapthoi as the chromogen. All steps were carried out with extensive shaking and, except for the overnight incubation, were performed at room temperature. RESULTS

Verification of early lesions The surgical method used here has been shown to give highly reproducible lesions of the SC 3o,39. Fig. 1 shows a case that received an SC lesion on PI and then was injected intraocularly with H R P on PI6 and perfused on PI7 (A and B) to visualize the extent of retinofugai projections. The lesion destroyed the superficial layers of the SC (Fig. IA), and is similar in extent and location to the other cases verified histologically (3 cases from 2 litters verified at PI7 and 2 cases from 2 litters verified as adults). A minor recrossing retinal projection can be observed along the medial edge of the intact SC. Dense patches of retinal hyperinnervation in the LGN and an abnormal projection to the LP can be seen in sections through the thalamus

~.

_.

I.QNd~~'inn,..,., 'gl~lgr~

h

Fig. I. Early lesions of the SC result in the formation of abnormal retinal projections. The right SC was lesioned on PI and HRP was injected into the eye contralateral to the damaged SC on PI6 (A and B) or in adult animals (C and D), and the hamsters were allowed to survive for 24 h. At the level of the SC, densely labeled retinal fibers course near the surface of the damaged area and a small projection can be observed crossing the midline into the intact SC, ipsilateral to the injected eye (A and C). In the diencephalon, dense patches of retinal terminals are seen in the dorsal LGN, and an anomalous retinal projection has formed in the LP, a nucleus which in normal animals receives little if any direct retinal input (B and D). LGNd, dorsal lateral geniculate nucleus; LP, lateral posterior nucleus of the thalamus; SC, superior colliculus. Bar = 100 ~m.

(Fig, 1B). Similarly, a case from a different litter that was allowed to survive more than 4 months shows a small recrossing retinal projection at the level of the SC (C), along with dense patches of retinal projections in the LP and LGN (D). Littermates of these animals were used to obtain the labeling of rapidly transported proteins as described above.

Visualization of rapidly transported proteins Among the proteins that are rapidly transported from the retinal ganglion cell bodies to the nerve terminals in the diencephalon is the growth-associated phosphoprotein GAP-43 (identified in our previous studies 2,~; also known as B-50, F1, pp46 and neuromodulin ~,,25), which is prominently labeled in the LGN of neonatal animals (Fig. 2A; see also ref. 29). Other transported proteins that are clearly evident at this time are indicated at 230 kDa~ 110 kDa, 100 kDa and 27 kDa. In the adult retinofugal pathway, GAP-43 and the 230 kDa species are markedly decreased, while the 27 kDa and the 64 kDa proteins show considerable increases (Fig. 2B). Also labeled prominently in the adult are proteins at 67 kDa and 94 kDa.

Quantitation of proteins rapidly transported in retinal axons Quantitative amdyses reveal that in normal animals, the labeling of various rapidly transported proteins in the diencephalon (Fig. 3, solid lines) is virtually identical to that reported previously in the SC 2,~. These rapidly transported proteins can be divided into three groups: those that show a continuous decline during development (including GAP-43 and a prominent 230 kDa, p l 4.9 protein; Fig. 3, top row), those that increase during maturation (Fig. 3, upper middle row), and those that increase and then decrease by adulthood (Fig. 3, lower middle row). These changes can be compared with some of the major morphogenetic events for retinal axons in the diencephalon (Fig. 3J). One-way analyses of variance revealed significant age-dependent changes for all proteins except for the 94 kDa species (4.6 < F < 19.9, df = 14, 0.025 > P > 0 . 0 0 0 1 ) . . Early lesions which induced the formation of abnormal connections altered the synthesis and transport of several proteins• The 230 kDa, p l 4.9 protein (see below), which normally declines during maturation, showed abnormally high levels at P17 in the animals with altered retinal projections (Fig. 3B). Whereas in normal intact axons, this protein accounts for 3•7% + 0.5 (mean + S.E.M.) of the rapidly transported radioactivity, in animals with early lesions its labeling is 5060% higher, (accounting for 5.8% + 0.2 of the total; P = 0.006, Student's t-test). An acidic 67 kDa (pl 4.7)

