ABSTRACT
There are no high resolution transmission electron microscope methods
available that can reliably image the microtubule-associated proteins (MAPs)
on the surface of microtubules, visualize lattice arrangements of MAPs or
measure their associated filament diameters. We have bypassed a low resolution
rotary replication method with ~4-5 nm resolution and implemented a vertical
Pt-C replication method with a resolution of 0.7-1.0 nm. Previously we studied
the MAP tau and found that it formed 2.1 nm triple-stranded left-hand helical
polymers in purified tau preparations that had elastic properties (Ruben et.
al., [1991] J. Biol. Chem., 266:22019-22027). This finding led to the
suggestion that these polymers could associate axially with the microtubule
wall protofilaments and could restore bent microtubules after axons crossing
knee and elbow joints were flexed. The 2.1 nm tau polymer filaments have been
documented in purified tau preparations but have never been seen in
association with microtubules. We have now directly imaged the freeze-dried
replicated microtubule surface in the presence and absence of MAPs. We present
here our first findings using a new method for visualizing heterogeneous MAP
preparations in association with taxol stabilized microtubules. These rat
brain MAPs were neither exposed to acid pH nor heating, to ensure their
assembly competence. Although we were unable to distinguish between MAP 1, MAP
2 or tau, it appears that very long (197-944 nm) MAP polymers (~2 nm diameter)
can be assembled axially along the microtubule surface. The finding is
consistent with the hypothesis that polymers of MAPs can associate with the
microtubule surface. This study will be extended in the future with purified
tau. Since the 2.1 nm tau polymer filaments have been found in Alzheimers
neurofibrillary tangles (Ruben et. al, [1992] Brain Res., 590:164-179) and
they may be precursor filaments of the paired helical filaments in tangles
(Ruben et. al, [1993] Brain Res., 602:1-13), we hope to establish in the
future a biological role for the triple-stranded left-hand helical 2.1 nm tau
polymer filaments. Our imaging technique employed here will make it possible
in future investigations to generally study the relationship of fibrous MAPs
and motor proteins, dyneins and kinesins to microtubules.
INTRODUCTION
Over the last 20 years, images of microtubules and the
microtubule-associated proteins (MAPs) have been reported using thin
sectioning, negative staining, or the quick-freeze deep-etch low angle rotary
replication technique for transmission electron microscopy (TEM) (Amos, 1979;
Sloboda and Rosenbaum, 1979; Hirokawa and Heuser, 1981). New methods that will
provide more structural information about MAPs relationship to microtubules
are needed. The thin sectioning method fixes, plastic embeds, and post stains
60-100 nm thick sectioned microtubules. The microtubules and MAPs are stained
electron dense structures, and are thus made visible in the TEM. With this
method the relationship of MAPs to the microtubule surface was unclear since
only MAP surface projections are visible or not visible identifying their
presence or absence on microtubules. This technique has been used to clearly
identify the presence or absence of MAP 2 projecting from microtubules (Sloboda
and Rosenbaum, 1979). Since the stain chemistry is variable, difficult to
control, and can greatly change the size of small structures and do it
nonuniformly (Ruben et al., 1992; Ruben and Telford, 1980), only the MAP
length projecting from the microtubule surface has been reported. The negative
staining method can provide higher resolution and more detailed images than
the thin sectioning method, but it also is flawed. Microtubules visualized by
negative staining appear less electron dense where the tubulin dimers of the
wall exclude the stain. The wall fine structure is dependent on tubulin dimer
surface roughness and stain penetration between wall dimers. The upper and
lower microtubule walls both exclude stain, and their images are superimposed
in TEM. The MAPs on these two surfaces are also superimposed, creating a
complex indecipherable image. The resolution of negative staining has been
shown to be 1-1.2 nm, but this resolution depends on the size of the stain
molecule and the beam sensitivity of the biological material. Extended MAP 1
& 2 and tau monomer filaments appear to be less than 2 nm thick, a size
range that is often difficult to image by negative staining due to sample
surface roughness, support film thickness, stain depth, and the problems
previously mentioned (Marx and Ruben, 1984). The quick-freeze deep-etch method
has the potential to surpass both the negative staining and the thin
sectioning technique. Hydrated microtubules can be quick frozen, avoiding the
non-physiological treatments of negative staining and thin sectioning. After
deep-etching, the microtubules can be coated with evaporated metal on just one
surface. This technique, in the past, has been performed in the presence of
~0.1 M assembly buffers. Despite rapid freezing, these salts still come out of
solution, forming eutectic structures that can bridge microtubules and
masquerade as MAPs. Salts suspended in solution and released during etching
can potentially mask microtubule surface fine structure. The specimen
temperature during replication, the replication method, and the vacuum around
the specimen influence the resolution of replication (Ruben, 1989; 1995). Low
angle (20° angle) rotary platinum-carbon deposition resolution depends
directly on the metal film thickness, and low angle replication, as it has
been performed, produces not better than 4-5 nm resolution. This technique, so
far, has only been used to identify MAPs such as tau as side arms projecting
18.7 ± 4.8 nm from microtubules (Hirokawa et al., 1988). It is our desire to
apply a more refined quick-freeze deep-etch technique to the study of MAPs
association with microtubules. We not only want to eliminate the presence of
high salts, but we also want to use a high resolution vertical platinum-carbon
(Pt-C) replication technique that is capable of 0.7-1.0 nm resolution (Ruben,
1989). We have previously applied an earlier version of this technique to
study the microtubule marginal band in erythrocytes of Bufo marinus (Centonze
et al., 1985;1986).
