University of Southern California Dissertations and Theses

Investigation of the molecular mechanisms underlying polarized trafficking of the potassium channels Kv4.2 and Kv1.3 in neurons

(USC Thesis Other)

Investigation of the molecular mechanisms underlying polarized trafficking of the potassium channels Kv4.2 and Kv1.3 in neurons

PDF

Download

Open document

Flip pages

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content INVESTIGATION OF THE MOLECULAR MECHANISMS UNDERLYING
POLARIZED TRAFFICKING OF THE POTASSIUM CHANNELS Kv4.2 AND
Kv1.3 IN NEURONS
by
Jacqueline Rivera
A Dissertation Presented to the
FACULTY OF THE GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(MOLECULAR BIOLOGY)
May 2007
Copyright 2007 Jacqueline Rivera
Acknowledgements
First of all I would like to thank my advisor, Don Arnold, who was not only a
mentor but also a role model. Thank you for all your guidance and encouragement
throughout my doctoral research. You have instilled in me all the research skills I
need to be a successful scientist.
I would also like to thank my dissertation committee, Dr. Michelle Arbeitman,
Dr. Christian Pike, Dr. Steve Goodman, and Dr. David McKemy. I appreciate all the
time and assistance you have given me.
Additionally, I would like to thank the administration of the molecular and
computational biology program: Eleni Yokas, Christina Tasulis, and the late Bill
Trusten for their help throughout the years. I would also like to thank Linda Bazilian
for all her support and patience.
There were days when things were just not working, many days, and I could
not have gotten through those days without my lab friends Tommy Lewis and Rudy
Mora. Thanks guys. I would also like to thank my fellow grad students Slav, Evan,
and Sudha for being great friends.
Now I would like to thank my mother, Lourdes Flores, who has sacrificed a
lot in life to see me come this far. I would also like to thank my sisters, Linda Flores,
Veronica Flores, and Brenda Flores and my brother Jaime Flores for always
believing in me. Two other people I would like to thank are my in-laws Gus Rivera
and Dalilah Rivera for all their encouragement and support throughout my doctoral
work.
ii
Lastly, I would like to thank my husband, Mark Rivera, who accompanied
me to lab late at night so many times and never complained. Without your love,
support and encouragement I would have never kept sane through all this. Of course
I have to thank my daughter, Serenity, who just made everything worthwhile.
iii
Table of Contents
Acknowledgements ii
List of Tables vii
List of Figures viii
Abbreviations x
Abstract xii
Chapter 1: Background
1.1 Polarized targeting of epithelial proteins 1
1.2 Protein motifs responsible for polarized targeting in neurons 2
1.3 Reasons to study voltage-gated potassium channels 3
1.4 Structure of voltage-gated potassium channels 3
1.5 Subcellular distribution of Kv1.4, Kv1.3 and Kv4.2 channels 4
1.6 Localization mechanisms in neurons 4
1.7 Additional mechanisms that may be involved in Kv channel 5
transport
1.8 The motor molecules: Kinesins 6
1.9 Cargo specificity of kinesins 7
Chapter 2: Methodology
2.1 DNA constructions 9
2.2 Induction of expression 12
2.3 Preparation of organotypic slice cultures 12
2.4 Biolistic Transfection 12
2.5 Preparation of dissociated neuronal cultures 13
2.6 Surface protein staining 14
2.7 Total protein staining 15
2.8 Immunocytochemistry of dissociated neuronal cultures 15
2.9 Antibody dilutions 15
2.10 Endocytosis 16
2.11 Image capture and analysis 17
2.12 Imaging of live neurons 20
Chapter 3: A dileucine-containing motif mediates somatodendritic targeting
of Kv4.2
3.1 Chimeric constructs between Kv4.2 and Kv1.4 22
3.2 Alignment of Kv4 channels identifies a 16 a.a. 22
dileucine-containing motif
iv
3.3 The 16 a.a. dileucine-containing motif is necessary for 25
somatodendritic targeting of Kv4.2
3.4 The 16 a.a. dileucine-containing motif is sufficient for 25
somatodendritic targeting
3.5 The 16 a.a. dileucine-containing motif functions in 28
an autonomous manner
3.6 The 16 a.a. dileucine-containing motif does not 33
function in a cell type specific manner
3.7 The 16 a.a. dileucine containing motif is not an 34
endocytosis signal
Chapter 4: Kif17 is the kinesin responsible for the transport of Kv4.2
4.1 Dominant negative Kif17 restricts Kv4.2 to the soma 37
4.2 Kif17 and Kv4.2 colocalize in dissociated neuronal 41
cultures
4.3 The 16 a.a. dileucine containing motif mediates 41
somatodendritic targeting independently of Kif17
4.4 Kv4.2 and Kif17 interact in coimmunoprecipitation 44
experiments
4.5 A dominant negative Kv4.2 blocks transport of Kv4.2 44
Chapter 5: The tetramerization (T1) domain of Kv1.3 mediates its axonal
targeting
5.1 The tetramerization domain (T1) of Kv1.3 is necessary 48
for its axonal localization
5.2 Vesicles containing Kv1.3 are transported selectively 51
to the axon
5.3 Somatodendritic vesicles containing TfR are also 54
directed to the axon when the T1 domain of Kv1.3
is added to TfR
Chapter 6: Kif5B is responsible for the axonal transport of Kv1 channels
6.1 Dominant negative Kif5B excludes Kv1.3 from the axon 60
6.2 Kif5B colocalizes with Kv1 channels in neuronal 66
dissociated cultures
6.3 Coimmunoprecipitation experiments show an interaction 67
between Kif5B and Kv1 channels
6.4 Dominant negative Kif5B blocks axonal targeting of 70
TfRT1-GFP
Chapter 7: Conclusion
7.1 Dendritic targeting of Kv4.2 is mediated by a 74
dileucine-containing motif
v
7.2 The possible role of the dileucine-containing motif in 75
the early secretory pathway
7.3 Possible role of the dileucine-containing motif in 76
the late secretory pathway
7.4 The tetramerization domain of Kv1.3 mediates its 77
axonal targeting
7.5 T1 interacting proteins that may play a role in 78
Kv1.3 targeting
7.6 Transport mechanism of Kv1.3 79
7.7 Identification of Kif17 as the kinesin that 79
transports Kv4.2 to the dendrites
7.8 Proteins known to interact with Kif17 that may 80
play a role in the transport of Kv4.2
7.9 Kif5B is the kinesin responsible for the axonal 82
transport of Kv1.3
7.10 Binding partners of Kif5B and possible mediators 82
in Kv1.3 transport
Bibliography 87
vi
List of Tables
Table 1: Localization of tagged and untagged proteins for dendritic motif 34
Table 2: Effect of dominant negative kinesins on the transport of Kv4.2 40
Table 3: Localization of tagged and untagged proteins for axonal motif 59
Table 4: Effect of dominant negative kinesins on the localization of Kv1.3 61
vii
List of Figures
Figure 1: Subcellular localization of K
+
channel constructs in cortical 23
pyramidal cells in slices.
Figure 2: Alignment of Shal K
+
channels reveals a conserved C-terminal 26
dileucine-containing motif that is necessary for dendritic
localization.
Figure 3: The dileucine-containing motif is sufficient to mediate dendritic 29
localization in cortical pyramidal neurons.
Figure 4: The dileucine-containing motif mediates dendritic localization 31
of CD8 in cortical pyramidal neurons.
Figure 5: The dileucine-containing motif functions in multiple cell types. 33
Figure 6: Mutation of the dileucine-containing motif does not affect the 35
rate of endocytosis.
Figure 7: A dominant negative variant of Kif17 blocks transport of Kv4.2. 38
Figure 8: Kv4.2 colocalizes with Kif17.
Figure 9: Dendritic targeting mediated by the dileucine-containing motif 43
does not require Kif-17.
Figure 10: Kv4.2 coprecipitates with Kif17. 45
Figure 11: The C-terminus of Kv4.2 acts as a dominant negative to block 47
transport of Kv4.2.
Figure 12: Surface Kv1.3 localized preferentially to the axon. 48
Figure 13: Axonal targeting signals were not found in the C-terminus or 49
the proximal N-terminus of Kv1.3.
Figure 14: A mutation in the T1 region caused Kv1.3 to localize 50
nonspecifically.
Figure 15: The T1 domain of Kv1.3 mediated axonal targeting of surface 51
protein.
viii
Figure 16: The T1 domain of Kv1.3 mediated axonal targeting of total 53
(intracellular + surface) protein.
Figure 17: Quantification of preferential targeting to the axon or 54
dendrite (calculated as axon to dendrite ratio, ADR)
for different constructs used in this study.
Figure 18: TfRT1-GFP total protein was expressed in axons, but 56
TfR-GFP was absent from axons in cortical neurons in
dissociated cultures.
Figure 19: The T1 domain mediated trafficking of transport vesicles 57
to axons in cortical neurons in dissociated cultures.
Figure 20: A dominant negative variant of Kif5B but not one of Kif17, 62
blocks axonal localization of introduced, tagged Kv1.3.
Figure 21: A dominant negative variant of Kif5B blocks axonal 64
localization of endogenous Kv1.1, Kv1.2 and Kv1.4.
Figure 22: Kif5B colocalizes withKv1.2 and Kv1.3. 66
Figure 23: Kv1 channels coimmunoprecipitate with Kif5B from brain 68
lysates when coexpressed in COS cells.
Figure 24: TfRT1-GFP colocalizes with Kif5B but TfR-GFP does not. 71
Figure 25: A dominant negative variant of Kif5B blocks axonal trafficking 72
mediated by the T1 domain.
Figure 26: Schematic model of Kv channel transport via kinesins. 85
ix
Abbreviations
ACR - Axon to cell body ratio
ADR- axon to dendrite ratio
ApoER2- apolipoprotein E receptor 2
APP- amyloid precursor protein
β-gal- β-galactosidase
CD8- cluster of differentiation 8- glycoprotein found on Tcells
Cos7 cells- cells derived from African green monkey kidney cells
CT- Carboxyl terminus
DN- Dominant negative
EAAT3- glutamate transporter excitatory amino acid transporter 3
GABA- Gamma-aminobutyric acid
GFP- Green fluorescent protein
GluR1- AMPA receptor subunit
HA- Hemagglutinin Tag
KHC- Kinesin Heavy Chain; also called Kif5B
Kif- Refers to kinesin molecule
Kv- Voltage-gated potassium channel
LDL- Low-density lipoprotein
1. Dileucine motif
MAP2- Microtubule associated protein
mGluR2- metabotropic glutamate receptor 2
x
mGluR7- metabotropic glutamate receptor 7
MVB- multivesicular body-like organelles
Myc- human p62
c-Myc
protein epitope tag
Na
v
- Voltage-gated sodium channel
NMDA- N-methyl-D-aspartate
pIND- Inducible vectors
RFP- red fluorescent protein
sADR- surface axon to dendrite ratio
T1- Tetramerization domain
tADR- total axon to dendrite ratio
TfR- Transferrin receptor
YFP- Yellow fluorescent protein
ΔCT- deleted carboxy terminus
ΔLL- deleted 16 amino acid dileucine motif
ΔNT- deleted amino terminus
xi
Abstract
Localization of voltage-gated potassium channels to specific subcellular
compartments in neurons allows them to perform specific functions that modulate
the overall electrophysiological properties of neurons. For instance, members of the
Shaker family are localized to the axon of cortical pyramidal neurons where they
contribute to the propagation of action potentials. Conversely, members of the Shal
family of K
+
channels are localized to the somatodendritic compartment of cortical
pyramidal neurons where they inhibit initiation and propagation of action potentials.
My research focuses on molecular mechanisms involved in targeting of these
specific voltage-gated potassium channels to different subcellular compartments in
neurons. By expressing chimeras between Kv1.4 and Kv4.2, two channels that are
targeted to different neuronal compartments, the axon and the dendrites, respectively,
as well as deletion mutants of Kv4.2 in cultured slices of rat cortical tissue we were
able to identify a sixteen amino acid dileucine-containing motif present on the C-
terminus that is necessary and sufficient for the somatodendritic targeting of Kv4.2.
This motif is conserved amongst all known members of the Shal family of K
+
channels and represents the first evolutionarily conserved dendritic targeting motif to
be identified. In addition, I used a similar approach to determine that the highly
conserved T1 domain found in Shaker K
+
channels is necessary and sufficient for
axonal targeting of Kv1.3. To further characterize the mechanisms responsible for
the differential targeting of Shal and Shaker channels I used a candidate approach to
identify kinesin isoforms that play a role in the transport of Kv1.3 and Kv4.2. By
xii
this method the kinesin Kif17 was identified as transporting Kv4.2 and the kinesin
Kif5B was identified as transporting Kv1.3. Overall, this research establishes that
like many other neuronal channels and receptors, the targeting of the two distinct
voltage-gated K
+
channels, Kv4.2 and Kv1.3, is modulated by motifs present on their
amino acid sequence; however the mechanisms involved have not been clearly
delineated. Investigating the interaction that these motifs might have with other
proteins in the transport pathway, including the kinesins identified, will shed some
light on this conundrum.
xiii
Chapter 1: Background
The subcellular localization of ion channels is crucial for the
electrophysiological properties of neurons yet the mechanism by which ion channels
are targeted to different subcellular compartments is not well understood. It was my
goal to further understand the mechanisms by which K
+
channels are targeted to their
specific subcellular locations in neurons.
1.1 Polarized targeting of epithelial proteins
Polarized targeting of proteins has been extensively studied in epithelial cells,
however less is known about polarized targeting in neurons. Epithelial cells have
two distinct membrane domains, an apical membrane and a basolateral membrane,
each of which expresses a distinct set of proteins. Apical targeting is controlled by
peptide motifs in the cytoplasmic domains and transmembrane domain, by N- or O-
