Dependence of Normal KZ4 Potassium Channel Expression on
Microtubule Stability in the Sf9 Cell Line
Benjamin Jacobson
Problems in Marine Biololgy
Professor W. Gilly
June 13,1995
Benjamin Jacobson
June 13,1995
Dependence of Normal KZ4 Potassium Channel Expression on Microtubule
Stability in the Sf9 Cell Line
ABSTRACT Microtubules are a cytoskeletal, filamentous polymer that have been shown
to play a role in fast axonal transport in nerve cells. Further it has been hypothesized that
microtubule stability is a factor in determining cell excitability. By culturing Spodoptera
frugiperda (Sf9) cells that express the KZ4, shaker potassium channel protein in
demecolcine, a microtubule destabilizing drug, it has been shown that there is a link
between microtubule stability and the proper expression of the channel through the use of
electrophysiology, immuno-flourescence, and immuno precipitation techniques.
Microtubules are long, cylindrical filamentous polymers composed of 13
protofilaments. Each protofilament is constructed from repeating subunits of a dimer
composed of o-tubulin and B-tubulin. The two similar polypeptide chains spontaneously
polymerize in the presence of GTP (reviewed by Kirschner 1986). The tubulin polymer is
à cytoskeletal element integral in the dynamic changes in cell morphology such as cell
division. Microtubules have also been recognized as an important element in
compartmentalizing the cytoplasm and holding organelles in place (Albert, et al.). More
recently, microtubules have been found to serve as the pathway for transport of both
organelles and proteins anterograde and retrograde in the axon. This mode of transport
along microtubule filaments forms the basis for for rapid axonal transport (Grafstein
1980).
It has been suggested that there is a physiological link between the excitability of
nerve cells and a structural element, tubulin (Matsumoto 1979). Conclusions concerning
the relationship between excitability and microtubules were reached through results
obtained by perfusing squid giant axons with chemicals like colchicine, vinblastine, and
taxol, known to either favor the polymerization or depolymerization of microtubules. In
these experiments, the axoplasm was extruded from the axons before perfusion, and
intracellular recordings were made to assess changes in action potential threshhold. The
chemical's action on the polymerization of microtubules was also measured in vitro by
analyzing the effects on solution turbidity. These experiments demonstrate that drugs
Benjamin Jacobson
June 13,1995
which favor polymerization of microtubules decreased the thres hhold for action potentials.
suggesting a link between K' channel exression and microtubule stability in squid axons
In order to further examine the possible role played by microtubules in determining
neuronal excitability, we treated living cells that overexpress a cloned squid potassium
channel with the microtubule destabilizing drug demecolcine and assessed whether these
cells showed normal pattem and level of K: channel expression as evaluated by immuno
flourescent microscopy and voltage clamp recordings. K channel protein was also
immuno precipitated from these same cells under conditions in which many protein-protein
interactions should be maintained. The material precipitated with the K: channel antibody
was evaluated for the presence of tubulin using SDS-polyacrylamide gel electrophoresis
followed by Westem blotting with an antibody that specifically recognizes q-tubulin.
The K' channel studied, KZ4, belongs to the KVI family of Kt channel and was
cloned from CDNA of giant fiber lobe neurons in the squid, Loligo opalescens (Perri, et al
1994). The giant fiber lobe (GFL) is a group of neurons whose axons are fused in the
neuropil to form the squid giant axon (Gilly, et al. 1990). In situ hybridization
experiments reveal a high level of KZ4 mRNA expression in the GFL, and physiological
properties of the KZ4 channel when expressed in frog oocytes (Rosenthal, et al 1995) and
in S19 cells (unpublished observations) have a great deal of similarity to potassium
channels present in the squid giant axon.
This study combines electrophysiological and immunological methods of studying
KZA expression. In order to perturb microtubule organization, infected Sf9 cells were
cultured in demecolcine (N-deactyl-N-methylcolchicine), a drug related to colchicine that
blocks microtubule based transport in the giant axon by destabilizing the structure of
microtubules (Brady et al. 1985). The results presented here suggest an interaction
between KZ4 and microtubules in Sf9 cells. Experiments such as these shed light on the
important role microtubules play in determining neuronal function.