We have had to solve number of experimental difficulties in engineering
this technique. In the beginning we found it difficult to maintain
microtubules with MAPs against depolymerization at a constant 37° C through
their preparation to freezing step. However, this was accomplished with a
combination of a large temperature bath, heat lamps, and an infrared
thermometer used to monitor the 50 µl specimen's temperature on a 13 mm
filter disc. All of the buffers, fixatives, and washes were held at 37° C in
the temperature bath. Microtubule preparations needed to be introduced into a
low salt environment before freezing, so their freeze-dried surfaces would be
free of salt. This step was solved only after numerous failures. The use of
taxol not only reduced the cold lability of microtubules, but allowed the full
complement of MAPs to bind at 37° C (Vallee, 1982) similar to microtubules
without taxol (Sloboda and Rosenbaum, 1979). The microtubules and MAPs were
also fixed for one minute in 4-5% glutaraldehyde in 0.1M cacodylate buffer,
pH. 7.2 (Cross and Williams, 1991), since longer fixations or fixation with
formaldehyde disrupted microtubule structure. These crucial steps allowed us
to remove the salts buffers and retain microtubule integrity. Since free
tubulin, salts, and fixatives had to be removed from the microtubules quickly
and without mechanical disruption, we avoided using sedimentation and
pelleting. Instead, we chose to spread the microtubules on a nonpolymeric
metallic filter so that the retained microtubules washed with buffer could be
fixed, in situ, for one minute, and washed free of salts, before
freezing. The metallic silver filter avoided introducing polycarbonate (Nucleopore
filters) , nylon, or cellulose (Millipore filters) fine filaments beneath the
sample that would interfere with the ability to identify MAPs. The method
utilizing silver filters, including their digestion, has been carefully worked
out and is reported here for the first time. Freeze-drying also has to be done
cautiously (-115° C) or the escaping water vapor can split microtubules.
Finally, we used a vertical Pt-C replication method which coats exposed
surfaces, followed by rotary carbon deposition, which glues the replica
together and can retain microtubules with MAPs suspended above a surface in a
three dimensional replica. The replication method has previously imaged
individual polymer chains, their side chains, as well as the DNA double helix
at a resolution of 0.7-1.0 nm (Ruben,1995; 1989; Ruben and Stockmayer, 1992).
Our efforts to visualize MAPs directly on microtubules have their roots in
the observation that the axonal MAP, tau, can form a triple-stranded left-hand
helical 2.1 nm polymer that has been observed to be 130-2088 nm long. Since
tau monomer can be as short as 32 nm and as long as 112 nm (Ruben et al.,
1991), and has been measured to be 56.1 ± 14.1 nm (Hirokawa et al., 1988) and
69-75 nm (Hagestedt et al., 1989), this protein was suggested to have elastic
properties. Based on this property, we suggested that the very long 2.1 nm tau
polymers would also be elastic and could restore axonal microtubules that
extend across knee and elbow joints. We hypothesized that tau polymers
probably lie parallel to the 13 axial rows of tubulin dimers and act as an
elastic sheath in restoring the microtubules after axonal bending. The
hypothesis that 2.1 nm tau polymers can associate longitudinally with
microtubules can only be tested by direct visualization of MAPs on the
microtubule surface. This is why the present study was originally undertaken.