glycans in the extracellular domain, and by domains that interact with lipid rafts.
Similarly, sequences found in the cytoplasmic domains, specifically the tyrosine-
based (YXX Φ) motif and the dileucine-based motif (containing a doublet of leucine
residues) have been found to be responsible for basolateral targeting (Winckler and
Mellman, 1999). Two examples of basolateral targeting sorting signals are found
within the LDL receptor and the Fc receptor sequences, which posses a tyrosine
motif and a dileucine motif, respectively (Matter et al., 1994). For apical sorting, a
39 amino acid sequence was identified in the protein rhodopsin, while a 32 amino
acid sequence was identified in the GABA transporter 3, both of which are present in
the cytoplasmic C-terminus (Chuang and Sung, 1998,
1
Muth et al., 1998). Protein localization studies conducted in epithelial cells serve as
good models for localization studies in neurons since both cell types are polarized.
1.2 Protein motifs responsible for polarized targeting in neurons
Several studies where proteins targeted to apical and basolateral membranes
were expressed in neurons demonstrated that apically targeted proteins preferentially
localized to the neuronal axon, while basolaterally targeted proteins preferentially
localized to the neuronal dendrites. This was the case for the LDL receptor and the
transferrin receptor, both of which localize to the somatodendritic compartment in
neurons and basolateral domain in epithelial cells (Jareb and Banker, 1998, West et
al., 1997). However, this is not universally true, since the glutamate transporter
excitatory amino acid transporter 3 (EAAT3) has been shown to be localized to the
apical compartment in epithelial cells and to the somatodendritic compartment in
neurons (Coco et al., 1997). Furthermore, the same 11 amino acid motif, found on
the C-terminus of EAAT3, was responsible for apical targeting in epithelial cells and
somatodendritic targeting in neurons (Cheng et al., 2002). Other trafficking studies
found additional examples of neuronal proteins that carried sorting motifs within
their sequences: (1) the 60 amino acid C-terminus of mGluR7 is the axonally
targeting signal which can redirect the somatodendritic mGluR2, (2) the 60 amino
acid C-terminus of mGluR2 excludes it from the axon, resulting in somatodendritic
localization (Stowell and Craig, 1999), (3) the 39 amino acids most proximal to the
transmembrane domain of the C-terminus of GluR1 constitute a dendritic sorting
motif (Ruberti and Dotti, 2000), (4) a 27 amino acid dileucine motif in the C-
2
terminus directs axonal localization of the sodium channel, Nav1.2 (Garrido et al.,
2001). These studies indicated that short motifs found on the amino acid sequence of
proteins might mediate axonal and dendritic targeting. Two of my projects involved
the identification of targeting motifs in two distinct voltage-gated potassium ion
channels.
1.3 Reasons to study voltage-gated potassium channels
For my thesis work I concentrated on voltage-gated potassium ion channels,
also known as Kv channels. Kv channels are a good model for studying protein
targeting in neurons for numerous reasons: (1) different isoforms of Kv channels
localize differently in neurons, (2) the localization of Kv channels is essential for the
electrophysiological function of neurons (Rudy, 1988), (3) they are the smallest
voltage-gated channels making them easy to manipulate utilizing molecular biology
techniques, (4) there is much known about their structure including X-ray
crystallography data.
1.4 Structure of voltage-gated potassium channels
Voltage-gated potassium channels are tetramers of Kv alpha subunits which
may associate with the accessory Kv beta subunits. The alpha subunits have a
cytoplasmic N-terminus, six transmembrane alpha-helical domains (S1-S6) and a
cytoplasmic C-terminus. The N-terminus contains the tetramerization (T1) domain
which is responsible for association of the alpha subunits. The S1-S4
transmembrane domains comprise the voltage sensing domain and the S5-S6
transmembrane domains form the pore domain. The voltage sensing domain is
responsible for detecting the voltage change across the membrane thereby triggering
3
a conformational change that leads to an open or closed conformation causing the
pore domain to allow or prevent the entry of potassium ions (Sands, 2005).
1.5 Subcellular distribution of Kv1.4, Kv1.3 and Kv4.2 channels
I conducted my experiments using two members of the Shaker Kv1 family,
Kv1.4 and Kv1.3, and a member of the Shal family, Kv4.2. These channels were
used because they target to different neuronal compartments; Kv1.4 and Kv1.3 are
targeted to the axon, while Kv4.2 is targeted to the dendrites. The axonal
localization of Kv1.4 and Kv1.3 is necessary for their participation in initiation and
propagation of the action potential down the axon. The Kv4.2 channels increase in
density from the proximal to the distal end of a dendrite. This characteristic
distribution is responsible for its role in reducing the amplitude of back propagating
action potentials along the apical dendrite (Hoffman et al., 1997, Jerng et al., 2004).
1.6 Localization mechanisms in neurons
At the beginning of my research there was nothing known about the
mechanisms responsible for the polarized targeting of different Kv channels.
However, mechanisms responsible for axonal or dendritic targeting had been
described for other neuronal proteins. Another one of my goals was to establish
whether these Kv channels were being transported by any of these known
mechanisms. One of these is compartment-specific endocytosis, which is when the
protein is transported throughout the neuron, but is selectively endocytosed from a
specific compartment, either the axon or the dendrites. The voltage-gated sodium
channel, Na
v
1.2, and synaptobrevin have been shown to be transported to both the
axon and dendrites and selectively endocytosed from the dendrites (Garrido et al.,
4
2003, Li et al., 1996). In compartment-specific vesicle docking vesicles containing a
protein will be transported to both axons and dendrites, but will only dock in axons
and only expressed on the surface of axons. Neuron-glia cell adhesion molecule,
Ng-CAM, has been shown to be axonally targeted via this mechanism (Burack et al.,
2000, Sampo et al., 2003). Another mechanism used is selective vesicular
trafficking where a protein gets transported directly to the compartment to which it is
localized. This mechanism is illustrated by the targeting of the transferrin receptor
which is loaded into vesicles that are specifically targeted only to dendrites. The last
mechanism I would like to mention is trancytosis which is where protein is targeted
to the dendrites, endocytosed then transported expressed on the axonal surface.
Some of the Ng-CAM protein was shown to localize to the axon via this mechanism
(Wisco et al., 2003).
1.7 Additional mechanisms that may be involved in Kv channel transport
Another mechanism that may be used to transport proteins is via lipid rafts.
Lipid rafts are composed of cholesterol and glycosphingolipids, their formation
occurs in the golgi, it is here that during vesicle formation membrane-spanning and
GPI anchored proteins may associate with these rafts and be transported to the cell
surface. Glycosylphosphatidylinositol (GPI)-anchored proteins have been shown to
be preferentially located in lipid rafts, and just like GPI-anchored proteins, proteins
containing palmitoylation motifs may be covalently bound to these lipid rafts, be
transported along with them from the golgi to the surface of a specific compartment
and in this way lipid rafts may be involved in protein sorting (Ikonen, 2001,
5
Rodriguez-Boulan et al., 2005). Both GAP43 and GAD65 are axonally targeted
proteins in which a palmitoylation motif is responsible for their targeting (El-
Husseini et al., 2001, Kannani et al., 2004). PSD95 also has a palmitoylation motif;
however this palmitoylation motif is necessary but not sufficient for its dendritic
targeting (El-Husseini et al., 2001). Lastly, I would like to discuss the role of
microtubules in transport mechanisms utilized by neurons. In both the axon and the
distal dendrites microtubules have their minus ends (slow growing) towards the cell
body and their plus ends (fast growing) toward the synapses, however in the
proximal dendrites the microtubules are running in both directions. This allows for
plus- and minus-end directed motors to transport to the dendrites, while only plus-
end directed motors target to the axon. Considering that (1) there are fewer motors
available for axonal transport and (2) the diameter of the axon is one tenth that of the
cell body, there must be some signals provided by microtubules for axonally targeted
proteins (Rodriguez-Boulan et al., 2005). In light of this, it has been shown that Kif5
along with EB1 have a high affinity for microtubules found at the initial segment of
the axon suggesting that microtubules may indeed provide cues for transport of cargo
by kinesins (Nakata and Hirokawa, 2003, Tirnauer and Bierer, 2000).
1.8 The motor molecules: Kinesins
Kinesins are motor proteins that transport vesicles along microtubules. They
are composed of three domains: (1) a globular head that is responsible for the
binding of the kinesin to the microtubules and for ATP hydrolysis that provides the
energy necessary for movement of the kinesin along the microtubules, (2) the stalk
6
region, which is where the kinesin monomers dimerize to form functional kinesin
molecules, (3) the tail region which is where the kinesin molecule will bind to its
specific cargo. There have been more than 60 kinesins identified including splice
variants. This large number of kinesins may be due to cargo specificity and cell type
specificity (Hirokawa and Takemura, 2004a, 2004b, 2005, Setou et al., 2004). To
further investigate the transport mechanisms of Kv1.3 and Kv4.2, I set out to identify
the kinesins responsible for their vesicular transport.
1.9 Cargo specificity of kinesins
There have been studies demonstrating that certain kinesins transport specific
cargos to the axons or the dendrites. In order to identify the kinesin isoform
responsible for transporting Kv4.2, I performed experiments with dominant negative
forms of five kinesins that are found specifically in dendrites or that have been
shown to transport dendritic proteins: Kif5B, Kif5A, Kif17, KifC2 and Kif21B.
KHC (or Kif5B) is a plus-end directed motor which was the first kinesin molecule
identified and described. Both amyloid precursor protein (APP) and apolipoprotein
E receptor 2 (ApoER2) have been identified as axonal cargo molecules for KHC,
while the GluR2 subunit has been identified as a dendritic cargo molecule for KHC
(Kamal et al., 2000, 2001, Verhey et al., 2001, Setou et al., 2002). Kif5A (or nKHC)
is a plus-end directed motor only found in neurons. Kif17 is a plus-end directed
motor that transports the NR2B subunit of NMDA receptors specifically along the
dendrites (Setou et al., 2000, Guillaud, 2003). KifC2 is a minus-end directed motor
responsible for the transport of multivesicular body-like organelles (MVB)
7
specifically along the dendrites (Saito, 1997). Finally, Kif21B is a plus-end directed
motor that is localized specifically to dendrites.
The research I conducted was pioneering work in the field of Kv channel
trafficking. I was able to identify the targeting motifs of Kv4.2 and Kv1.3 along
with the kinesins responsible for their transport to their respective compartments.
This research sets down the foundation needed to elucidate the transport mechanisms
of these channels.
8
Chapter 2: Methodology
2.1 DNA constructions
Kv4.2, Kv1.3 and Kv1.4 were each labeled with a double-Myc tag inserted
after aa residues 221, 315 and 433, respectively. All new restriction sites were
generated using PCR mutagenesis with the Quickchange protocol (Stratagene, Cedar
Creek, Texas). In Kv1.3 −Myc, Asn237 and Asn238 were mutated to Gly and Ser,
respectively, to remove glycosylation sites. Kv4.2 LL was generated by eliminating
aa residues 501 −516 of Kv4.2−Myc. Chimera 1.4_1.4_4.2 consisted of residues
1−561 of Kv1.4 ligated to residues 405 −630 of Kv4.2. Chimera 4.2_1.4_4.2
consisted of residues 1 −183 of Kv4.2 ligated to residues 308 −787 of 1.4_1.4_4.2.
Kv4.2LL/AV contains the substitutions L480A and L481V. I used four different
targeting tags: LL- FETQHHHLLHCLEKTT, AG- GAGAGA, LL5 (5’ of the LL
tag) –FETQHHHLLH, LL3 (3’ of the LL tag) – QHHHLLHCLEKTT. The tags LL,
AG, LL5 and LL3 were each inserted after residue 521 of Kv1.3 −Myc to create
Kv1.3-LL, Kv1.3-AG, Kv1.3-LL5 and Kv1.3-LL3, respectively. Kv1.4-AG and
Kv1.4-LL contain LL and AG tags after residue 607. CD8 constructs were made
from pJPA5 −CD8 −YFP. CD8-LL and CD8-AG had their respective motifs inserted
after aa residue 218. The last 19 amino acids were truncated from both constructs.
Later, the Myc-tagged rat Kv1.3 construct was modified by adding a double
hemagglutinin (HA) tag (YPYDVPDYA) following amino acid 315 in place of the
Myc. In addition, the pore mutation W386F, which has been shown to eliminate
currents in the Shaker K+ channel (Perozo et al., 1993), was introduced to Kv1.3.
9
This construct was used to generate Kv1.3DCT1, Kv1.3DCT2 and Kv1.3DNT by
deleting amino acids 448–487, 486– 525 and 3–45, respectively. All deletion
