1949
Journal of Cell Science 107, 1949-1957 (1994)
Printed in Great Britain © The Company of Biologists Limited 1994
Dephosphorylation of the largest neurofilament subunit protein influences
the structure of crossbridges in reassembled neurofilaments
Takahiro Gotow1,*, Toshihisa Tanaka2, Yu Nakamura2 and Masatoshi Takeda2
Departments of 1Anatomy and 2Neuropsychiatry, Osaka University Medical School, Suita, Osaka 565, Japan
*Author for correspondence
SUMMARY
Phosphorylation-dependent change in electrophoretic
mobility is the most unique characteristic of NF-H, the
largest molecular mass subunit of the neurofilament. We
dephosphorylated NF-H using Escherichia coli alkaline
phosphatase, then reassembled it into neurofilaments with
NF-M and NF-L, and into NF-H filaments with NF-H
alone. We compared these dephosphorylated filaments
with control: projections by low-angle rotary-shadow,
crossbridges by quick-freeze deep-etch, and core filament
packing density by thin-section electron microscopy. Projections in the dephosphorylated filaments were basically
similar in structure to those in control, although there was
a tendency for them to be wider and less dense, especially
in NF-H filaments. Dephosphorylated filaments were still
able to form crossbridges between core filaments, but their
crossbridges were significantly wider, less dense, more
branched and more irregular than crossbridges in control,
and core filaments were more densely packed. These structural differences may be brought about by the removal of
phosphate groups from NF-H tail and consequent
reduction of electrostatic repulsion between adjacent crossbridges extending from the same core filament. The results
indicate that phosphorylation of NF-H is necessary for
forming well developed crossbridges, straight and at
constant intervals, like those of in vivo axonal neurofilaments.
INTRODUCTION
1985; Glicksman et al., 1987; Lee et al., 1987; Gotow and
Tanaka, 1994a). Morphological organization of NFs is also
different (Hirokawa et al., 1984; Tytell et al., 1988; Gotow and
Tanaka, 1994a): core filaments of axonal NFs align regularly
and are extensively crossbridged while those of perikaryal or
dendritic NFs are oriented randomly, are less crossbridged, and
are sometimes fascicled (Hirokawa et al., 1984). We think that
the most significant structural difference between axonal NFs
and perikaryal or dendritic NFs is the development of the
crossbridge. Well-developed crossbridges connect core
filaments and may align them parallel to each other. Lessdeveloped or irregular crossbridges, may be unable to align
core filaments parallel to each other. Conversely, where NF
parallel alignment is required crossbridges must be frequent,
where parallel alignment is not required crossbridges must be
rare. If crossbridges do not appear, core filaments contact each
other directly, as is seen in fascicled NFs in perikaryon or
dendrite (Hirokawa et al., 1984).
For the present study, to observe how NF ability to form
crossbridges is altered by dephosphorylation, we reconstructed
NFs from triplet proteins and from NF-H alone. We examined
these with low-angle rotary shadowing and with quick-freeze
deep etching. We obtained the triplet proteins from myelinated
axonal NFs because the NF-H of these NFs is heavily phosphorylated (Watson, 1991). Further, we believe that our quickfreeze deep-etch electron microscopy, with reassembled NFs
Neurofilaments (NFs) are composed structurally of two distinct
portions, a central 10-12 nm wide filament portion, called core
filament, and a 4-5 nm wide fluffy portion extending laterally
from the core filament called projections (Hisanaga and
Hirokawa, 1988; Mulligan et al., 1991; Gotow et al., 1992).
Projections form crossbridges when two or more NFs are close
enough to each other to interact (Gotow et al., 1992). The core
filament is considered to be constructed from the aminoterminal head and central helical rod domains of all subunit
proteins, and the projection from the carboxy-terminal tail
domains of NF-H and NF-M (Hisanaga and Hirokawa, 1988;
Mulligan et al., 1991; Shaw, 1991; Gotow et al., 1992). Of the
NF subunits, the two with larger molecular mass, NF-M and
NF-H, are provided with considerable amounts of phosphates
in their extremely extended carboxy-terminal tail domains
(Geisler et al., 1983, 1984; Julien and Mushynski, 1983;
Carden et al., 1985). Major phosphorylation sites are considered to be located at serine residue in the characteristically
repeated sequence of amino acids (Lys-Ser-Pro) of these tail
domains (Geisler et al., 1987; Lee et al., 1988).