Our study has only progressed to the stage of using a mixed MAP test
preparation that contains MAP 1 & 2 and tau (Rozdzial et al., 1990). This
mixed MAP preparation will demonstrate that fine filaments in association with
the microtubule surface can be visualized. In spite of the mixed MAP
preparation used in these experiments, a few interesting and important new
observations can be made about the relationship of MAPs to microtubules.
MATERIALS AND METHODS
1. Isolation of Rat Brain tubulin and microtubule
associated proteins, MAPs.
The rats were kept in an animal colony at the NYS Institute for Basic
Research where they had unlimited access to food and water. All of the NIH
guidelines for the care of these animals were followed.
Rat brain microtubules were isolated by performing two
temperature-dependent cycles of microtubule polymerization-depolymerization (Shelanski
et al., 1973) and pelleting the microtubules. The microtubule pellet was
stored in glycerol at -75° C until used. At the time of use, a third
polymerization-depolymerization step was performed and pelleted to select for
active tubulin. The supernatant was discarded. The tubulin from the pelleted
microtubules was purified by phosphocellulose ion-exchange column
chromatography (Sloboda and Rosenbaum, 1979).
Microtubule-associated proteins, MAPs, were selected by three
polymerization-depolymerization steps similar to that in Grundke-Iqbal et al.
(1986) and pelleting the MAP microtubule polymer after each step.
The tubulin polymerization buffer was 100 mM MES, 1 mM EGTA, 1 mM MgCl2,
1 mM GTP, pH 6.7. To this buffer, 20 µM taxol and 1-2 mg/ml tubulin was added
to make taxol stabilized microtubules at 37° C. Taxol was purchased from the
Sigma Chemical Co of St. Louis, MO. In Figs. 1a and 1b the microtubules were
assembled with taxol in the absence of MAPs. In Figs. 2 and 3 a depolymerized
MAP microtubule mixture was depolymerized and repolymerized in the presence of
added taxol at 37° C. MAPs from this preparation were detached from the taxol
stabilized microtubules by treatment with 0.5 M NaCl for 15 minutes (Vallee,
1982).
The MAP fraction isolated from bovine brain is similar to that described in
Rozdzial et al., (1990) and was mostly MAP 1, 2 and tau.
2. Specimen Preparation for Transmission Electron
Microscopy (TEM)
a) Negative Staining of
Microtubules
The solution containing
microtubules and taxol was diluted by a factor of three with buffer, and a 5
µl drop of the solution was placed on a 10 nm thick indirectly evaporated
carbon film on 300 mesh Cu grids prepared as previously described (Ruben and
Marx, 1984). The grids were treated ~3 hrs with glutaraldehyde vapor to make
them hydrophilic and bind proteins as described in Ruben et. al (1988). The
microtubules were lightly fixed on the grids with a 4% glutaraldehyde 0.1 M
sodium cacodylate (pH 7.2) for one minute and were then negatively stained
with two applications of 2% uranyl acetate (pH 3.8).
b) Freeze-Drying and
Vertical Pt-C Replication
A 13 mm silver filter (0.2
µm pore size, Poretics, Cat. # 50514) was placed in a Boyden chamber filter
unit (Ruben et al., 1992) on a platform in contact with a water bath heated
to ~37° C. A 200 watt lamp heated the filter from above, and the
temperature was adjusted to 37° C and monitored with an infrared
thermometer (Raynger PM, Raytek, Inc, model # WWPRGM3CF). Fifty microliters
covered the 13 mm 0.2 µm silver filter, and the liquid was pulled through
the filter; the filter was then washed with polymerization buffer. A 100 µl
of 4% glutaraldehyde in 0.1 M sodium cacodylate (pH 7.2) was placed on the
microtubule coated filter for one minute and the liquid was then pulled
through the filter by a weak vacuum. The filter was then washed 3 to 4 times
with distilled water (37° C), removed from the Boyden Chamber and plunge
frozen by hand in liquid propane cooled with liquid nitrogen -195.8° C. The
sample was freeze-dried conventionally in a modified Balzers 300 at -115° C
for three hours and vertically replicated with 0.93 nm of Pt-C and ~11.7 nm
of rotary deposited carbon applied in two steps (Ruben, 1989). The replica
was removed from the silver filter by floating the filter on the surface of
a concentrated KCN solution for 1-2 days. The silver filter was oxidized and
the silver oxide dissolved in the following chemical reaction: 4Ag + O2
+ 2H2O + 8CN-
--------> 4Ag(CN-)2 + 4OH-.