268 molecular weight proteins visualized in this study are NCAM, we used a monospecific antibody which was raised against the purified protein, a , d which recognizes all major for".~ nf this molecule 19. On a 2-dimensional Western blot of neona+.:: hamster brain proteins this antibody uniquely recognized a protein migrating with an apparent molecular weight of 230 kDa, pl 4.9 (Fig. 4). This coincides precisely with one of the proteins whose developmental regulation was affected by the lesions.

protein was also significantly affected by the early lesions (Fig. 31). Labeling of this protein normally rises between P2 and P5, remains high into the second postnatal week, and begins to decline aft'r PI2. However, in animals with early lesions, lew.ls of this protein remained about 4-fold higher at PI7 (5.0% + 0.4 for lesioned; 1.2% + 0.7 for control; P - 0.005). A more basic 230 kDa protein, pl 5.1, appeared to be elevated in the early lesion group at P12 and P17 (P = 0.07 at P17). A 64 kDa, pl 5.2, protein appeared somewhat retarded in its rise in P12 and P17 animals with abnormal retinal cc,nnections. However, these last two differences between the experimental and control groups did not reach significance. By adulthood, no protein labeling differences persisted between the experimental and control animals.

DISCUSSION

We examined the extent to which the developmental regulation of rapidly transported membrane proteins in the mammalian optic pathway is influenced by interactions between the advancing nerve endings and the target cells they encounter. Previous studies ~7.2s.2,~.44 demonstrated a striking pattern of changes in this group of proteins during the course of development, some of which occur at the time when optic axons reach their targets, cease elongating and commence

Identification of NCAM The neural-cell adhesion molecule, NCAM, is known to be transported down axons in the rapid phase of axonal transport and to be developmentally regulated ~s..~7. To investigate whether any of the high

B

•

~dt

41

208-

116"

117-

45. !

F

216.4

4,3

Fig. 2. Retinal protein~ ~apidly transported to the diencephalon of normal and early lesioned hamsters. Proteins were labeled by intraocular injections of ['~SS]methi~)nine.After tallowing 4 h for the newly synthesized proteins to be rapidly transported in ganglion cell axons, animals were sacrificed, the diencephalon removed and the radiolabeled particulate proteins analyzed by 2-dimensional gel elcctrophoresis and fluorography. Apparent molecular weight in kiloDaltons is indicated to the left and i~electric point along the bottom. A: fluorogram of rapidly transported proteins in the dience~alon in ,ormal P5 hamstel ~howing the typical pattern o[' n~onatal proteins. At this stage, prominently labeled rapidly transported proteins include GAP-43 and the 230 kDa NCAM. as indicated. Also visualized are actin (A) and jg-tubulin (fiT) which reflect background labeling (see text). B: fluorogram from the adult lesioned group shows that GAP-43 and NCAM are reduced while proteins ~f 64 kDa and 27 kDa are increased. Arrowheads indicate other rapidly transported proteins identified in previous studies-~s.2~ whlch were also quantified here (see text).

269 elaborating synaptic contacts in the LGN and SC. This raised the question of whether changes that occur at this time might be triggered by signals generated by target interactions, and the related question of the identity and functions of the specific proteins involved. To investigate these issues we prevented optic axons from encountering their normal targets in the SC and found that this alters the synthesis and rapid axonal transport of several proteins destined for the nerve terminal membrane. These changes cannot be accounted for by a global shift in the timing of axonal growth since the developmental timecourse of other proteins was not affected. One of the proteins that showed a significant change in its synthesis and axonal transport after our experi-

11"

0

"1

'i,

mental manipulations coincides with the neural cell adhesion molecule, NCAM, by immunologic criteria and by its electrophoretic migration properties. This integral membrane glycoprotein, involved in neuronal adhesion, neurite fasciculation, axon guidance and junctional communication ]l,20,3s,4s, has previously been shown to be developmentally regulated and conveyed down axons in the rapid phase of axonal transport 18.3"~ The high levels of the molecule in hamster optic fibers are associated with the stage of axon elongation when the fibers grow rapidly in compact fascicles 29, consistent with the known role of NCAM in axon fasciculation. During normal development, the decline in NCAM in retinal fibers may contribute to the defasciculation

:,,,o

f

'

.,o

F ---

- -

.

.