The degree of phosphorylation of NF-M and NF-H, especially of NF-H, is different in each neuronal compartment,
being high in the axon, but extremely low in the perikaryon
and dendrite (Sternberger and Sternberger, 1983; Carden et al.,
Key words: neurofilament, phosphorylation, projection, crossbridge,
quick-freeze deep-etch electron microscopy
1950 T. Gotow and others
attached to mica (Heuser, 1983), provides a more accurate
image of the interactions of these molecules than any other
morphological approach.
MATERIALS AND METHODS
Isolation of crude NFs
Bovine spinal cords were obtained from a local slaughterhouse. The
white matter of the spinal cords was dissected, minced, and homogenized with a Dounce homogenizer in 0.85 M sucrose in 10 mM BisTris buffer, pH 6.5, containing 0.2 M KCl, 10 mM MgCl2, 2 mM
EGTA and 1% Triton X-100. Myelinated axons were obtained by the
axonal floatation method as described by Yen et al. (1976). Myelin
sheaths were removed by osmotic shock and vigorous homogenization in the Bis-Tris buffer, and the pellet was centrifuged at 100,000
g for 30 minutes at 4°C, and recentrifuged in 0.85 M sucrose in the
Bis-Tris buffer (~1.0 mg protein/ml).
Purification of NF triplet proteins
Crude NFs were dissoved in 10 mM Bis-Tris buffer, pH 6.5, containing 8 M urea, 1 mM EGTA and 1% β-mercaptoethanol, and the
NF subunits isolated by gradient chromatography (0-0.25 M NaCl)
with DEAE cellulose (DE-52; Whatman) (Liem, 1982; Tokutake et
al., 1984). After glial fibrillary acidic protein passed through the
column, the first protein peak obtained was NF-H. The proteins of the
second broad peak obtained were NF-M and NF-L. Each purified NF
Fig. 1. SDS-PAGE of NFs (A) and NF-H filaments (B). (A) NFs
reassembled from purified triplet proteins. (B, lanes 1-3) NF-H
filaments reassembled from purified NF-H alone. (B, lane 4)
Untreated NF-L and NF-M for comparison. Lanes 1-3, incubated at
37°C for 5 hours in 50 mM Tris buffer, pH 8.5. Lane 1, control. Lane
2, NF-H dephosphorylated with 2 units of alkaline phosphatase
(ALP). Lane 3, NF-H dephosphorylated with 20 units of ALP.
(H,M,L) Positions for NF-H, NF-M, NF-L. Partial
dephosphorylation of NF-H is demonstrated by mobility in lane 2
and almost complete dephosphorylation by increasing mobility in
lane 3.
Fig. 2. Low-angle rotary-shadow electron micrographs of reassembled NFs. (A) Control. (B) Dephosphorylated with 20 units of ALP. There
appears to be no distinct difference in structure between control and dephosphorylated NFs. Most are composed of long core filaments and
numerous projections, but there are also many with short or fragmented core filaments. Projections are 4-5 nm wide. Wider ones (~5 nm)
(arrows) appear especially in dephosphorylated NFs. Bar, 100 nm.
Phosphorylation-dependent NF structure 1951
Fig. 3. Quick-freeze deep-etch electron micrographs of reassembled NFs. (A) Control. (B) Dephosphorylated with 2 units of ALP.
(C) Dephosphorylated with 20 units of ALP. (A) NFs consist of core filaments (arrowheads) and crossbridges (arrows). Where core filaments
are close together, crossbridges are thin (4-5 nm), straight, and relatively regular. (B and C) Crossbridges are less conspicuous and less regular.
Although some are as thin as those in control, most are wider and branched (arrows). They appear the same in both 2 units and 20 units ALPtreated samples. In (C), two or more core filaments are close enough to be almost attached to each other without the intervention of
crossbridges (arrowheads). Bar, 100 nm.
subunit was collected and concentrated to a protein concentration of
1.0 mg/ml. They were then dialysed in 50 mM Tris-HCl (pH 8.5) containing 2 mM MgCl2, 1 mM EGTA and 0.1 mM DTT.
