Expression of neurofilament proteins during development of the nervous system in the squid Loligo...

THE .JOURNAL OF COMPARATIVE NEUROLOGY 356~311-326 (1995)

Expression of Neurofilament Proteins During Development of the Nervous

System in the Squid LoZigo peaZei

PHILIP GRANT, DANIEL TSENG, ROBERT M. GOULD, HAROLD GAINER, .4ND HAKISH C. PANT

Laboratory of Neurochemistry, National Institute of Neurological Disorders and Stroke, National Institutes of Health, Bethesda, Maryland 20892 (P.G., H.G., H.C.P.); New York

State Institute for Basic Research, Staten Island, New York 10314 (R.M.G.); Marine Biological Laboratory, Woods Hole, Massachusetts 02543 (P.G., D.T., R.M.G., H.G., H.C.P.)

ABSTRACT The squid nervous system includes various brain ganglia, optic lobes (the visual center),

and the stellate ganglia, the system of @ant motor fibers responsible for rapid jet-propelled escape behavior. The large caliber of giant fibers is due, in part, to the accumulation of squid-specific neurofilaments (NFs) made up of a heavily phosphorylated NF 220 protein together with NF 70 and NF 60 subunits. Using antibodies prepared against known peptide sequences in these proteins, together with a mammalian-derived antibody that specifically recognizes phosphorylated squid N F 220, we studied the localization of NFs in adult tissues and during neural development. Immunoblot and immunohistochemical analyses showed that NFs were present in adult neural tissues, primarily in selected fibers, with giant axons showing the most robust expression. After the first neurons differentiated at stage 22, immunoblots showed NF 60- and NF 70-immunoreactive proteins at all stages. The NF 220 subunit, however, was not detected in immunoblots at any developmental stage. Phosphorylated N F 220 immunoreactivity, although absent in immunoblots, was first seen in selected fibers of the stellate ganglia at stage 25, increasing thereafter in all giant fibers until hatching (stage 30). The stellate ganglion is the first neural tissue to acquire a mature neurofilament complement (i.e., phosphorylated NF 2201, shortly before the onset of jet-propelled escape behavior. The temporal pattern of expression of the NFs during development resembled that seen in vertebrates; i.e., the smaller NFs appeared before the larger subunit in most neural tissues. In the squid, the expression pattern seems to depend upon the post-transcriptional regulation of a single gene rather than upon transcriptional regulation of three independent genes as in vertebrates. c 199s Wiiey-Lib>, Inc.*

Indexing terms: cephalopod, immunohistochemistry, stellate ganglion, giant axons

The cephalopods possess one of the most complex of invertebrate nervous systems, with elaborately lobated brains and large vertebrate-like eyes connected to an intri- cately organized visual center, the optic lobe (Young, 1974). In decapodan species such as Loligo pealei, there is also a unique peripheral giant fiber system in the stellate ganglion that controls rapid swimming and escape behavior (Young, 1938, 1939). Although the development of these and other neural systems have been studied both descriptively and experimentally (Korschelt, 1892; Sacarrao, 1956; Martin, 1965; Martin and Rungger, 1966; Meister, 1972; Marthy, 1973, 1987; Meinertzhagen, 1990; Gilly et al., 19911, the cellular patterns of neuronal development and differentiation in central and peripheral ganglia have not been eluci- dated. The two systems that have been examined most are

u 1995 WILEY-LISS, INC. *This article is a US Govern- ment work and, as such, is in the puhlic domain in the United States of America.

the retinaioptic lobe (Meinertzhagen, 1990) and the system of @ant fibers and stellate ganglia that control the jet- propelled escape responses (Young, 1939; Martin, 1965; Gilly et al., 1991).

Jet-propelled escape behavior is mediated centrally from two g a n t coordinating cells connected via fused axons (the first-order “command” neurons), located in the subesopha- geal magnocellular lobe of the brain. Receiving inputs from cerebral, visual, tactile, and other sense modalities, their s o n s (first-order giant axons) make synaptic connections

Accepted October 27, 1994 Address reprint requests to Hansh C. Pant, LNC, NINDS, Bldg. 36, Rm

4D20, Kational Institutes of‘Health, 36 Convent Drive. MSC 4130, Bethesda, MD 20892-4130

312

with small numbers of large second-order neurons in the palliovisceral lobe of the brain. These neurons, in turn, coalesce to form a limited set of giant axons (second-order axons) within the pallial nerves that course to and enter the ipsilateral stellate ganglia in the dorsal mantle, where they make synaptic contact (giant synapse) with the classic squid giant motor axons (i.e., third-order giant fibers). The latter arise from the fusion of many axons of large motor neurons within the giant fiber lobe (GFL) of the stellate ganglion. Together with bundles of smaller axons, they constitute the stellar nerves that radiate from the stellate ganglia to innervate the circular muscles of the mantle. Sudden contraction of these muscles contributes to the escape response.

The rapidity of jet-propelled escape behavior is due to rapid impulse conduction in the large-caliber axons of the giant fiber system. In Loligo vulgaris, third-order axons increase in diameter from 1 pm, shortly after they appear in the embryo, to 8 bm at hatching (Marthy, 1987). They continue to increase during posthatching growth, as GFL neurons enlarge and their axons fuse, to achieve diameters of 300-800 pm, as in adult Loligo pealei (Adelman and Gilbert, 1990). It is important to note that at the time of hatching, the behavioral response is well established (Gilly et al., 1991).

Large axon caliber may be attributed, in part, to the accumulation and assembly of abundant neurofilaments (NFs) in the axon cytoskeleton (Hirokawa, 1991). In squid giant axons, they are the predominant cytoskeletal element in the inner axoplasm, and include one large (220 kD) and two small (60 kD and 70 kD) proteins (Pant et al., 1986). These proteins, particularly the NF 220, are highly phosphorylated post-translationally within the axon by associated protein kinases (Pant et al., 1978; Floyd et al., 1991), a process that is believed to contribute to the stability of large axon caliber. A single squid gene has been cloned which codes for all three NF proteins; these different subunits arise by alternative splicing from the single gene rather than from three independent genes, as in mammals (Szaro et al., 1991; Way et al., 1992).

During development of the vertebrate nervous system, the expression of NF subunit proteins in most, if not all, neurons of the central nervous system (CNS), follows a characteristic temporal order (Carden et al., 1987). As neurons differentiate and axons first appear, the smaller subunits, NF-L and NF-M, are coordinately expressed at low levels. These assemble into relatively immature neurofilaments, presumably to maintain plasticity during axonal growth (Garden et al., 1987). Later, as axons make synaptic contacts with target cells, the large NF-H begins to appear in cell bodies and axons. The triplet neurofilaments can be detected at this time. Subsequently, the NF-M and NF-H subunits become highly phosphorylated, and axons increase in caliber and are myelinated. Phosphorylation of neurofilaments is the hallmark of a mature axonal cytoskeleton.

In the squid, with a single gene coding for all three NF proteins (Way et al., 19921, the question arises whether NF subunit expression during neural development follows a succession pattern similar to that seen in vertebrate development. If so, this would suggest that alternative splicing of a single gene leads to tissue-specific and/or spatio-temporal patterns of NF subunit expression during development. Such patterns may correlate with the relative states of

P. GRANT ET AL.

differentiation, maturation, and function of different neural regions.

Using Western blots and immunohistochemical techniques, we have examined the localization of NFs in the visual (retina and optic lobe) and giant fiber systems (stellate ganglion) since they are important behavioral integration centers that differentiate early during development (Marthy, 1987; Meinertzhagen, 1990; Gilly et al., 1991). Our data suggest that expression of NFs during development of the squid nervous system follows a similar temporal pattern to that seen in vertebrate embryos, probably as a result of post-transcriptional regulation of a single gene.

MATERIALS AND METHODS Adult squid and egg strings were obtained from the

Marine Resources Center of the Marine Biological Labora- tory, Woods Hole, during the summer months 1991/1992. Adult tissues such as optic lobe and stellate ganglion were dissected and fixed immediately. Egg strings were separated into small bunches, tied with string, and suspended in a tank with running seawater to develop. Seawater temperature varied between 19°C and 22°C.