mutants were made by introducing restriction sites flanking the region to be deleted
with polymerase-chain reaction (PCR)-based mutagenesis using the Quickchange
protocol (Stratagene, Cedar Creek, TX, USA), followed by excision and ligation.
Kv1.3T67V was made with PCR-based mutagenesis as above. CD8T1 was generated
by inserting amino acids 55–185 from Kv1.3 after amino acid 227 on the C-terminus
of CD8 (gift from Gary Banker). Transferrin receptor (TfR)T1 was constructed by
adding amino acids 55–185 from Kv1.3 to the N-terminus of TfR– green fluorescent
protein (GFP), a gift from Gary Banker. All constructs were verified by restriction
digest and sequencing. Kif17 was amplified from mouse cDNA using PCR. The
dominant negative kinesin constructs were generated by replacing the motor domains
of the respective kinesin molecules with GFP or YFP and placing the resulting
constructs in the mammalian expression vector GW. For DNKif21B amino acids 2-
530 were deleted, for DNKif5B, 2-366, for DNKif5A, 2-373 and for DNKif17, 2-
323, respectively. Because KifC2 has its motor region on the C-terminus, amino
acids 370-792 were deleted and replaced with YFP. Two tagged constructs were
made for wild-type kinesins, one with GFP or YFP and the other with a double HA
epitope tag. All of the tags were on the N-terminus except those for KifC2, which
was tagged on the C-terminus. Kv4.2 was tagged with GFP on the N-terminus to
give GFP-Kv4.2. GFP-Kv4.2ΔC was made from GFP-Kv4.2 by deleting 30 amino
acids at the extreme C-terminus. GFP-Kv4.2C consists of the last 30 amino acids
10
(601-630) of Kv4.2 tagged at the N-terminus with GFP. GFPKv4.2ΔLL consists of a
GFP-tagged version of Kv4.2 with amino acids 474-489 deleted.
GFPKv4.2TΔLLΔC consists of a GFP-tagged version of the cytoplasmic tail of
Kv4.2 with amino acids 474-489 and 601-630 deleted. YFP-LL was made by tagging
the amino acids 474-489 of Kv4.2 with YFP at the N-terminus. The construct
Kv4.2ΔN was made by deleting amino acids 2-31 of Kv4.2. The inducible Kv4.2,
CD8 and CD8-LL constructs were generated by inserting Kv4.2-MYC, CD8, and
CD8-LL (10) into the pIND vector (Invitrogen). pJPA5_TFR_GFP had the GFP at
the C-terminus replaced with an HA tag (CYPYDVPDYASL) creating
pJPA5_TFR_HA. TFR_HA was digested out and ligated into the pIND vector
(Invitrogen) to create pIND_TFR_HA. pJPA5_TI_TFR_GFP had the GFP at the C-
terminus replaced with an HA tag (CYPYDVPDYASL) creating
pJPA5_TI_TFR_HA. TI_TFR_HA was digested out and ligated into the pIND
vector to create pIND_TI_TFR_HA. Other plasmids used were the β-galactosidase
expression plasmid pCMV β (Stratagene) and VgRxR (Invitrogen). All mutations
and insertions were confirmed by sequencing. In order to verify that constructs were
properly folded and trafficked to the surface, each was expressed in COS cells and
stained using a surface staining protocol (see below). COS cells were cotransfected
with 0.3 lg of the construct to be tested and 0.1 lg of GW GFP using Effectene
reagent (Qiagen, Valencia, CA, USA). Only constructs that expressed on the surface
of COS cells were expressed in neurons in slices.
11
2.2 Induction of expression
Constructs driven by the pIND promoter (Invitrogen) were cotransfected into
neurons using biolistic Transfection (see below) in cortical slices (see organotypic
slice cultures below) with the pVgRXR vector and incubated at 37°C and 5% CO
2
for 24 hours. They were subsequently induced using ponasterone A (Invitrogen),
which was added to a final concentration of 5 μM in the slice culture medium.
Immunocytochemisty was performed after an additional 48 hours of incubation.
2.3 Preparation of organotypic slice cultures
Culturing of slices was previously described (Arnold & Clapham, 1999). In
brief, 200 lm-thick slices of parietal cortex were cut from postnatal day 12–14 rats
that were killed by exposure to CO2. The slices were cultured in medium composed
of 10% Earle’s salts (Invitrogen, Carlsbad, CA, USA), Eagle’s basal medium (BME)
amino acids (Sigma, St Louis, MO, USA), BME vitamins (Invitrogen), 20 mm
NaHCO3, 10 mg/ mL bovine serum albumin (Sigma), 32 mm sucrose (Sigma), 5 μg /
mL insulin, 5 μg / mL transferrin, 5 ng / mL sodium selenite, 20 units / mL each
penicillin and streptomycin, and 2 mm l glutamine (Invitrogen).
2.4 Biolistic Transfection
Biolistic transfection refers to transfection using the gene gun. The tubing
containing the plasmid to be transfected was prepared in the following way. I placed
0.2 mg of gold beads (0.8 -1 μm in diameter) in a 1.5 ml eppendorf tube, washed
three times with 70% ethanol. I precipitated the plasmid/s (up to 21 μg) of interest
onto the gold beads by adding 50 ul of 0.1 M spermidine to the beads, vortexing,
adding the plasmid, vortxing and finally adding 50 ul of 1M CaCl
2
. Incubate for 10
12
minutes at room temperature. The solution was spun in a microfuge for one minute
to pellet the gold, the supernatant was discarded. The gold was washed three times
with 100% ethanol. Resuspend gold in 500 μl of 100% ethanol and add this to a
15ml conical tube containing 3ml of 100% ethanol plus polyvinylpyrrolidone. This
solution was used to coat tefzel tubing (Biorad) by inserting the solution into the
tubing using a syringe and allowing DNA coated gold to settle to the bottom of the
tubing, removing the liquid and turning the tubing upside down for 20 seconds then
allowing it to rotate for 3 minutes while N
2
gas was flowing through it to dry the
inside of the tubing. The tubing was cut to approximately ¾ inch sections for use in
the Helios gene gun (Biorad, Richmond, CA, USA). To transfect organotypic slices,
tubing was inserted into the gene gun which was connected to the He
2
tank. The He
2
gas was what propelled the beads out of the tubing and onto the slices. Following
incubation (37°C, 5% CO
2
) for 4 h the slices were transfected using the Helios gene
gun system. Pyramidal cells from all layers were transfected although cells from
layers 2 and 3 predominated.
2.5 Preparation of dissociated neuronal cultures
Briefly, cortices of embryonic day 18 Sprague-Dawley rat fetuses, whose
mother was killed by exposure to CO
2
, were dissected in Hank’s balanced salt
solution (Invitrogen). Cortices were dissociated by incubating in papain enzyme
solution [100 mm CaCl
2
, 50 mm EDTA (pH 7.0), 0.1% β-mercaptoethanol (ICN),
100 U papain (Sigma), in Earle’s balanced salt solution, final pH 7.4] for 30 min.
Dissociated cortical neurons were plated on polylysine-coated glass coverslips at a
13
density of 1 x 10
5
neurons / well in neurobasal medium supplemented with 10 mL / L
glutamax, 1 μg / mL gentamicin, 20 mL / L B-27 supplement and 50 mL / L fetal
bovine serum (all from Invitrogen). The medium was changed after 1 h to neurobasal
medium without fetal bovine serum and every 4 days thereafter. Dissociated cell
cultures 11–15 days in vitro were transfected using the Calphos transfection system
(BD Biosciences, San Jose, CA, USA) using procedures suggested by the
manufacturer. All experiments were performed according to the guidelines of the
University of Southern California Animal Care and Use Committee and the NIH.
2.6 Surface protein staining
After incubation for 40 h, the slices were bathed in medium containing 2
μg/ml anti-Myc monoclonal antibody (Covance, Berkeley, California) for 30 min, or
1:50 anti-CD8 mononclonal antibody (Dako, Carpinteria, California) for 5 min, 2 μg
/ mL anti-HA monoclonal antibody (Covance, Berkeley, CA, USA) for 30 min, then
fixed for 30 min in 2.5% paraformaldehyde and 4% sucrose, permeabilized with
0.2% triton, blocked and incubated in anti-GFP polyclonal antibody (Clontech, Palo
Alto, California) at 1:2,000. They were then incubated in 4 μg/ml Alexa 594 goat
anti-mouse (Molecular Probes, Eugene, Oregon) and 4 μg/ml Alexa 488 goat anti-
rabbit (Molecular Probes) for 1 h. Slices were then cleared with xylene and mounted.
COS-7 cells were incubated in primary antibody in medium for 30 min. They were
then fixed, permeabilized and incubated in secondary antibody for 1 h, and then
mounted onto slides.
14
2.7 Total protein staining
Slices were fixed with 2.5% paraformaldehyde and 4% sucrose for 30 min,
followed by permeabilization and incubation with primary antibody for 1 h.
Subsequent steps were as described above.
2.8 Immunocytochemistry of dissociated neuronal cultures
Cells were fixed with 4% paraformaldehyde for 5 min and washed with PBS.
This was followed by a 30-min blocking step with PBS containing 1% bovine serum
albumin, 5% normal goat serum and 0.1% Triton X-100. After blocking, primary
antibody was diluted in blocker and added for 1 h. Secondary antibody was diluted in
blocker and added for 30 min in the dark.
2.9 Antibody dilutions
Primary antibody dilutions for slices were as follows: mouse anti-HA 1:500
(Covance) rabbit anti-GFP 1: 2000 (BD Biosciences), chicken anti- βGal 1:1000
(ICL), mouse anti-MYC 1:500 (Covance), anti-Kv4.2 antibody (Sigma, St. Louis,
Missouri) at 1:200,and mouse anti-CD8 1:50 (Dako). Antibody labeling was
visualized by Alexa 594, Alexa 488, and Alexa 647 conjugated secondary antibodies
(Molecular Probes). Primary antibody dilutions for dissociated cultures were as
follows: chicken anti-mitogen activated protein-2 (MAP2), 1 : 6000 (AbCam,
Cambridge, MA, USA), polyclonal anti-RFP (anti-red fluorescent protein) 1 : 8000
(BD Biosciences) and monoclonal anti-GFP 1 : 200 (Molecular Probes, Portland, OR,
USA), rabbit anti-GFP 1:2000 (BD Biosciences), mouse anti- MYC 1:1000
(Covance), mouse anti-HA 1:1000 (Covance), rabbit anti-Kv4.2 1:2000 (Alomone
15
Labs), rabbit anti-GluR2 1:200 (Chemicon), and rabbit anti-Kif17 (Abcam).
Antibody labeling was then visualized by incubating cells with Alexa 488, Alexa 594,
and Alexa 647-conjugated secondary antibodies (Molecular Probes).
2.10 Endocytosis
Experiments were done in a manner described previously (Whitworth et al.,
2001). Cos7 cells were washed 2 times with phosphate-buffered saline/Ca
2+
/Mg
2+
.
The cells were then incubated with biotinylation reagent at 4°C then the temperature
was raised to 24°C in 20 minute increments from 0 to 120 minutes. The cells were
placed immediately on ice after the incubation. Surface bound biotin was removed
using buffer (150 mM NaCl, 1mM EDTA, 0.2% bovine serum albumin, 20 mM Tris)
containing freshly dissolved Mes-Na
+
. Cells were rinsed three times to remove Mes-
Na
+
. The cells were lysed using 1 ml of radioimmune precipitation buffer (100mM
Tris-Cl, pH 7.4, 150 mM NaCl, 1 mM EDTA, 1% Triton X-100, 1% sodium
depxycholate, 0.1% SDS, 1 μg/ml leupeptin, 1 μg/ml aprotinin, 250 μM
phenylmethanesulfonyl fluoride) at 4°C for 60 min. The lysates was centrifuged at
20,000 x g at 4°C for 60 min. The supernatant was taken and incubated with an
equal volume of Immunopure immobilized monomeric avidin beads for 60 min. The
beads were washed three times with radioimmune precipitation buffer, and proteins
were eluded with SDS sample buffer at RT for 30 min. The samples were run on an
SDS-Gel, blotted onto nitrocellulose paper and Kv4.2 channels were detected using
myc antibodies. Data was obtained using densitometry measurements from three
different experiments.
16
2.11 Image capture and analysis
All imaging was done on a Biorad MRC-1024 confocal microscope. Each
cell was imaged as a stack of optical sections, 1 μm in depth apart. The cell body was
approximately in the center of each image. All calculations were performed on
compressed images. For each cell, an initial image of the expressed construct was
taken using the 568 nm laser line at the same setting. Additional images were taken
of the cell at settings that were more optimal for visualization. Each cell was also
imaged using the 488 nm laser line for GFP. Each construct was imaged in at least
five different cells that were taken from at least three different slices. Cells that had
clearly identifiable axons, overall healthy morphology and were not obscured by
neighboring cells were chosen on the basis of GFP staining. An axon was identified
as a single process that (i) projected in a direction opposite to that of the cortical
surface and the apical dendrite and (ii) was clearly longer than any dendrite.
To quantify the degree of polarization in the distribution of a particular
protein, we defined two ratios, the surface axon-to-dendrite ratio (sADR) and the
total protein axon-to-dendrite ratio (tADR). sADR represents the relative density of a
particular surface protein in the axon versus that in the dendrite and is defined as
follows:
17
where F
P1,axon/dendrite
is the relative amount of surface protein P1 in the axon or
dendrite and SA
axon/dendrite
is the surface area of the axon or dendrite.
We calculated sADR from images obtained using surface staining procedures.
To calculate the ratio of the surface areas of the two compartments, we assumed that
all processes were circular in cross-section. In this case, the ratio of the two surface
areas is approximately equal to the ratio of the areas of silhouettes of the two
compartments. Note that for a cylinder of radius r and height h, the area of the
curved surface is given by 2 rh, which is proportional to the silhouette area, 2rh.
We defined tADR as follows:
where tP1
axon/dendrite
is the relative amount of the total protein (surface + intracellular)
P1 in the axon or dendrite, and GFP
axon/dendrite
is the relative amount of GFP in the
axon or dendrite. The relative amount of total protein in the axon or dendrite was