Benjamin Jacobsor
June 13,1995
Materials and Methods
Sf9 Cell Culture
Sf9 cells were obtained from the ATCC and maintained at 28° as described elsewhere
(Summers, et al 1987.) supplemented with 10% fetal calf serum, 10 ug/ml penicillin/
streptomyocin, and 5 ug/ml Fungizone (Detomaso, et al. 1993) Sf9 cells were maintained
at 28° C at a density of 1-4 X 10° cells / ml in volumes ranging from 50-250 ml.
Uninfected cells were split every 3 days. For infections, log phase high viability cells
(98%, as determined by trypan blue exclusion) were centrifuged for 10 minutes and
resuspended at a density of 1X10 cell/mi in medium at room temperature. Viral stock was
added at a multipliticity of infection of 5-10. After 1 hour incubation, cells were dilluted to
1.5-2 X10° cells/ml with medium.
It was found that optimal expression of potassium current in the KZA infected cells was
produced by incubating the cells at 28°C for 1 day after infection and moving the cells to
18°C. Maximum current was found between days 5 and 8 at 18°C (unpublished
observations).
Electrophysiology
Whole cell voltage clamp was performed using fire-polished glass pipettes (1-3 MQ) filled
with an internal solution containing 50 mM KCl, ImM MgCl,, 5 mM HEPES, 50 mM K¬
glutamate, 5 mM EGTA, 20 mM K- aspartate, 60 mM KF, at a pH of 7.5 and osmolarity
of 320 mOsm. Experiments were performed at a temperature of 20° C. A List EPC-7
voltage clamp (Medical Systems, Garden City, NY) operating at low gain with series
resistance compensation was used. Linear ionic and residual capacity currents were
subtracted using a p/4 method deliverered at the holding potential of -100 mV. The data
Benjamin Jacobson
June 13,1995
was filtered at 10 kflz using an 8-pole bessel filter and sampled at a rate of 2-20 kklz. The
external solution contained 20 mM KCI, 130 mM Nacl, 10 mM Cacl, 10 mM MgCl., 5
mM HEPES, 5 mM MgSO,, and 5 mM Glucose at a ph of 7.2 and osmolarity equal to the
internal solution.
Data Analysis
Recordings were first qualitatively judged to possess Kt current by observing the presence
of normal" channel kinetics, increased rate of activation and inactivation with increased
voltage steps . For our purpose, cells that did not display these properties were assigned a
K conductance value of zero. In cells that showed the presence of normal current, tail
currents were measured to assess the level of K: conductance (Gy). Voltage was first
stepped 420mV for 5 ms to fully activate the K: channels and then stepped to a series of
levels (140,-120,-100, and-80 mV) for 73 ms. The amplitude of this inward tail current
immediately following the pulse was measured and plotted as a function of the
repolarization voltage in order to estimate the K: conductance of the cell (G. - %y in a lV
plot). Gx was normalized to cell surface area dividing by the capacitance of the cell (C).
which was measured by integrating the area under the current trace for a-10 mV step for s
ms. Values for G C (conductance density) thus serve as an indicator of functional Kt
channel expression at the plasma membrane..
Immuno-flourescence
Cells were cultured on 12 mmn diameter circular glass coverslips in a 24 well culture dish.
At the end of the 8 day course of the experiment, the cells were fixed 100% methanol at O
C for 6 minutes. They were then treated for 30 minutes in 0. 1% TritonX-100 in PBS
(C.O8M Na, HPO, 0025M Nakl, PO, and O. IM NaCl, pH 7.5), and blocked for 1 hour
in PBS containing 5% dry milk at 4°C. The primary antibody was anti-KZ4, a rabbit
polyclonal antiserum directed against a KZ4 fusion protein (yet unpublished) and used at
1.100 dilution in O.1% milk-PBS solution. The cells were incubated in the primary
Benjamin Jacobson
June 13,1995
antibody overnight at 4°C. Cells were then washed 3X in PBS and incubated for 1 hour at
4°C in 1:500 goat anti rabbit antibody conjugated to Texas Red (Molecular Probes, Inc.,
Eugene, OR). Coverslips were finally mounted with SlowFade Component B (Molecular
Probes). The slides were viualized and photographed on a Zeiss Axioskop flourescent
microscope using a 40X and 100X oil immersion objective.