Antibodies A diagram showing the structure of the three neurofila-

ments is seen in Figure 1 (See Szaro et al., 1991; Way et al., 1992). Note that all share identical 5’ N-terminal and rod sequences but differ significantly in the 3’ C-terminal regions, except for a short region shared by the NF 60 and NF 70 proteins. An antibody (NF-NT) was prepared in rabbits to a common N-terminal sequence EISTTT- TYEGESRPSS (AAs 9-24; Immunodynamics). Another antibody (NF 601 70) was prepared against a peptide EAEV- LSTILTRSEGG ( U s 465-479, Peptide Technologies) shared by both the NF 60 and NF 70 proteins. A third antibody (NF 70) was prepared against a peptide KGEDKA- NYTQNTWQ (AAs 601-615, Peptide Technologies) specific for the NF 70 protein. An attempt to make an antibody specific for the N F 220 was unsuccessful. Prebleed sera were also taken from donor rabbits for controls. All antibodies were prepared as 1:500 dilutions in 10% normal goat serum (NGS) in phosphate-buffered saline (PBS), pH 7.4, 0.1% azide for immunohistochemical assays and 1:1,000 in 0.2% Tween 20-Tris-buffered saline, pH 7.5, 0.1% azide (TTBS) for immunoblots.

Two mammalian-derived antibodies were also used, a commercially obtained monoclonal to (1-tubulin at 1: 1,000 dilution in 10% NGSiPBS or in TTBS for immunohistochemical or immunoblot assay, respectively, and Sternberg- ers’ SMI 31 (the monoclonal antibody to phosphorylated mammalian NFH; Sternberger and Sternberger, 1983) at a 1: 1,000 dilution in the same diluants. This antibody (here called P-NFH) was used to detect the phosphorylated form of squid NF 220. The Western blots and immunohistochemical data obtained using this antibody were similar to a monoclonal antibody prepared against the phosphorylated epitope of the squid NF 220 (Cohen et al., 1987).

As secondary antibodies, biotin-labeled goat anti-rabbit (GAR-BT) or anti-mouse (GAM-BT; Kirkegaard and Perry) at a 1:300 dilution in PBS was used for immunohistochemical assay. Secondary antibody for the immunoblot was a 1:2,000 dilution of alkaline phosphatase-conjugated goat

NEUROFILAMENTS IN SQUID DEVELOPMENT 313

NEUROFILAMENT PEPTIDES FOR ANTIBODY PRODUCTION

3’ UT - ROD - 5’ UT __.

N f -60 w-

1-L \ \

‘A? d IEMkAJdAI m i P , r12.:

Fig 1 Structure of squid neurofilamrnt proteins showing the regons from which mdividual peptide sequences were obtained for the production of polyclonal antibodies The symbols (dots, hatching, etc ) represent regons of identical amino acid sequence (for details about gene structure, see Way et a1 , 1992)

anti-rabbit or goat anti-mouse (Kirkegaard and Perry) in Fixation TTBS. Embryos at different stages were mechanically separated

from the jelly and chorions and fixed in one of a variety of fixatives. which included methanol (at -20°C). Bouin’s. Immunoblots

Tissues from adult squid were dissected and homogenized in 0.05% sodium dodecyl sulfate (SDS), 20 mM Tris, pH 7.4, centrifuged at 10,000 rpm for 10 minutes, and the supernatant was pipetted off for gel electrophoresis. Approxi- mately 200-400 embryos at different stages were homogenized in the same buffer, centrifuged, and the superna- tants were collected. Aliquots were taken for determination of protein by the Bradford procedure (1976). The same standardized amounts of protein were applied for gels; i.e., for each curtain gel, 200 kgitissue protein was used, while 10 kg of protein was used for each lane. The higher concentrations of protein failed to resolve the protein bands in the minigels. Nine percent gels were run in all cases with appropriate standards. The proteins were then transferred to polyvinylikdene Auoride (PVDF) membranes, air-dried, and stored. For the immunoblot, 3-mm strips were cut from each membrane, placed in wells and washed in TTBS for 1 hr at room temperature (RT) with shaking. This was followed by incubation in the primary antibody 1:1,000 dilution in TTBS for 1-3 hours at RT or overnight at 4°C. Control strips were incubated in diluted prebleed sera or in 10% normal goat serum. Strips were then washed three times in TTBS for 10 minutes each, whereupon strips were incubated with the secondary antibody for 1 hour at room temperature (RT). The strips were washed three times, 10 minutes each, with TTBS. Color was developed in the alkaline phosphatase substrate reagent BCIPINBT. After appropriate development, the strips were washed with water and dried.

Smith’s, and a special squid fixative prepared as follows: 4% paraformaldehyde, in 0.6 M sucrose, 0.1 M cacodylate acid, 10 mM MgClz, and 1 mM EGTA, pH 7.2. Ten to 50 embryos or whole adult tissues fixed in Bouin’s were washed several times in water to remove picric acid, and stored in 70% alcohol. Embryos fixed in methanol were washed and stored in methanol. Embryos fixed in Smith’s were washed and dehydrated to be stored in 70% alcohol. Paraformaldehyde- fixed embryos were washed and stored in cacodylate fixation buffer. All fixed material was stored at 4°C.

Embedding and sectioning Adult tissues were dehydrated then placed in xylene,

passed through xy1ene:paraffin (Tissuemat, 56°C) mixtures to two changes of paraffin, and embedded. Embryos were dehydrated and then placed in absolute alcoho1:toluene mixtures (3:1,1:1, and 1:3) for 15-30 minutes then through toluene:paraf€in mixtures (3: 1, 1:1, and 1:3, respectively, 15-30 minutes each), and finally two changes of paraffin (30 mins each) before embedding.

Ten-micron serial sections of adult tissues and embryos were cut, mounted on gelatin-coated slides, and stored. Serial sections from an embryo or adult tissue were mounted sequentially on several slides so that all antibodies could be tested on the same tissue under identical conditions.

Immunohistochemical assay Slides were passed through several changes of xylene, 15

minutes each, to remove paraffin and were rehydrated

314

through a series of alcohols, placed in PBS and washed 3 times, 5-10 minutes each. They were then incubated in 10% NGS/PBS/O.2% Triton X-100 for 1 hour at RT, followed by incubation for 30 minutes at RT in 10% NGS/PBS. To each slide approximately 100 pl of diluted primary antibody were added and incubated overnight at 4°C. After removing the primary antibody, slides were washed three times in PBS, 5-10 minutes each and incubated in the appropriate secondary antibody, followed by three washes in PBS. Finally, slides were incubated in ABC reagent (Vector Labs) 1 : l O O in PBS for 1 hour at R T and washed three times in PBS, 5 minutes each. The diamino- benzidine reagent (DAB) was prepared, and the slides were incubated in the reagent for 15-30 minutes until the color reaction was completed. The sections were counterstained with 0.2% methyl green, dehydrated, and mounted in Permount.