measured in the same manner as for surface protein.
To test the method for calculating ADR, we expressed pCA −GAP −GFP, a
membrane-associated GFP (mGFP), and RFP in pyramidal cells. We calculated a
simulated sADR by determining the ratio of axonal to dendritic mGFP as if it were a
18
surface-labeled protein. We calculated the ratio of the surface areas of the axon and
dendrite using the RFP image (equation 1). Expression levels (in arbitrary units)
were calculated by summing the fluorescence in both the axon and the dendrite. Only
expression levels of constructs that were detected with the same antibodies (K
+
channels or CD8 constructs) using the same immunocytochemistry procedures can
be compared. All measurements were done using ImageJ (Rasband, 1997-2006). All
values of ADR were expressed as standard error of the mean (s.e.m.). Comparisons
of ADRs were made with the Wilcoxon-Mann-Whitney test. Images shown in
figures had only brightness and contrast adjusted. Fixed neurons in dissociated
cultures were also imaged with the same confocal microscope as above. However,
because the cultures were essentially flat, only a single optical section was taken for
each cell. GFP-tagged constructs were imaged with the 488 nm line, red fluorescent
protein (RFP) was imaged with the 568-nm line, and the 647-nm line was used to
image MAP2 staining. Axons were identified as being RFP-positive but MAP2-
negative. The ratio of the expression level of Kv4.2 in the dendrites vs. the cell body
was calculated as follows: The average intensity of fluorescence associated with
Kv4.2 staining was measured in the cell body and at points 25μm from the cell body
on three dendrites for cells in dissociated culture and on the apical dendrite alone for
cells in slices. The ratio of the dendritic value to that of the cell body was calculated
for individual neurons. All analyses were performed by blinded observers, in other
words, images were labeled anonymously, scrambled and analyzed by both myself
and individuals who had never seen the images.
19
2.12 Imaging of live neurons
Cultures expressing GFP-tagged TfR constructs and RFP were imaged at 8 h
following transfection using a Deltavision deconvolution microscope. Images were
taken using a 60x oil-immersion objective. Cultures were kept at 30 °C using a
Warner confocal imaging chamber (Warner Instruments, Hamden, CT, USA).
Images were taken at 0.5-s intervals to capture vesicle movement.
20
Chapter 3: A dileucine-containing motif mediates somatodendritic
targeting of Kv4.2
In order to visualize the potassium channels Kv4.2, Kv1.3 and Kv1.4 in our
model system I had to tag these channels with a double Myc or HA epitope. The
process of tagging involves the insertion of a coding sequence for each tag in frame
with the rest of the coding sequence so as to fuse the epitope tag to the N-terminus,
C-terminus or anywhere within the protein sequence. For my experiments I was
interested in visualizing only channels present on the cell surface of the neurons,
therefore the Myc tag was placed in an extracellular loop of the channels. The first
step was to confirm the subcellular localization of Kv4.2myc, Kv1.3myc and
Kv1.4myc using our experimental system. Kv4.2myc, Kv1.3myc, or Kv1.4myc
were cotransfected along with GFP into cortical pyramidal cells of slices from P12-
P14 rats. As previously shown, surface Kv4.2myc was localized to the dendrites,
while surface Kv1.3myc and surface Kv1.4myc preferentially localized to the axon
(Figs. 1a, 1b, 1c). To quantify the amount of localized surface protein between
constructs I calculated the surface axon-to-dendrite ratio (sADR). The sADR is the
ratio of the density of protein in the axon versus that in the dendrite. If the value of
this ratio is one there are equal amounts of protein in the axon and the dendrites. A
value above one identifies a preferentially localized axonal protein, whereas a value
less than one identifies a preferentially localized dendritic protein. The sADR for
Kv4.2 was 0.06 ± 0.01 and for Kv1.3 and Kv1.4 it was 34 ± 12 and 13 ± 4,
respectively (Fig 1e) (Rivera et al., 2003).
21
3.1 Chimeric constructs between Kv4.2 and Kv1.4
Chimeric constructs between homologous proteins have been extensively used to
identify regions within these proteins important for their specific targeting (Stowell
and Craig, 1999). To identify the amino acid motif that targets Kv4.2 to the
somatodendritic domain, chimeric channels composed of the somatodendritic Kv4.2
channel and the axonal Kv1.4 channel were expressed in slices of cultured rat cortex.
After incubation for 40 hours, the subcellular localization of those channels was
examined using immunocytochemistry. The most informative chimeric construct,
1.4_1.4_4.2, contained the N-terminus of Kv1.4, the transmembrane domains of
Kv1.4 and the C-terminus of Kv4.2. This construct localized to the somatodendritic
domain (Fig. 1d) and had an sADR of 0.08 ± 0.01 (Fig 1e) suggesting that a targeting
sequence was present in the C-terminus of Kv4.2 (Rivera et al., 2003).
3.2 Alignment of Kv4 channels identifies a 16 a.a. dileucine-containing motif
Deletion mutants of the C-terminus of Kv4.2 were constructed to determine
where in the C-terminus the somatodendritic targeting motif was present. The
deletion mutants did not express on the surface of the neurons, indicating that there
were likely defects in folding and/or subunit assembly. As an alternative strategy, the
amino acid sequences in the C-termini of different Shal K
+
channels that localized to
the somatodendritic domain were compared to determine whether there were
conserved sequences (Fig. 2a) (Rivera et al., 2003). There was a common 16 amino
acid motif conserved from humans to C. elegans that contains a dileucine motif, a
sequence containing two sequential leucines, which has been implicated in
subcellular targeting in nonneuronal cell types (Miranda, et al. 2001).
22
Figure 1. Subcellular localization of K
+
channel constructs in cortical pyramidal
cells in slices. (a) GFP and Kv4.2 were coexpressed in a cortical pyramidal cell for
40 h. Anti-Myc surface staining revealed that Kv4.2 was expressed preferentially in
the dendrites. (b) Surface Kv1.3 localized predominantly to the axon, as did surface
Kv1.4 (c). (d) The chimera 1.4_1.4_4.2 localized preferentially to the dendrites. (e)
sADR (surface axonal-to-dendrite ratio) is shown for individual cells. Expression
levels shown were averaged over all cells that expressed the same construct and are
in arbitrary units. Error bars represent standard deviation (s.d.). Filled arrowheads
point to the axon; open arrowheads point to the dendrite. Note that the axons in (c)
and (d) have multiple branches and that arrowheads do not point to all minor
branches. Scale bars, 10 m.
23
24
3.3 The 16 a.a. dileucine-containing motif is necessary for somatodendritic
targeting of Kv4.2
To determine whether the 16 amino acid motif is necessary for
somatodendritic localization of Kv4.2, I constructed two different deletion mutants
and expressed them in slices of rat cultured cortex. In the first I deleted the 16 amino
acid motif from wild-type Kv4.2, this caused the Kv4.2 channel to localize
nonspecifically to the proximal axon and proximal dendrites with an sADR 3 ± 0.8.
This result indicates that the amino motif is necessary for somatodendritic
localization of Kv4.2. I also mutated the two leucines contained within the motif to
an alanine and a valine and examined the effect on localization. The mutant,
Kv4.2LL/AV, was also nonspecifically localized to the proximal axon and proximal
dendrites with a sADR of 2 ± 0.2 (Figs. 2b, 2c). To confirm these results I
conducted total protein staining of Kv4.2myc and Kv4.2LL/AV. Kv4.2 was
restricted to the dendrites while the two dileucine mutants were nonspecifically
localized with a tADR 0.8 ± 0.07 and 0.9 ± 0.08, respectively (Fig 2d) (Rivera et al.,
2003). I concluded that this dileucine-containing motif is necessary for
somatodendritic targeting.
3.4 The 16 a.a. dileucine-containing motif is sufficient for somatodendritic
targeting
To determine whether the amino acid motif is sufficient to direct an axonal
K
+
channel to the somatodendritic compartment, I added the dileucine-containing
motif to the axonal channel Kv1.3 and expressed it in slices of rat cultured cortex.
The resulting channel, Kv1.3-LL, expressed exclusively in the somatodendritic
compartment with a sADR of 0.05 ± 0.01. To ensure that the reason for the change
25
Figure 2. Alignment of Shal K
+
channels reveals a conserved C-terminal
dileucine-containing motif that is necessary for dendritic localization. (a)
Comparison of the C termini of different Shal K
+
channels. Letters shown in bold are
conserved in at least five out of the six sequences. The C. elegans gene
Y73B6BL.19.p is likely a Shal gene, as it shares 61% identity with the P. interuptus
Shal gene. Accession numbers in order: NM_008423, AAB19939, AAC52695,
2207310A, P17971, NM068574. (b, c) Kv4.2 LL and Kv4.2LL/AV localized
nonspecifically on the surface of cortical pyramidal neurons. (d) Kv4.2 total protein
is expressed preferentially in the dendrites. (e) Total protein for Kv4.2LL/AV is
distributed nonspecifically. (f) Comparisons of the ADRs of Kv4.2 LL and
Kv4.2LL/AV with that of Kv4.2 (Fig. 2) show a dramatic shift in localization with
the elimination or mutation of the dileucine-containing motif. Expression levels
shown were averaged over all cells that expressed the same constructs and are in
arbitrary units. Error bars represent s.d. Filled arrowheads point to the axon; open
arrowheads point to the dendrite. Scale bars, 10 m. The prefix 's' (b, c) refers to
surface protein; the prefix 't' (d, e) refers to total protein.
26
A
27
Kv4.1 mouse EDSGSGD---GQMLCVRSRSAFEQQHHHLLHCLEKTTCHEFTDELTFSEALGAVSLGGR-TS
Kv4.2 rat SNQLQSS-EDEPAFVSKSGSSFETQHHHLLHCLEKTTNHEFVDEQVFEES-CMEVATVNRPSS
Kv4.3 rat NEALELTGTPEEEHMGKTTSLIESQHHHLLHCLEKTTNHEFIDEQMFEQN-CMESSMQNYPST
Shal P.interuptus RLAAQESG-LEMDEFTKEEDIFEMQHHHLLRCLEKTTDREFVELEVPYNGQPNRPGSASPPQS
Shal D. melanogaster RWAAQESG-IELDDNYRDEDIFELQHHHLLRCLEKTTM
Shal C. elegans RMLAFEQG-HLSFDALRDEDIFEIQHHHLLQCLEKATEREFVESEVMFEG-----GRNTPPPS
Consensus FEXQHHHLLHCLEKTT
in localization was due to insertion of the dileucine motif and not due to interruption
of an axonal localizing motif at site of insertion, I inserted an AG tag, GAGAGA, at
the same place I had inserted the dileucine motif. The insertion of the AGs did not
cause Kv1.3 to change its localization pattern; it retained its axonal localization with
a sADR of 18 ± 2 (Figs. 3a, 3b). I confirmed these results by performing total
protein staining of Kv1.3 and Kv1.3-LL that showed total protein for the dileucine
construct mainly in the dendrites 9 ± 3 and 0.2 ± 0.02, respectively (Fig 3c, 3d)
(Rivera et al., 2003). In addition, deletion of either the 5' (LL3) or 3' (LL5) amino
acids of the dileucine-containing motif severely impairs its ability to mediate
dendritic localization of Kv1.3 (Fig 3e). From these results we can conclude that the
dileucine motif is sufficient for somatodendritic localization.
3.5 The 16 a.a. dileucine-containing motif functions in an autonomous manner
To determine whether the amino acid motif functions in an autonomous
manner I added the dileucine-containing motif to the C-terminus of CD8 and
expressed it in slices of rat cultured cortex. CD8 is a T-cell receptor that when
expressed in cortical pyramidal neurons is nonspecifically localized (Fig 4a). When
I added the dileucine-containing motif to CD8, (CD8-LL), it localized preferentially
to the somatodendritic compartment and was absent from the axon with a sADR of
0.1 ± 0.01 (Fig. 4b). As a control I also added an insert of AGs at the same place
that I had added the dileucine motif, CD8-AG. Total protein staining confirmed that
CD8-LL was preferentially localized to the somatodendritic compartment, whereas
CD8-AG was localized nonspecifically, 0.02 ± 0.005 and 2 ± 0.2, respectively (Figs.
28
Figure 3. The dileucine-containing motif is sufficient to mediate dendritic
localization in cortical pyramidal neurons. (a) Surface Kv1.3-LL localized
preferentially to the dendrites. (b) Surface Kv1.3-AG localized preferentially to the
axon. (c) Total Kv1.3 localized preferentially to the axon (d) Total Kv1.3-LL
localized preferentially to the dendrites. (e) ADR calculations verify that addition of
the dileucine-containing motif to the axonal channels Kv1.3 and Kv1.4 caused them
to localize preferentially to the dendrites over the axon, whereas addition of the AG
motif did not affect localization. In addition, deletion of either the 5' (LL3) or 3'
(LL5) amino acids of the dileucine-containing motif severely impairs its ability to
mediate dendritic localization. Total protein ADRs reflect trends that are similar to
the surface ADRs. Expression levels shown were averaged over all cells that
expressed the same construct and are in arbitrary units. Error bars represent s.d.
Filled arrowheads point to the axon; open arrowheads to the dendrite. Scale bars, 10
m.
29
30
Figure 4. The dileucine-containing motif mediates dendritic localization of CD8
in cortical pyramidal neurons. (a) Surface CD8, identified with an anti-CD8
antibody, localized nonspecifically, although there was bias toward the axon. (b)
Surface CD8-LL localized preferentially to the dendrites. (c) Total CD8-AG
localized preferentially to the axon. (d) Total CD8-LL localized preferentially to the