Enzymic dephosphorylation of NF-H and reassembly of
NFs and NF-H filaments
NF-H was dephosphorylated by incubation with 2 units or 20 units of
E. coli alkaline phosphatase (ALP) (type III; Sigma Chemicals) per
milligram of protein in 50 mM Tris buffer, pH 8.5, containing 2 mM
MgCl2, 1 mM EGTA, 0.1 mM DTT, 0.1 mM phenylmethylsulfonyl
fluoride and 0.1 µg/ml leupeptin for 5 hours at 37°C. The samples were
applied to a Sephacryl S-200 (Pharmacia) gel filtration column for
removing ALP. Control NF-H was incubated in the same way without
ALP. Dephosphorylated NF-H and control NF-H (~1 mg/ml) were
polymerized in 0.1 M MES buffer (pH 6.5) containing 0.17 M NaCl
and 1 mM EDTA at 37°C for 5 hours for NF-H filaments, and in combination with NF-M and NF-L (~1 mg/ml), at volume ratio 1:1, for NFs.
Some NFs and some NF-H filaments were analyzed by SDS-PAGE
1952 T. Gotow and others
on 7.5% gels (Laemmli, 1970) and stained with Coomassie Blue. The
remainder were used for electron microscopy.
Quick-freeze deep etching
Quick-freeze deep-etch electron microscopic samples of reassembled
NFs and NF-H filaments were prepared according to the method of
Heuser (1983) (Gotow et al., 1992). A fixed tissue block was placed on
an aluminum freezing disc and the surface covered with a thin layer of
homogenized mica flakes. Then a small amount (~10 µl) of NF solution
(0.5 mg protein/ml) was dropped on the thin layer of mica flakes and
the specimen quick frozen in a Polaron E7200 by slamming against a
copper block cooled by liquid helium. Frozen sample was mounted on
a Balzers BAF 400D apparatus, and only the surface fractured carefully
in a vacuum of more than 2× 10 mbar. The specimen holder was at
−120°C at fracture, and warmed to −95°C for 3-4 minute etching. The
sample was then rotary replicated with platinum/carbon at an angle of
25°. The thickness of replica (~2.5 nm) was controlled with a quartz
crystal monitor. Replicas were placed in household bleach, treated with
hydrofluoric acid, picked up on grids and examined in a JEOL 100CX
electron microscope operated at 80 kV.
Low-angle rotary shadowing
Low-angle rotary-shadow electron microscopic samples were
prepared according to the method of Tyler and Branton (1980). A
small amount of the reassembled filaments in the buffer solution was
mixed with glycerol to adjust the concentration of NF protein to 50100 µg/ml and of glycerol to 50%. Then a small sample of the mixture
(100 µl) was sprayed onto freshly cleaved mica attached to the
specimen holder of Balzers freeze-fracture apparatus. The sample was
mounted in this freeze-fracture apparatus at room temperature, then
rotary shadowed at an angle of 6°, in a way similar to the above quickfreeze deep-etch method. Replicas in mica were floated off onto water,
picked up on grids and examined in the electron microscope.
Thin section
Control and dephosphorylated NFs were precipitated by centrifugation at 15,000 g for 30 minutes and then fixed with 2% glutaraldehyde in 50 mM Bis-Tris buffer, pH 6.5, containing 0.17 M NaCl and
1 mM EDTA. The samples were then post-fixed with 2% OsO4, and
stained en bloc with 2% uranyl acetate. After dehydration with
ethanols, the samples were embedded in Luveak 812. Thin sections
were cut, stained with uranyl acetate and lead citrate, and examined
in the electron microscope.
Quantification
Projection and crossbridge structure of control and dephosphorylated
NFs and NF-H filaments were compared quantitatively using morphometric analyses of quick-freeze deep-etch and low-angle rotaryshadow preparations. Length and density of projections were
measured in rotary-shadow samples, and width, density, and length
of crossbridges from core filament wall to the first branching, in deepetch samples. Packing density of core filaments was also compared in
thin sections. Measurements were made using a pair of vernier
calipers on micrographs at final magnification of more than ×150,000
for density, length and packing density, and more than ×300,000 for
width. Density of projections and crossbridges was calculated along
one side of core filaments. Magnification was calibrated by phyrophyllite (0.45 nm). All structural parameters measured were statistically compared by Student’s t-test.