P. GRANT ET AL.

RESULTS Expression of neurofilament proteins in axoplasm: Validation of the antibodies

To determine the antigenic specificities of each antibody, immunoblots were prepared from extracts of axoplasm isolated from giant axons. Five different primary antibodies were evaluated, and results are shown in Figure 2 . Each of the antibodies stained proteins of appropriate size. The a-tubulin Ab reacted strongly (lane 4) and a robust reaction was also obtained with the NF 60170 antibody, revealing a band at MW 70 and a broad band at MW 60. The specific antibody to NF 70 also revealed the NF 70 neurofilament epitope. The strong immunoreactive band between NF 60 and NF 70 (lane 2) with the NF 60170 antibody and the smaller reactive bands (< 70 kD) seen in lane 3 with the NF 70 antibody may be due to proteolytic fragments of the NF 70 protein. These antibodies were made to the C-terminal domain of the NF 60 and NF 70 neurofilament proteins. Since the NF 70 antibody also showed strong immunohistochemical reactivity with non-neuronal tissues (Fig. 8C), the smaller fragments stained in lane 3 of Figure 2 could be non-neuronal proteins sharing a similar epitope. Moreover, it is unlikely that these smaller immunoreactive bands seen in lanes 2 and 3 could be due to phospho-isoforms of NF 60 and NF 70 since in previous studies we found no changes in mobility of NF 60 and NF 70 after phosphorylation in vitro (Pant et al., 1982; Cohen et al., 1987). The NF-NT antibody showed a predicted strong reaction with the NF 220 neurofilament protein as well as with the NF 60 and NF 70 antigens. Finally, the axoplasm reacted strongly with SMI- 31, an antibody to mammalian phosphorylated NFH (Stern- berger and Sternberger, 19831, here identified as P-NFH. Two well-defined bands were seen, a strong band at MW 220 and another at a higher molecular weight (HMW), provisionally suggested to be a crosslinked product of NF 220 (Cohen et al., 1987). The pattern of NF expression in axoplasm with the P-NFH was similar to that obtained with a monoclonal antibody specifically prepared against phosphorylated squid NF 220 protein (Cohen et al., 1987). Unfortunately, this antibody was unavailable for the cur- rent study.

Similar immunoblots were obtained with extracts from other neural tissues such as the optic lobe, giant fiber lobe of the stellate ganglion, and fin nerve. All adult neural tissues examined showed the presence of all three NF proteins, although the levels of expression of each NF

Fig. 2. Immunoblots of squid axoplasm from the giant axon showing patterns of neurofilament (NF) protein expression with antibodies prepared as describcd in Figure 1 and the text. Lane 1: NF-N-terminal. Lane 2: NF 60170. Lane 3: N F 70. Lane 4: u-T’ubulin. Lane 5: P-NFH. Molecular weight markers are also shown.

protein varied with the tissue (data not shown). For example, the expression of phosphorylated NF-H was high- est in axoplasm and barely detectable in the optic lobe.

Immunohistochemical localization of neurofilament epitopes in adult tissues: The

stellate ganglion The levels of NF expression in mammalian tissues as

studied with a variety of monoclonal and polyclonal NF antibodies has been found to be quite heterogeneous and to depend, in part, upon the fixatives and tissue processing procedures used (Hickey et al., 1983; Trojanowski et al., 1985). We tested various fixatives for the squid analysis and found that Bouin’s fixative and cold 100% methanol gave the most reproducible, interpretable expression patterns in adult and embryonic tissues, respectively. Immunocyto- chemical localization of NF antigens was seen in sections of the adult stellate ganglion and optic lobe (Figs. 3 ,4 ) .

The differential expression of NF epitopes in the stellate ganglion is shown in Figure 3A-D. Note that the giant axon (g.f.3) and smaller axons within the stellar nerves (s.n.1 robustly expressed the NF 60170 and NF-NT epitopes (Fig. 3A) and displayed an intense expression of the phosphorylated NFH (Fig. 3 0 . Fibers within the stellate ganglion neuropils and those entering the neuropil in the mantle connective (containing second-order giant fibers), as well as the adjacent fin nerve, also revealed heavy staining with

NEUROFILAMENTS IN SQIJID DEVELOPMENT 315

Fig. 3. Immunohistochemistry of the stellate ganglion. Sections of adult stellate ganglion incubated in different primary antibodies. A$: Low-power sections of ganglion and stellar nerve with the principal third-order giant fiber. B,D: High-power sections of the giant fiber lobe region of the ganglion. A,B: Sections incubated in the N-terminal antibody. Similar patterns were obtained with the NF 60170 antibody

these three antibodies (data not shown). By contrast, the NF 70 antibody disclosed virtually no fiber staining in the giant axon and smaller stellar nerve axons, whereas many nonneuronal tissues stained intensely in the adult and embryo (see Fig. 8C). Because of this cross reactivity with nonneurofilament epitopes, the NF 70 antibody was consid- ered inappropriate for immunohistochemistry.

Selective staining was also obtained within cell bodies of the giant fiber lobe (Figs. 3B,D). The large cell bodies stained intensely with the NF 60170 and NF-NT antibodies (Fig. 3B) suggesting the presence of at least one, and possibly all three NF proteins. Staining was particularly strong in the axon hillock region. Smaller neuron cell bodies within the ganglion also expressed these antigens (Fig. 3A). Note that the P-NFH antibody reacted only with a few small perikarya in the ganglion (Fig. 3Cj and the axon hillock region of some of the larger perikarya of the GFL (Fig. 3D). This result is consistent with the suggestion that phosphorylation of the NF 220 protein may begin at the axon hillock region of giant GFL neurons and continue in the axons that fuse to become third-order giant fibers (Cohen et al., 1987; Tytell et al., 1990).

(not illustrated). C,D: Sections showing pattern of P-NFH staining of phosphorylated SF 220. Open arrows in B and D indicate intense rcactivity n f NF-NT antigens and modest expression of P-NFH epitope in the axon hillock region. g.f.3, third-order giant fiber; s.g., stellate ganglion; s.n., stellar nerve.

Immunohistochemistry of the optic lobe The architecture of the optic lobe has been described in

detail by Young (1974) (see Fig. 4A-E). Briefly, optic fibers from the retina enter the lateral surface of the lobe and pass through the outer granule cell layer to penetrate into the underlying laminated plexiform layer. Their terminals synapse with processes from large and small second-order visual cells located in the outer and inner granule cell layers, above and below the plexiform zone. The plexiform zone is organized into transverse neuropil layers. Axons from second-order visual cells, after synapsing in the plexiform laminae, turn 180" and run perpendicularly through these layers penetrating deeply into the lower medulla region. There, they encounter islands of smaller neurons, neuropils, and many transverse giant cell bodies. Deeper in the medulla, numerous large bundles of optic tract fibers assemble before coursing to the cerebral lobes.

The antibody staining pattern in whole lobe sections exhibited a distinct heterogeneity. First, optic fibers from the retina expressed all NF epitopes including the phosphorylated NFH (Fig. 4B,C). The NF 60/70, the N-terminal

Fig. 4. Immunohistochemistry of the optic lobe. Sections of the adult optic lobe treated with different antibodies. A Control section incubated in a prebleed serum from a rabbit producing the N-terminal antibody. B: Section treated with the NF fiOi70 antibody. A similar pattern of expression is obtained with the N-terminal antibody and with anti-a-tubulin (not illustrated). C: Section incubated in the P-NFH antibody. Open arrows show limited reactivity in fibers in plexiform layer and a few radial bundles in the medulla. D: Section of the deep medulla region incubated in the N-terminal antibody showing

some stained large visual cells (solid arrows) and the many positive fiber bundles of the optic tract (open arrows). A similar pattern is obtained with the N F fi0/70 antibody (not illustrated). E: Section of the deep medulla region incubated in the P-NFH antibody showing intense reactivity in the optic tract fiber bundles (open arrows). No cell bodies stain, however. i.gran.c.l., inner granular cell layer; med., medulla; op.fib., optic fibers; o.gran.c.l., outer granular cell layer; plex.l., plexiform layer. Magnification of all as in D.

NEUROFILAMENTS IN SQUID DEVELOPMENT

antibodies, and the antibody to a-tubulin demonstrated similar staining patterns (Fig. 4B). Several fiber layers of the plexiform zone, regions of synaptic neuropil, stained intensely as did most radial bundles of axons descending deeply into the medulla. Most cell bodies in the outer and inner granule cell layers, however, showed no NF expression except for a few giant second-order visual cells in the innermost layer of the inner granule cell layer which stained intensely with the N-terminal antibody (not shown). and less so with the NF 60170 antibody. Differential staining was clearly exhibited by the P-NFH antibody; only the retinal fibers, some fiber layers within the plexiform zone, and a few radial fiber bundles entering the medulla expressed the phosphorylated NF 220 epitope (Fig. 40.