dendrites. (e) ADR calculations confirm that the dileucine-containing motif is
sufficient to direct the non-channel protein CD8 preferentially to the dendrites in
cortical pyramidal cells. Expression levels shown were averaged over all cells that
expressed the same construct and are in arbitrary units. Error bars represent s.d.
Filled arrowheads point to the axon; open arrowheads point to the dendrite. Scale
bars, 10 m.
31
32
4c, 4d), showing that the dileucine-containing motif functions in an autonomous
manner (Rivera et al., 2003).
3.6 The 16 a.a. dileucine-containing motif does not function in a cell type
specific manner
These results suggest that we have identified a dileucine-containing motif in
the C-terminus of Kv4.2 that is necessary and sufficient for mediating
somatodendritic localization. I further wanted to investigate if this dileucine-
containing motif was cell-type specific. To do this I expressed CD8-LL in both
dentate gyrus cells of the hippocampus and purkinje cells of the cerebellum in
organotypic slices. CD8-LL was expressed specifically to the somatodendritic
compartment, while CD8 was nonspecifically localized (Figs. 5a, 5b) (Rivera et al.,
2003). A summary of the localization of the different tagged and untagged proteins
is listed in table 1.
Figure 5. The dileucine-containing motif functions in multiple cell types. (a, b)
Surface CD8-LL localized preferentially to the dendrites in Purkinje cells of the
cerebellum (a) and dentate granule cells of the hippocampus (b). Filled arrowheads
point to the axon. Scale bars, 10 m. The prefix 's' refers to surface protein.
33
Protein Localization
Kv1.4 axonal
Kv1.3 axonal
Kv4.2 somatodendritic
1.4_1.4_4.2 somatodendritic
Kv4.2ΔLL nonspecific
Kv4.2LL/AV nonspecific
Kv1.3LL somatodendritic
CD8 nonspecific
CD8LL somatodendritic
CD8AG nonspecific
Table 1. Localization of tagged and untagged proteins for dendritic motif. A
list of the localization of untagged proteins, and the corresponding proteins tagged
with the dileucine motif (LL) or the control AG insert.
3.7 The 16 a.a. dileucine containing motif is not an endocytosis signal
It has been reported that dileucine motifs act as sorting signals by mediating
location-restricted endocytosis (Matter et al., 1994, Haft et al., 1994) . To test if the
dileucine motif on Kv4.2 is acting as an endocytosis signal, a collaborator, Dr.
Michael Quick, compared the rate of endocytosis between Kv4.2, Kv4.2ΔLL, and
Kv4.2LL/AV. The rate of endocytosis between the wild-type and mutant Kv4.2
constructs was not significantly different indicating that the dileucine motif is most
likely not acting as an endocytosis signal (Fig. 6) (Rivera et al., 2003).
34
Figure 6. Mutation of the dileucine-containing motif does not affect the rate of
endocytosis. The amount of internalized protein at each time point is plotted as the
percentage relative to surface Kv4.2 protein at the start of the assay. Internalization
was performed at 24 °C; the non-internalizing control experiment was performed at 4
°C. Data are densitometry measurements from three separate experiments.
Representative immunoblots from one experiment are shown for wild-type Kv4.2
(upper blot) and for Kv4.2LL/AV (lower blot). Values under each blot refer to the
time permitted for internalization. 'S' denotes the total amount of surface Kv4.2
protein at the start of the assay.
35
36
Chapter 4: Kif17 is the kinesin responsible for the transport of Kv4.2
4.1 Dominant negative Kif17 restricts Kv4.2 to the soma
To determine which kinesin is responsible for the targeting of Kv4.2, I made
constructs to express dominant negative variants of kinesins, which would be able to
bind to the vesicles but not be able to transport them. These included Kif17, Kif5B,
Kif5A, KifC2, and Kif21b. The variants had the motor replaced with a GFP
molecule (Fig. 7a). This dominant negative kinesin would still be able to form a
heterodimer with a wild-type kinesin molecule, but it would not be able to move
along the microtubules. One limitation that I encountered was that if I cotransfect
the dominant negative kinesin along with the Kv4.2myc, there would be some time
needed for the dominant negative to be expressed and take effect. To surmount this
limitation, I used an inducible vector system, pIND. pIND-Kv4.2myc was
cotransfected with the dominant negative kinesin and the dominant negative kinesin
was allowed to express for 48 hours in slices of rat cultured cortex. This time
allowed most of functional kinesin molecules to turnover and be replaced by the
dominant negative form. After the 48 hours, pIND-Kv4.2myc was induced and its
localization was investigated by immunocytochemistry (Fig. 7b). If one of the
dominant negative kinesin molecules caused localization of induced Kv4.2 to be
restricted to the cell body, I would deduce that the kinesin was involved in the
transport of the potassium channel. In order to quantify the effect of particular DN-
kinesins on localization of Kv4.2, we calculated the ratio of the expression level of
the tagged channel in the dendrites vs. that in the cell body. Lower ratios indicate
37
Figure 7. A dominant negative variant of Kif17 blocks transport of Kv4.2. (A)
Dominant negative variants of different kinesin isoforms (GFP-DNKinesin) were
made by replacing the motor domains of wild-type kinesins (WT Kinesin) with GFP.
(B) One of five dominant negative kinesin constructs was initially expressed in
neurons in cortical slices for 24 hours, allowing time to dimerize with endogenous
kinesins in these cells. Subsequently, tagged Kv4.2 was expressed for 48 hours in the
same cells. ( β-Galactosidase was also expressed in the same cells as a counterstain.)
Slices were then fixed and stained for Kv4.2 to determine whether transport of the
channel had been blocked. (C) Cells expressing Kv4.2 and either GFP-DNKif17 or
GFP-DNKif5A (and β-Galactosidase as a counterstain). GFP-DNKif17 completely
blocked transport of Kv4.2, whereas GFP-DNKif5A did not. (D) Expression of
GFP-DNKif17 blocked transport of Kv4.2 in 80% of cells (n=25). Each of GFP-
DNKif21B, GFP-DNKif5A, GFP-DNKif5B, and GFP-DNKifC2 blocked transport
of Kv4.2 in less than 20% of cells. Average expression levels of Kv4.2 and of the
dominant negative (DN) construct did not vary dramatically for different DN
constructs. Transport of Kv4.2 was considered to be blocked when there was no
expression of the channel observed more than 25 μm into processes. (E) GFP-
DNKif5B blocked transport of endogenous GluR2 in a cortical neuron in dissociated
culture. When only GFP is expressed in a similar cell endogenous GluR2 is found
throughout the dendrites. Scale bars are 10 μm.
38
39
that a protein is localized preferentially to the cell body, consistent with blocking of
transport to the dendrites of that protein. Immunocytochemistry results demonstrated
that GFP-DNKif17 was able to restrict expression of Kv4.2 to the cell body with
ratio of 0.06 ± 0.003, whereas all the other dominant negative kinesins had no effect
on transport of Kv4.2 (Fig. 7c). I show the result with GFP-DNKif5B as a
representative result of the dominant negative kinesins with no effect on transport of
Kv4.2 (Fig. 7c). These results are quantified on Figure 7d. It has been established
that Kif5B is the motor responsible for the transport of GluR2, to test if the dominant
negative strategy was functioning as I predicted, I transfected GFP-DNKif5B into
neuronal dissociated cultures and performed immunocytochemistry for endogenous
GluR2. The GFP-DNKif5B restricted GluR2 expression to the cell body of these
cultures (Fig. 7e). As a control for the dominant negative Kif17, the nonspecifically
localized protein CD8 was transfected with the GFP-DNKif17 to show that its
localization was not affected 0.64 ± 0.07. Table 2 lists the dominant negative
kinesins and their effect on Kv4.2 localization. Since the dominant negative Kif17
Dominant negative
kinesin
Potassium
Channel Effect on Kv channel localization
GFP (control) Kv4.2 no effect (remained somatodendritic)
GFP-KHC Kv4.2 no effect (remained somatodendritic)
GFP-nKHC Kv4.2 no effect (remained somatodendritic)
GFP-Kif21b Kv4.2 no effect (remained somatodendritic)
GFP-KifC2 Kv4.2 no effect (remained somatodendritic)
GFP-Kif17 Kv4.2 restricted to cell body
Table 2. Effect of dominant negative kinesins on the transport of Kv4.2.
40
was the only one that restricted Kv4.2 to the cell body it provided proof that this was
the kinesin responsible for its transport (Chu et al., 2006).
4.2 Kif17 and Kv4.2 colocalize in dissociated neuronal cultures
To corroborate these studies I tested if Kif17 and Kv4.2 are associated. Here,
I transfected YFP-Kif17 and Kv4.2myc into neuronal dissociated cultures, an in vitro
system of the rat cortical slices, and found that indeed puncta staining for YFP-Kif17
and Kv4.2myc colocalized (Fig. 8b). To support this result, Dr. Po-Ju Chu in our
laboratory used neuronal dissociated cultures to perform colocalization studies with
endogenous Kv4.2. Transfection of HA-Kif17 and immunocytochemistry for both
HA and endogenous Kv4.2 showed colocalization of puncta staining for both HA
and Kv4.2 (Fig. 8a) (Chu et al., 2006). These colocalization studies further
supported that Kif17 was the motor responsible for transporting Kv4.2.
4.3 The 16 a.a. dileucine containing motif mediates somatodendritic targeting
idependently of Kif17
I had previously showed that the dileucine-containing motif found in Kv4.2
was necessary and sufficient for somatodendritic localization (Figures 2 and 3).
When I inserted the dileucine containing motif to CD8, making CD8LL, I was able
to show that this conferred somatodendritic localization to CD8. Now that there was
sufficient evidence that Kif17 was responsible for transporting Kv4.2, I wanted to
see what effect expressing GFP-DNKif17 with CD8LL in slices of rat cultured
cortex would have on the localization of CD8LL. I was surprised to find that
dominant negative Kif17 had no effect on the localization of CD8LL which retained
its somatodendritic localization with ADR of 0.12 ± 0.04 (Fig 9), which was
41
Figure 8. Kv4.2 colocalizes with Kif17. (A) A cortical neuron in dissociated culture
transfected with HA-Kif17 and stained for HA (green) and endogenous Kv4.2
(purple). The two proteins show a high degree of colocalization (white). (B) A
cortical neuron expressing Kv4.2-MYC (purple) and YFP-Kif17 (green). The two
proteins show a high degree of colocalization. Regions inside white boxes on low
power images are shown in high power images. Scale bars are 10 μm.
42
Figure 9. Dendritic targeting mediated by the dileucine-containing motif does
not require Kif-17. (A) GFP-DNKif17 was initially expressed in neurons in cortical
slices. After 24 hours CD8-LL (CD8 with the dileucine-containing motif added to
the C-terminus) was expressed in the same cells for 48 hours. CD8-LL was targeted
specifically to dendrites despite the presence of GFP-DNKif17, indicating that Kif17
was likely not necessary for transport of CD8-LL. Scale bar is 10 μm. Unfilled
arrows point to the axon. Filled arrows point to dendrites. (B) The average axon to
dendrite ratio (ADR) of CD8-LL was less than 1 when expressed in the presence of
either GFP-DNKif17 or GFP, indicating that GFP-DNKif17 did not block dendritic
targeting. In contrast, CD8 expressed in a nonspecific manner with an average ADR
of approximately 1. The expression level of CD8LL was not significantly different in
cells expressing GFP vs. cells expressing GFP-DNKif17. DN refers to dominant
negative construct.
comparable to the control where CD8LL was expressed with GFP, ADR of 0.14 ±
0.04 (Chu et al., 2006). This suggested that the dileucine-containing motif on Kv4.2
was mediating somatodendritic localization independently of Kif17.
43
4.4 Kv4.2 and Kif17 interact in coimmunoprecipitation experiments
Coimmunoprecipitation studies performed by Dr. Chu in Cos7 cells
confirmed that Kv4.2 was interacting with Kif17. These studies further showed that
indeed the dileucine-containing motif was not interacting with Kif17 because GFP-
Kv4.2ΔLL did not immunoprecipitate HA-Kif17. Dr. Chu’s results showed that the
C-terminus of Kv4.2, GFP-Kv4.2CT, was able to precipitate HA-Kif17 (Fig 10)
(Chu et al., 2006). These results suggested that it was the C-terminus of Kv4.2 that
was directly or indirectly interacting with Kif17.
4.5 A dominant negative Kv4.2 blocks transport of Kv4.2
To verify that the C-terminus of Kv4.2 interacts with Kif17, I generated a
dominant negative of Kv4.2. This dominant negative was a fusion protein between
GFP and the last 30 amino acids of Kv4.2, GFP-Kv4.2CT, when cotransfected into
slices of rat cultured cortex with inducible Kv4.2 it is expected to restrict Kv4.2
localization to the cell body because it will compete with Kv4.2 for binding with
Kif17. As anticipated, cotransfection of the GFP-Kv4.2CT construct and Kv4.2
resulted in Kv4.2 expression restricted to the cell body (Fig 11) (Chu et al., 2006).
44
Figure 10. Kv4.2 coprecipitates with Kif17. HA-Kif17 coimmunoprecipitated with
full-length GFP-Kv4.2, with the 30 amino acids at the distal C-terminus of Kv4.2
(GFP-Kv4.2C), with Kv4.2 lacking the 16 amino acid dileucine motif (Kv4.2ΔLL),
and with Kv4.2 lacking amino acids 2-31 from the N-terminus (GFP-Kv4.2ΔN). HA-
Kif17 did not coprecipitate with the channel lacking the distal 30 amino acids of the
C-terminus (GFP-Kv4.2ΔC), or with the 16 amino acid dileucine-containing motif of
