Regulation of axonal development by the cGMP signaling pathway

PDF

Download

Open document

Flip pages

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content REGULATION OF AXONAL DEVELOPMENT BY
THE cGMP SIGNALING PATHWAY

By
Zhen Zhao

A Dissertation Presented to the
FACULTY OF THE USC GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirement for the Degree
DOCTOR OF PHILOSOPHY
(NEUROSCIENCE)
December 2009

Copyright 2009 Zhen Zhao

ii
Acknowledgements
I would like to express my gratitude to all those who assisted me to complete my work
and this thesis. I want to thank University of Southern California, Neuroscience Graduate
Program and Zilkha Neurogenic Institute for giving me the opportunity to pursue my
doctoral degree.
I am deeply indebted to my supervisor Prof. Dr. Le Ma. It was his guidance,
encouragement, suggestions and full support that made this work possible. I also have to
thank my committee members, Dr. Samantha Buttler, Dr. James Knowles, Dr. Emily
Liman, Dr. David Mckemy, Dr. Zuo-zhong Wang and Dr. Qilong Ying for their valuable
help, and Dr. Li Zhang’s lab, Dr. Jonah Chan and Dr, Zuo-zhong Wang’s lab for sharing
their reagents and equipments.
I have furthermore to thank all my colleagues in Dr. Ma’s lab who have worked on my
side, assisted and inspired me.
I would like to give my special thanks to my parents, my wife Yan Liu and our baby boy
Jiaming, whose love is my strength to complete this work.

iii
Table of Contents
Acknowledgements ii
List of Tables v
List of Figures vi
Abstract viii
Introduction 1
Chapter 1: cGMP Regulates Axon Branching via PrkG1 5
1.1 cGMP Increases Formation of Axonal Branches in Cultured 6
Sensory Neurons
1.2 Activation of PrkG1 Is both Necessary and Sufficient for 13
DRG Axon Branching in Culture
Chapter 2: GSk3 Is a Novel Substrate for PrkG1 19
2.1 cGMP Activation Leads to Phosphorylation of GSK3 in 20
both DRG Neurons and COS Cells
2.2 PrkG1 Binds and Phosphorylates GSK3 Directly 24
2.3 A Dominant Active GSK3β Phospho-mutant Inhibits 28
cGMP-induced Branching in DRG Neurons
Chapter 3: Natriuretic Peptides Can Regulate Axonal Development 32
3. 1 The CNP Precursor Is Expressed in the Developing 33
Spinal Cord
3.2 CNP Promotes Axon Branching of Dissociated DRG 35
Neurons in culture
3. 3 CNP Stimulates Axon Outgrowth from DRG Explants 38
3.4 A Point Source of CNP Attracts the Growth Cones of 42
DRG Neurons
3.5 The Specificity of NPs in Axonal Development 45

iv
Chapter 4: cGMP Pathway Controls Sensory Axon Bifurcation 46
4. 1 Developmental Requirement of PrkG1 in DRG Axon 47
Bifurcation in Mice
4.2 CNP Is the Environmental Cue for the Sensory Neurons to 51
Bifurcate
4.3 CNP Is Expressed in a Gradient Pattern in Dorsal Part of 55
the Spinal Cord
4.4 Peripheral Sensory Projections in the Absence of CNP 58
4.5 Bifurcation Pattern in GSK3α/β 21A/21A/9A/9A Double 60
Knock-in Mice
Conclusion 65
Discussion 67
References 74
Appendix: Materials and Methods 81

v
List of Tables
Table 1 Quantification of Sensory Axons with or without Bifurcation 50
in PrkG1 Knockout Mouse Embryos
Table 2 Analysis of Sensory Axon Phenotypes from Individual Nppc 54
Mutant Embryos
Table 3 Quantification of Sensory Axons with or without Bifurcation 54
in Nppc Mutant Mouse Embryos
Table 4 Quantification of Sensory Axons with or without Bifurcation in 62
GSK3α/β 21A/21A/9A/9A Double Knock-in Mouse Embryos

