Axotomy-Induced Neurofilament Phosphorylation Is ... - Cell Press

Neuron,

Vol. 10,613-625,

April,

1993, Copyright

@ 1993 by Cell Press

Axotomy-Induced Neurofilament Phosphorylation inhibited In Situ by Microinjection of PKA and PKC Inhibitors into Identified lamprey Neurons Garth F. Hall* and Kenneth S. Kosik+ *Department of Neurology (Neuroscience) Harvard Medical School and Department of Neurology Children’s Hospital Boston, Massachusetts 02115 +Department of Neurology (Neuroscience) Harvard Medical School and Center for Neurologic Diseases Department of Medicine (Division of Neurology) Brigham and Womens Hospital Boston, Massachusetts 02115

Summary Close axotomy of identified lamprey neurons induces phosphorylation of somatodendritic neurofilaments (NFs), followed by ectopic regeneration of neurofilamentous sprouts from the dendrites. We used in situ intracellular microinjection to study the mechanism of axotomy induced NF phosphorylation. We found that inhibitors of protein kinase C (PKC) and protein kinase A (PKA) block somatodendritic NF phosphorylation for up to 15 days when injected at the time of axotomy. Injection of PKA catalytic subunit, diacylglycerol, or okadaic acid induces somatodendritic NF phosphorylation in intact neurons with the same time course as close axotomy. These results suggest that transient activation of PKC, PKA, and/or serine phosphatase inhibition by axotomy triggers persistent intracellular changes that may be related to polarity loss in these neurons.

Introduction Axonal transection results in a variety of early cellular changes in the affected neuron, some of which may play roles in converting the cellular metabolism from a homeostatic to a regenerative state, and thus may be necessary preliminaries to axonal regeneration. A prominent consequence of axotomy is the reorganization of the neuronal cytoskeleton, frequently involving an increase in the number of somatic neurofilaments (NFs) (Lieberman, 1971; Torvik, 1976) and changes in the phosphorylation state and spacing of NFs (Dragerand Hofbauer, 1984; Goldstein et al., 1987; Shaw et al., 1988; Hall et al., 1989, 1991). This is sometimes accompanied by the destabilization, loss, or redistribution of somatic microtubules (Hall et al., 1989, 1991). Axotomy also induces an increase in the synthesis of the tubulins and actin (Heacock and Agranoff, 1976; Sinicropi and McElwain, 1983; Oblinger et al., 1989) and a decrease in synthesis of the NF triplet proteins (Hoffman and Lasek, 1980; Hoffman et al., 1987; Oblinger et al., 1989). Axotomy at a point very close to the soma results in the loss of aspects of

Is

cellularpolarityinawidevarietyof neurontypes(Murphy and Kater, 1980; Hall and Cohen, 1983; Roederer and Cohen, 1983; Schacher and Proshansky, 1983; Linda et al., 1985). In these instances, the axon regenerates from the soma and/or the dendritic tips, as well as from the cut axon stump, indicating that the cell has lost the ability to funnel axonal growth specifically into the appropriate process. It is likely that some of the cytoskeletal changes mentioned above play important roles both in axonal regeneration and in determining whether normal cellular polarity is maintained or lost following axotomy. Both actin and the tubulins play major roles in normal axonal growth (Wessells et al., 1978; Bentley and Toroian-Raymond, 1986; Letourneau et al., 1987), whereas the loss of stable (acetylated) microtubules from the dendrites has been correlated with the loss of normal polarity following close axotomy (Hall et al., 1991). Microtubule-associated proteins such as tau have been shown to be both necessary (Caceres and Kosik, 1990) and sufficient (Knops et al., 1991) to initiate process outgrowth in the presence of tubulin under certain conditions and are important in establishing initial polarity in neurons (Caceres and Kosik, 1990). Although NFs have traditionally been thought to play no dynamic role in axonal outgrowth, recent studies have shown that they are present in growing sprout tips at early stages of axonal regeneration (Lanners and Grafstein, 1980; Hall et al., 1991) and are transported more rapidly than normal in actively regenerating axons (McKerracher et al., 1990). It has also been shown recently that selectively inactivating NFs during development can block axonal development in vivo (Szaro et al., 1991). Progress in elucidating the cellular mechanisms underlying the response to axotomy has been hampered by the relative inaccessibility of in situ neuronal systems, particularly in the mammalian CNS. In this study, we have attempted to circumvent this difficulty by using an unusually accessible in situ system consisting of a set of identified giant motor interneurons, anterior bulbar cells (ABCs)and MauthnercelIs(MCs), in the hindbrain of the larval sea lamprey to study the cellular responses to axotomy on a single-cell level in the vertebrate CNS. Axotomy of either ABCs or MCs results in major cytoskeletal changes in the soma and dendrites at early times postaxotomy (Hall et al., 1991). By 6 days after axotomy at a site 500 urn from the soma (close axotomy), phosphorylated NFs have begun to appear in the soma and dendrites; this is accompanied by a destabilization of somatodendritic microtubules. By IO-15 days after close axotomy, sprouts begin to appear at the end of the axon stump and at the dendritic tips; these then elongate and take on the appearance of regenerating axons (Hall and Cohen, 1983, 1988; Hall et al., 1991). The cytoskeleton of these sprouts