No small perikarya expressed any NF epitopes in the medulla, but many large perikarya of transverse neurons and their axons did stain intensely with the N-terminal antibody (Fig. 4D), and less so with the NF 60170 antibody. Tangles of optic tract fibers were also positive for the N-terminal (Fig. 4D) as well as for the NF 60170 and a-tubulin antibodies (not shown). Finally, only the optic tract fibers within the medulla strongly expressed the phosphorylated NF 220 antigen (Pig. 4E).

317

Immunoblot analysis of the developmental expression of neurofilament epitopes

The staging series for the development of the nervous system was based on studies by Arnold (19651, Meister (1972), and Marthy (1987). At stages 16-17, morphogenesis begins and various ectodermal placodes appear that mark the anlage of each brain region. The retinal anlage appears first, followed soon by the anlage of the optic lobe and cerebral ganglia. By the time the eye vesicle closes at stage 18/19, the anlage of the pedal and basal ganglia of the brain have been established, and shortly thereafter at stage 20, the thickened anlage of the stellate ganglia appear in the future mantle region. About 12 hours later at stage 21, axons and fibers appear in virtually all brain and stellate anlage, suggesting that neurogenesis has begun. The mantle muscle shows slow contractions at this stage, implying some functional connections. In subsequent development (stages 22-26), as the brain and stellate ganglia enlarge, and as new peripheral ganglia arise (brachial and subradu- lar ganglia), major fiber tracts and connectives are laid down, establishing central and peripheral connections. Jet-controlled escape behavior is detectable by stage 26 (Gilly et a]., 19911, and the first giant axon responses can be recorded at stage 27 (Gilly et al., 1991). According to Marthy (1987), by stage 27 the ganglia and tracts of the nervous system are essentially similar to those seen in hatchlings (stage 30). Further development after hatching is characterized by rapid growth of the numerous ganglia with an accumulation of fibers and fiber tracts. It is during this post-hatching period that the giant fibers undergo exponential growth to their adult dimensions.

To determine the time at which various neurofilament antigens are expressed during development, whole embryos at different stages were homogenized and Western blots were prepared &h all (l?ig. 5). Immunoblots of axoplasm from the adult giant axon were included for

lanes (except the axoplasm) contained identical amounts of total protein (20 pg) and were run under identical conditions.

Fig. 5. Developmental expression of neurofilament antigens. Immu- noblots of extracts of whole embryos at different developmental stages incubated with NF antibodies. Stage numbers above each lane are based on Arnold’s developmental serics (1965). Axoplasm NF expression is seen in lane marked “zo.”

318 P. GRANT ET AL.

The NF 60170 epitope was expressed throughout development from stage 22, when the first fibers emerge from the various ganglia, to hatching at stage 30 (Fig. 5). Of the two bands expressing the NF 60 epitope, the upper one that is intensely expressed in axoplasm showed evidence of developmental regulation; it progressively increased in reactivity through development. The NF 70 band seemed to be fully expressed from the earliest stages and was less intensely expressed than NF 60 in axoplasm. This result suggests that both NFs are synthesized in a coordinate fashion from the time axons differentiate in the various ganglia. This observation was supported by the expression pattern of the NF-NT antibody. With this antibody, three bands migrated in the region of the NF 60 epitope throughout development. At stage 30, however, only the upper band persisted, while the lower two bands were missing or greatly reduced. The upper NF 60 band was also strongly expressed in mature axoplasm. Here, too, with this antibody, strong expression of the NF 70 epitope was seen throughout development, similar to its expression pattern with the NF 60i70 antibody. NF 220, which also could react with the N-terminal antibody in axoplasm, was not detected at any embryonic stage of development. It is possible that the level of this antigen was insufficient for detection in immunoblots. Failure to detect NF 220 does not necessarily mean that the NF 220 was not present during development; we shall see later that the phosphorylated NF 220 was detected in embryonic axons by immunohistochemical techniques with the P-NFH antibody (see below). Finally, the specific NF epitope detected by the NF 70 antibody was expressed at each stage and in the axoplasm as a single band similar to its pattern with the NF 60170 and NF-NT antibodies (Fig. 5). Although the gel from which this immunoblot was obtained did not run as well as the others (migration of the NF band in several lanes was distorted), it is included here since it was run under identical conditions. Although the Western blots with this antibody show specific reactivity at the correct molecular weight for the NF 70 epitope, it also displayed extensive immunohistochemical expression in nonneuronal tissues of the embryo. Accordingly, the developmental results with this antibody are not presented in this study, except for one figure (Fig. 8C) to illustrate the overall staining pattern.

As we have already seen, the P-NFH antibody reacted strongly with the phosphorylated NF 220 and with a higher molecular weight crosslinked neurofilament derivative (HMW) in axoplasm. During development, however, there was no evidence of the phosphorylated NFH in immunoblots, probably for the same reason mentioned above; i.e., levels of antigen too low to be detected against a background of total embryonic and yolk protein (Fig. 5). Instead, several low molecular weight protein bands were seen a t each developmental stage. Evidently some non-NF phosphorylated proteins from embryonic tissues share homologous sequences with the phosphorylated NF 220 during development. None of these proteins was detected in mature axoplasm.

Alpha-tubulin immunoreactivity appeared throughout development as a doublet compared with only a single band in the axoplasm (Fig. 5). Because whole embryos were extracted, nonneuronal tubulin was undoubtedly included to account for the additional band.

Immunohistochemical analysis of the developmental expression of NF epitopes in

neuronal tissues In this investigation we focused on the development of

two neuronal systems, the retina/optic lobe and the stellate ganglion, one sensory and the other, motor. It was of interest to compare patterns of neurofilament expression in the cell bodies and axons within these two neural systems which exhibit different rates of morphogenesis. The retina and optic lobe anlage appear early, at stage 18, whereas the stellate ganglia anlage appear later at stage 20.

As noted above, by stage 22, most brain anlage as well as those of the stellate ganglia have appeared and the first fibers have emerged. With the a-tubulin antibody, the first fibers in the stellate ganglia and the optic fibers coursing from the presumptive retina to the optic lobe anlage were readily detected (Fig. 6A). An almost identical pattern of staining was seen with the antibodies to the NF-NT and NF 60170 epitopes.

Although the optic lobe is immature a t stage 22 (it lacks the plexiform layers), optic axons from the retina can be seen entering the lobe region after reacting with the NF 60/70, NF-NT, and a-tubulin antibodies (Fig. 6A,B). Some fibers within the lobe medulla also stain with these antibodies (Fig. 6B). Fibers in other brain anlage also react with these antibodies (data not shown). No cell bodies expressed any XF epitopes, however. The P-NFH antibody also failed to detect any NF epitopes in fibers in the retina or optic lobe at this early stage, but did react with most embryonic nuclei at this stage and in later stages until hatching (stage 30). The antigens responsible for this nuclear staining may be the low MW proteins in the gel (Fig. 5). It suggests that proteins in nuclei share homologous epitopes with phosphorylated NF 220.

Important morphological changes had occurred in the nervous system between stages 22 and 25 (Fig. 6 0 . Not only had most brain ganglia increased in size, they had become more complex as numerous fiber tracts were established between central and peripheral ganglia. The fiber patterns as revealed by NF antibodies recorded this extensive maturation process.

The retina and optic lobe displayed significant morphological development at stage 25; incoming optic fibers and a differentiating plexiform layer reacted strongly with the a-tubulin, NF 601 70, and N-terminal antibodies (Fig. 6D,E). Neurofilament-rich optic fibers entering the lobe were seen descending into and joining with the presumptive plexiform (Fig. 6D). Bundles of stained fibers were also seen within the medulla and optic tract. On the other hand, the P-NFH antibody did not react with any fibers, either in the retina or the optic lobe; if NF 220 was present, it was not phosphorylated. Again large numbers of nuclei in neuronal and nonneuronal tissues were stained intensely with this antibody.

The expression patterns for all antibodies in the optic lobe and retina at stage 27 were similar to those seen at stage 25 except for the presence of more fibers and an enlarging plexiform layer. Even a t this stage, no expression of the P-NFH epitope in any fibers was seen.