RESULTS
NFs reassembled from triplet proteins with
dephosphorylated NF-H
When these NFs, with NF-H dephosphorylated either by 2
units or 20 units of alkaline phosphatase (ALP), were checked
by SDS-PAGE, dephosphorylation of NF-H was found to be
almost complete with 20 units of ALP, increasing in mobility
to the level of the upper portion of NF-M (Fig. 1A), and partial
with 2 units of ALP, intermediate in mobility between control
and 20 units of ALP (Fig. 1A).
Low-angle rotary shadowing
Low-angle rotary shadowing showed no distinct difference in
structure between control and dephosphorylated NFs (Fig. 2).
Core filaments were 20-25 nm wide, smooth and usually up to
1 µm long. Projections were 80-90 nm long, and extended
regularly from the core filaments. Projections were 4-5 nm
wide, with a few 5-6 nm wide notably in dephosphorylated NFs
(Fig. 2B). The density of projections appeared to be
unchanged. Both control and dephosphorylated NFs were fundamentally the same in profile as untreated NFs that were
reassembled without incubation at 37°C (Gotow et al., 1992),
although short fragmented or rod-like core filaments were more
conspicuous in the incubated samples.
Quick-freeze deep etching
Deep etching showed conspicuous differences in structure
between control and dephosphorylated NFs. Core filaments
looked the same, elongated and 12-14 nm wide, but crossbridges, 20-70 nm long, were considerably different. Core
filaments were narrower in quick-freeze deep etching than in
low-angle rotary shadowing due to the absence of glycerol decoration, and thus were closer to their natural size. Crossbridges
formed from projections were much shorter than two projection lengths or even one. After treatment with ALP, most crossbridges were wider, less dense, less regular in interval and
more branching than those in control (Fig. 3A-C). In dephosTable 1. Differences in width, density and length of NF
and NF-H filament crossbridges and packing density of
NF core filaments in control, 2 units ALP- and 20 units
ALP-treated samples
Crossbridge
Treatment
NF
NF-H filaments
Width (nm)
Control
2 units ALP
20 units ALP
4.7±1.0
5.9±1.2
6.3±1.5
4.7±1.1
6.1±1.3
6.2±1.4
Density
(no./µm)
Control
2 units ALP
20 units ALP
30.4±7.2
22.3±7.0
17.3±6.8
34.3±10.1
22.0±8.9
18.6±8.7
Length (nm)
Control
2 units ALP
20 units ALP
30.8±10.2
23.2±8.4
20.9±6.9
31.6±9.6
18.3±6.8
16.0±7.1
Packing density
(no./µm)
Control
2 units ALP
20 units ALP
NF core filaments
312.6±20.9
373.5±31.5
416.9±35.8
Values are mean ± s.d., and were obtained from three independent samples.
Width, density and length were from quick-freeze deep-etch samples, and
packing density was from thin section samples. Length of crossbridges means
the distance from core filament wall to the first branching. By treatment with
2 units of ALP, all these morphological parameters are significantly changed
(P<0.001) in both filaments. There are also significant differences in NF
crossbridge density (P<0.01) and core filament packing density (P<0.001)
between 2 units ALP- and 20 units ALP-treated samples.
Phosphorylation-dependent NF structure 1953
Fig. 4. Thin-section electron micrographs of reassembled NFs in the centrifuged precipitates. (A) Control. (B) 2 units ALP-treated sample.
(C) 20 units ALP-treated sample. Core filaments align randomly. Packing density of core filaments increases with ALP concentration. Bar, 200 nm.
phorylated samples, even where the core filaments were close
and relatively parallel to each other, crossbridges were less
frequent and more branched. Sometimes there were hardly any
crossbridges and core filaments were in almost direct contact
with each other (Fig. 3C).
Quantification showed significantly that the structural differences were dependent on ALP concentrations (Table 1, Fig. 7).
The width of crossbridges was less than 5 nm in control but
around 6 nm in dephosphorylated samples (Table 1, Fig. 7A).
The density of crossbridges was reduced to little more than half
(57%) with 20 units of ALP. The standard deviations for
densities were almost unchanged even when the means were
reduced drastically (Table 1, Fig. 7B), indicating that variations
of the density become higher in dephosphorylated samples,
which means that crossbridges are more irregular in interval.
The criterion for branching was length of crossbridge portions
from the core filament wall to the first branch. With dephos-
phorylation, those lengths were reduced gradually down to two
thirds of the control with 20 units of ALP (Table 1, Fig. 7C).