vi
List of Figures
Figure 1 cGMP Promotes Axon Branching in Cultured DRG Neurons 7
Figure 2 cGMP Promotes Branch Formation Instead of Accelerating Axon 11
Growth
Figure 3 cGMP Promotes Axon Branching in A Time Dependent Manner 12
Figure 4 KT-5823 Blocks cGMP-induced Axon Branching in Culture 14
Figure 5 PrkG1 Is Necessary for DRG Axon Branching in Culture 15
Figure 6 PrkG1 Is Sufficient for DRG Axon Branching in Culture 18
Figure 7 Conserved Sequences around the GSK3 Phosphorylation Site 19
Figure 8 cGMP Activation Leads to Phosphorylation of GSK3 in DRG 21
Neurons and COS Cells
Figure 9 PrkG1 Is Required for Phosphorylation of GSK3 by cGMP 22
Activation
Figure 10 Biochemical Interaction of GSK3β with PrkG1α 26
Figure 11 GSK3β Can Be Phosphorylated by PrkG1α in vitro 27
Figure 12 A Dominant Active GSK3β Mutant Blocks Branch Formation 29
Stimulated by 8-Br-cGMP in DRG Neurons
Figure 13 Summary of the Common Targets for the cGMP and NGF 31
Pathways and the Inhibition Studies
Figure 14 Expression of CNP and Npr2 in Mouse Embryos 34
Figure 15 NPs Promote Axon Branching of Dissociated DRG Neurons 36
in Culture
Figure 16 NPs Promote Axon Branching via PrkG1α 37

vii
Figure 17 NPs Induce Axon Outgrowth from DRG Explants 40
Figure 18 NPs Induce Axon Outgrowth via PrkG1α 41
Figure 19 Attractive Turning Responses of DRG Growth Cones in CNP 44
Gradient
Figure 20 Misprojection of Sensory Afferents in PrkG1 Null Mice 48
Figure 21 Requirement of PrkG1 for DRG Axon Bifurcation during Early 49
Embryonic Development.
Figure 22 CNP Is Required for Sensory Axon Bifurcation 53
Figure 23 Gradient Expression of Nppc along the Dorsal-ventral Axis in 56
the Spinal Cord.
Figure 24 CNP Is Not Required for the Formation of DREZ or Collateral 57
Branches
Figure 25 Peripheral Sensory Projections in lbab/lbab Mutant Embryos 59
Figure 26 No Bifurcation Defect in GSK3 Double Knock-in Mice 61
Figure 27 No Cortical Development Defect in GSK3 Double Knock-in 64
Mice
Figure 28 A Working Model of cGMP signaling in Axon Branching 66

viii
Abstract
The involvement of cyclic guanosine-3’, 5’-monophosphate (cGMP) signaling in axonal
development has been proposed. However, the underlying molecular and cellular
mechanism is not well understood yet. In this study, we used rodent embryonic sensory
neurons from the dorsal root ganglion (DRG) as a model system, provided
pharmacological, genetic and biochemical evidences, demonstrated the role of cGMP
signaling in axon branching, further identified GSK3 as the downstream kinase, and CNP
as upstream activator, thus established a functional signaling pathway mediating this
regulation.
First, we reported cGMP can regulate sensory neuron branching. Pharmacologically, a
specific cGMP analog, 8-Br-cGMP, promotes DRG axon branching in cultured DRG
neurons, and this can be mimicked by activating the endogenous soluble guanylyl cyclase
that produces cGMP. This unique activity is mediated by cGMP-dependent protein kinase
1 (PrkG1), because pharmacological inhibitor KT-5823 or genetic ablation of PrkG1 can
abolish the cGMP induced axon branching, whereas overexpression of a dominant active
mutant of PrkG1 is sufficient to promote it. Furthermore, PrkG1 is required for axon
branching in vivo as well, because mouse lacking PrkG1 shows loss of sensory axon
bifurcation in spinal cord. In addition, we provided biochemical evidences that PrkG1 can
directly interact with, further phosphorylate and inactivate glycogen synthase kinase 3
(GSK3), which is a protein that normally suppresses branching. More importantly,

ix
overexpression of a dominant active form of GSK3, which can no longer be
phosphorylated, can block axon branching induced by cGMP in DRG neurons. Finally,
we traced upstream and indentified C-type natriuretic peptide (CNP) as the activator. It is
expressed in a restricted area of the dorsal spinal cord providing a cue that is necessary
for bifurcation of central sensory afferents, and able to mimic cGMP in the cultured
embryonic DRG neurons in stimulating branch formation. And interestingly, the family of
natriuretic peptides (NPs), although initially identified as cardiac- and vascular-derived
hormones that regulate blood pressure, has influence on axonal development in aspects
like inducing axon outgrowth and attracting growth cones as well. Thus,
NP-Npr-cGMP-PrkG-GSK3 signaling pathway could provide a general mechanism to
regulate axonal development in the nervous system.
Keywords: axon branching, cGMP, PrkG1, GSK3, natriuretic peptide, bifurcation,
dorsal root ganglion, sensory neuron