Once the silver filter was dissolved, the replica was floated on a saturated
KOH solution to remove the microtubules for 1 day and then floated on water
for another day to remove any salts. The replica was then picked up on 300
mesh grids without a support film as described in Ruben (1989).
3. Transmission electron Microscopy methods
A JEM 100 CX TEM was used with a 400 µm condenser and a 40 µm objective
aperture at 80 kV with a LaB6
filament. Images were recorded at direct magnifications of 27,600 x to
96,800x. The film plates of negatively stained microtubules were printed with
a Durst 1200 point source enlarger on Ilford multicontrast fiber based paper.
Reversal negatives of the replicated microtubules and MAPs were made as
previously described (Ruben, 1989) and the reversal negatives were enlarged
2.5x as 8" x 10" prints on Ilford Multicontrast paper. Sections of
these prints were scanned by a LaCie Silver Scanner III at 600 dpi using Adobe
Photoshop and were also labeled and collaged using Photoshop 3.0 and saved in
final form as 150dpi JPEG files.
RESULTS
We wanted to visualize fully formed microtubules and their MAPs, using a
freeze-drying technique, that were vertically Pt-C replicated and glued
together by deposition of a rotary carbon film. This technique has been shown
to be able to resolve structures on surfaces to 0.7 nm resolution with no more
than ~ 4% shrinkage (Ruben, 1989; Ruben and Stockmayer, 1992). We used the
negative staining technique in Fig.
1a to characterize microtubules assembled from tubulin dimers in the
presence of taxol. These images confirm that our microtubules are normal with
tubulin protofilaments in their wall. In a freeze-etched vertically replicated
image of a similar microtubule, we see a smooth cylindrical surface without
MAPs (Fig.
1b). The tubulin dimers in the wall of the microtubule are not visible.
|
|
| Fig. 1a. Negatively stained microtubules
produced with taxol and without MAPs Rat tubulin was polymerized in the
presence of taxol without microtublule associated proteins (MAPs) at
37° C. When part of the microtubule wall surface is removed, the linear
tubulin filaments in the microtubule wall were visible. If the wall is
unbroken, the lines in the microtubule wall are not visible (see MT
& arrows). |
|
|
| Fig. 1b. Freeze-dried and platinum-carbon
replicated microtubules produced with taxol and without MAPs Rat tubulin
was polymerized in the presence of taxol without MAPs at 37° C. The
single microtubule shows no prominent features other than a diameter of
25 nm and the absence of fine MAP-like filaments. |
In the following studies, we polymerized tubulin dimers with taxol at 37°
C in the presence or absence of a mixture of MAPs. In Fig.
2a we see three microtubles, two of which are labeled MT that are ~25 ±
1.5 nm where they are undamaged. In the upper left, arrows point to where two
~2nm filaments join the microtubule surface. The lower ~2 nm filament appears
to have come off the MT's lower side, snagging a structure and then rejoining
the parent microtubule after extending 197 nm. Another ~2 nm filament on the
lowest MT reaches across a break in the surface, merging with the surface and
reemerging and extending straight for 251 nm as the MT bends away at the lower
right.
|
|
| Fig. 2a. Freeze-dried and platinum-carbon
replicated microtubules produced with taxol and Maps Rat tubulin was
polymerized in the presence of taxol with MAPs at 37° C. The
microtubules (MT) show fine filaments (arrows) ~2 nm in diamter that
have come off the microtubule surfaces in the top, middle and bottom of
the figure. The MAP (2 arrows) in the middle region leaves and rejoins
the surface of the bent microtubule after extending about 197 nm. This
MAP-like filament extends axially along the microtubule's surface since
such a long section of this MAP could not have come off and rejoined the
surface if it had been associated differently. The MAP marked by 2
arrows at the bottom is 251 nm long although part of it is not shown. At
the lower left a ~2 nm MAP-like filament emerges and rejoins the
microtubule surface, and then leaves the surface at the lower right. |
In Fig.
2b are two roughly parallel MT with ~2 nm filaments emerging from their
surfaces. The ~2 nm filament at the upper left extends ~297 nm suspended above
the surface before joining another structure not seen in the figure. The
filament at the lower right extends 944 nm suspended above the filter surface
before it joins another structure also not included in the image.
|
|
| Fig. 2b. Freeze-dried and platinum-carbon
replicated microtubules produced with taxol and Maps These damaged
microtubules (MT) have ~2 nm diameter MAP-like filaments that extend off
their surface. These MAP-like filaments also appear to extend along the
MT's long axis. The MAP-like filament at the top left extends 297 nm and
the one at the lower right extends 944 nm although visible in the
original image. The top and middle MAP-like filaments appear to extend
along the MT's axis before merging with the surface. |
In Fig.