N~

mmm~TLL

IVl

POITT~LTL~

i

i~|,i

i

i

I iiI

tD41

|VEil OPIN

t~

&IWH

Fig. 3. Ouantitation of rapidly transported proteins in hamster retinal axons during normakdevelopment and after early lesions. Each data point represents the mean (± S.E.M.) from 3 to 4 gels in different experiments. In some cases error bars are about the size of the symbol (e). Data from normal animals are depicted by solid lines, and broken lines represent data from animals with early SC lesions. Ast,=risks (*) indicate statistically significant differences between groups. Top row: among the proteins which normallyshow a developmental decline in expression are GAP-43, NCAM and a 230 kDa species, pl 5.1 (A-C). NCAM is significanlyelevated at P17 after early lesions, while the more basic 230 kDa protein is slightlyelevated at PI2 and PI7. Upper middle row: proteins which show a general increase during normal development include the 27 kDa protein and a protein of about 64 kDa which is somewhat delayed in its increase after the experimental manipulations(D-F). Lower middle row: proteins which show an increase in levels of labeling followed by a subsequent decrease include a 67 kDa species which is significantly elevated at P17 (G-l). The normal curves include data for P2 animals while the lesioned group does not, due to the low survival rate of newborn animals receiving surgery twice in 2 days. J: schematic summaryof the normal developmentalevents of retinal axons in the diencephalon (based on refs. 9, 12, 23, 24).

270 rat sciatic nerve which undergoes a lesion-dependent decrease (see Fig. 5 in ref. 36). A second protein, 27 kDa, p l 4.8, has been noted in a number of other studies 29.3~,.44,and may be related to the developmentally regulated synaptosomal protein SNAP-25 (25-27 kDa, p l 4.5; ref. 32). Whatever their identity, the 64 kDa and 27 kDa rapidly transported proteins correlate well with the maturation of retinal axons, and may serve as useful markers for the formation cf adult-like terminals in other systems as well. Also among the proteins whose developmental time course was not affected by early SC lesions, was GAP43. Studies in regenerating goldfish optic pathway likewise indicate that if axons are allowed to encounter their general target environment, GAP-43 synthesis and transport decline nearly normally 5,sl (however, see ref. 34). The latter study, while likewise finding little or no effect of tectal removal on GAP-43 during the early stages of regeneration, did report significantly elevated levels of the protein at later stages 34. The down-regulation of this protein, which normally occurs after synaptic relationships form, also can be altered by more extreme manipulations that prevent nerve endings from encountering any target cells at all ~o, or the

of the axons and the subsequent elaboration of terminal arbors in target regions, in the experimental animals, however, the elevated levels of the protein at PI7 may reflect a continued growth of retinal axons in a fasciculated state when they do not encounter their normal target cells. This would be consistent with previous studies which show fasciculated bundles and the more rapid elongation mode of growth at later ages following SC removal 13.42.It is interesting to note that in the goldfish rctinotectal pathway, preventing retinal axons from encountering the optic tectum (i.e. the "i~u.~C,logt:e ef the SC) likewise prolonged the expression of a 200 kDa protein 4, although whether this protein is homologous to NCAM has not been investigated. The acidic 67 kDa protein (pl 4.7), which normally increases during terminal arborization and then decreases, remained significantly elevated in animals with early SC lesions, in biochemical studies, this protein has been shown to comigrate with a phosphorylated and myristoylated protein "~"~'~, however, the precise identity of the protein remains unknown. Two other proteins of interest in these studies include a 64 kDa species which resembles a protein from

306-

tl6-

93,

61-

4S-

6.4

4A

Fig. 4, A monospecific antibody which reacts with all forms of the neural-cell adhesion molecule, NCAM, uniquely recognizes one of the rapidly Iransported, development-,,'Iv regulated proteins, A: a 2-dimensional fluorogram of rapidly transported proteins at !)5 shows the normal pattern, including the 230 kDa, pl ~.9 ~pecies, B: the antibody against NCAM uniquely recognizes a 230 kDa protein on a 2-dimensional Western blot of neonatal hara~ter brain proteins with identical electrophoretic properties as the rapidly transported protein,