1
Introduction
The most distinguished feature of neurons is the highly specialized morphology, which
provides the structural basis for its functional complexity. Axon branching is one of the
essential processes in establishing neuronal circuits. However, mechanisms regarding
axon branching are not well understood yet. To date, several categories of factors have
been documented: extracellularly, growth factors like NGF, axon guidance molecules like
Slit, adhesion molecules like Dscam; intracellularly, signaling molecules like calcium,
cytoskeletal proteins and protein kinases like GSK3. However, whether there are more of
such molecules and how those factors actually work together at specific developmental
windows and anatomical sites to create precise and functional branching patterns are still
unknown to us.
cGMP is a second messenger involved in a variety of biological regulation. It is also
known to play a significant role in the nervous systems (Feil et al., 2005). In the visual
and olfactory systems, it controls cGMP-gated ion channels and thus modulates
neurotransmitter release. It is implicated to be involved in neuronal survival and
regeneration as well (Thippeswamy et al., 1997). There are pioneering works about its
critical role in neuronal plasticity: first, experiments of cultured Xenopus spinal neurons
using pharmacological reagents have shown that cGMP level affects growth cone turning
behaviors (Song and Poo, 2001); second, several groups have demonstrated the
involvement of cGMP in LTP through hippocampal slice culture. These studies have

2
pointed to a possible role of cGMP in modulating neuronal morphogenesis. However
direct in vivo and in vitro evidences were far from complete.
cGMP function is mediated by a number of immediate targets (Hofmann et al., 2000).
The key enzymes in cGMP metabolism and signaling are expressed in brain regions, such
as cerebellum and hippocampus, which are highly plastic during development, (Lein et al.,
2007). Pharmacological studies in intact animals or slice cultures have shown that
manipulation of a number of enzymes that regulate cGMP levels perturbs the growth of
both axons and dendrites (Cogen and Cohen-Cory, 2000; Gibbs and Truman, 1998;
Polleux et al., 2000; Seidel and Bicker, 2000; Xiong et al., 2007). On the other hand
cGMP-dependent protein kinase (PrkG), as one of the main downstream target, has also
been suggested to be involved in regulating axonal development: mouse lacking
cGMP-dependent protein kinase 1 (PrkG1) has a defect in the central axonal projection of
the dorsal root ganglion (DRG) sensory neurons (Schmidt et al., 2002); and PrkG1 is
synaptically localized during potentiation and PrKG2 can interact with the Carboxyl
terminal of GluR1 and modulate AMPA receptor trafficking (Serulle et al. 2007). Thus,
PrkG is a potential candidate for mediating cGMP signaling in regulating axonal
development, but this hypothesis still needs to be confirmed. Moreover, the establishment
of downstream signaling mediators will further help us answer the fundamental question
on how cGMP signaling can actually lead to change in cellular processes, such as
cytoskeleton remodeling and branch formation, during axonal development.

3
cGMP is converted from GTP by guanylyl cycalses, which contain 2 families:
membrane-bound (type 1) and soluble (type 2) form of guanylyl cyclases. Soluble
guanylyl cyclases (sGC) can be activated by nitric oxide, whereas membrane-bound
guanylyl cyclases are single transmembrane receptor proteins that can binds to
extracellular factors, including natriuretic peptides. The family of natriuretic peptide
hormones contains 3 members (22-32 amino acid residues): atrial NP (ANP), brain NP
(BNP) and c-type NP (CNP). They are structurally related, and known to regulate
homeostasis of body water and salt (Feil et al., 2005), by binding to two
membrane-bound guanylyl cyclases, Npr1 and Npr2 (also known as NPR-A and NPR-B),
that catalyze the production of cGMP. Since the first discovery of NPs in the heart nearly
thirty years ago, it has been found that both NPs and Npr receptors are also expressed in
the brain, with CNP and Npr2 being most abundant (Feil et al., 2005). Cell culture studies
have implicated NPs in regulating transmitter release, synaptic transmission and neuronal
survival (Feil et al., 2003, Thippeswamy et al. 1997, Song et al. 2001), and suggested
their potential involvement in neural development, but their precise physiological roles in
the nervous system remain to be established.
In this study, we are able to address these questions by investigating cGMP signaling
pathway in a three-dimensional culture system of embryonic DRG neurons and several
mouse models. First, we found that activation of the cGMP pathway indeed can promote
axonal branch formation in culture, which provided strong evidence to establish its role in