Neuron 614

appears to consist almost exclusively of phosphorylated NFs, with tubulin being scarce and often undetectable. This distinctive cytoskeletal composition is seen from the earliest stages of axonal regeneration (Hall et al., 1991) and is still apparent many weeks after theonsetof sprouting, bywhich timethesprouts have elongated many hundreds of micrometers following typically axonal trajectories (Hall et al., 1989; Hall, Lee, and Kosik, 1990, J. Cell Biol., abstract). Although the tips of these sprouts are usually swollen and spindle shaped, thus resembling the growth cones of developing axons and dendrites in other systems (Skoff and Hamburger, 1974; Nordlander and Singer, 1982), they contain phosphorylated NFs instead of the microtubules and actin filaments normally found in growth cones (Fine and Bray, 1971; Letourneau, 1981; Bridgman and Dailey, 1989). These findings raise the possibilitythat early changes in the pattern of NF phosphorylation induced by axotomy may be a contributing factor in the mechanism of axonal regeneration, as well as in the loss of normal cellular polarity following close axotomy in the lamprey (Hall et al., 1991). by In this study, we focused on the mechanism which axotomy causes the early phosphorylation of somatodendritic NFs in lamprey ABCs and MCs. Specifically, we asked whether NF phosphorylation following axotomy is a consequence of the activation of cellular protein kinases by second messengers. We used intracellular microinjection into identified lamprey central neurons in situ to determine whether NF phosphorylation following axotomycould be blocked or reduced by agents known to inhibit second messenger activation of specific protein kinases. In this paper, we show that it is possible to manipulate NF phosphorylation in single cells in vivo using intracel-

lular microinjection of specifically acting agents; that both a protein kinase C (PKC) and a protein kinase A (PKA)-like kinase are involved in the sequence of events leading to NF phosphorylation following axotomy; and that a protein phosphatase activity plays a role in preventing the phosphorylation of somatodendritic NFs in intact neurons. Results Effects of Second Messenger-Activated Kinase Inhibitors on NF Phosphorylation in ABCs and MCs Axotomy of ABCs and MCs at a point 508 pm or less from their somata results in the phosphorylation of somatodendritic NFs in some cells by 2-4 days and in virtually all cells within 6 days (Hall et al., 1991). We examined the effects of injected kinase inhibitors on NF phosphorylation in ABCs and MCs (Figure 1)6days after combined close axotomy and microinjection by cutting serial transverse sections through the hindbrain near the level of the eighth nerve and performing immunocytochemical analysis with three monoclonal antibodies (RM034, RM062, and SMl31) that recognize the phosphorylated multiphosphorylation site in the NF carboxy-terminal domain on sections containing both injected and noninjected cells (see Experimental Procedures). We found that microinjection of K252-a, an agent that has been shown to block process outgrowth in response to nerve growth factor in PC12 cells via a broad spectrum inhibition of second messenger-dependent kinases (Koizumi et al., 1988), was effective in reducing axotomy-induced phosphorylation of somatodendritic NFs when injected at the time of axotomy, even at low doses (10

Figurel. and MCs

In Situ

Microinjection

of ABCs

Dorsal view of a lamprey head showing the hindbrain exposed and illuminated as it would be for microinjection, and the position of the neurons used in this study. The somata of two ABCs and the MC on the right side of the brain have been filled with 0.5% Fast green via a microelectrode (shown pointingattheMCsoma). mb, midbrain; oc, otic capsule. Bar, 500 urn.

PKA and PKC Inhibitors 615

Block

NF Phosphorylation

In Situ

PM in the electrode tip). Microinjection of sphingosine (0.5 mM), which is most effective as an inhibitor of PKC, also caused a reduction in staining for phosphorylated NFs by 6 days postaxotomy in the injected cells, relative to adjacent noninjected cells and to control injectionscontaining rabbit IgGand/oran inactive peptide (Table 1). However, the dose of sphingosine necessary to achieve a significant effect was too high for us to be certain that the response to sphingosine was the result of a specific inhibition of PKC alone. We therefore proceeded to examine the effects of microinjecting PKC-specific peptide inhibitors on NF phosphorylation. inhibition of NF Phosphorylation by Peptides That Specifically Block the Activation of Protein Kinase C In a series of experiments similar to those described above, we microinjected peptides that have been shown to inhibit PKC into ABCs and MCs at the time of axotomy (see Experimental Procedures for details). One of these peptides (19-36) is a pseudosubstrate peptide inhibitor; the other peptide employed is a specific substrate for PKC derived from the sequence of the phosphorylation site of the mammalian epidermal growth factor receptor (Heasley and Johnson, 1989). Both of these peptides strongly inhibited NF phosphorylation when they were injected at the time of axotomy (Figure 2; Table 1). The 19-36 peptide was tested at concentrations of 1,5,20, and 500 PM (Table 1). At each of these concentrations, it effectively inhib-

Table

1. Effects

Agents Injected Close Axotomy

of PKC and

at Time

Second messenger activated kinase PKC-specific inhibitory

peptides

PKA-specific inhibitory

peptide

PKA Inhibitors

on Axotomy

Induced

of

Sphingosine K252a

Somatodendritic

Days Post Axotomy

Doss

inhibitors

ited NF phosphorylation, indicating that it was acting via a specific inhibition of PKC (see Smith et al., 1990). Controls for these experiments consisted of microinjetting either rabbit IgG alone (injection control) or the control injection mixture plus an inactive analog of the pseudosubstrate peptide (Figure 3). Injection of IgG alone had no effect on NF phosphorylation by 6 days following injection and axotomy, and was not significantly different from the slight inhibitory effects of injecting the inactive 19-36 analog at 500 PM (Table 1). All concentrations of the active 19-36 peptide tested produced stronger effects than did the inactive peptide, although the 1 and 5 PM concentrations produced less significant differences than the higher concentrations (p = 0.1 versus p< 0.05, respectively). All concentrations of the active peptide produced highly significant inhibition of phosphorylation when compared with the injection controls (p < 0.01). Wealsoexamined theeffectsof PKC inhibitorypeptides at 15 days following injection and axotomy to determine whether the inhibition of NF phosphorylation persists until the onset of ectopic axonal regeneration from the dendrites (Hall and Cohen, 1988). We found that NF phosphorylation was still clearly inhibited by the pseudosubstrate peptide at this time (Figures 2E and 2F; Table 1). In some of these cells, the inhibition was even more pronounced than that seen at 6 days, possibly because somatodendritic NFs are often more heavily phosphorylated at 15 days than at 6 days following close axotomy. The results of control