Thin section
As close apposition of core filaments appeared rather frequently
in deep-etch samples, we compared packing density of control
and 2 units and 20 units ALP-treated NFs using thin sections
from corresponding levels of the three centrifuge-precipitated
samples (Fig. 4). Core filaments looked shorter than in rotary
shadowing and deep etching, probably due to various treatments
for thin section, such as chemical fixation, dehydration and
embedding. Since they aligned randomly, density was calculated where they were roughly perpendicular to the plane of
section. As expected, the packing density significantly
increased in dephosphorylated NFs and there was also significant difference between the two ALP-treated samples (Figs 4,
7D, Table 1).
1954 T. Gotow and others
Fig. 5. Low-angle rotary-shadow images of NF-H filaments. In rotary shadowing filaments are structurally similar in control (A) and
dephosphorylated (B) samples. Projections (arrows), however, are wider and less frequent in B, although their lengths are unchanged. Bar, 100 nm.
NF-H filaments reconstructed from NF-H alone
Since differences in structure of crossbridges and ability to
form bridges were observed in NFs reassembled with dephosphorylated NF-H, and since NF-H is the most essential
component for forming crossbridges (Hirokawa, 1984;
Leterrier and Eyer, 1987; Hisanaga and Hirokawa, 1988;
Gotow et al., 1992), we expected such difference to be
enhanced in filaments reassembled from dephosphorylated NFH alone (Fig. 1B). Further, since filaments have core filaments,
projections and crossbridges almost the same in appearance
and density as those of NFs, even though their core filaments
are much shorter than those of NFs, we considered NF-H
filaments comparable to NFs for examining characteristics of
crossbridges, and better than NFs for showing the effect of
ALP on projections or crossbridges.
Low-angle rotary shadowing
Both control and dephosphorylated samples showed rod-like
short core filaments ~20 nm wide, 50-300 nm long, and
numerous projections 4-5 nm wide, 70-80 nm long (Fig. 5A,B).
Core filaments were much shorter than those in NFs constructed from triplet proteins, and still shorter in dephosphorylated samples than in control. Projections were clearer in
dephosphorylated samples, wider (~5 nm as compared with ~4
nm in control) and less frequent, ~15/100 nm in 20 units of
ALP, ~20/100 nm in control (Fig. 5A,B).
Quick-freeze deep etching
Control NF-H core filaments were usually narrower, ~10 nm,
than in low-angle rotary shadowing, and also longer, 100-400
nm, probably because they were mechanically broken during
vigorous spraying onto the mica. Control crossbridges were
numerous, straight, 4-5 nm wide and 20-70 nm long (Fig. 6A).
Dephosphorylated crossbridges were wider, less frequent, less
regular and more branched (Fig. 6B,C). These structural
changes increased in proportion to increase of ALP concentration (Table 1, Fig. 7), and were very similar in both quality
and quantity to those observed in NFs. This is a strong indication that the effects of ALP on the crossbridges are reproducible. The width of crossbridges was less than 5.0 nm in
control but 6.0 nm and more with dephosphorylation (Table 1,
Fig. 7A). Density declined with increasing ALP concentration
(Table 1, Fig. 7B). Although mean density was reduced by
almost half, to 54% in 20 units of ALP, standard deviations
were little changed, indicating an increase of irregularity in
interval by dephosphorylation. The incease in width and
decrease in density of crossbridges observed in the deep
etching appeared to correspond to the increase in width and
decrease in density of projections observed in the rotary
Phosphorylation-dependent NF structure 1955
Fig. 6. Quick-freeze deep-etch images of NF-H filaments. In control (A), the core filaments (arrowheads) are extensively crossbridged. Even where
they are far apart, there are relatively many bridges (arrows). In dephosphorylated samples (B and C), crossbridges are wider, more irregular and
more branched. The branchings are more conspicuous than those of NFs (Fig. 3) and form honeycombed pattern (arrows). Bar, 100 nm.
shadowing. Length of crossbridge portions from core filament
to the first branch declined gradually to about half, 51%, of
control (Table 1, Fig. 7C). Branching was much more notice-
able in dephosphorylated NF-H filaments than in dephosphorylated NFs, even where the core filaments were close in a
honeycomb-like pattern (Table 1, Fig. 6B,C).