4
axon morphogenesis, also revealed that the kinase activity of PrkG1 is both necessary and
sufficient for this cGMP induced axon branching activity. In addition, we provided
biochemical and functional evidence to link this novel activity to the regulation of GSK3,
a kinase involved in growth factor-dependent axonal growth and branching (Kim et al.,
2006). Furthermore, we traced natriuretic peptide hormones as one of the endogenous
upstream activator of cGMP signaling in promoting axon branching, and showed that
they are able to regulate axon outgrowth and growth cone pathfinding as well. Finally, we
were among the first to illustrate the physiological significance of the
CNP-Npr2-cGMP-PrkG1 signaling pathway in vivo by demonstrating the requirement of
this pathway in controlling bifurcation of DRG sensory axons in spinal cord. Thus, our
study establishes a signaling cascade based on cGMP signaling that regulates axonal
development.

5
Chapter 1:
cGMP Regulates Axon Branching via PrkG1
cGMP is a second messenger involved in a variety of biological regulation (Feil et al.,
2003). In the nervous system, it has been shown to modulate neurotransmitter release,
neuronal survival, and synaptic plasticity ( Feil et al 2003, Barnstable et al 2004, Fiscus et
al 2002, Kleppisch et al 2003). However as for axonal development, only a few
observations have been documented. One of the most interesting finding is that cGMP
signaling is involved in modulating the growth cone response to extracellular guidance
cues. High level of cGMP promotes growth cone attraction while low level promotes
repulsion in response to Sema3A (Song et al. 2001). Also, a few pharmacological studies
in intact animals or slice cultures demonstrated that manipulation of enzymes which
perturbs cGMP levels actually affects axonal and dendritic growth (Cogen et al. 2000,
Gibbs et al. 1998, Polleux et al. 2000, Seidel et al. 2000, Xiong et al. 2007). These data
suggest a potential role of cGMP signaling pathway in axon morphogenesis. On the other
hand, the key enzymes in cGMP metabolism and signaling, like sGC, PrkG, are expressed
in brain regions known for their plasticity, such as cerebellum and hippocampus (Lein et
al. 2007). Mice lacking PrkG1, have abnormal central axon projection of the dorsal root
ganglion (DRG) sensory neurons (Schmidt et al. 2002, Schmidt et al. 2007). Thus, the
cGMP signaling pathway may provide a unique and interesting target for studying the
regulation of axon morphogenesis.

6
1.1 cGMP Increases Formation of Axonal Branches in Cultured Sensory Neurons
In order to characterize the effect of cGMP on axonal development, we utilized a well
established in vitro culture system of dissociated E14 rat DRG neurons. As previously
described (Wang et al., 1999), DRG neurons at this age normally exhibit simple
morphology when cultured in collagen gels, with a single axon extending out from the
cell body (Fig. 1 A-C) after 2 days in culture. As shown by distribution, more than 40%
of neurons had no branch, about half made one, and only less than 10% generated more
than two branches (Fig. 1 M), and the number of average branching points per neuron is
<1 (Fig.1 M, P). Interestingly however, when treated with 8-bromoguanosine-3’,5’-cyclic
monophosphate (8-Br-cGMP), a commonly used membrane-permeable analog of cGMP
(Schwarzschild and Zigmond, 1991) for just one day, these neurons made significantly
more branches from the single axon, with some having more than four per axon (Fig. 1
D-F). This effect can be seen at an 8-Br-cGMP concentration as low as 2 μM, and with
the increasing analog concentration, the distribution shifted toward the right, indicating
an increase in the percentage of neurons with higher number of branches (Fig. 1 M). At
the highest concentration tested (200 μM), >50% neurons had more than two branches,
while only <10% had no branches (Fig. 1 M), and the number of average branching
points per cell reached 2.8 fold of that in the control neurons (Fig. 1 P).

7
Figure 1 cGMP Promotes Axon Branching in Cultured DRG Neurons

A-L) DRG neurons from E14 rat embryos were dissociated and cultured in collagen gels
in the presence of NGF for 24 hours and then treated with buffer (A-C), 8-Br-cGMP (50
μM, D-F), 8-Br-cAMP (50 μM, G-I), or YC-1 (20 μM, J-L). After cultured for another
day, they were fixed and stained with antibody against neurofilament. Regions of the
cultures are shown at low magnification in A, D, G, and J, while individual cell
morphologies are shown at high magnification in B, C, E, F, H, I, K, and J. Neurons
treated with 8-Br-cGMP or YC-1 but not 8-Br-cAMP showed a significant increase in
branches formed from a single axon extending from the cell body. Bars: 100 μm. (M-O)
Comparison of axon branching by the distribution of neurons with different number of
branches in the above cultures treated with different concentrations of each chemical. (P)
Comparison of the average number of branching points measured from neurons cultured
under different chemical concentrations.