1 mM 10 uM

NF Phosphorylation Stamina Noni&ted Weaker

Relative to Cells Similar

Stronger

Statistical Significanceb

6 6

6 8

3 6

0 1

p < 0.05 p < 0.1

ECF peptide 19-36 peptide

1 500 500 20 5 1

mM pM pM pM PM PM

6 6 15 6

9 34 8 8 10 22

1 17 3 3 7 15

0 1 0 0 1 1

p p p p p p

6-22

peptide

500 PM 20 uM 1 WM

6

15 21 5

a 9 2

0 1 0

p = 0.025 p < 0.05 p < 0.05

PKA, PKC inhibitory peptide mixture

6-22

+ 19-36

6

16

a

0

p < 0.01

Controls

IgG alone 19-36 control

6 6 15

4 12 4

20 24 12

5 5 1

p = 0.99d 0.8 > p > 0.2 0.8 > p > 0.2

20uM 500 PM

+20uM

PKC, protein kinase C; 19-36 peptide, PKC pseudosubstrate peptide inhibitor; 19-36 control, inactivated pseudosubstrate 6-22, PKA inhibitory peptide; EGF peptide, epidermal growth factor receptor substrate sequence for PKC. a Concentrations in the electrode tip. b Compared with all 6 day controls. Significance determined by chi-square analysis. c Compared with 15 day 19-36 control. d Compared with the expected distribution if no artifacts induced by injection of control substances are present.

< < < = <
p > 0.2 p < 0.05 p < 0.01

2,000 U/ml

6

1

7

p < 0.005

10 uM 50 PM 150 pM 150 uM

6

6 3 0 5 0 1

4 5 6 5 6 5

p < 0.05 p < 0.005 p < 0.005 All p < 0.005 vs controls p < 0.05 3 days vs 6-9 daysb

IgG alone DMSO Racemized

6

15 7 5

1 0 1

All p = 0.9’ vs other controls

1 mM

3 6 9

DiC8”

DMSO, dimethylsulfoxide; Dick, di(octoyl)glycerol. a Compared with all controls. Significance determined by chi-square analysis. b Compared in a 2 x 3 chi-square analysis after 6 and 9 day data were combined. c Compared in a 2 x 3 chi-square analysis. d Stock solution allowed to stand at 4T for 24 h before use.

Significance’

Neuron 620

C


Block

NF Phosphorylation

In Situ

suggest that the dynamic removal of phosphate may normallyplayan important (though probably indirect) role in controlling the phosphorylation state of somatodendritic NFs. We therefore suggest that changes in the activities of PKC, PKA, and protein phosphatases might play significant roles in thechain of events leading from close axotomy via somatodendritic NF phosphorylation to axonal regeneration itself and/or the loss of cellular polarity that accompanies it in these neurons. Specificity of PKA and PKC Inhibitory Peptides We used the 19-36 and 6-22 peptides at a range of concentrations to test their specificity for PKC and PKA, respectively. A recent study (Smith et al., 1990) directly tested the ability of these peptides (6-22 is a derivative of the protein kinase I peptide) to inhibit a variety of mammalian serine/threonine kinases in vitro. Smith and coworkers found that both of these peptides were highly specific inhibitors of their respective kinases if moderately low doses were used, but that the 19-36 peptide was less specific at higher concentrations (5 PM or more). We have made two assumptions in asserting that microinjection of these peptides into ABCs results in specific inhibition of the target kinases. First, the amount of material injected into a given ABC could have been at most 10% of its volume, so that the effective intracellular concentrations of these peptides were at least IO-fold lower than the concentration in the electrode tip. Hence, all but the highest concentrations used in our experiments should have had specific effects by the criteria given by Smith et al. (1990). Second, we assumed that species differences between lamprey and mammals were not important in determining inhibitor specificity, given the highly conserved nature of PKC and PKA between species as diverse as humans and yeast. Activation or Redistribution of Kinase Activities by Axotomy Our results suggest that PKA- and PKC-like activities are generated in the soma and dendrites as an early consequence of close axotomy. One mechanism by which axotomy might activate PKC is by an increase in the level of intracellular Ca*+. This might occur as aconsequence of either a direct influx of extracellular Ca2+ at the lesion site, or the release of bound Caz+ from intracellular stores as the result of local depolarization (Holliday et al., 1991). Axotomy of lamprey gi-