1956 T. Gotow and others
Fig. 7. Graphic representation of the influence of dephosphorylation
on the morphological parameters. Differences in width (A), density
(B) and length (C) of NF and NF-H filament crossbridges, and
packing density (D) of NF core filaments in control, 2 units ALPand 20 units ALP-treated samples. Values in these graphs were
obtained from Table 1.
DISCUSSION
The drastic shift in electropholetic mobility of NF-H on SDSPAGE after dephosphorylation (Carden et al., 1985; Glicksman
et al., 1987; Lee et al., 1987) suggests that there must be a phosphorylation-dependent change in function and conformation of
the NF-H tail domain (Shaw, 1986; Glicksman et al., 1987;
Eyer and Leterrier, 1988; de Waegh et al., 1992; Gotow and
Tanaka, 1994a). However, no substantial evidence has yet been
produced that the carboxyl-terminal tail domain of NF-H is
structurally alterable by phosphorylation (cf. Matus, 1988;
Nixon and Sihag, 1991; Nixon and Shea, 1992). Using lowangle rotary shadowing and quick-freeze deep etching,
Hisanaga and Hirokawa (1989) have tried to see in crude NFs
the relationship between phosphorylation and change in projection structure, or ability to form crossbridges. They found no
morphological influence of NF phosphorylation on projections
or crossbridges, and concluded that change in phosphorylation
state did not affect the ability to form crossbridges.
For the present study we followed that of Hisanaga and
Hirokawa (1989) with slightly modified methods. We used NFs
polymerized from NF-L, NF-M and dephosphorylated NF-H,
and NF-H filaments polymerized from dephosphorylated NF-H
alone. We also used the mica technique for quick-freeze deep
etching designed by Heuser (1983) to prevent NF proteins from
becoming tightly condensed by centrifugation, and allow them
to be freely suspended in the physiological solution and lightly
attached on mica surface. In our specimens, although changes
were detectable in the rotary-shadow sample, especially in the
NF-H filament, changes were remarkable in the deep-etch
sample: in both kinds of reassembled filaments crossbridges
were more branched and less regular in interval in dephosphorylated samples than in phosphorylated (control) ones. And
following our observations in the deep-etch sample, the increase
in width of projections was noticeable in the rotary-shadow
sample. In the rotary-shadow samples of Hisanaga and
Hirokawa (1989) also, although they did not mention it, the
profile of projections is more evident in dephosphorylated than
in phosphorylated sample. Their rotary-shadow projections,
however, as well as ours, appeared the same in extension in both
control and dephosphorylated conditions. We believe this
image may be related to processes in the preparation, such as
heavy decoration of glycerol or strong spraying.
It may be thought that the structural alteration of crossbridges we observed is brought about by proteolysis with endogeneous proteases, because dephosphorylation makes NF-H
degrade easily (Goldstein et al., 1987; Pant, 1988). We reject
this as a possibility, however, because the projections we
observed, a possible morphological representation of polypeptide chains, were not influenced in length, and the SDS-PAGE
showed no detectable band for degradation of NF-H in dephosphorylated samples. We found a similar relationship between
NF-H phosphorylation and crossbridge development in our
recent study using aluminum-intoxicated rabbit brain (Gotow
and Tanaka, 1994b): NFs accumulated by aluminum in
dendrite or perikaryon, with core filaments less regularly
oriented but similar in density to those in axon, had significantly less phosphorylated NF-H than axonal NFs, and fewer,
shorter and more branched crossbridges.
Increase in width and branching of crossbridges may come
from a common cause. When their highly charged phosphate
groups are removed by ALP, NF-H tail polypeptide chains
constituting crossbridges are unable to repel adjacent bridges
electrostatically (Carden et al., 1987; Glicksman et al., 1987;
Matus, 1988; de Weagh et al., 1992), and tend to be close to
or attached to each other, resulting in greater width or more
branching. Thus, phosphorylation may be necessary for the
crossbridges to electrostatically repel adjacent bridges deriving
from the same core filaments and to extend straight and
regularly, so that neighboring core filaments may align parallel
to each other. The charge of these phosphates may be influenced by adjacent highly charged amino acids, such as
glutamic acid and lysine residues (Geisler et al., 1985; Myers
et al., 1987; Lees et al., 1988). Where phosphate groups are
considerably reduced, narrower space between core filaments
may be due to the inability of projections to extend straight
from the filament. The decrease in frequency of crossbridges
also appears to be related to this inability of NF-H tail amino
acids to extend, and a consequent difficulty of NF-H projections to interact with opposed NF-H projections.