8
This cGMP effect on DRG axon branching is specific, because 8-bromoadenosine
-3’,5’-cyclic monophosphate (8-Br-cAMP), an equivalent analog to activate the cAMP
pathway, did not elicit any similar effect (Fig. 1 G-I): neither the distribution nor the
average number of branching points changed significantly in the 8-Br-cAMP-treated
culture as compared to the untreated one (Fig. 1 M-P). Furthermore, YC-1, a small
molecule that activates soluble guanylyl cyclase (sGC) to produce cGMP (Galle et al.,
1999), also induced branch formation in a similar dose dependent manner (Fig. 1 J-L, O,
P). Conversely, Zaprinast, an inhibitor that prevents cGMP degradation by
phosphodiesterase-5 (Dundore et al., 1993), also induced branching (data not shown).
We further examined several parameters to determine if the in vitro effect reflects the role
of cGMP signaling in axon growth and branching. We defined the length of the primary
axon as L
p
; length of each branch as L
i
(i = 1~n, n is the total number of branches, L
1
is
the branch closest to soma.); the distance from soma to the first branch as L
o
; B
t
is the
total number of branches; B
1
is the number of primary branches which grow out from the
primary axon; B
2
is the number of secondary branches which come out from primary
branches; and so on. So L
p
reflects the axonal growth, L
b
=
∑
Li reflects the growth on
branches, L
0
/ L
p
indicates the relative position of first branch; (L
p
+ L
b
) is the total
neuritic length.
First, neurons from the control cultures had a total length (L
t
) of 254±14 μm, but neurons

9
treated with 8-Br-cGMP had a nearly two-fold increase, reaching 467±20 μm per cell
(Fig. 3 F). The 80% increase in L
t
is not contributed solely by the elongation of axons, as
L
p
increased only modestly by 30%. Instead, it is a result of the increase in length of
branches, as the total branch length (L
b
) per cell had a more than two-fold increase (Fig. 3
F). This further strengthens our data on comparing L
0
/ L
p
(Fig.3 D) which indicates first
branch was formed earlier from the axon after cGMP treatment. Furthermore, we found
no linear correlation between the primary axon length (L
p
) and B
t
for all the neurons
analyzed (Fig. 3 H-I), indicating cGMP-stimulated branch formation is independent of
the growth of primary axons. However our analysis revealed a strong linear correlation
between (L
t
– L
0
) and B
t
, and the slopes in cGMP treated condition (88.7, Fig. 3 K) was
very similar as that in the control condition (86.4, Fig. 3 J), which indicates the average
neuritic length between two adjacent branches ((L
t
– L
0
) / B
t
) is somehow independent of
cGMP, so the increase in branch growth is correlated well with the increase in branching
points (Fig. 3 G). Thus, the effect of cGMP on axon branching is mediated mainly by
promoting branch formation and elongation, instead of accelerated axon growth.
In order to determine whether cGMP treatment induces transcriptional modification, we
performed a pulse-chase experiment. In this experiment, dissociated DRG neurons were
treated with 0, 2, 5, 10 or 24 hours of 8-Br-cGMP followed by excessive washes on the
second day of culture (Fig. 4 A-F). The quantification of average branching points from
those cultures demonstrated that the longer the treatment, the more branches were formed

10
in those neurons (Fig. 4 G). This result is not in favor of the hypothesis that the addition
of axonal branches is a result of modification in gene expressing by cGMP signaling.
We also examined the possible role of cGMP signaling in regulating neuronal survival
and axongenesis as previously suggested (Fiscus, 2002; Seidel and Bicker, 2000;
Thippeswamy and Morris, 1997). As shown by TUNEL staining, the cell death did not
change significantly in the cultures treated with 8-Br-cGMP as compared to the untreated
control (Fig. 3 H). In addition, these neurons usually made only one axon per cell body in
collagen gels, and this number did not change after cGMP stimulation, indicating that the
pathway does not promote axonogenesis (data not shown). This conclusion is further
supported by our study of BAX null DRG neurons (Lentz et al., 1999), which survived in
the absence of NGF but did not make more axons in the presence of 8-Br-cGMP (Fig. 3
I).
Taken together, these analyses have revealed that the cGMP pathway can provide a
specific intracellular signaling mechanism that promotes branch formation in culture.