Figure

5. Effects

of Microinjecting

OA

into

ABCs

and

ant neurons in thespinal cord has, in fact, been shown to generate a flux of Ca*+ and other cations into the axon stump from the extracellular fluid at the site of axotomy (Borgens et al., 1980; Strautman et al., 1990). This flux is transient, persisting for 1 day or less followingaxotomy, and is localized to theareawithin several hundred micrometers of the lesion. It is therefore reasonable to assume that following axotomy very close to the soma, somatodendritic Ca*+ levels might be elevated for a short period of time, permitting the activation of PKC either directly or via a “cascade” mechanism involving other kinases. Axotomy very close to the soma might also cause the redistribution of the kinase(s) responsible for NF phosphorylation within the cell, rather than (or as well as) its activation in the somatodendritic compartment. The results of a number of studies indicate that most NF protein does not become heavily phosphorylated until it enters the axon (Bennett and DiLullo, 1985; Black et al., 1986), suggesting that the kinases responsible for heavy carboxy-terminal phosphorylation of NFs may reside primarily in the axon. This raises the possibility that axotomydisrupts the localization of these kinases to the axon, thereby causing them to phosphorylate NFs in the somatodendritic compartment. Roles of PKC, PKA, and Protein Phosphatases in NF Phosphorylation Our results suggest that the roles played by PKA, PKC, and OA-sensitive phosphatases in NF phosphorylation are indirect. One indication of this is thediscrepancy between the time course of NF phosphorylation induced by either axotomy or OA and DiC8 injection (which occurs over several days) and the likely time course of the effectiveness of injected OA and DiC8. Both OA and DiC8 possess large hydrophobic moieties and are consequently capable of crossing membranes relatively easily. OA has been shown to enter cultured and isolated cells within minutes of its addition to the external medium (Klumpp et al., 1990; for review see Cohen et al., 1990). Thus, in a situation in which OA is microinjected into a single cell surrounded by OA-free extracellular space, one would expect an effective dose of OA to be present in the cell for a few hours at most. This is equally true for DiC8, which in addition tends to racemize relatively rapidly to an inactive form in aqueous solution. It is therefore hard to see how either of these agents could be directly responsible for the NF phosphorylation

MCs

Intact ABCs and MCs were injected with the serineithreonine phosphatase inhibitor OA at concentrations of 10 PM (C and D), 50 uM (E and 0, and 150 PM (C and H) in the electrode tip and examined 6 days later. Some preparations were also examined at 3 days (A and B) and 9 days (data not shown) following injection of 150 pM OA. (A), (C), (E), and (G) were stained for rabbit IgG and identify the injected cells. Adjacent sections (B, D, F, and H) were stained with RM034. Note that staining for phosphoryiated NFs is faint or nonexistent (as shown in [B]) for the first few days postinjection but becomes prominent by 6 days (H). Note also that the response to OA injection at 6 days is dose dependent (D, F, and H), with the 10 gM dose eliciting barely detectable staining in some cells (D; Table 2) and the 150 KM dose eliciting strong staining in virtually all cells (H; Table 2). The locations of noninjected cells are marked by asterisks. Bar, 100 urn.

Neuron 622

seen in our study, which was strongest and most consistent at 6 or more days postinjection of either DiC8 or OA. These considerations suggest that the PKC-like activity and/or phosphatase inhibition induced by axotomy is itself transient, but causes a persistent alteration in the environment of the somatodendritic compartment. This alteration may include either the constitutive activation of a NF kinase or a change in the balance between kinase and phosphatase activities acting on the NF multiphosphorylation site. The second indication that PKC and PKA do not act directly on the multiphosphorylation site of lamprey NFs comes from recent studies indicating that NF phosphorylation at the homologous sites in mammalian NF-L and NF-M are PKC and PKA independent (Sihag and Nixon, 1989,199O; Bennett, Rodriguez, and Quintana, 1991, J. Cell Biol., abstract). This is the phosphorylation site recognized by the monoclonal antibodies used in the present study, which have been shown to recognize the corresponding site on the carboxyl sidearm region of the lamprey NF protein (Lee et al., 1988; Pleasure et al., 1989). Therefore, either NF phosphorylation at this site following axotomy is accomplished directly by PKC in the lamprey (unlike the situation in mammals),or (more likely) PKCactivity in the lamprey acts by some indirect mechanism to accomplish NF phosphorylation. Studies of NF phosphorylation in mammals have shown that NF proteins contain both PKA and PKC phosphorylation sites near their amino termini and that phosphorylation at these sites appears to prevent the assembly and transport of NF subunits or oligomers into the axon (Hisanaga et al., 1990; for review see Nixon and Sihag, 1991). These phosphorylation sites are not recognized bythe antibodies used in this study. Furthermore, this phosphorylation is normally transient and reversible by the action of cellular protein phosphatases. It is therefore possible that increased phosphorylation at these sites by PKC and/ or PKA activation or phosphatase inhibition might indirectly induce the accumulation of phosphorylated NFs in the soma by detaining newly synthesized NF protein in the soma long enough for it to be phosphorylated at the multiphosphorylation site (and further immobilized) by other kinases. Such a mechanism would also be consistent with the nonadditive effects of inhibiting both PKA and PKC on NF phosphorylation observed in this study if phosphorylation at one of these sites were sufficient to block NF assembly and transport. PKC and PKA activation might also result in somatodendritic NF phosphorylation without interacting directly with NFs. This might occur via a cascade of kinase phosphorylations culminating in the activation of a NF-associated kinase and phosphorylation of the carboxyl domain of the NF protein. Such phosphorylation cascades have been found to operate in other situations in which a singleevent ortransient extracellular signal results in a switching of developmental or metabolic programs within the cell (for review see