3a, four microtubules are present, and three are labeled with MT. At the
left, an arrow points to a ~2 nm filament joining the microtubule. A diagonal
microtubule reaching from the upper left to the lower right, has two ~2 nm
filaments (see arrows) extending ~50 nm parallel to its surface with neither
end extending off the microtubule as in Fig.
2a or 2b.
There is also a fine filament on the right microtubule (MT) which appears to
be extending along its surface. This ~2 nm filament extends ~33 nm along the
microtubule surface before rejoining it on the other side of the microtubule
break. This MAP like filament appears to extend axially along the microtubule
without projecting off its side similar to the two on the diagonal
microtubule.
|
|
| Fig. 3a. Freeze-dried and platinum-carbon
replicated microtubules produced with taxol and Maps MAP-like filaments
(arrows) appear to join (left side) or extend axially along microtubules
(MT) at the lower and upper right in the figure. In the lower right two
~2 nm filaments can be seen extending 62 nm and 38 nm along the
microtubule with a center to center distance of ~5.2 nm which is about
the width of a tubulin protofilament. In the upper right a single
MAP-like filament extends 30 nm axially along a MT surface before
merging with it. The four microtubule diameters range from 24.5-26.4 nm
in this image. |
In Fig.
3b, two microtubules enter the image from the left. The upper one shows a
~2 nm filament extending 75 nm axially along its surface across a break in the
microtubule. On the microtubule just below it are two parallel slightly
elevated ~2 nm filaments above the microtubule surface extending 70 nm and 45
nm axially along it.
|
|
| Fig. 3b. Freeze-dried and platinum-carbon
replicated microtubules produced with taxol and Maps Two microtubules
show ~2 nm MAP-like filaments extending axially along their surface. The
arrows on the top most microtubule define a MAP-like filament extending
75 nm axially along the filament. The lower MT shows two MAP-like
filament extending axially just before the Pt-C coating of the MT ends.
These MAP-like filaments are visible for 70 nm and 45 nm along the lower
microtubule. |
It is well known that MAPs can be removed from taxol stabilized
microtubules in high salt. We also performed this experiment and removed the
MAPs in high salt. We found many ~2 nm filaments on the specimen surface not
associated with microtubules. Two of these MAP-like filaments are shown in Fig.
4. These filaments were glutaraldehyde fixed which may be why their
structure is not well defined. Nonetheless, when MAPs are absent (Fig.
1b), the ~2 nm filaments on the microtubules are absent, and when the MAPs
are removed with high salt, they are not found on the microtubules but on the
sample surface.
|
|
| Fig. 4. Freeze-dried and platinum-carbon
replicated ~2 nm MAP-like filaments released after 0.5 M NaCl treatment
of taxol stabilized microtubules Two fine filaments, ~2 nm in diameter,
were found in this image free of microtubules and many more were found
in other images of this preparation. These fine filaments ranged in
length, 95 nm, 178 nm, and 277 nm. These MAP-like filaments were
associated with the microtubules before 0.5 M NaCl removal and these
MAP-like filaments were not present on microtububules polymerized in the
absence of MAPs. |
DISCUSSION
The literature has identified MAP 1, MAP 2, and tau as thin flexible
rod-like proteins. These MAPs have been identified as cross-bridges between
microtubules or microtubule bundling factors, and each contains a microtubule
binding region of 3-4 repeat sequences located in its carboxy terminal region.