11
Figure 2 cGMP Promotes Branch Formation Instead of Accelerating Axon
Growth

(A-B) Example pictures of single neuron morphology in culture treated with 8-Br-cGMP
(B) or not (A). (C) A diagram showing a model neuron, defining length of primary axon
as L
p
; length from soma to first branch as L
0
; primary branch as B
1
; secondary branch as
B
2
; length of primary branch as L
1
; length of secondary branch as L
2
. (D) Comparison of
L
0
/L
p
between neurons from culture treated with or without Br-cGMP. (E) Comparison of
(L
p
- L
0
) between neurons from culture treated with or without Br-cGMP. (F)
Quantification of total neuritic length (L
t
), length of all branches (L
b
) and L
p
in cultures
treated with or without Br-cGMP. (G) Quantification of total number of branches (B
t
),
number of primary branches (B
1
) and secondary branches (B
2
) in cultures treated with or
without Br-cGMP. (H-I) Scattered plot showing no linear relationship between L
p
(Y-axis)
and B
t
(X-axis) in condition treated with (I) or without Br-cGMP (H). (J-K) Scattered plot
showing strong linear relationship between (L
t
- L
0
) (Y-axis) and B
t
(X-axis) in both
conditions.

12
Figure 3 cGMP Promotes Axon Branching in a Time Dependent Manner

A) A diagram showing experimental design for pulse-chasing cGMP treatment.
Dissociated E14 rat DRG neurons were cultured for 1 day in collagen, and then treated
with 8-Br-cGMP for a certain amount of time as indicated (0-24hours) followed by
excessive wash with growth medium. The cells were fixed after 2 days in culture and
stained with antibodies against neurofilament. Regions of the cultures are shown at low
magnification in (B-F). (G) Quantification of average branching points per neuron from
above cultures. Note the longer treatment of cGMP, the more branches formed in those
neurons. (H) Cell death rate was examined by TUNEL staining in dissociated E14 rat
DRG neuron cultures, and no difference in cell death (p=0.88, t-test) was found with
Br-cGMP treatment. (I) DRG neurons isolated from Bax -/- embryos were cultured in the
absence of NGF. cGMP had no significant effect on axonogenesis (p>0.05) as indicated
by the similar number of neurons with neurites from cultures with or without Br-cGMP.

13
1.2 Activation of PrkG1 Is both Necessary and Sufficient for DRG Axon Branching
in Culture
cGMP dependent protein kinase is one of the most important mediators of cGMP
signaling. Only PrkG1α is expressed in DRG neurons during early embryonic
development as reported by in situ hybridization (Fig. 21 F-O) or immunohistochemical
studies (Qian et al., 1996; Schmidt et al., 2002). PrkG1α transcript appears in DRG as
early as E10.5, peaks at E12.5 and becomes more restricted between E15.5 and E18.5
(Fig. 21 F-I), while PrkG1β or PrkG2 is not expressed (Fig. 21 J-O). In addition genetic
ablation of this gene leads to abnormal sensory projection to spinal cord from DRG (Fig.
20 A-E, and Schmidt et al., 2002). Thus we focused our study on PrkG1. Interestingly as
we expected, KT-5823, a specific inhibitor that interferes with the ATP binding site of the
catalytic domain of PrkG (Hidaka and Kobayashi, 1992), blocked the 8-Br-cGMP
induced axon branching in culture (Fig. 4 A-F): at 2μM concentration, it reduced the
cGMP’s effect almost by half (Fig. 4 C-D), and at 10μM, it completely abolished that
(Fig. 4 E-F), as shown by both the distribution of percentage of neurons with different
number of branches (Fig. 4 G-H) and quantification of average number of branching
points per neuron (Fig. 4 I). To further confirm the requirement of this kinase, we
isolated DRG neurons from E12.5 PrkG1 null embryos (Wegener et al., 2002) and
compared their response to 8-Br-cGMP in culture with neurons from wild-type
littermates. E12.5 mouse DRG neurons are equivalent to that from E14 rat, with simple
morphology in collagen culture, but able to make more branches in response to Br-cGMP

14
Figure 4 KT-5823 Blocks cGMP-induced Axon Branching in Culture.

(A-F) PrkG1 inhibitor KT-5823 abolished cGMP induced axon branching in cultured
DRG neurons. (A-F) E14 rat DRG neurons were dissociated and cultured for two days in
collagen gels and treated with buffer (A, B), 2μM (C, D) or 10μM KT-5823 (E, F), in the
absence (A, C, E) or presence of 50 μM 8-Br-cGMP (B, D, F) on the second day. They
are visualized by antibody staining for neurofilament. Bars, 100 μm. (G-H) Comparison
of axon branching by the distribution of neurons with different number of branches in the
cultures treated with different combination of chemicals. (I)The difference in the above
culture is shown by the change in the average number of branching points. Note that
KT-5823 had al most no effect on basal branching, but was able to reduce cGMP induced
branching at 2μM, and even abolished it at 10μM.