Kv4.2 (YFP-LL). Neither HA-Kif5A nor HA-Kif5B coimmunoprecipitated with
Kv4.2C. All pairs of proteins were first coexpressed in COS cells, then
immunoprecipitated with anti-GFP antibodies from lysates and probed with anti-HA
antibodies to detect kinesin isoforms. To ensure that immunoprecipitation was
specific to anti-GFP, aliquots of the same lysates were also immunoprecipitated with
rabbit IgG antibodies and probed with anti-HA. To ensure that in each case, the anti-
GFP antibody precipitated the GFP-tagged protein, aliquots of each lysate that had
been immunoprecipitated with anti-GFP were probed either with anti-GFP (GFP-
Kv4.2C and YFP-LL) or with anti-Kv4.2 (GFP-Kv4.2, GFP-Kv4.2ΔC, GFP-
Kv4.2ΔN, GFP-KV4.2ΔLL).
45
46
Figure 11. The C-terminus of Kv4.2 acts as a dominant negative to block
transport of Kv4.2. (A) Either a protein consisting of the C-terminal 30 amino acids
of Kv4.2 fused with GFP (GFP-Kv4.2C) or GFP alone was expressed in neurons in
cortical slices and, 24 hours later, tagged Kv4.2 was expressed in the same cells for
48 hours. GFP-Kv4.2C completely blocked transport of Kv4.2, while GFP did not
affect transport. Scale bar 10 μm. (B) Expression of GFP-Kv4.2C prior to expression
of Kv4.2 blocked transport of the channel in 90% of cells. Expression of GFP prior
to Kv4.2 blocked transport of the channel in less than 10% cells. Expression levels of
Kv4.2 were not significantly different in cells expressing GFP vs. cells expressing
GFP-Kv4.2C. Transport was considered to be blocked when Kv4.2
immunoreactivity was not present above background more than 25 μm into any
processes.
47
Chapter 5: The tetramerization (T1) domain of Kv1.3 mediates its axonal
targeting
5.1 The tetramerization domain (T1) of Kv1.3 is necessary for axonal
localization
To determine the targeting motif of the axonally targeted Kv1.3 channel (Fig
12), I began by constructing deletion mutants and examining them in slices of rat
cultured cortex. I deleted the C-terminus of Kv1.3 and did so by deleting it in two
separate parts. First I deleted the most proximal C-terminus of Kv1.3, making
Kv1.3ΔCT1, this construct still localized to the axon (Fig 13a). I then deleted the
most distal part of the C-terminus, making Kv1.3ΔCT2, but this construct targeted to
the axon as well (Fig 13b). The construct with the deleted N-terminus, Kv1.3ΔNT,
remained localized to the axon (Fig 13c).
Figure 12. Surface Kv1.3 localized preferentially to the axon. (A) Schematic of
Kv1.3 construct containing an HA tag on the extracellular loop between S3 and S4
and the pore mutant W386F. The T1 domain is marked in black. (B) Images of
expressed GFP and surface Kv1.3. Kv1.3 expressed predominantly on the axonal
surface. Arrows point to the axon, which projects to the top of the figure. Scale bar,
10 μm.
48
Figure 13. Axonal targeting signals were not found in the C-terminus or the
proximal N-terminus of Kv1.3. (A–C) Surface Kv1.3DCT1, Kv1.3DCT2 and
Kv1.3DNT were localized specifically to the axon. (D) Schematics of the deletion
mutants. The T1 domain is marked in black. Arrows point to the axon, which
projects to the top of the figure. Scale bar, 10 μm.
49
When I deleted the N-terminus of Kv1.3 I left the tetramerization domain
intact because others had shown that deleting it would result in nontetramerization of
Kv1.3 subunits and retention in the ER (Zerangue et. al., 2000). To circumvent this
problem, I mutated a threonine to a valine at position 67 which is part of the T1
domain, Kv1.3T67V. This mutation caused Kv1.3 to be nonspecifically localized, to
the proximal dendrites and the proximal axon, suggesting that the T1 domain might
be critical for axonal localization of Kv1.3 (Fig 14). I wanted to know if the T1
domain from Kv1.3 would cause a nonspecifically localized protein to be targeted
only to the axon. For this experiment, I fused the T1 domain from Kv1.3 to the
cytoplasmic N-terminus of CD8. The CD8T1 construct showed surface expression
mainly in the axon suggesting that the T1 domain of Kv1.3 is sufficient for axonal
localization (Fig 15) (Rivera et al., 2005).
50
Figure 14. A mutation
in the T1 region caused
Kv1.3 to localize
nonspecifically.
(A) Schematic of
Kv1.3T67V, which has a
mutation at the N-
terminal end of the T1
domain. (B) Surface
Kv1.3T67V localized to
the soma and dendrite in
addition to the proximal
axon. Arrows point to
the axon, which projects
to the top of the figure.
Scale bar, 10 μm.
Figure 15. The T1 domain of Kv1.3 mediated axonal targeting of surface
protein. (A) Surface CD8 localized nonspecifically to both the axon and the
dendrites. (B) Surface CD8T1 localized specifically to the axon and was largely
absent from dendrites. (C) Schematic of CD8T1. CD8 is marked in grey and the T1
domain is marked in black. Arrows point to the axon, which projects to the top of the
figure. Scale bar, 10 μm.
5.2 Vesicles containing Kv1.3 are transported selectively to the axon
Now that I had identified the targeting motif in Kv1.3, I wanted to investigate
the mechanism used to transport Kv1.3 to the axon. One straight forward
mechanism is that vesicles containing Kv1.3 are transported directly to and only to
the axon. Another mechanism is that the vesicles containing Kv1.3 could be
transported to both the axon and dendrites but be selectively endocytosed from the
dendrites therefore only expressed on the surface of the axon. To determine which
51
of these mechanisms was responsible for axonal expression of Kv1.3, I performed
total (intracellular +surface) protein immunocytochemistry of CD8TI in slices of rat
cultured cortex. Total protein staining is informative because surface staining of
CD8T1 shows that it is only in the axon, but total protein staining will also show
where intracellular protein is being transported, only to the axon or to both the axon
and the dendrites. I showed that CD8T1 was transported only to the axon, in contrast
to CD8 which was transported to both the axon and dendrites (Fig 16a,b). I had
previously shown (Figure 5) that the dileucine-containing motif of Kv4.2 was
functional in cerebellar purkinje cells and dentate granules cells. I performed a
similar experiment in cerebellar slices with CD8T1, but CD8T1 was nonspecifically
localized in cerebellar purkinje cells (Fig 16c). This suggested that the function of
the T1 domain was cell-type specific. All ADR values for Kv1.3 and CD8 constructs
used for the determination of the T1 domain as the axonal targeting signal are
summarized in Figure 17 (Rivera et al., 2005).
52
Figure 16. The T1 domain of Kv1.3 mediated axonal targeting of total
(intracellular + surface) protein. (A) CD8 total protein labeled under
permeabilized conditions localized to both the dendrite and the axon of a cortical
pyramidal neuron. (B) CD8T1 total protein was expressed specifically in the axon of
a cortical pyramidal neuron. (C) CD8T1 total protein localized nonspecifically in
Purkinje cells. Arrows point to the axon, which projects to the top of the figure.
Scale bar, 10 μm.
53
Figure 17. Quantification of preferential targeting to the axon or dendrite
(calculated as axon to dendrite ratio, ADR) for different constructs used in this
study. The quantification of ADRs for either surface proteins (present on plasma
membrane) or total proteins (present on membrane and intracellularly) are as
indicated. Expression level of each construct is measured by fluorescence intensity in
arbitrary units and can only be compared between constructs that have been stained
with the same antibody and using the same procedures. ‘Purk. CD8T1’ refers to
CD8T1 expressed in Purkinje cells. All other results were obtained from cortical
pyramidal cells. Error bars represent SEM.
5.3 Somatodendritic vesicles containing TfR are also directed to the axon when
the T1 domain of Kv1.3 is added to TfR
To further characterize the mechanism by which the T1 domain mediated
axonal targeting of vesicles, I performed live imaging experiments with CD8 and
CD8T1 tagged to GFP. Unfortunately the vesicles were difficult to discern due to
their small diameter. To circumvent this problem I used the transferrin receptor (TfR)
which has been shown to be transported in much larger vesicles. TfR-GFP is
targeted to dendrites and excluded from the axon, whereas TfRT1-GFP is targeted to
both the axon and the dendrites (Fig 18). Even though this system is not as clear cut
as the CD8 system since TfRT1-GFP does not get targeted solely to the axon, it
works well because the vesicles are large enough to be visualized by live imaging
54
technology. For these experiments, I had to perform live imaging experiments
followed by immunocytochemistry of those same cells to determine the identity of
the processes. I cotransfected RFP and TfR-GFP or RFP and TfRT1-GFP into
neuronal dissociated cultures and did immunocytochemistry against MAP2, RFP,
and GFP. The MAP2 and RFP staining served to identify the axon of the neuron,
while GFP localized the vesicles carrying TfR or TfR-T1. TfRT1 vesicles were
found to travel into a process which was identified via immunocytochemistry as the
axon, while TfR containing vesicles were only found in the dendrites (Fig 19)
(Rivera et al., 2005). Using live imaging technology in combination with
immunocytochemistry, I was able to show that the T1 domain of Kv1.3 directed
TfRT1 containing vesicles not only to the dendrites, but to the axon as well. Table 3
is a summary of the localization of proteins used to determine the role of T1.
55
Figure 18. TfRT1-GFP total protein was expressed in axons, but TfR-GFP was
absent from axons in cortical neurons in dissociated cultures. (A) Neuron stained
for introduced RFP and TfRT1-GFP and for endogenous MAP2. Filled arrows point
to the dendrites, which are stained for both RFP and for MAP2. Unfilled arrows
point to the axon, which is stained for RFP but not for MAP2. TfRT1-GFP was
present in both axons and in dendrites. In the schematic diagram the axon is shown
in grey and the dendrites are black. (B) Neuron stained for introduced RFP and TfR-
GFP and for endogenous MAP2. TfR-GFP labelling was confined to the dendrites.
(C) Axon-to-dendrite ratios for total protein (tADR) and expression levels for
neurons expressing TfR-GFP and TfRT1-GFP. TfR-GFP localized to the dendrite
(tADR, 0.08 ± 0.08), while TfRT1-GFP localizes nonspecifically (tADR, 1.3 ± 0.7).
The expression levels of the two proteins are similar (6 ± 5 for TfRGFP vs. 8 ± 7 for
TfRT1-GFP). Scale bars, 10 μm.
56
Figure 19. The T1 domain mediated trafficking of transport vesicles to axons in
cortical neurons in dissociated cultures. (A–C) A fixed neuron in dissociated
culture stained for introduced RFP and TfR-GFP and for endogenous MAP2. (D)
Schematic showing the axon in grey and dendrites in black. (E) Live image of the
same neuron as in A–D showing vesicles in the dendrites but not in the axon. The
region shown is outlined in the box in (D). Filled arrows point to the dendrites,
which were stained both for RFP and for MAP2. Unfilled arrows point to the axon,
which was stained for RFP but not for MAP2. (F,G) A fixed neuron in dissociated
culture stained for introduced TfRT1-GFP and for endogenous MAP2. (H)
Schematic showing the axon in grey and dendrites in black. (I) Live image of the
same neuron as in F–H showing vesicles in the axon. The region shown is outlined in
the box in H. Scale bars, 10 μm.
57
58
Construct Localization
Kv1.3ΔCT1 axonal
Kv1.3ΔCT2 axonal
Kv1.3ΔNT axonal
Kv1.3T67V nonspecific
CD8T1 axonal
TfR-GFP somatodendritic
TfRT1-GFP axonal-somatodendritic
Table 3. Localization of tagged and untagged proteins for axonal motif. A list of
the protein localization of untagged and tagged proteins used to determine the role of
the T1 domain.
59
Chapter 6: Kif5B is responsible for the axonal transport of Kv1 channels
6.1 Dominant negative Kif5B excludes Kv1.3 from the axon
It had previously been shown that Kv1 channels in the giant squid axon were
transported by Kif5B (Clay and Kuzirian, 2001). At this time I wanted to identify
the kinesin responsible for transport of Kv1.3 in mammalian cells. I took the same
dominant negative approach as I did for Kv4.2 and Kif17. I used dominant negatives
of Kif5B, Kif17, or Kif3A cotransfected with Kv1.3 in slices of rat cultured cortex.
To determine whether transport of the channel had been blocked I calculated the
axon to cell body ratio (ACR) which is the ratio of the expression level of the
induced channel at a location 25 μm into the axon with the expression level of the
same channel in the cell body. DN-GFPKif5B was the only kinesin to exclude
Kv1.3 from the axon, ACR 0.02 ± 0.007 (Fig 20a). Kv1.3 targeting was unchanged
when coexpressed with DN-GFPKif17 and DN-GFPKif3A, ACR 0.33 ±0.07 and
0.29 ±0.05, respectively (Fig 20b,c). This suggested that Kif5B was responsible for
the transport of Kv1.3. To investigate whether Kif5B is responsible for the transport
of other Kv1 channels I performed endogenous immunocytochemistry for Kv1.1,
Kv1.2 and Kv1.4 in neuronal dissociated cultures transfected with GFP-DNKif5B.
As controls I performed similar experiments but transfected either GFP or GFP-
DNKif17. Neurons transfected with GFP-DNKif5B excluded the Kv1 channels from
the axon and restricted expression of the Kv1 channels mostly to the cell body, for
Kv1.1 ACR was 0.03 ± 0.01 (Fig 21a,b,c). Transfection with either GFP or GFP-
DNKif17 had no effect on axonal localization of the Kv1 channels, for Kv1.1 the
60