This work was supported in part by Grant-in-Aid no. 03807001,
from the Ministry of Education, Science and Culture, of Japan. We
thank Dr Gen Kanayama for help in preparing SDS-PAGE, Mr
Nobukazu Komori for some technical assistance and Miss Constance
Purser for linguistic help.
Phosphorylation-dependent NF structure 1957
REFERENCES
Carden, M. J., Schlaepfer, W. W. and Lee, V. M.-Y. (1985). The structure,
biochemical properties, and immunogenicity of neurofilament peripheral
regions are determined by phosphorylation state. J. Biol. Chem. 260, 98059817.
Carden, M. J., Trojanowski, J. Q, Schlaepfer, W. W. and Lee, V. M.-Y.
(1987). Two-stage expression of neurofilament polypeptides during rat
neurogenesis with early establishment of adult phosphorylation patterns. J.
Neurosci. 7, 3489-3504.
de Waegh, S. M., Lee, V. M.-Y. and Brady, S. T. (1992). Local modulation of
neurofilament phosphorylation, axon caliber, and slow axonal transport by
myelinating Schwann cells. Cell 68, 451-453.
Eyer, J. and Leterrier, J. F. (1988) Influence of the phosphorylation state of
neurofilament proteins on the interactions between purified filaments in vitro.
Biochem. J. 252, 655-660.
Geisler, N., Kaufmann, E, Fischer, S., Plessmann, U. and Weber, K. (1983).
Neurofilament architecture combines structural principles of intermediate
filaments with carboxy-terminal extensions increasing in size between triplet
proteins. EMBO J. 2, 1295-1302.
Geisler, N., Fischer, S., Vandekerckhove, Plessmann, J. U. and Weber, K.
(1984). Hybrid character of a large neurofilament protein (NF-M):
intermediate filament type sequence followed by a long and acidic carboxyterminal extension. EMBO J. 3, 2701-2706.
Geisler, N., Fischer, S., Vandekerckhove, J., Damme, J. V., Plessmann, U.
and Weber, K. (1985). Protein-chemical characterization of NF-H, the
largest mammalian neurofilament component; intermediate filament-type
sequences followed by a unique carboxy-terminal extension. EMBO J. 4, 5763.
Geisler, N., Vandekerckhove, J. and Weber, K. (1987). Location and
sequence characterization of the high molecular mass neurofilament proteins
M and H. FEBS Lett. 221, 403-407.
Glicksman, M. A., Soppet, D. and Willard, M. B. (1987). Posttranslational
modification of neurofilament polypeptides in rabbit retina. J. Neurobiol. 18,
167-196.
Goldstein, M. E., Sternberger, N. H. and Sternberger, L. A. (1987).
Phosphorylation protects neurofilaments against proteolysis. J.
Neuroimmunol. 14, 149-160.
Gotow, T., Takeda, M., Tanaka, T. and Hashimoto, P. H. (1992).
Macromolecular structure of reassembled neurofilaments as revealed by the
quick-freeze deep-etch mica method: difference between NF-M and NF-H
subunits in the ability to form cross-bridge. Eur. J. Cell Biol. 58, 331-345.
Gotow, T. and Tanaka, J. (1994a). Phosphorylation of neurofilament H
subunit as related to arrangement of neurofilaments. J. Neurosci. Res. 37,
691-713.
Gotow, T. and Tanaka, J. (1994b). The organization of neurofilaments
accumulated in perikaryon following aluminium administration: relationship
between structure and phosphorylation of neurofilaments. Neuroscience (in
press).
Heuser, J. E. (1983). Procedure for freeze-drying molecules adsorbed to mica
flakes. J. Mol. Biol. 169, 155-195.
Hirokawa, N., Glicksman, M. A. and Willard, M. B. (1984). Organization of
mammalian neurofilament polypeptides within the neuronal cytoskeleton. J.
Cell Biol. 98, 1523-1536.
Hisanaga, S. and Hirokawa, N. (1988). Structure of the peripheral domains of
neurofilaments revealed by low angle rotary shadowing. J. Mol. Biol. 202,
297-305.
Hisanaga, S. and Hirokawa, N. (1989). The effects of dephosphorylation in
the structure of the projections of neurofilament. J. Neurosci. 9, 959-966.