Immuno-Precipitation
Antibody labeling
A solution containing Sf9 cells infected with KZ4 was centrifuged for 1 minute at 13,000
rpm, and the pellet was mechanically lysed by grinding it into solution. Denser portions of
the lysate, including cell membranes, were isolated by spinning the solution again for 30s
at 13,000 rpm and removing the supernatant. Resuspension of the pellet was performed in
200 ul HBS solution (25mM HEPES, 150 mM NaCl, pH 7.4) of 1% CHAPS. CHAPS
was used because it is a detergent that effectively solubilizes lipids in the lysate while
maintaining strong protein interactions present in the cell. Primary anti KZA antibody was
added in solution at 1/250 and solution was incubated in a LABQUAKE rotator at 4° C
overnight. IX 10° beads of sheep anti-rabbit-coated magnetic Dynabeads (Dynal Corp.
Great Neck, NY) were added and incubated for 1 hour in lysate solution. The magnetic
beads were isolated by holding the tube next to a magnet and aspirating the supernatant.
Beads were then washed 3X with 1 ml of 1% CHAPS in HBS. Precipitated protein was
eluted by suspending the washed Dynabeads in sample buffer and incubating for 10 min at
80° C. The eluted protein was seperated by 7.5% SDS-PAGE, transferred to nylon
blotting paper, and immunoblotted for tubulin with monoclonal anti o-tubulin N. 356
developed in mouse (Amersham Corp.).
Radioactive labeling
Benjamin Jacobson
June 13,1995
A portion of the sames cells used for the precipitation above were removed from the normal
medium and placed in methionine free Grace’s medium (Life Technologies, Inc. Grand
Island, NY) for 1 hour at room temperature. 150uCi of PSmethionine was added and
the cells incubated for 3 hours and chased in the normal medium.
Results
Electrophysiology
It was found by clamping 12 control cells that 58.3% of the cells showed outward K
current similar to that seen in naturally occuring KVI-type channels (table 1, figure la). Of
the 16 demecolcine treated cells, only 38.25% showed current (table 1). Three of the
demecolcine treated cells showing current displayed kinetics similar to those in untreated
cells. The rest of the cells showed abnormal channel kinetics defined by lack of delay in
activation and the absence of outward current during 250 ms pulse (figure lb.). In the 7
control cells that showed outward current, the average density (G./C) of the channels was
1.01120.533 nS/pF, whereas the average density in demecolcine treated cells was
0.39320.2 nS/pF. When subjected to Student's t-test, the difference between these values
was significant at the 95% confidence interval.
It is important to note that in control cells showing low channel density, abnormal
current traces similar to those seen in demecolcine treated cells were also observed. Only
three treated cells, however, had normal current, and none showed G/C values above
626 nS/pF.
Immuno-flourescence
Results obtained from immuno-flourescent microscopy of the treated cells showed an
obvious differences in labeling intensity relative to control controls. Further it was seen
Benjamin Jacobson
June 13,1995
that the labeling of cells cultured for 3 days in demecolcine (12844183demecdne) was
brighter than that of cells treated with demecolcine for 7 days (1287demecoldne)
In addition to the differences in total level of immunoreactivity, more subtle
differences in protein distribution were observed between the experimental and control
cells. Tight, bright rings were seen outlining the control cell, (figure 2a.), whereas cells
incubated in demecolcine showed less dense staining around the periphery. In the treated
cells, the stain is far more diffuse throughout the cell (figure 2b)
Immuno-precipitation
Results from the immuno-precipitation experiment show that a-tubulin is precipitated with
the KZA channel protein. The protein composing the visible band (figure 3a) has been
labeled with anti o-tubulin monoclonal anti-mouse primary antibody. A protein within the
correct size range for tubulin, 53-57 kDa (figure 3b) is observed. Two controls were
performed in order to assure that first, KZ4 was precipitated, and second that tubulin was
not precipitated independent of KZ4. The first control, a radioactively labeled immuno
precipitation prepared from the same cells used for the antibody probed precipitation,
shows that KZ4 is indeed precipitated by the anti KZA antibody (figure 3c). The second
control, à non-infected Sf9 culture, immunoprecipitated, and blotted alongside the
experimental cultures, showed no indication that tubulin was immuno-precipitated. This
suggests that only KZ4, and proteins connected to KZA were precipitated with the primary
antibody (figure 3a).