It is also generally acknowledged that the carboxy terminal region binds the
microtubule, while the amino terminal region can form the cross-bridge between
microtubules (Wiche et al., 1991). MAP 1B protein has been shown to be a long
flexible rod, 186 ± 38 nm long with a small spherical portion at one end and
has also been identified as a cross-bridge between microtubules in neurons
(Sato-Yoshitake et al., 1989). MAP 1A has also been identified as a long
flexible rod shaped protein which forms cross-bridges between microtubules in
dendrites (Shiomura and Hirokawa, 1987). On SDS gels, the MAP 1 proteins are
banded at a higher molecular weight than MAP 2. The MAP 2 proteins, MAP 2A,
2B, & 2C, appear to be a family of alternately spliced proteins from the
same gene ranging from 199 -70 kilodaltons (Wiche et al., 1991). MAP 2 has
been shown to be a long flexible filament ~185 nm long that when attached to
microtubules can protrude up to 90 nm from the polymer surface (Voter and
Ericson, 1982). In another study the MAP 2 was found to extend 30 nm (Zingsheim
et al., 1979) with the reported range 30-90 nm. MAP 2 monomer is thought to be
a long thin flexible rod 1.6 nm in diameter (Voter and Erickson, 1982) or that
of an alpha helix sized coil of ~1 nm in diameter (Ruben et al., 1991). MAPs 2
has also been identified as a cross-bridge or bundling factor between
microtubules (Centonze et al., 1985; Chen et al., 1992). There has also been a
great deal of effort to categorize the frequency of MAP 2 projections from the
side of the microtubule with either the 12 or 6 tubulin dimer repeat lattice
(Amos, 1979; Jensen and Smaill, 1986). The bovine and human tau proteins form
families of six alternately spliced proteins from the same gene with amino
acid sequences ranging from 448-302 and 441-352, respectively. These monomers
have been identified as long thin flexible rods 32-112 nm long about ~1 nm in
diameter (Ruben et al., 1991). The tau monomer has also been measured to be
56.1 ± 14.1 nm long (Hirokawa et al., 1988) projecting 18.7 ± 4.8 nm from
the microtubule surface. It has also been identified as a bundling factor in
axonal microtubules (Chen et al., 1992).
The observation in the references above of frequent MAP 2 side arms
projecting from microtubules is not confirmed by either Figs. 2 & 3. Our
observation, however, should not be seriously considered because the MAP 1,
MAP 2, and tau tubulin binding appears to be competitive (Coffey and Purich,
1995; Kim et al., 1986), and we do not know the ratio of MAPs to the number of
tubulin monomers in the images. One should take seriously the fact that taxol
stabilized microtubules without MAPs are smooth walled cylinders of ~25 nm
with no fine filaments present (Fig.
1b). In a conventional negatively stained image (Fig.
1a), these taxol stabilized microtubules are clearly composed of axial
tubulin protofilaments which are not visible in the freeze-dried vertically
replicated microtubule in Fig.
1b. In Figs.
2a and 2b,
very long ~2 nm filaments join the microtubule and appear to associate with it
axially. In Fig.
2a this is especially evident where an axially associated 197 nm length of
a ~2 nm diameter filament has come off a section of bent microtubule wall.
This thin filament section is longer than any of the reported monomer lengths
for the MAPs. This is also true of the filament reaching into the image from
the lower right which is 251 nm long. In Fig.
2b, this same situation is repeated with ~2 nm diameter filaments that are
297 nm and 944 nm long. In Fig.
3a, there is one filament cross-bridge evident at the left side of the
image. The other fine filaments appear to extend axially along the
microtubules for 62 nm and 38nm with a center to center distance of ~5.2 nm,
close to the distance between tubulin protofilaments (Amos, 1979). There is
also a third axial filament at the right that extends 30 nm axially along the
filament. In Fig.
3b, three fine filaments (see arrows) can be seen to extend axially along
the surface of the microtubules. These filaments are ~2 nm in diameter and
extend 75 nm, 70 nm, and 45 nm along the microtubule surface. What we found
surprising in our images is that the ~2 nm MAP filaments disappear into the
microtubule surface. We hope to be able to solve this problem by varying the
concentration of glutaraldehyde fix but holding the time of fixation constant.
Its not clear whether too much or too little fixation has obscured the 2 nm
filaments. In Fig. 4, taxol stabilized microtubules with MAPs were washed with
0.5 M NaCl salt which removes the MAPs and preserves the microtubules (Vallee,
1982). Many fine filament like the two seen in Fig. 4 were suspended above or
on the silver filter surface. Some of these filaments were 95 nm, 178 nm, and
277 nm in length, where the last ~2 nm fine filament mentioned is also longer
than any of the reported MAP monomers.
Only two estimates for MAP monomer filament diameters have been published,
and only the one of tau comes from a high resolution image with an established
replica correction factor (Ruben et al., 1991). Tau's monomer diameter of 1 nm
also correlates with its linearly distributed secondary structure of ~22 %
alpha helix and ~78% beta spiral (Ruben et al., 1991). Although we do not have
any high resolution images of MAP 2, its predicted secondary structure is also
27% alpha helix and ~73% beta spiral suggesting it too has a diameter of ~1
nm.