Julien, J.-P. and Mushynski, W. E. (1983). The distribution of
phosphorylation sites among identified proteolytic fragments of mammalian
neurofilaments. J. Biol. Chem. 258, 4019-4025.
Laemmli, U. K. (1970). Cleavage of structural proteins during the assembly of
the head of bacteriophage T4. Nature 227, 680-685.
Lee, V. M.-Y., Carden, M. J., Schlaepfer, W. W. and Trojanowski, J. Q.
(1987). Monoclonal antibodies distinguish several differentially
phosphorylated states of the two largest rat neurofilament subunits (NF-H
and NF-M) and demonstrate their existence in the normal nervous system of
adult rats. J. Neurosci. 7, 3474-3488.
Lee, V. M.-Y., Otovos Jr, L., Carden, M. J., Hollosi, M., Dietzschold, B. and
Lazzarini, R. A. (1988). Identification of the major multiphosphorylation
site in mammalian neurofilaments. Proc. Nat. Acad. Sci. USA 85, 1998-2002.
Lees, J. F., Shneidman, P. S, Skuntz, S. F., Carden, M. J. and Lazzarini, R.
A. (1988). The structure and organization of the human heavy neurofilament
subunit (NF-H) and the gene encoding it. EMBO J. 7, 1947-1955.
Leterrier, J.-F. and Eyer, J. (1987). Properties of highly viscous gels formed
by neurofilaments in vitro. A possible consequence of a specific interfilament cross-bridging. Biochem. J. 245, 93-101.
Liem, R. (1982). Simultaneous separation and purification of neurofilament
and glial filament proteins from brain. J. Neurochem. 38, 142-150.
Matus, A. (1988). Neurofilament protein phosphorylation - where, when and
why. Trends Neurosci. 11, 291-292.
Mulligan, L., Balin, B., Lee, V. M.-Y. and Ip, W. (1991). Antibody labeling
of bovine neurofilaments: implication on the structure of neurofilament
sidearms. J. Struct. Biol. 106, 145-160.
Myers, M. W., Lazzarini, R. A., Lee, V. M.-Y., Schlaepfer, W. W. and
Nelson, D. L. (1987). The human mid-size neurofilament subunit: a repeated
protein sequence and the relationship of its gene to the intermediate filament
gene family. EMBO J. 6, 1617-1626.
Nixon, R. A. and Sihag, R. K. (1991). Neurofilament phosphorylation: a new
look at regulation and function. Trends Neurosci. 14, 501-506.
Nixon, R. A. and Shea, T. B. (1992). Dynamics of neuronal intermediate
filaments: a developmental perspective. Cell Motil. Cytoskel. 22, 81-91.
Pant, H. C. (1988). Dephosphorylation of neurofilament proteins enhances
their susceptibility to degradation by calpain. Biochem. J. 256, 665-668.
Shaw, G. (1986). Neurofilaments: abundant but mysterious neuronal
structures. BioEssays 4, 161-166.
Shaw, G. (1991). Neurofilament protein. In The Neuronal Cytoskeleton (ed. R.
D. Bourgoyne), pp. 185-214. Wiley-Liss, Inc. USA, New York.
Sternberger, L. A. and Sternberger, N. H. (1983). Monoclonal antibodies
distinguish phosphorylated forms of neurofilaments in situ. Proc. Nat. Acad.
Sci. USA 80, 6126-6130.
Tokutake, S., Liem, R. K. H. and Shelanski, M. L. (1984). Each component
of neurofilament assembles itself to make component-specific filament.
Biomed. Res. 5, 235-238.
Tyler, J. M. and Branton, D. (1980). Rotary shadowing of extended
molecules dried from glycerol. J. Ultrastruct. Res. 71, 95-102.
Tytell, M., Zackroff, R. V. and Hill, W. D. (1988). Axonal neurofilaments
differ in composition and morphology from those in the soma of the squid
stellate ganglion. Cell Motil. Cytoskel. 9, 349-360.
Watson, D. (1991) Regional variation in the abundance of axonal cytoskeletal
proteins. J. Neurosci. Res. 30, 226-231.
Yen, S. H., Dahl, D., Schachner, M. and Shelanski, M. L. (1976).
Biochemistry of the filaments of brain. Proc. Nat. Acad. Sci. USA 73, 529-533.
(Received 25 January 1994 - Accepted 31 March 1994)