The anlage of the stellate ganglia which, a t stage 22, do not contain identifiable giant axons, expressed a strong positive reaction with the NF 60/70, NF-NT, and a-tubulin antibodies (Fig. 7A,B), suggesting the presence of some, if not all neurofilaments in the small

The retina and optic lobe.

The stellate ganglia.

Fig. 6. Immunohistochemical analysis of NF expression in the eye and optic lobe anlage in horizontal sections of stage 22 and 25 embryos. A Low-power horizontal section of a stage 22 embryo incubated in NF 60170 antibody with labeled fibers in stellate ganglion, eye, and optic lobe anlage. B: An adjacent section at higher magnification incubated in the antibody to NF 60170 displaying heavily stained optic fibers (open arrows) entering the optic lobe anlage (01) and fibers of the developing optic tract (black arrows). C: Horizontal section of a stage 25 embryo at low power reacted with the NF-NT antibody showing labeled fibers in

the optic lobe and intense labeling in the stellate ganglion. D: Horizon- tal section of a stage 25 retinaand optic lobe stained with the NF 60/70 antibody exhibiting strong expression in retinal fibers (black arrows) entering the stained plexiform layer (open arrows) of the optic lobe. The NF-NT antibody yields a similar reaction pattern. E: Section incubated in a-tubulin antibody showing many stained retinal fibers entering the plcxiform layer (open arrows) of the lobe and stained fibers in the medulla. ret., retina; tent., tentacle. Scale bar = 100 pm in A bar in E also applies to B and D.

320 P. GRANT ET AL.

axons arising from ganglion cell bodies. Fibers in the mantle connective containing second-order fibers leading to the stellate ganglion were also stained as they entered the ganglion neuropil (data not shown).

At this early stage there was no expression of P-NFH in any region of the nervous system. If NF 220 was present in neural tissues at this stage, these results suggest it was not phosphorylated. This is consistent with the Western blot data (Fig. 5).

With respect to NF expression, maturation of the stellate ganglion was significantly more advanced than the optic lobe at stage 25 (Fig. 7C). The neuropil, the incoming presynaptic second-order fibers within the mantle connectives, and the fibers in the stellar nerves reacted strongly with the NF 60170 and N-terminal antibodies (Fig. 7C). Some cell bodies in the ganglion expressed the NF 60170 and N-terminal epitopes within axon hillock regions and perinuclear cytoplasm.

Significantly, a select group of fiber bundles and axons entering the stellate ganglion and within the neuropil expressed the phosphorylated NF 220 antigen for the first time at stage 25 (Fig. 7D, open arrows). This result implies that within the stellate ganglion, a t least two (NF 60 and NF 2201, if not all NFs, were present including the phosphorylated form of NFH, a sign suggestive of maturation. No other brain ganglia contained P-NFH-positive fibers. Although the future giant axons are present in these ganglia at this stage (Marthy, 19871, they were not distin- guished by size from the axons destined to become the smaller stellar fibers.

The rapidly maturing stellate ganglia at stage 27 exhibited a striking pattern of expression of NF epitopes within neuropils, stellar nerves, adjacent fin nerves, mantle connectives, and interganglionic connectives (Fig. 7E,F). Fibers within and leading to and from the ganglion also expressed the P-NFH epitope in patterns similar to those seen at stage 25 (see Fig. 7D). Likewise, cross sections of the stellar nerves within the mantle were also deeply stained with the NF 60170, NF-NT, and P-NFH antibodies (Fig. 7F), suggesting that all NFs were present, including the phosphorylated NF 220. The third-order giant axons, which are well established and functional at this stage (Gilly et al., 1991), seem to be expressing all NFs.

The hatchling: Stage 30 At hatching (stage 301, a differential pattern of' neuronal

tissue expression of NF epitopes was most striking (Fig. 8A-E). The variable extent and intensity of staining in any section was most likely correlated with abundance of neuropils and fiber bundles in the section. At low magnification, one can easily see tissue-specific expression patterns for each antibody. All neuropils and tracts deep within the brain ganglia, stellate ganglia, and optic lobe stained robustly with the a-tubulin and NF 60170 antibodies (Fig. 8A,E). A more selective population of fiber tracts and neuropils seemed to be labeled by the NF-NT epitope, but this depended on the plane of section (Fig. 8B). In general, the NF 60170 and N-terminal epitopes displayed similar tissue-specific patterns of expression, but with different intensities.

The NF 70 antibody displayed dramatically different patterns from NF 60170 and NF-NT. Only the neuropils and a few fibers within the brain ganglia expressed low levels of the NF 70 epitope in contrast to the strong reactivity in most nonneuronal tissues, particularly mantle

muscle (Fig. 8C). The NF 70 staining pattern seen here is typical of its nonspecific reactivity.

The P-NFH antibody demonstrated a highly restricted pattern of NF expression (Fig. 8D). Compared to the staining with a-tubulin, NF 60170, and NF-NT antibodies, brain neuropils were only lightly reactive with the P-NFH antibody. Several selective tracts within the brain ganglia reacted vigorously with this antibody as did the stellate ganglion fibers and stellar nerves. Few, if any fibers within the optic lobe were stained, however. This suggests that only limited populations of central axons contain the phosphorylated NF 220.

The optic lobe at hatching, stage 30, morphologically resembled the adult pattern; it had developed a plexiform layer with lamina organization whose fibers stained intensely with a-tubulin, NF 60170, and the NF-NT antibodies (Figs. 8, 9A,C). In addition, the numerous islands of neuropil within the medulla also showed strong expression of these epitopes. Nevertheless, the optic lobe was still immature with regard to its expression of phosphorylated NF 220. Though some retinal fibers and a few fine axons within the medulla were stained with the P-NFH antibody, the plexiform layers and neuropils were negative (Fig. 9B).

At stage 30, the stellate ganglion, second-order fibers in pallial nerves, the stellar nerves, i.e., those regions containing axom of the giant fiber system, strongly expressed all NF epitopes, including the P-NFH (Fig. 10A,B). The ganglia in the dorsal mantle stained with the N-terminal antibody demonstrated a mature adult pattern of fiber connections and neuropils (Fig. 10A). Along with the fibers in the mantle connectives, stellate connectives and stellar nerves, the second-order giant axons could be seen entering the giant synapse region of the ganglion (g.f.2) as strongly reactive third-order giant axons (g.f.3) emerge from the ganglion and enter the stellar nerves (s.n.1. In the same ganglion, most cell perikarya were unstained, whereas the adjacent ganglion contained a few large cell bodies with NFs in the axon hillock and perinuclear cytoplasm.

The pattern of expression of the phosphorylated NF 220 was seen in a different ganglion section (Fig. 10B). The large caliber third-order giant axons forming within the ganglion neuropil displayed an intense reactivity as did the fibers in the fin nerve and stellar nerves coursing to the mantle muscles. No cell bodies within the ganglion expressed the phosphorylated epitope. At hatching, the fibers of the giant fiber system contain a mature neurofilament cytoskeleton, with robust expression of the phosphorylated NF 220.