Benjamin Jacobson
June 13,1995
Discussion
Treatment with demecolcine appears to have a detrimental effect on the normal expression
of squid K24 channels in 8f9 cells. Patch clamp experiments show a significant decrease
of functional channels in the membrane. The mean conductance density of control cells
was 0618 nS/pF greater than the mean channel density for cells treated with demecolcine.
Abnormal currents described are not specific to demecolcine treated cells, but are observed
in K24 infected Sf9 cells that have been incubated continously at 28? until the time of their
recordings, a treatment that also greatly reduces the probability of observations of so called
normal' K' channels (WF Gilly, unpublished observations). It has also been observed in
control K2Z4 infected Sf9 cells, that these characteristic abnormal currents are seen when
pulse are stepped to high voltages.
in addition, the intensity of the immuno-flourescence staining was far greater in the
control cells than in those cultured in demecolcine. Decreased production of the channel as
observed by a decreased intensity in immuno flourescence experiment is indicative of a
possible effect demecolcine has on a cell's synthetic pathway
A hypothesis can be proposed that accounts for these observations. If cells are
transporting the protein to the membrane along microtubules from the golgi apparatus as
suggested by the tubulin-KZ4 interaction shown by immuno-precipitation, then the absence
of microtubules would eliminate the proper pathway for the channels to be inserted into the
membrane. This may not entirely prevent the proteins from reaching the membrane by
alternative non-microtubule-based mechanisms. In the absence of microtubules, vesicles
containing the channel protein may continue to bleb from the surface of the golgi and
diffuse throughout the cell. This would acount for the diffuse nature of the staining
observed in treated cells. Some of these vesicles may eventually reach the cell membrane
where the K' channel protein is not inserted into the membrane normally. The vesicles that
Benjamin Jacobson
June 13,1995
do not make it to the membrane may be "reabsorbed" by the cell, perhaps in endosomes,
where the K' channel protein is broken down. This could result in a lower detectable level
of protein in immuno-labeling.
The above hypothesis may also account for the presence of abnormal current seen
in most of the control cells recorded from. It is possible that whether the microtubules are
intact or not, à population of KZA containing vesicles simply diffuse to the surface
membrane, where KZ4 protein might be inserted into the membrane in a manner different
than those channels that are delivered by microtubules as described above. Therefore many
untreated cells expressing KZ4, have a low, underlying level of abnormal current
expression normally masked by the current produced by properly inserted channel protein.
In demecolcine treated cells, very few channels, if any, are transported to the membrane
and inserted properly, resulting in the expression of small, abnormal currents.
Even though the above hypothesis is questionable, and can only be validated with
further research, a clear result still emerges. Cells treated with demecolcine express a
lower number of functional KZ4 channels. A link between ion channel expression and
microtubule stability has been suggested by the experiments presented in this paper.
supporting the hypothesis that microtubules have a function in determining neuronal
excitability. The dynamic nature of microtubules, their ability to quickly form complex
networks, and change distribution might have implications on the way in which
information is processed by each individual neuron, and in turn, in the entire central
nervous system.
10
Benjamin Jacobson
June 13,1995
References
Alberts, Bray. Lewis, Raff. Roberts, and Watson. Molecula- Biology of the Cell. And ed.
New York: Garland Publishing, Inc. 1989.
De lomaso, Anthony W., Zi Jian Xie, Guoquan Liu, and Robert W. Mercer. (1993)
Expression, Targetin, and Assembly of Functional Na, K-ATPase Polypeptides in
Baculovirus insected Insect Cells. The Journal of Biological Chemistry, Vol. 268. No 2.
1470-1478.
Gilly, William F., Mary T. Lucero, and Frank T. Horrigan. (1990) Control of the Spatial
Distribution of Sodium Channels in Giant Fiber Lobe Neurons of the Squid. Neuron. Vol.
S, 663-674.