271

appropriate target cells in culture 3. In our studies, the hamster SC was lesioned on P I when retinal axons have begun collateralization in the LGN (Fig. 3J; refs. 9, 24), thus leaving collateral endings in close proximity to target cells. The possibility that the expression of GAP-43 is negatively regulated by a factor transported from the nerve endings is supported by the finding that the levels of GAP-43 mRNA in dorsal root ganglion cells increase markedly after the blockade of bi-directional axonai transport in peripheral nerves 8'49. The ~.ature of the mechanisms that could signal local interactions between the growing axon tips and the central targets might include physiological acitivity, which has been shown in other systems to affect several rapidly transported proteins t,7, or external factors which originate in the surrounding cells and ~europil (e.g. NGF; see refs. 16 and 50 for the localization of NGF and NGF receptor in the developing visual system). Another possible mechanism may involve proteins that are transported from the cell bodies to the distal regions, and then turn around and move back to the soma perhaps after posttranslational modification in the terminal region 2,14,26,41,47. In any event, the present study and others 3,5.lt),34,4b provide compelling evidence that various target manipulations can selectively alter the expression of some proteins but not others, and this in turn suggests that there may be a multiplicity of signals governing the pattern of proteins in developing axons. Finally, it is interesting to note that in adult animals no differences persisted, suggesting that the abnormal connections formed in the present study are sufficient for the emergence of a normal mature pattern of nerve terminal proteins. AcknowkodRements.We would like to thank Dr. Urs Rutishauser for the gift of the NCAM antibody. This work was supported by an NSF Minority Graduate Student Fellowship, NIH Grants EY05690, EY00126, and EY02621 and rAssociation Claude Bernard of France.

REFERENCES I Anlonian, E., Perry, G.W. and Grafstein, B., Fast axonally transported proteins in regenerating goldfish optic nerve: effect of abolishing electrophysiological activity with T r x , Brain Res., 400 (1987) 403-408. 2 Aquino, D.A., Bisby, M.A. and Ledeen, R.W., Bidirectional transport of gangliosides, glycoproteins and neutral glycosphingolipids in the sensory neurons of the rat sciatic nerve, Neuro. science, 20 (19871 1023-1029. 3 Baizer, L. and Fishman, M.C., Recognition of specific targets by cultured dorsal root ganglion neurons, J. Neurosci., 7 (19871 2305-231 I. 4 Benowitz, L.i., Target-dependent and target-independent changes in rapid axonal transport during regeneration of the goldfish retinotectal pathway, in J.S. Elam and P. Cancalon (Eds.), ,,Ix. o,al Transport in Neuronal Growth and Regeneration, Plenum, New York, 1984, pp. 145-169. 5 Benowitz, L.I., Yoon, M.G. and Lewis, E.R., Transported proteins in the regenerating optic nerve: regulation by interactions with the optic tectum, Science, 222 (19831 185-188.