Cohen, 1988); a well-described example is glycogenolysis. Such a mechanism might involve a family of serinelthreonine kinases (extracellular signal-regulated kinases; Boulton et al., 1991). These kinases have been shown to transduce extracellular signals and appear to participate in phosphorylation cascades that include the activation of PKC (Nel et al., 1990), the strong control by OA-sensitive protein phosphatases (Haystead et al., 1990), and the phosphorylation of cytoskeletal proteins. Extracellular signal-regulated kinases have also been shown to phosphorylate proline-rich sequences similar to those found at the carboxy1 domain multiphosphorylation site on the lamprey NF protein, which has recently been sequenced (A. Jacobs and M. Selzer, personal communication). Possible Roles of NF Phosphorylation in Response to Axotomy We have shown (Hall et al., 1991) that phosphorylation of somatodendritic NFs iswell correlated with the loss of normal polarity following axotomy at a site very close to the cell body. This raises the possibility that the phosphorylation of NFs normally serves as a targeting signal for the sorting of NFs into the axon during axonal regeneration, which is disrupted by abnormal NF phosphorylation in the somatodendritic compartment following close axotomy. However, such NF phosphorylation might also occur as a consequence of the loss of polarity, rather than as a cause of it. It is not yet clear whether the massive presence of phosphorylated NFs in sprouting neurites following close axotomy plays any active role in axonal regeneration. Although neurite elongation has traditionally been thought to depend on a combination of growth cone-mediated “pull” on an adhesive substrate and microtubule polymerization, it has been demonstrated that the extrusive “pushing” forces generated by microtubule polymerization alone are capable of causing process elongation in the absence of growth cones or an adhesive substrate (Marsh and Letourneau, 1984; Letourneau et al., 1987; Knops et al., 1991). Thus, it is conceivable that extrusive forces produced by NF polymerization might be sufficient to cause sprout elongation in the absence of growth conelike specializations and that such a mechanism might operate in axonal regeneration in the lamprey. It should be possible to address this question directly in situ in the lamprey system by microinjection experiments involving various cytoskeletal inhibitors and/ or purified specific antibodies such as those used by Szaro et al. (1991). We have used the technique of intracellular microinjection to study the cellular response to axotomy in situ in identified vertebrate central neurons. Our results suggest that the activity of PKA, PKC, and OAsensitive protein phosphatases play important but indirect roles in controlling the phosphorylation of somatodendritic NFs following axotomy in this system, possibly via some type of phosphorylation cascade


Block

NF Phosphorylation

In Situ

set in motion by axotomy. It is not yet known whether NF phosphorylation or the activities of the kinases and phosphatases that control it play important roles in the loss of cellular polarity following close axotomy or in the mechanisms underlying axonal regeneration. However, we feel that the ability to employ microinjection to study cellular mechanisms in this system offers a significant opportunity for the future study of the response of vertebrate central neurons to injury in situ on a single-cell level. Experimental

Procedures

Agents Injected A variety of agents were found to be effective in modulating NF phosphorylation when microinjected into intact and axotomized ABCs and MCs. We used four different agents to block NF phosphorylation following axotomy. Two of these have relatively broad specificity: K-252a (from Nocardiopsis; obtained from CalBiochem), which acts on CAMP-activated kinase (K = 8.818 uM) and cCMP-activated protein kinase (K, = 0.020 uM) as well as PKC (K, = 0.025 uM); and sphingosine (Sigma), which is most effective in inhibiting PKC (Hannun et al., 1986), but also has effects on other kinases. Two other agents are inhibitory pep tides that appear to be specific to PKC, at least under certain conditions (Smith et al., 1998). One of these peptides (obtained from GIBCO) is a 17-mer containing residues 19-36 of rat PKC (ArgPheAlaArgLysGlyAlaLeuArgGInLysAsnValHisGluValLysAsn). This sequence comprises a pseudosubstrate domain that normally fits in the active site of the kinase and prevents its constitutive activation (House and Kemp, 1987). A control peptide (inactivated by the substitution of Clu for Arg at position 27) was also obtained from GIBCO. The other PKC inhibitory peptide used (ob tained from Research Products Inc.) was a nonapeptide (ValArgLysArgThrLeuArgArgLeu) containing a sequence derived from that of the phosphorylation site of the epidermal growth factor receptor, which has been shown to bea specific substrate for PKC (Heasley and Johnson, 1989). Peptide 6-22 (GIBCO), which was derived from the highly specific protein kinase I inhibitory peptide for PKA (Smith et al., 1998), was used as a specific inhibitor of PKA. DiC8, kindly provided by Dr. Mark Warner, was used as a specific activator of PKC. The purified catalytic subunit of PKA (from beef heart) was obtained from Sigma. OA, a polyether fatty acid used as a specific inhibitor of protein phosphatases 1 and 2A (Cohen et al., 1998), was obtained from CIBCO. Surgery and Microinjection Techniques Larval lampreys (18-F cm long) were anesthetized by immersion in saturated aqueous benzocaine. ABCs and MCs were axotomized by bilateral transection of the medial motor tracts in the hindbrain at a point between 250 and 588 urn from their somata (at the level of the eighth nerve) with a#11 scalpel blade. Microinjection of peptides, rabbit IgC, and other substances into ABCs in anesthetized lampreys was accomplished using a Picospritzer 2 pressure injection setup (General Valve Corp.) through prebroken microelectrodes pulled on a Sutter Brown-Flaming electrode puller. Electrode tips were filled by capillary action with the injection mixture and then backfilled with 150 mM KCI. The cells were visualized for microinjection by opening the skull at the level of the midbrain and reflecting back the cartilaginous flap of skull covering the hindbrain. Blood was directed away from the cells by a constant flow of chilled lamprey saline (Wickelgren, 1977) onto the hindbrain. The quality of the injections was monitored bycontinuous measurementof the resting membrane potential, with cells possessing 20 mV or more resting membrane potential upon withdrawal of the electrode surviving 80% of the time. Agents to be microinjected were diluted in 188 mM PIPES buffer (pH 7.4). Fast green (0.5% final concentration) and purified rabbit IgG (5 mg/ml; Sigma) were added to identify cells during microinjection (Figure 1) and tissue processing, re-