Abstract (if available)

Abstract Localization of voltage-gated potassium channels to specific subcellular compartments in neurons allows them to perform specific functions that modulate the overall electrophysiological properties of neurons. For instance, members of the Shaker family are localized to the axon of cortical pyramidal neurons where they contribute to the propagation of action potentials. Conversely, members of the Shal family of K+ channels are localized to the somatodendritic compartment of cortical pyramidal neurons where they inhibit initiation and propagation of action potentials. My research focuses on molecular mechanisms involved in targeting of these specific voltage-gated potassium channels to different subcellular compartments in neurons. By expressing chimeras between Kv1.4 and Kv4.2, two channels that are targeted to different neuronal compartments, the axon and the dendrites, respectively, as well as deletion mutants of Kv4.2 in cultured slices of rat cortical tissue we were able to identify a sixteen amino acid dileucine-containing motif present on the C-terminus that is necessary and sufficient for the somatodendritic targeting of Kv4.2. This motif is conserved amongst all known members of the Shal family of K+ channels and represents the first evolutionarily conserved dendritic targeting motif to be identified. In addition, I used a similar approach to determine that the highly conserved T1 domain found in Shaker K+ channels is necessary and sufficient for axonal targeting of Kv1.3. To further characterize the mechanisms responsible for the differential targeting of Shal and Shaker channels I used a candidate approach to identify kinesin isoforms that play a role in the transport of Kv1.3 and Kv4.2. By this method the kinesin Kif17 was identified as transporting Kv4.2 and the kinesin Kif5B was identified as transporting Kv1.3.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