15
Figure 5 PrkG1 Is Necessary for DRG Axon Branching in Culture.

(A-D) DRG neurons from PrkG1 mutant mice do not support axon branching after cGMP
activation. E12.5 mouse DRG neurons from wild type (+/+, A-B) or PrkG1 mutant (-/-,
C-D) littermates were cultured for two days in collagen gels and treated with (B, D) or
without (A, C) 50 μM 8-Br-cGMP on the second day. They are visualized by antibody
staining for neurofilament. Note that wild type neurons (B) generated more branches in
the presence of Br-cGMP, but not the PrkG1
-/-
neurons (D). Bars, 100 μm. (E-F) The
difference in the above culture is shown by the change in the distribution of cells with
different branch numbers (E) and the average number of branching points (F).

16
(Fig. 5 A-B). DRG neurons from E12.5 PrkG1 null embryos grew normally (Fig. 5 C)
and had no defect in survival (data not shown). However, they apparently failed to
respond to 8-Br-cGMP (Fig. 5 D) as shown by both the distribution of percentage of
neurons with different number of branches (Fig. 5 E) and quantification of average
number of branching points per neuron (Fig. 5 F). Thus, PrkG1 is indeed required for
cGMP-stimulated axon branching in culture.
To further characterize the function of PrkG1, We overexpressed the full length wild type
mouse PrkG1α in E14 rat DRG neurons by electroporation. However, only a modest
increase in basal branching (by 20%) as well as in cGMP-stimulated branching (by 20%)
was observed, compared to neurons expressing EGFP from the same culture (Fig. 6 A-D,
I-K). This is consistent with the auto inhibitory feature of this kinase. So, we next asked if
perturbation of PrkG1α activity at the molecular level could actually modulate axon
branching by generating and testing several PrkG1α mutants. Strikingly, neurons
expressing an S64A mutant, which has twice the basal kinase activity of the wild type and
10-fold higher affinity for cGMP (Busch et al., 2002), displayed excessive branching
even in the absence of 8-Br-cGMP (Fig. 6 E-F, I-K). The number of branching points
increased by nearly 1.5-fold, indicating this mutant can promote branching with the
endogenous cGMP level (Fig. 6 E, I-K). 8-Br-cGMP did not further increase the
branching number (Fig. 6 F), which is not surprisingly considering the mutant was
already fully activated and the downstream branching program might also limit. On the

17
other hand, the amino-terminal fragment (-N) containing only the regulatory domain of
PrkG1α, which was reported to be dominant inhibitory (Browning et al., 2001), blocked
branching in neurons treated with 8-Br-cGMP (Fig. 6 G-H, I-K), but had little effect on
basal axon growth and branching at all. Thus, these results demonstrate that PrkG1 is the
key mediator for the cGMP signaling in axon branching.

18
Figure 6 PrkG1 Is Sufficient for DRG Axon Branching in Culture.

(A-H) Overexpression of PrkG1α mutants modulates axon branching in DRG neurons.
Dissociated rat E14 DRG neurons transfected with either EGFP (A-B), the full length
(-FL, C-D), the S64A mutant (E-F) or the amino-terminal fragment (-N, G-H) of PrkG1α
were cultured for two days in collagen gels and treated with (B, D, F, H) or without (A, C,
E, G) 50 μM 8-Br-cGMP. They are visualized based on the EGFP fluorescence (G-H) or
antibody staining against the FLAG tag fused to the PrkG1α proteins (C-H). The S64A
mutant induced the formation of multiple branches (arrows) in the absence of 8-Br-cGMP
(E), while the N-terminal fragment blocked 8-Br-cGMP-induced branching (H). Arrows
point to the end of axonal branches. Bars: 100 μm. (O-Q) The difference in the above
culture is shown by the change in the distribution of cells with different branch numbers
(I-J) and the average number of branching points (K).