spectively. DMSO (2.5%) was used to make up stocks of OA, DiC8, and sphingosine. Injection controls mentioned in Results consisted of PIPES buffer, Fast green, and rabbit IgG in theabove concentrations with or without DMSO as appropriate. It should be noted that all drug and peptide concentrations mentioned here and in Results are concentrations in the electrode tip, not intracellular concentrations following microinjection, which are probably 28 to 188fold lower than the concentrations given. lmmunocytochemistrywas performed on 6 urn transverse sections of paraffin-embedded, Bouins-fixed lamprey heads. Sections were deparaffinized in xylene or Histoclear and then examined to determine whether they contained sections through ABC somata and dendrites. They were then rehydrated, quenched in excess HzOv washed, and incubated overnight with primary at 4°C. Antibody staining was revealed using an biotinylated secondary Extravidin-HRP kit (Sigma) with diaminobenzidine as the chromagen.

Antibodies Recognizing Phosphorylated lamprey NFs Lampreys have only one NF subunit, as opposed to the three found in most mammalian NFs. The lamprey NF protein possesses immunological and functional characteristics of all three mammalian NFsubunits (Pleasureetal., 1989).Theantibodyused to measure NF phosphorylation in most of the experiments described in this study (RM034; generously provided by Dr. Virginia Lee) has been extensively characterized by Dr. Lee and her colleagues and has been shown to recognize specifically the multiphosphorylation site on the carboxy terminal sidearm domains of mammalian NF-M and NF-H onlywhen this site is heavily phosphorylated (Pleasure et al., 1989; Lee et al., 1986, 1988). The site with which this monoclonal antibody cross-reacts in the lamprey has also been characterized biochemically and appears to be highly homologous to the mammalian site (Pleasure et al., 1989). Two other antibodies known to interact with this site in mammals (SMl31, obtained from Sternberg-Myer, and RM062, provided by Dr. Virginia Lee) were also used in some experiments and gave results similar to those obtained with RM034. Staining with RM062 is shown in Figure 28; staining with SMl31 is not shown. Identification of Injected Cells Serial transverse sections were cut through the region of the hindbrain containing the ABC and MC somata and dendrites. Every third section was taken and stained using biotinylated antirabbit IgG and then processed as described above to reveal the injected cells (examples are shown in Figures 2-5). Thus, every section stained for phosphorylated NFs was adjacent to a section stained for IgC, making it possible to identify injected cells unambiguously by their location and outline within the section. Most (295%) of the injected cells identified in sections appeared to be in good condition, with no beading of processes or vacuolization of the cytoplasm. It is likely that any cells seriously injured by microinjection died soon afterward and became undetectable by the time of analysis. Only cells that were indistinguishable from adjacent noninjected cells by the above criteria were selected for further analysis. Data Analysis Injected cells were scored by comparing them with adjacent noninjected cells in the same section of the same type (i.e., MCs were compared only with MCs and ABCs only with other ABCs). This allowed us to avoid artifacts caused by slide to slide and section to section variations in staining intensity due to processing. A total of 472 out of a total of 710 surviving filled cells in approximately 588 lampreys were scored. Each experiment was scored independently by two investigators, neither of whom knew the identity of the experiment or whether it involved a control or experimental substance. In experiments in which close axotomy had been performed on all of the MCs and ABCs (and all cells were expected to have phosphotylated somatodendritic NFs), injected cells were scored as staining more intensely than any adjacent noninjected cells; staining within the range

Neuron 624

of noninjected cells (i.e., either a noninjected cell stained with the same intensity or some noninjected cells stained both more and less intensely than the injected cell); or staining less intensely than any noninjected cell. In experiments in which somatodendritic NF phosphorylation was induced by microinjection of intact cells, injected cells were scored as not staining or staining at the same background level as adjacent noninjected cells; or staining detectably more than noninjected cells for phosphorylated NFs. The statistical significance of differences between the results of experimental and control injections was determined using the chi-square test. Acknowledgments We thank Dr. Jun Yao and Ms. Marian Slaneyfor technical assistance and Dr. Gloria Lee for her critical review of the manuscript. Wealsothank Dr.VirginiaLeefor hergenerousgiftof antibodies RM034 and RM062 and Dr. Mark Warner for his gift of DiC8. This work was supported by NIH grants NS 29281 (C. F. H.), NS 29031, and AC 86601 (K. S. K.) and by American Paralysis Association grant HA-9107 to G. F. H. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertiserned in accordance with 18 USC Section 1734 solely to indicate this fact. Received

May

20,1992;

revised

January

G. S., and DiLullo, of a neurofilament

C. (1985). Slow posttranslational modprotein. J. Cell Biol. 700,1799-1804.

Bentley, D. R., and Toroian-Raymond, finding by pioneer neurone growth by cytochalasin treatment. Nature

A. (1986). Disoriented pathcones deprived of filopodia 323, 712-715.

Black, M. M., Keyser, P., and Sobel, E. (1986). Interval between the synthesis and assembly of cytoskeletal proteins in cultured neurons. J. Neurosci. 6, 1004-1012. Borgens, R. B., Jaffe, L. F., and Cohen, M. J. (1980). Large and persistent electrical currents enterthetransected lamprey spinal cord. Proc. Natl. Acad. Sci. USA 77,1209-1213. Boulton, T. G., Nye, S., Robbins, D., Ip, N. Y., Radziejewska, E., Morgenbesser, S., DePinho, R., Panayotatos, N., Cobb, M., and Yancopoulos, G. (1991). ERKs: a family of protein-serine/threo nine kinases that are activated and tyrosine phosphorylated in response to insulin and NGF. Cell 65, 663-675. Bridgman, P. C., and Dailey, M. E. (1989). The organization of myosin and actin in rapidly frozen nerve growth cones. J. Cell Biol. 708, 95-109. Caceres, A., and Kosik, K. S. (1990). by tau antisense oligonucleotides rones. Nature 343, 461-463.