A novel pulse-chase paradigm to visualize the trafficking of transport vesicles in neurons

PDF

Molecular mechanisms underlying polarized distributions of mRNA and transmembrane proteins in neurons

PDF

Biophysical and molecular identification of ion channels in sour taste

PDF

Transient receptor potential A1 channel in nociceptive sensation of weak acids and carbon dioxide and its regulation

PDF

The role of the cofilin/Limk1 signaling pathway in axon growth during development and regeneration

PDF

Molecular mechanisms underlying the effects of CB1 antagonism in canine model of obesity

PDF

The structure of loop 2 is important for agonist and ethanol sensitivity in glycine and GABA Alpha receptors

PDF

The expected and unexpected roles of TRPM8: cold pain and metabolism

PDF

Investigating the role of ion channel activity of coronaviruses envelope protein in CD1d regulation

PDF

A million-plus neuron model of the hippocampal dentate gyrus: role of topography, inhibitory interneurons, and excitatory associational circuitry in determining spatio-temporal dynamics of granul...

Asset Metadata

Creator Rivera, Jacqueline (author)

Core Title Investigation of the molecular mechanisms underlying polarized trafficking of the potassium channels Kv4.2 and Kv1.3 in neurons

School College of Letters, Arts and Sciences

Degree Doctor of Philosophy

Degree Program Molecular Biology

Publication Date 04/25/2007

Defense Date 10/19/2006

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag kinesin,OAI-PMH Harvest,potassium channels,targeting motif

Language English

Advisor Arnold, Donald B. (committee chair), Arbeitman, Michelle N. (committee member), Goodman, Steve (committee member), McKemy, David D. (committee member), Pike, Christian (committee member)

Creator Email jacquelf@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-m467

Unique identifier UC1187197

Identifier etd-Rivera-20070425 (filename),usctheses-m40 (legacy collection record id),usctheses-c127-493788 (legacy record id),usctheses-m467 (legacy record id)

Legacy Identifier etd-Rivera-20070425.pdf

Dmrecord 493788

Document Type Dissertation

Rights Rivera, Jacqueline

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Repository Name Libraries, University of Southern California

Repository Location Los Angeles, California

Repository Email cisadmin@lib.usc.edu