Gilly, William F. M.T. Lucero, M. Perri, and J. Rosenthal (1993), Control of the spatial
distribution of sodium channels in the squid giant axon and its cell bodies. Cephalopod
Neurobiology. N.J. Abbott, R. Williamson, and L. Maddock, eds. Oxford. Oxford
University Press.
Grafstein, Bernice and David S. Forman. (1980) Intracellular Transport in Neurons,
Physiological Reviews. Vol 60, No. 4.
Kirschner, Marc and Tim Mitchison. (1986) Beyond Self-Assembly. From Microtubules to
Morphogenesis. Cell, Vol. 45, 329-342.
Matsumoto, Gen and Kikoichi Sakai. (1979) Microtubules inside the Plasma Membrane of
Squid Giant Axons and their Possible Physiological Function. J. Membrane Biol. 50. 1-
Perri, MA., J. Rosenthal, Liu, and W. F. Gilly (1994) Cloning and sequencing of KVI.
type potassium channel sequences in the squid stellate ganglion. Biophys. J. 66. A1OS.
Kosenthal, J.C., MA Perri, and WF Gilly. (1995) Abstract. Biophysical Journal, Vol 68.
No. 2, part 2.
Summers, M.D., and G.E. Smith. (1987) Tex. Agric. Exp. Stn. Bull. Vol. SS. 1-56.
Figl.
a. Control



-100 —

GIN9SF
11
Benjamin Jacobson
June 13,1995
b. Demecolcine-treated
cells


g


oo
1 nA
50 m
Table 1
Cx(S)
—
6.38
39.25
13.45
20.75
34.75
4.96
1.065
10.825
1.7
8.527
13.93
4.25
12
mV)
I
Conrols
43.4
-29.1
-38.66
13.8
23.64
17.4
-30.48
22.9
-17.55
16.75
—
-61.5
18.95
331.8
15.2
Control mean + standard error
287demecolcine
-27.3
17.6
50.88
35.4
-18.2
2347
3demecolcine
128 4418
22.2
-25.865
-37.9
12.48
Experimental mean 4 standard error
CiPE)
Benjamin Jacobson
June 13,1995
LC(nSPE
147
2.84
773
2.07
262
07
1.011 +0.533
615
048
335
626
341
0.393 +0.2
13
Benjamin Jacobson
June 13,1995
Figure 2.
a. KZ4 protein is concentrated around the membrane in control cells.
b. KZA infected Sf9 cells, cultured for the final 3 days of an 8 day period in SOuM
demecolcine, show less dense labeling and decreased intensity
14
Benjamin Jacobson
June 13,1995
a. A7.5% SDS-PAGE seperating proteins immuno-precipitated with KZA was blotted
with with anti a-tubulin monoclonal antibody.
K21
ol
54
b. A 10% SDS-PAGE seperating proteins from a lysate of KZA infected Sf9 cells was
probed with anti a-tubulin monoclonal antibody,
c. Radiolabeled immunoprecipitation of KZA shows solid band of precipitated KZA
protein.
4u
ees
25
300
e
1123
17.2
ko.
15
Benjamin Jacobson
June 13.1995
Current Traces Seen in Demecolcine Treated Cells. a. Shows the current traces
resulting from 250 ms voltage steps (—40,0, 10,20, and 40 mV) in a KZA expressing Sf9
cell kept in culture for 8 days at 18°C. b. Shows Sf9 cells also expressing KZ4, but
incubated for the last three of the eight days in 50 uM demecolcine.
Table 1. Conductance, Resting Potential, Capacitance, and Channel density
for all cells recorded. Gy and V, were estimated as the slope and X-intercept
respectively of the I/V curve described in Materials and Methods. Cells under the heading
1e  7aemeoleme were incubated for one day aster infection at 28.C and treated with 50 ub
demecolcine for 7 days. Similarily, the cells headed 128 44183demeedene were incubated
after infection at 28°C, but were kept in normal medium at 18?C for 4 days, and then
treated with demecolcine at 18°C for only 3 days.
Figure 2. Localization and intensity of KZ4 channel protein. KZA protein was
identified with anti KZ4 polyclonal antibody, followed by Texas Red-conjugated anti-
rabbit secondary antibod
figure 3. Results that a-tubulin is immuno-precipitated with KZA channel
protein