Inhibition of neurite polarity in primary cerebellar neu-

Cohen, P. (1988). Protein phosphorylation Proc. R. Sot. Lond. (B) 234, 115-144. Cohen, P., Holmes, C. F. B., and Tsukitani, acid: a new probe for the study of cellular Biochem. Sci. 5, 98-102.

and hormone

action.

Y. (1990). regulation.

Okadaic Trends

Drager, U., and Hofbauer, A. (1984). Antibodies to heavy neurofilamentsubunitdetect asubpopulation of ganglion cells in retina. Nature 309, 624-626. Fine, R. E., and Bray, Nature 234, 115-118.

D. (1971).

Actin

in growing

nerve

cells.

Goldstein, M. E., Cooper, H. S., Bruce, J., Carden, M. I., Lee, V. M.-Y., and Schlaepfer, W. W. (1987). Phospholylation of neurofilament proteins and chromatolysis following transection of rat sciatic nerve. J. Neurosci. 7, X86-1594. Hall, C. F., and Cohen, M. J. (1983). Extensive dendritic sprouting induced by close axotomy of central neurons in the lamprey. Science 222, 518-521. Hall,

C.

F., and

Cohen,

M. J. (1988).

The

pattern

and retraction induced by axotomy 1. Neurosci. 8, 3584-3597.

of lamprey

central

Hall, G. F., Poulos, A., and Cohen, M. 1. (1989). Sprouts emerging from the dendrites of axotomized lamprey central neurons have axonlike ultrastructure. J. Neurosci. 9, 588-599. Hall, G. F., Lee, V. M.-Y., and Kosik, K. S. (1991). destabilization and neurofilament phosphorylation ndritic sprouting after close axotomy of lamprey rons. Proc. Natl. Acad. Sci. USA 88, 5016-5020. Hannun, Y.A., Loomis, Sphingosine inhibition bol dibutyrate binding Chem. 267,12604-12609.

Microtubule precede decentral neu-

C. R., Merrill,A. H., and Bell, R. M. (1986). of protein kinase C activity and of phorin vitro and in human platelets. J. Biol.

Haystead,T. A., Weiel,J., Litchfield, J. E.,Tsukitani, D. W., Fischer, Y., and Krebs, E. G. (1990). Okadaic acid mimics the action of insulin in stimulating protein kinase activity in isolated adipocytes: the role of phosphatase 2a in attenuation of the signal. J. Biol. Chem. 265, 16571-16580. Heacock, A. M., and Agranoff, B. W. (1976). Enhanced labelling of a retinal protein during regeneration of optic nerve in goldfish. Proc. Natl. Acad. Sci. USA 73, 828-832. Heasley, L. E., and Johnson, G. L. (1989). Regulation of protein kinase C by nerve growth factor, epidermal growth factor and phorbol esters in PC12 cells. J. Biol. Chem. 264, 8646-8652. Hisanaga, S., Gonda, Y., Inagaki, M., Ikai, A., and Hirokawa, N. (1990). Effects of phosphorylation of the neurofilament L protein on filamentous structures. Cell Reg. 7, 237-248.

25,1993.

References Bennett, ification

sprouting neurons.

of dendritic

Hoffman, P. N., Cleveland, D. W., Griffin, J. W., Landes, P. W., Cowan, N. J., and Price, D. L. (1987). Neurofilament gene expression: a major determinant of axonal caliber. Proc. Natl. Acad. Sci. USA 84, 3472-3476. Hoffmann, P. N., and Lasek, R. J. (1980). Axonal cytoskeleton in regenerating motor neurons: change. Brain Res. 202, 317-333. Holliday, Calcium cultured

transport constancy

of the and

J.,Adams, R. J., Sejnowski,T. J., and Spitzer, N. C. (1991). induced release of calcium regulates differentiation of spinal neurons. Neuron 7, 787-796.

House, C., and pseudosubstrate 238, 1726-1728.

Kemp, 6. E. (1987). Protein kinase C contains a prototype in its regulatory domain. Science

Klumpp, S., Cohen, P., and Schultz, J. (1990). Okadaic acid, an inhibitor of protein phosphatase 1 in Paramecium, causes sustained Ca++ dependent backward swimming in response to de polarizing stimuli. EMBO J. 9, 685-689. Knops, J., Kosik, and McConlogue, ronal cell induces 733.

K. S., Lee, C., Pardee, J. D., Cohen-Gould, L., L. (1991). Overexpression of tau in a nonneulong cellular processes. J. Cell Biol. 774,725

Koizumi, S., Contreras, M. L., Matsuda, Y., Hama, P., and Curoff, C. (1988). K-252a, a specific inhibitor of nerve growth factor on PC12 cells. J. Neurosci.

T., Lazarovici, of the actions 8, 715-721.