Finally, equilibrium binding of MAP 2 to taxol stabilized microtubules
showed positive cooperativity (Wallis et al., 1995) which was absent in the
equilibrium binding of cloned MAP 2 microtubule binding region. Interestingly,
one of the cloned microtubule binding region constructs bound as a dimer but
still showed no positive cooperativity (Coffey and Purich, 1995).
With the evidence from four separate observations, we argue that the MAPs
associated with our taxol stabilized microtubules can be MAP polymers because
(1) the MAP filaments are longer than any of the reported monomer lengths, (2)
MAP filament diameters of ~2 nm appear to be twice as large as either tau or
MAP 2 monomer, (3) MAP 2 shows positive cooperativity in binding microtubules,
while a cloned microtubule binding domain formed dimers before binding to
microtubules, and (4) stretches longer than any monomer microtubule binding
region (197 nm, 62 nm, 70 nm and 75 nm) were found to be axially associated
with microtubles.
We believe that MAPs may self associate as polymers with microtubules and
that many of the MAPs we have observed are associated parallel with the
tubulin protofilaments, which represents a new finding. We plan to do similar
experiments in the future with purified tau.
ACKNOWLEDGMENTS
These studies were supported in part by NIH grants AG11054 (G.C. Ruben),
NS18105 (I. Grundke-Iqbal) and AG05892, AG08076 and TW00507 (K. Iqbal). The
authors wish to thank Roger Sloboda for critically reading the manuscript and
the Rippel Electron Microscopy lab at Dartmouth for the use of their
equipment.
REFERENCES
Amos, L. (1979) Structure of microtubules. In: Microtubules, K. Roberts and
J.S. Hyams, eds., Academic Press, NY and London, pp. 1-64.
Centonze, V.E., Ruben, G.C. and Sloboda, R.D. (1985) Structure and
Composition of the Cytoskeleton of Nucleated Erythrocytes II. Immunogold
Labelled Microtubules and Crossbridges in Platinum-Carbon Replicas of the
Marginal Band of Buto marinus Erythrocyte cytoskeletons. Eur. J. Cell Biol.,
39:190-197.
Centonze, V.E., Ruben, G.C. and Sloboda, R.D. (1986) Structure and
Composition of the cytoskeleton of Nucleated Erthrocytes. III. Organization of
the Cytoskeleton of Bufo marinus Erythrocytes as Revealed by Freeze-Dried
Platinum-Carbon Replicas and Immunofluorescence Microscopy. Cell Motil. and
Cytoskel., 6:376-388.
Chen, J., Kanai, Y., Cowan, N.J., and Hirokawa, N. (1992) Projection
domains of MAP-2 and tau determine spacings between microtubules in dendrites
and axons. Nature, 360:674-677.
Coffey, R.L., and Purich, D.L. (1995) Non-cooperative binding of the MAP-2
microtubule-binding region to microtubules. J. Biol. Chem., 270:1035-1040.
Grundke-Iqbal, I., Iqbal, K., Quinlan, M., Tung, Y.-C., Zaidi, M.S., &
Wisniewski, H.M. (1986) Microtubule-associated protein tau: a component of
alzheimer paired helical filaments. J. Biol. Chem. 261, 6084-6089.
Hagestedt, T., Lichtenberg, B., Wille, H., Mandelkow, E.-M., and Mandelkow,
E. (1989) Tau protein becomes long and stiff upon phosphorylation: correlation
between paracrystalline structure and degree of phosphorylation. J. Cell
Biol., 109:1643-1651.
Hirokawa, N., Heuser, J.E. (1981) Quick-freeze, deep-etch visualization of
the cytoskeleton beneath surface differentiations of intestinal epithelial
cells. J. Cell Biol., 91:399-409.
Hirokawa, N., Shiomura, Y., and Okabe, S. (1988) Tau proteins: The
molecular structure and mode of binding on microtubules. J. Cell Biol.,
107:1449-1459.
Jensen, C.G. and Smaill, B.H. (1986) Analysis of the spacial organization
of microtubule associated proteins. J. Cell Biol., 103:559-569.
Kim, H., Jensen, C.G., and Rebhun, L.I. (1986) The binding of MAP-2 and Tau
on brain microtubules in vitro: Implications for microtubule structure. In:
Dynamic Aspects of Microtubule Biology, D. Soifer, ed., Ann. NY Acad. Sci.,
466:218-239.
Marx, K. A. and Ruben, G.C. (1984) Studies of DNA orgnization in hydrated
spermidine-condensed DNA toruses and spermidine-DNA fibres. J. Biomolec.