6 Benowitz, L.I. and Routtenberg, A., A membrane phosphoprorein associated with neural develonment, axonal regeneration, phospholipid metabolism, and synaptic plasticity, Trends Neurosci., 10 (1987) 527-532. 7 Benowitz, L.I. and Schmidt, J.T., Activity-dependent sharpening of the regenerating retinotectal projection in goldfish: relationship to the expression of growth-associated proteins, Brain Res., 417 (19871 118-126. 8 Benowitz, L.I., O'Brien, C., Perrone-Bizzozero, N.I., Irwin, N. and Woolf, C.J., The expression of GAP-43 mRNA is regulated by an axonally-transported factor, Neurosci. Abstr., 16 (19901338. 9 Bhide, P.G. and Frost, D.O., Stages of growth of hamster retinofugal axons: implications for developing axonal pathways with multiple targets, J. Neurosci., 11 (19911 485-504. 10 Bisby, M.A,, Dependence of GAP-43 (B-50, FI) trans,=c:'t on axonal regeneration in rat dorsal root ganglion neuro.~t.~. ~rahz Res., 458 (1988) 157-161. 11 Buskirk, D.R., Thiery, J.-P., Rutishauser, U. and Edelman, G.M., Antibodies to a neural cell adhesion molecule disrupt histogenesis in cultured chick retinae, Nature, 285 (19801 488-489. 12 Campbell, G., So, K.-F. and Lieberman, A.R., Normal post-natal development of retinogeniculate axons and terminals and identification of inappropriately-located transient synapses, Neuroscience, 13 (19841 743-759. 13 Carman, L.S., Jhaveri, S. and Schneider, G.E., Some growth characteristics of axons in a re-growing optic tract in neonatal hamsters, Neurosci. Abstr., 13 (19871 257. 14 Couraud, J.-Y, and Di Giamberardino, L., Axonal transport of the molecular tbrms of acetylcholinesterase: its reversal at a nerve transection. In: D.G. Weiss (Ed.), Axoplasmic Transport, Springer, Berlin, 1982, pp. 144-152. 15 Crain, B.J. and Hall, W.C., The normal organization of the lateral posterior nucleus of the golden hamster, J. Comp. Neurol., 193 (1980) 350-370. 16 Finn, P.J., Ferguson, i.A., Wilson, P.A., Vahavolos, J. and Rush, R.A., Immunohistochemical evidence for the distribution of nerve growth factor in the embryonic mouse, J. Neurocytol., 16 (19871 639-647. 17 Freeman, J.A., Book, S., Deaton, M., McGuire, B,, Norden, J.J. and Snipes, G.J,, Axonal and glial proteins associated with development and response to injury in the rat and goldfish optic nerve, E,~7~.Bra#l Res., Suppl. 13 (19861 34-47. 18 Garner, J., Watanabe, M. and Rutishauser, U.S., R~q)id itxontli transport of the neural cell adhesion molecule, J. Nem'osci., ~ ( ! 986) 3242-3249. 19 Hall, A, and Rutishauser, U.S., Phylogeny of a neural cell lldhesion molecule, Dec. Biol., II0 (1985)39-46. 20 Hoffman, S., Sorkin, B.C., White, P.C., BrLickenbury, R., Mail. hammer, R., Rutishauser, U.S., Cunningham, B.A. and Edelman, G.M., Chemical characterization of a neural cell adhesion molecule purified from embryonic brain membranes, J. Biol. Ciwm., 257 (1982) 7720-7729. 21 Jhaveri, S, Edwards, M. and Schneider, G.E., Two stages of growth during development of the hamster's optic tract, Anat. Rec., 205 (1983) 225A. 22 Jhaveri, S., Carman, L. and Hahm, J.-o., Visualizing anterogradely transported HRP by use of TMB histochemistry: comparison of the TMB-SNF and TMB-AHM methods, J. Histochem. Cytochem., 36 (19881 103- i(15. 23 Jhaveri, S., Edwards, M.A. and Schneider, G.E., Initial stages of retinofugal axon development in hamster: evidence for two distinct modes of growth, Exp. Brain Res., 87 (19911 371-382. 24 Langdon, R.B. and Frost, D.O., Transient retinal axons collaterals to visual and somatosensory thalamus in neonatal hamsters, J. Comp. Neurol., 310 (19911 200-214. 25 Liu, Y. and Storm, D.R., Dephosphorylation of neuromodulin by calcineurin, J. Biol. Owm., 264 (19891 12800-12804, 26 Martz, D., Garner, J. and Lasek, R.J., Protein changes during anterograde-to-retrograde conversion of axonally transported vescicles, Brain Res., 476 (19891 199-203. 27 Moya, K.L., Axonal Growth: Modev Molecules and Mechanisms., Doctoral dissertation, Department of Brain and Cognitive Sciences, Massachusetts Institute of Technology, 1989.