DISCUSSION Neurofilament localization in adult

neural tissues The immunoblot analyses demonstrated the presence of

the NF 60, NF 70, and NF 220 neurofilament epitopes in adult neural tissues, confirming previously published studies (Pant et al., 1978, 1982; Tytell et al., 1990; Szaro et al., 1991). In the axoplasm, giant fiber lobe (GFL) of the stellate ganglion, fin nerve, and optic lobe, each antibody detected its respective NF antigens, albeit at different tissue-specific levels of expression. For example, the phosphorylated NF 220 subunit was most intensely expressed in axoplasm where neurofilament proteins are abundant but much less so in optic lobe and the GFL. This was consistent with the hypothesis that large-caliber axons are presumed to be

Fig. 7. Immunohistochemical analysis of NF expression in the stellate ganglion in horizontal sections of stage 22, 25, and 27 embryos. A Low-magnification view of the stage 22 embryo incubated in a-tubulin antibody with labeled fibers in the eye, optic lobe, and stellate ganglion. B: Higher magnification of a section showing labeled fibers in the stage 22 stellate ganglion expressing NF-NT antigen (arrow). C: Section of a stage 25 embryo (see Fig. 6 0 showing the labeled fibers in the ganglion neuropil, the cells within the ganglion labeled within the axon hillock region (arrowheads), and the hcavily reactive stellar nerve with a third-order giant fiber (large arrow) entering the mantle. D:

Section of the stage 25 stellate ganglion and mantle incubated in the P-NFH antibody showing the cells of the stellate ganglion (sgl, the labeled fibers of the mantle connective (open arrows), and the labeled third-order giant fibers within the ganglion (arrowheads). Note also the numerous labeled nuclei. E: A horizontal section of the stage 27 stellate ganglion displaying very strong expression of the NF-NT antigens in the neuropil (n) , the stellar nerves (sn) leaving the ganglion and entering the mantle (m) and the fin nerve (fn) F: Section of the mantle showing stellar nerve cross sections (arrows) expressing the NF 60170 antigen. Scale bar in C also applies to D and E; bar in B also applies to F.

322 P. GRANT ET AL.

Fig. 8. Developmental expression of neurofilament antigens. Immu- nohistochemical analysis of NF expression in sagittal sections of a stage 30 hatchling reveals diverse expression with different antibodies. A: A section incubated in the NF 60i70 antibody expressing strong reactivity in central and peripheral ganglia, nerve tracts, and neuropils. R: An adjacent section exhibiting a similar pattern of reactivity to the NF-NT antihody in neuropils, nerve tracts, and ganglia. C: A section incubated in the NF 70 antibody reacts most strongly with mantle muscle and non-neuronal tissues and basement membranes. NF 70 antigen is expressed in a few fiber bundles in the pedal and cerebral ganglia. D In

stabilized by the phosphorylated tails of the large NF subunit (Hirokawa, 1991). NF 220 expression was lower in the GFL because the large motor neuron cell bodies accumu- late little NF 220 (Cohen et al., 1987; Tytell et al., 19901.

Evidently, fibers and fiber bundles were the major source of NF proteins expressed in the tissue immunoblots. Both the NF 60170 and NF-NT antibodies displayed immunohistochemical patterns in axons and neuropils similar to those

this section incubated in the P-NFH antibody, phosphorylated 220 neurofilammts are restricted to a few fiber tracts in central ganglia and the fibers of the stellate ganglia. Note the absence of expression in the optic lobe. E: An adjacent section incubated in a-tubulin displays strong expression of tubulin in all brain ganglia neuropils, the stellate ganglia, fiber tracts, and optic lobe. F: A control section incubated in the NF-NT prebleed. b g., buccal ganglion; br.g., brachial ganglion; c.g., cerebral ganglion; mt., mantle; mt.c., mantle connective; o.L, optic lobe; p.g., pedal ganglion; pv.g., palliovisceral ganglion; s.g., stellate ganglion; s.n., stellar nerve; st., statocvst; ten., tentacle.

obtained with the wtubulin antibody, which is known to be specific for axonal and dendritic microtubules. In the stellate ganglion, these antibodies produced strongly positive reactions in the giant axons, in small fibers in the neuropil, and associated stellar and fin nerves. In general, a limited number of small perikarya in the ganglia showed NF expression with NF 60170 and NF-NT antibodies. Most large perikarya of the GFL contributing to the giant


Fig. 9. Immunohistochemical demonstration of NF expression in the optic lobe of the stage 30 hatchling. A: Section displaying a pattern of expression of NF 60170 antigens in the optic fibers (arrowheads) penetrating the outer granular cell layer to the plexiform layer [open arrows) which shows a laminated organization. B: An adjacent section

incubated in the P-NFH antibody exhibits no strong reactivity in the optic fibers, the plexiform layer (arrowheads), or in the neuropils within the medulla (n.1. C: Expression ofa-tubulin in a similar section. Optic fibers (small arrowheads), plexiform layer (large open arrows); and medulla neuropils (n) exhibit strong reactivity.

third-order axons displayed much less reactivity than the giant axons and neuropil fibers.

Likewise, within the optic lobe, the retinal fibers entering the lobe, fibers in the plexiform layers, the bundles of large radial fibers descending into the medulla, and optic tract fiber bundles all expressed these N F antigens. Neurofila- ments were seen only in the large cell bodies of a few second-order visual cells in the innermost layer of the inner granule cell layer, with the most intense reactivity found in small numbers of large transverse visual cells within the medulla.

Phosphorylated NF 220 was most abundant in the giant fibers and neuropils of the stellate ganglia and in the stellar and fin nerves. The expression pattern of this mammalian- derived antibody was identical with that seen previously with an antibody prepared against phosphorylated squid NF 220 (Cohen et d., 1987). We were confident, therefore, that the P-NFH antibody was reacting with phosphorylated epitopes on squid NF-H within nerve fibers. Only in the axon hillock region of the large motor cell bodies of the GFL was the phosphorylated epitope of NF 220 detected. Re- stricted localization of phosphorylated NF 220 to this region and the giant axons is consistent with the suggestion that neurofilament proteins synthesized in the giant perikarya of the GFL are not phosphorylated until they enter the hillock prior to transport into the axons (Cohen et al., 1987; ‘I’ytell et al., 1990).

In the optic lobe, however, a much more restricted distribution of P-NFH expression was displayed. It was most apparent in retinal and optic tract fibers, but many fewer fibers in the plexiform layers or radial bundles converging into the medulla contained the phosphorylated form of NF 220. Unlike the situation in the GFL, no phosphorylated NF 220 was detected in any large optic lobe neurons.

The detection of neurofilaments in axons but not in cell bodies proper may reflect the fact that NFs are transported

prior to posttranslational modifications at the hillock region or within the axon. The nature of this process of synthesis, posttranslational modification, assembly, and transport of NFs within neurons is poorly understood. What is evident from the immunohistochemical expression patterns in adult tissues, however, is a considerable variabil- ity in regulation of each step of this process in different neurons. In summary, the expression patterns of the NF 60/70, NF-NT, and P-NFH antibodies in adult tissues were entirely consistent at the immunoblot and immunohistochemical levels and could, therefore, be used to explore the spatio-temporal pattern of NF expression in the developing embryo.

Spatio-temporal NF expression during development

In this study, we find that the temporal pattern of NF expression during development is very similar to that seen in the vertebrate embryo, where an early coordinate expression of the low molecular weight subunits (NF-L and NF-M) precedes the appearance of NF-H in maturing neural tissues (Shaw and Weber, 1982; Pachter and Liem, 1984; Cochard and Paulin, 1984; Carden et al., 1987). We base this conclusion largely on the immunoblot analysis. Early in squid development, as the first axons differentiate in various central and peripheral ganglia, the two smaller NF subunits, NF 60 and NF 70, are expressed coordinately in the axons at stage 22. The immunoblots with the NF 60/70, NF-NT, and NF 70 antibodies suggest the presence of both antigens from this stage onward. Moreover, the immunohistochemical observations consistently show fiber pattern staining similar to that obtained with the a-tubulin antibody from stage 22 to hatching. At no stage in development, however, was the NF 220 detected in immunoblots. Even the P-NFH antibody, which reacts so robustly in immunoblots and immunohistochemically, failed to detect

Fig. 10. A A horizontal section of a stage 30 hatchling in the region ofboth stellate ganglia incubated in the NF-NT antibody. Fiber bundles and giant fibers show the most intense expression ofNF antigens in the mantle connective (mx.), the stellate connective (s.c.) linking both stellate ganglia (s.g.1, second- and third-order giant fibers (g.f.2., g.f.3), the stellar nerves (s.n.), and the fin nerve (f.n.1. Some perikarya in the

giant fiber lobe of the upper ganglion are also reactive. B: A horizontal section of another stage 30 hatchling showing the expression pattern of the phosphorylated NF 220 in the stellate ganglion (s.g.1, third-order giant fibers (g.f.31, the stellar nerves (s.n.) leaving the ganglion and entering the mantle (m.), and the fin nerve (f.n.). Scale bar in A also applies to B.


any NF 220 during development. Instead, several unknown low molecular weight phosphorylated epitopes were expressed throughout development. One or more of these could account for the intense nuclear staining seen at each stage with the P-NFH antibody. Cross reactivity of this antibody with phosphorylated histones may explain this nuclear expression (Wood et al., 19851, for the expression correlates with cell proliferation in developing embryos; i.e., the extent of nuclear P-NFH staining diminished in progressively older embryos, presumably as proliferation ceased during histogenesis. No nuclear staining was seen in adult tissues.