Lanners, H. N., and Grafstein, B. (1980). Early stages of axonal regeneration in the goldfish optic tract: an electron microscopic study. J. Neurocytol. 9, 733-751. Lee, V. M.-Y., Otvos, L., Carden, M. J., Hollosi, M., Dietzschold, B., and Lazzarini, R. A. (1988). Identification of the major multiphosphorylation site in mammalian neurofilaments. Proc. Natl. Acad. Sci. USA 85,1998-2002. Lee, V. M.-Y., Carden, M. J., and Trojanowski, J. Q. (1986). Novel monoclonal antibodies provideevidence forthe in situ existence of a non phosphorylated form of the largest neurofilament subunit. 1. Neurosci. 6, 850-858. Letourneau, P. C. (1981). lmmunocytochemical evidence for colocalization in neurite growth cones of actin and myosin and their relationship to cell-substratum adhesions. Dev. Biol. 85, 113-122. Letourneau, and “push”

P. C., Shattuck,T.A., in neurite elongation:

and Ressler,A. observations

H. (1987).“Pull” on the effects of


different Cytoskel.

concentrations 8, 193-289.

Block

NF Phosphorylation

of cytochalasin

In Situ

B and taxol.

Lieberman, A. R. (1971). The axon reaction: a review pal features of perikaryal responses to axon injury. robiol. 74, 49-124.

Cell Motil.

of the princiInt. Rev. Neu-

Linda, H., Risling, M., and Cullheim, S. (1985). “Dendraxons” regenerating motoneurons in the cat: do dendrites grow axons after central axotomy? Brain Res. 358, 329-333.

in new

McKerracher, L., Vidal-Sanz, M., and Aguayo, A. J. (1998). Slow transport rates of cytoskeletal proteins change during regeneration of axotomized retinal neurons in adult rats. J. Neurosci. 70, 641-648. Marsh, L., and Letourneau, P. C. (1984). Growth of neurites without filopodial or lamellopodial activity in the presence of cytochalasin B. J. Cell Biol. 99, 2841-2847. Murphy, A. D., and Kater, S. B. (1988). Sprouting regeneration of an identified neuron in Helisoma. 251-272.

and functional Brain Res. 186,

Nel,A., Hanekom, C., Rheeder,A., Williams, K., Pollack, S., Katz, R., and Landreth, G. (1998). Stimulation of MAP2 kinase activity in T lymphocytes by anti-CD3 anti-Ti monoclonal antibody is partially dependent on protein kinase C. J. Immunol. 744,26832689. Nixon, R. A., and Sihag, R. K. (1991). Neurofilament phosphorylation: a new look at regulation and function. Trends Neurosci. 74, 501-505. Nordlander, R. H., and Singer, M. (1982). tion of growth cones in the developing Dev. Brain Res. 4, 181-193.

Morphology Xenopus

and posispinal cord.

Oblinger, M. M., Szumlas, R. A., Wong, J., and Liuzzi, F. J. (1989). Changes in cytoskeletal gene expression affect the composition of regenerating axonal sprouts elaborated by dorsal root ganglion neurons in vivo. J. Neurosci. 9, 2645-2653. Pleasure, S. J., Selzer, M. E., and Lee, V. M.-Y. (1989). Lamprey neurofilaments combine in one subunit the features of each mammalian neurofilament protein, but are highly phosphorylated only in large axons. 1. Neurosci. 9, 698-709. Roederer, E., and Cohen, M. J. (1983). Regeneration of an identified neuron in the cricket: control of sprouting from dendrites, soma and axon. J. Neurosci. 3, 1835-1847. Schacher, by Aplysia hemolymph Neurosci.

S., and Proshansky, neurons in dissociated and the presence 3, 2403-2413.

Shaw, G., Winialski, and deafferentation of neurofilaments 460, 227-234.

E. (1983). Neurite regeneration tissue culture: modulation by of the initial axon segment. J.

D., and Reier, P. (1988). The effect of axotomy on phosphorylation dependent antigenicity in rat superior cervical ganglion. Brain Res.

Sihag, R. K., and Nixon, R. A. (1989). In vivo phosphorylation of distinct domains of the 70 kilodalton neurofilament subunit involves different protein kinases. J. Biol. Chem. 264, 457-464. Sihag, R. K., and Nixon, R. A. (1990). Phosphorylation of the amino-terminal head domain of the middle molecular mass 145 kda subunit of neurofilaments. J. Biol. Chem. 265,4166-4171. Sinicropi, D. V., and McElwain, amounts of cytoskeletal proteins of regenerating frog motoneurons.

D. L. (1983). Changes in the within the perikaryaand axons J. Cell Biol. 96, 248-247.

Skoff, R. P., and Hamburger, V. (1974). Finestructureof dendritic and axonal growth cones in embryonic chick spinal cord. J. Comp. Neurol. 753, 107-148. Smith, M. K., Colbran, R. J., and Soderling, T. R. (1998). Specificities of autoinhibitory domain peptides for four protein kinases. J. Biol. Chem. 265, 1837-1848. Strautman, distribution modulation

A. F., Cork, R. J., and Robinson, K. R. (1990). The of free calcium in transected spinal axons and its by applied electric fields. J. Neurosci. 70,3564-3575.

Szaro, B. G., Grant, P., Lee,V. M.-Y., and Gainer, H. (1991). Inhibition of axonal development after injection of neurofilament anti-

bodies 585.

into a Xenopus

laevis

Torvik, A. (1976). Central a reappraisal. Neuropathol.

embryo.

1. Comp.

Neurol.

308,576-

chromatolysis and the axon reaction: Appl. Neurobiol. 2, 423-432.

Wessells, N. K., Johnson, S. R., and Nutall, R. P. (1978). Axon initiation and growth cone regeneration in cultured motor neurons. Exp. Cell Res. 777, 335-345. Wickelgren, W. 0. (1977). Physiological and anatomical characteristics of reticulospinal neurones in lamprey. J. Physiol. 270, 89-116.