272 28 Moya, K.L,, Benowitz. L.i., Jhaveri, S. and Schneider, G.E., Enhanced visualization of axonally transported proteins in the immature CNS by suppression of systemic labeling, Dee,. Brmn Res., 32 (19871 183-191, 29 Moya, K.L,, Benowitz, L.I., Jhaveri, S. and Schneider, G.E., Changes in rapidly transported proteins in developing hamster retinofu~al axons. J. Neurosci., 8 (19881 4445-4454. 30 Moya, K,L., iknowitz, L.I. and Schneider, G.E., Abnormal retinal projections alter GAP-43 patterns in the diencephalon, Brain Res., 527 (1990) 259-265. 31 O'Farrell, P.H., High resolution two-dimensional electrophoresis of proteins, J. Biol. Chem., 250 (19751 4007-4021. 32 Oyler, G.A., Polli, J.W., Wilson, M.C. and Billingsley, M.L., Developmental expression of the 25-kDa synaptosomal-associated protein (SNAP-25) in rat brain, Proc. Natl. Acad. Sci. U.S.A., 88 ( 1991) 5247-525 I. 33 Perrone-Bizzozero, N.I., Benowitz. L.I., Apostolides. P.J., Franck, E.R., Finklestein, S.P. and Bizzozero, O.A.. Protein fatty acid acylation in developing cortical neurons, J. Neurochem., 52 (1989) 1149-1155. 34 Perry, G.W., Burraeister, D.W. and Grafstein, B., Effect of target removal on goldfish optic nerve regeneration: analysis of fast axonally transported proteins. J. Neurosci.. 10 (19901 3439-3448. 35 Perry. V.H. and Cowey. A., A sensitive period for ganglion cell degeneration and the formation of aberrant retino-fugal connections fidlowing tectal lesions in rats, Neurosci.. 7 (1982) 583-594. 36 Redshaw. J.D. and Bisby, M.A.. Proteins of fast axonal transport in the regenerating hypoglossal nerve of the rat. Can. J. PhysioL Pharmacol., 62 (19841 1387-1393. 37 Rutishauser, U.S., Differential adhesion through spatial and temporal variations of NCAM, Tr,,nds N,,urosci., 6 (19861 374-378. 38 Rutishauser, U.S., Acheson. A., Hall, A.K., Mann, D.M. and Sunshine. J.. The neural cell adhesion molecule (NCAM) as a regulator of cell.cell interactions, Science, 240 (1988) 53-57. 39 Sabel, B.A. and Schneider, G.E., The principle of "conservation of total axonal arborisations": massive compensatory sprouting in the hamster subcortical visual system after eariy tectal lesions, Exp, Brahl Res., 73 (1988) 5(18=518. 40 Sachs, G.M. and Schneider, G.E., The m~rphology of optic tract axons ttrborizit~g i, the: superior colliculus of the hamster, J, ('crop. Neurol., 230 ( 19841 155- Its7.

41 Schmidt, R.E. and Modert, C.A., Orthograde, retrograde, and turnaround axonal transport of dopamine-/3-hydroxylase: response to axonal injury, J. Neurochem., 43 (1984) 865-870. 42 Schneider, G.E., Early lesions of the superior colliculus: factors affecting the formation of abnormal retinotectal projections, Brain Behat'. Et'ol., 8 (1973) 73-109. 43 Schneider, G.E., Jhaveri, S., Edwards, M. and So, K.-F., Regeneration, re-routing, and redistribution of axons after early lesions: changes with age, and functional impact. In: J.C. Eccles and M.R. Dimitrijevic (Eds.), Recent Achierements in Restorative Neurology, Vol. 1. Upper Motor Function and Dysfunction, Karger, Basel, 1985, pp. 291-310. 44 Skene, J.H.P. and Willard, M., Axonally transported proteins associated with axon growth in rabbit central and peripheral nervous system, J. Cell Biol., 89 (19811 96-103. 45 So, K.-F. and Schneider, G.E., Abnormal recrossing retinotectal projections after early lesions in Syrian hamsters: age-related effects, Brain Res., 147 (19781 277-295. 46 Sonderegger, P., Fishman, M.C., Bokoum, M., Bauer, M.C., Neale, E. and Nelson, P.G,, Axonal proteins of presynaptic neurons during synaptogenesis, Science, 221 (! 983) ! 294-1297. 47 Tedcschi, B.W. and Wilson, D.L., Modification of a rapidly-transported protein in regenerating nerve, J. Neurosci., 3 (1983) 17281734. 48 Thanos, S., Bonhoeffer, F. and Rutishauser, U.S., Fiber-fiber interaction and tectal cues influence the development of the chicken retinotectal projection, Proc. Natl. Acad. ScL U.S.A., 81 (1984) 1906-1910. 49 Woolf, CJ., Reynolds, M.L., Molander, C,, O'Brien, C,, Lindsay, R.M. and Benowitz, L.I,, The growth-associated protein GAP-43 appears in dorsal root ganglion cells and in the dorsal horn of the rat spinal cord following peripheral nerve injury, Neuroscience, 34 (199(I) 465-478. 50 Yah, O. and Johnson, E.M., An immunohistochemical study of the nerve growth factor receptor in developing rats, l, Neur¢~ci,, 8 (1988) 3418-3498. 51 Yoon, M.G., Benowitz, L.I, and Baker, F., Tectal regulation of transported proteins in regenerating optic fibers of goldfish, Bnai, Re,v., 382 (1~86) 339=351,

Note added ht proof Loewy el al. have identified the prominently labeled rapidly transported 27 kDa protein visualized here as the synaptic terminal protein, SNAP-25 (1 NeurmcL, il (1991)3412-34211,