It appears that the NF 220, at least in the phosphorylated state, arises later in development, primarily in the afferent and efferent fibers of the stage 25 stellate ganglion, principal synaptic center of the giant fiber system. Subsequently, as different brain ganglia mature, only select bundles of axons contain the phosphorylated NF 220.

Because of the absence of a specific antibody to NF 220, we cannot preclude the possibility that the non-phosphorylated NF 220 is also present before stage 25. Failure to detect NF 220 a t any developmental stage with both NF-NT and P-NFH antibodies in Western blots suggests either it is 1) relatively unstable compared to the smaller NFPs, particularly if non-phosphorylated, or 2) that in contrast to the NF 60 and NF 70 subunits, expression of the NF 220 is specifically regulated during development; i.e., compared to the NF 60 and NF 70 neurofilament proteins, lower levels of this subunit are synthesized, too low to be expressed against a background of total embryo and yolk protein. We suggest that the latter alternative of specific regulation of the NF 220 subunit is probably correct.

The absence of any NF 220 expression in embryo immunoblots may relate to the differences between vertebrates and squid in the number of genes coding for neurofilament proteins (three in the former and one in the latter) and in the manner in which these genes are regulated. In vertebrate development, each NF gene is independently regulated in a stage-specific and neuronal-spccific manner at the level of transcription. In fact, intragenic cis-acting elements have been detected in the NF-L gene (Beaudet et al., 1992). Moreover, DNA-binding proteins have been identified which participate in the regulation of human NF-H gene expression (Elder et al., 1992). Finally, in transgenic mice, there is evidence that the NF-L gene and not the NF-H gene is also regulated at the translational level (Charron et al., 1993). For the most part, it appears that the temporal and tissue-specific expression of neurofilament genes during vertebrate development primarily reflects regulation at the level of transcription of the three independent genes.

The temporal order of NF expression during vertebrate development is presumed to correlate with stages in axonal outgrowth and synaptogenesis (Carden et al., 1987; Nixon and Shea, 1992). A very early expression of such neuronal intermediate filaments as nestin, internexin, and periph- erin occurs as s o n s first emerge. Soon thereafter, the downregulation of these proteins is correlated with the coordinate appearance of NF-L and NF-M. These assemble into an “immature” filament that retains sufficient plasticity to permit axonal outgrowth (Carden et al., 1987; Scott et al., 1985). As targets are innervated and synaptogenesis begins, axonal growth ceases and a more stable cytoskeleton is presumably established; i.e., NF-H is synthesized and the mature triplet neurofilament is assembled (Willard and Simon, 1983; Carden et al., 1987). NF-H subunits

mediate the formation of a stable cytoskeleton essential for synapse formation and functional expression. Later, as NF-H and NF-M are phosphorylated, axon caliber increases accompanied by myelination (Hoffman and Lasek, 1975; Foster et al., 1987; Dahl et al., 1986; Carden et al., 1987; deWaegh et al., 1992).

This same developmental scenario may be true for the squid. Early formation of immature neurofilaments might be essential to support plasticity of axon growth before synaptogenesis. Accordingly, no, or very little, NF-H should be produced along with the NF 60 and NF 70, thereby preventing formation of the more stable triplet neurofilament. Because the three subunits are transcribed from the same gene, we would expect all three gene products to be detected at similar levels in the immunoblots unless the NF 220 is specifically prevented from expression by posttranscriptional mechanisms to insure the formation of immature neurofilaments. Later, after axons have formed syn- apses, the posttranscriptional mechanism is modified so that NF 220 is expressed and the mature triplet neurofilament is established. Finally, as axon caliber increases, particularly in the giant fiber system of the stellate ganglion, the NF 220 subunit is phosphorylated.

Our data showing no expression of NF 220 in immunoblots during development, while the NF 60 and NF 70 epitopes are expressed at all stages from the beginning of neurogenesis, is at least consistent with the developmental model mentioned above. The operation of a posttranscriptional regulatory mechanism could prevent the synthesis of the NF 220 subunit, or inhibit its synthesis so that it was expressed at very low, undetectable levels in Western blots at all developmental stages. The fact that the phosphorylated NF 220 could be detected immunohistochemically in the stellate ganglion at stage 25 but not in immunoblots may be attributed to differences in sensitivity of the methods.

Early maturation of the giant fiber system Axons first appear at stage 22 in the stellate ganglia and

immediately express a-tubulin, and NF 60170 epitopes. At stage 25, the fibers leading to and within the stellate ganglia are the first to express the phosphorylated NF 220 subunit. The precursors of second- and third-order giant fibers interact in this region to form the giant synapse. It appears that mature cytoskeletal elements (phosphorylated neurofilaments) first appear within the giant fiber system well before the jet-propelled escape responses are detected at stage 26 (Gilly et al., 1991). As development continues, (stages 25 to hatching), the level of expression of the P-NF220 increases conspicuously within second-order axons to the stellate ganglion, the neuropils (giant synapse region), and stellar nerves to the mantle; i.e., the major elements of the giant fiber system responsible for the rapid jet-propelled escape responses. At hatching, the giant fiber system of the stellate ganglion resembles that of the adult; all second- and third-order giant axons contain high levels of neurofilament protein, including phosphorylated NF 220, suggesting the presence of a mature cytoskeleton.

In a similar manner, elements of the visual system, such as the retinal fibers, also show a-tubulin, NF 60, and NF 70 expression as soon as they arise at stage 22. Later, as the optic lobe matures, the cytoskelctal proteins appear in plexiform fiber layers and in some optic tract fibers in the medulla. No axons, however, express the phosphorylated NF 220 epitope before hatching in the visual system and

326 P. GRANT ET AL.

even at this stage, expression is limited to a few retinal and optic tract fibers. Clearly, the optic lobe matures later than the giant fiber system.

Although the retina and optic lobe arise before the stellate ganglion, at the time of hatching, the axonal cytoarchitecture is still relatively immature since the phosphorylated N F 220 is, for the most part, not present in sufficient amounts to be detected. A functioning escape response system has fundamental survival value, and it is not surprising that it is the first to mature prior to hatching.

LITERATURE CITED Adelman, W., and D.L. Gilbert (1990) ElectrophysioloE and biophysics of

the squid giant axon. In D.L. Gilbert, W.J. Adelman, Jr., and J.M. Arnold (eds): Squid as Experimental Animals. New York: Plenum, pp. 93-132.

Arnold, JX. i1965j Normal embryonic stages of the squid Luligu pealei (Lesueurj. Biol. Bull. 128:24-32.

Beaudet, L.G., G. Charron, I. Tretiakoff, and J.P. Julien (1992) Intragenic regulatory elements contribute to transcriptional control of the ncurofila- ment light gene. Gene, 116i205-215.

Beaudet, L., F. Cote, D. Houle, and J.P. Julien (1993) Different posttranscriptional controls for the human neurofilament light and heavy genes in transgenic mice. Mol. Brain Res. 18::23-31.

Bradford, M. (1976) A rapid and sensitive method for the quantitation o f microgram quantities of protein utilizing the principle of protein-dye binding. Anal. Binchem. 72.248-254.

Carden, M.J., J.Q. Trojanowski, W.W. Schlaepfer, and V.M.-Y. Lee (1987) Two-stage expression of neurofilament polypeptides during rat neurogenesis with early establishment of adult phosphorylation patterns. J. Neurosci. 7:3489-3504.