Struct. & Dyn., 1: 1109-1132.
Rozdzial, M.M., Neighbors, B. W., & McIntosh, J.R. (1990) Blot overlay
identification of microtubule-binding peptides from bovine brain. Eur. J. Cell
Biol., 52: 27-35.
Ruben, G.C. , Iqbal, K., Grundke-Iqbal, I, Wisniewski, H.M., Ciardelli, T.L.
and Johnson Jr., J.E. (1991) The microtubule associated protein tau forms a
triple-stranded left-hand helical polymer, J. Biol. Chem., 266: 22019-22027.
Ruben, G.C. Iqbal, K., Wisniewski, H.M., Johnson, Jr., J.E., and
Grundke-Iqbal, I. (1992) Alzheimer neurofibrillary tangles contain 2.1 nm
filaments structurally identical to the microtubule associated protein tau: A
high resolution transmission electron microscopy study of tangles and senile
plaque core amyloid. Brain Research, 590:164-179.
Ruben, G.C., Iqbal, K., Grundke-Iqbal, I., and Johnson, Jr., J.E. ( 1993)
The organization of the microtubule associated protein tau in Alzheimer paired
helical filaments. Brain Research, 602:1-13.
Ruben, G.C.(1989) Ultrathin (1nm) vertically shadowed Pt-C replicas for
imaging individual molecules in freeze-etched biological DNA and material
science metal and plastic specimens. J. Electr. Microsc. Tech., 13:335-354.
Ruben, G.C., Harris, Jr., E.D., and Nagase, H. (1988) Electron microscope
studies of free and proteinase-bound duck ovostatins (ovomacroglobins) : model
of ovostatin structure and its transformation upon proteolysis, J. Biol.
Chem., 263: 2861-2869.
Ruben, G.C. and K.A. Marx (1984) Parallax measurements on stereomicrographs
of hydrated single molecules, their accuracy and precision at high
magnification. J. Elect. Micros. Tech., 1: 373-385.
Ruben, G.C. and Stockmayer, W. H. (1992) Evidence for helical structures in
poly(1-olefin sulfones) by transmission electron microscopy. Proc. Natl. Acad.
Sci. USA 89: 11645-11649.
Ruben, G.C. and Telford, J.N. (1980) Dimensions of active cytochrome c
oxidase in reconstituted liposomes using a gold ball shadow width standard: A
Freeze-etch Electron Microscopy Study. J. Microsc. 118: 191-216.
Sato-Yoshitake, R., Shiomura, Y., Miyasaka, H, and Hirokawa, N. (1989)
Microtubule-associated protein 1B: molecular structure, localization, and
phosphorylation-dependent expression in developing neurons. Neuron, 3:229-238.
Shiomura, Y. and Hirokawa, N. (1987) The molecular structure of
microtubule-associated protein 1A (MAP 1A) in vivo and in vitro . An
immunoelectron microscopy and quick-freeze, deep-etch study. J. Neuroscience,
7:1461-1469.
Shelanski, M.L., Gaskin, F., and Cantor, C.R. (1973) Microtubule assembly
in the absence of added nucleotides. Proc. Nat. Acad. Sci. USA 70: 765-768.
Sloboda, R.D., and Rosenbaum, J.L. (1979) Decoration and stabilization of
intact, smooth-walled microtubules with microtubule-associated proteins.
Biochemistry 18: 48-55.
Vallee, R.B. (1982) A taxol-dependent procedure for the isolation of
microtubules and microtubule-associated proteins (MAPs). J. Cell Biol.,
92:435-442.
Voter, W.A., and Erickson, H.P.(1982) Electron microscopy of MAP-2. J.
Ultrastruct. Res. 80:374-382.
Wallis, K.T., Azhar, S., Rho, M.B., Lewis, S.A., Cowan, N.J., and Murphy,
D.B. (1993) The mechanism of equilibrium binding of microtubule-associated
protein 2 to microtubules. J. Biol. Chem., 268:15158-15167.
Wiche, G., Oberkanins, C., and Himmler, A. (1991) Molecular structure and
function of microtubule-associated proteins. Internat. Rev. Cytol.,
124:217-273.
Zingsheim, H.-P., Herzog, W., and Weber, K. (1979) Differences in surface
morphology of microtubules reconstituted from pure brain tubulin using two
different microtubule-associated proteins: The high molecular weight MAP 2
proteins and tau proteins. Eur. J. Cell Biol., 19:175-183.