Cochard, P., and D. Paulin (1984) lnitial expression of neurofilaments and vimentin in the central and peripheral nervous system of the mouse embryo sn uiuo. J. Neurosci. 42080-2094.

Cohen, R.S., H.C. Pant, S. House, and H. Gaincir (1987) Biochemical and immunocytocheniical characterization and distiihution of phosphow- lated and nonphosphorylated subunits of neurofilaments in squid giant axon and stellate ganglion. J. Neurosci. 7:2056-2074.

Dahl, D., C.J. Crosby, E.E. Gardner, and A. Bignami (1986) Delayed phosphorylation of the largest neurofilament protein in rat optic nerve development. J. Neurosci. Res. 15:513-519.

dcWaegh, S.M., VM-Y. Lee, and S.T. Brady (1992) Local modulation of neurofilament phosphorylation, axonal caliber and slow axon transport by myelinating Schwann cells. Cell 68:451-463.

Elder, G.A., Z. Liang, N. Lee, V L. Friedrich, and R.A. LaLzarini (1992) Novel DNA binding proteins participate in the regulation of human neurofilament H gene expression. Mol. Brain. Res. 1585-98.

Floyd, C.C., P. Grant, P.E. Gallant, and H.C. Pant (1991) Principal neurofilament-associated protein kinase in squid axoplasm is related to casein kinase I. J. Biol. Chem. 266r4987-4994.

Foster, G.A., D. Dahl, and V.M.-Y. Lee (1987) Temporal and topographic relationships between the phosphorylated and non-phospborylatcd epitopes of the 200 kDa neurofilament protein during development in uitro. J. Neurosci. 7:2651-2663.

Gilly, W.F., B. Hopkins, and G.O. Mackie (1991) Development of giant motor axons and neural control of escape responses in squid embryos and hatchlings. Biol. Bull. 150:209-220.

Hickey, W.F., V.M.-Y. Lee, J.Q. Trojanowski, L.J. McMillan, T.J. McKearn, J. Gonatas, and N.K. Gonatas (1983) Immunochemical application of monoclonal antibodies against myelin basic protein and neurofilament triple protein subunits: advantages over antisera and technical limita- tions. J. Histochem. Cytochem. 31:1126-1135.

Hirokawa: N. (1991) Molecular architecture and dynamics of the neuronal cytoskclcton. In R.D. Burgoyne (ed): The Neuronal Cytoskeleton: New York John Wiley and Sons, pp. 5-74.

Hoffman, P.N., and R.J. Lasek 11975) The slow component of axonal transport: Identification of the major structural polypeptides of the axon and their generality among mammalian neurons. J. Cell Biol. 66:351-366.

Korschelt, E. (1892) Beitrage zur Entwicklungsgeschichte der Cepha- lopoden. I. Die Entstehung des Darmkanals und Nervensystems in

Bezichung zur Keimblatterfrage. Verh. d. 2001. Bot. Ges. (Leipzig) Festschr. Leukart. 345-373.

Marthy, H . 4 . (1973) An experimental study of eye development in the cephalopod LoLtgo uulgnris: determination and regulation during formation of the primary optic vesicle. J. Embryol. Exp. Morphol. 29:347-361.

Marthy, H . J . (1987) Ontogenesis of the nervous system in ccphalopods. In M.A. Ali (ed): Nervous Systems in Invertebrates. New York Plenum, pp. 443459.

Martin, R. (1965) On the structure and embryonic development of the giant fibre system ofthe squid Loligo uulgaris. Z. Zellforsch. 67:77-85.

Martin, R., and D. Rungger (1966) Zur structur und Entwicklung des Riesenfasersystems erster Ordnung von Sepia oficinalzs. Z. ZellCorsch. 74:454-463.

Meinertzhagen. I.A. (1990) Development uC the squid’s visual system. In D.L. Gilbert, W.J. Adelman, Jr , and J.M. Arnold (edsj: Squid as Experimental Animals. New York: Plenum, pp. 399-419.

Meister, G. (1972) Organogenese van Loligo uulguris Lam. Zool. Jarb. Anat. 8.9247-300.

Nixon, R.A., and T.B. Shea (1992) Dynamics of ncuronal intermediate filaments: a developmental perspcctive. Cell Motil. Cfloskeleton 22:81- 91.

Pachter, J.S., and R.K.H. Liem (1984) Thc differential appearance of neurofilament triplet polypeptides in the developing rat optic nerve. Dev. Bid. 103200-210.

Pant, H.C., G. Shecket, H. Gainer, and R.J. Lasek (1978) Neurofilament protein is phosphorylated in the squid giant axon. J. Cell. Biol. 78:R23- R27.

Pant, H.C.. P.E. Gallant, R. Gould, and H. Gainer (19821 Distribution of calcium-activated protease activity and endogenous substrates in the squid nervous system. J. Neurosci 2:1578-1587.

Pant, H.C., P.E. Gallant, and H. Gainer (1986) Characterization of a cyclic-nucleotide- and calcium-independent neurofilament protein kinase activity in axoplasm from thr squid giant axon. J. Bid. Chem. 261:296%2977.

Sacarrao, G.F. 11956) Contribution a I’ktude du developpement embryon- naire du ganglion stellaire et de la gland endocrine des Cephalopodes. Arqu. Mus. Bocage, Lisboa27:137-152.

Scott, D., K.E. Smith, B. O’Brien, and K.J. Angelides (1985) Characteriza- tion o f mammalian neurofilament triplet proteins: subunit stoichiom- etry and morphology of mature and reconstituted filaments. J. Biol. Chem. 260:10736-10797.

Shaw, G., and K. Weber (1982) Differential expression of neurofilament triplet proteins in brain development. Nature 298.277-279,

Sternberger, L.A., and N.H. Sternberger (19831 Monoclonal antibodies distinguish phosphorylated and nonphosphorylated forms of neurofila- rnents in nitu. Proc. Natl. Acad. Sei. U.S.A. 80:6126-6130.

Szaro, B.G., H.C. Pant, J. Way, and J. Battey (1991) Squid low molecular weight neurofilament proteins are a novel class of neurofilament protein. d. Bid. Chem. 266:15035-15041.

Trqjanowski, J.Q., M.A. Obrocka, and V.M.-Y. Lee (1985) Distribution of neurofilament subunits in neurons and neuronal processes: immunohistochemical studies of bovine cerebellum with subunit-specific monoclonal antibodies. J. Histochem. Cytochem. 3.?:557-563.

Tytell. M., H.C. Pant, H. Gainer, and W.D. Hill (19901 Characterization of the distinctive neurofilament subunits of the soma and axon initial segments in the squid stellate ganglion. J. Neuroscience Res. 25:153- 161.

Way, J., M.R. Ilellmich, H. Jaffe, B. Szaro, H.C. Pant, H. Gainer, and J. Battey (1992) A high-molecular-weight squid neurofilament protein contains a lamin-like rod domain and a tail domain with Lys-Ser-Pro repeats. Proc. Natl. Acad. Sci. U S A . 89:69634967.

Willard, M.B., and C. Simon (1983) Modulations of neurofilament axonal transport during the development of rabbit retinal ganglion cells. Cell 35551-559.

Wood, J.N., N.B. Lathangue, D.R. McLachlan, B.J. Smith, B.H. Anderton, and A.J. Dowding (1985) Chromatin proteins share antigenic determi- nants with neurofilaments. J. Neurochem. 44:149-154.

Young, J.Z. (1938) The runctioning of the giant nerve fibres o f the squid. J. Exp. Biol. 15170-185.

Young, J.Z. (1939) Fused neurons and synaptic contacts in the giant nerve fibers of cephalopods. Philos. Trans. R. Soc. Lond. [Biol.l229:465-503.

Young, J.Z. (19741 The central nervous system of Loligo. I. The optic lobe. Philos. Trans. R. Soc. Lond IBiol.l267:263-302.

Expression of neurofilament proteins during development of the nervous system in the squid Loligo...

Documents

Transcript of Expression of neurofilament proteins during development of the nervous system in the squid Loligo...