19
Chapter 2:
GSk3 Is a Novel Substrate for PrkG1
To understand how PrkG1 activation by cGMP controls branching, our next endeavor
was to identify the downstream pathways that mediate this novel activity in axonal
development. Since most of the well known substrates of PrkG1 haven’t been reported
with any effect on branch formation in axons, we decided to look for proteins that are
known with a role in branching and can be phosphorylated on Ser/Thr sites. On the other
hand, our DRG neurons were cultured in the presence of NGF, so the existing TrkA
signaling pathway can be a potential target of PrkG1 regulation. Therefore, GSK3
became one of our candidates, because: it is has been shown to negatively regulate axon
growth and branching but can be inhibited by NGF signaling in DRG neurons (Kim et al.,
2006); the sequence surrounding the phosphorylation site on the amino terminal is highly
conserved among different species, and the adjacent residues share high similarity with
the consensus sequence (-Arg-Arg-X-Ser/Thr-) in well-known PrkG substrates (Fig. 7 A).
Figure 7 Conserved Sequences around GSK3 Phosphorylation Site

A) Comparison of the sequences surrounding the amino-terminal phosphorylation site in
GSK3α and GSK3β from different species. (∗) positive charge; (♦)phosphorylation site.

20
2.1 cGMP Activation Leads to Phosphorylation of GSK3 in both DRG Neurons and
COS Cells
GSK3 has two isoforms, GSK3α and GSK3β. We first examined their phosphorylation in
cultured DRG neurons by western blot using phosphor-specific antibodies that can
recognize phosphor-serine at both position 21 in GSK3α and position 9 in GSK3β (Fig. 8
A). Within five minutes of 8-Br-cGMP treatment, the phosphorylation level of both
isoforms started to increase, and at 30 min, reached at about 1.5 fold of the basal level
(Fig. 8 B). As a comparison, their phosphorylation in response to NGF reached nearly
2-fold of the basal level within 5 min, similar to what was reported recently (Zhou et al.,
2004). In addition, this cGMP-stimulated phosphorylation is dependent on PrkG1,
because no such change in phosphorylation level was detected in neurons isolated from
PrkG null mouse embryos, whereas NGF-dependent phosphorylation was still retained
(Fig. 9 A-B). As comparison, we also examined other known downstream molecules
involved in NGF signaling, including Akt, the kinase that is activated by
phosphatidylinositol 3-kinase (PI3-K) to phosphorylate GSK3, and ERK, the kinase that
mediates a separate Raf signaling pathway (Fig. 13 A and Markus et al., 2002; Zhong et
al., 2007; Zhou et al., 2004). Both kinases responded to NGF stimulation well in DRG
neurons; however neither of them showed any noticeable change in phosphorylation level
after 8-Br-cGMP treatment (Fig. 9 C). Thus, cGMP activation indeed phosphorylates
GSK3 in DRG neurons via PrkG1α.

21
Figure 8 cGMP Activation Leads to Phosphorylation of GSK3 in DRG Neurons
and COS Cells

A) Western blot of phospho-GSK3α (Ser21) and phospho-GSK3β (ser9) in extracts from
dissociated E14 rat DRG neurons cultured for 24 hrs. These cells were starved for 3 hours
by removing NGF and then treated with 8-Br-cGMP (50 μM) or NGF (50 ng/ml) for the
time periods indicated. Tubulin was blotted as a loading control. (B) The relative level of
phosphorylation on the above western blot was calculated based on the intensity of each
band, normalized to the expression of tubulin on the same membrane, and compared with
the control sample. Note the increase in phosphorylation in extracts treated with
8-Br-cGMP or NGF. (C) HA-tagged GSK3β was expressed in COS cells along with
EGFP, the full length (FT) or the S64A mutant of PrkG1α. Its phosphorylation was
detected by a GSK3 phospho-antibody, while its expression was probed by HA antibody.
PrkG1α expression was revealed by antibodies against the FLAG tag fused to the protein.
Tubulin staining was used as a loading control. (D) Quantification of the GSK3β
phosphorylation level from the above western blots and the band intensity is normalized
sequentially to that from blots for HA for GSK3β and FLAG for PrkG1α. The number in
the parentheses indicates the fold change as compared to the control sample expressing
GSK3β and EGFP.

22
Figure 9 PrkG1 Is Required for Phosphorylation of GSK3 by cGMP Activation

A) GSK3 phosphorylation remained the same in neurons isolated from PrkG1 mutant
after cGMP treatment. E12.5 mouse DRG neurons from wild type (+/+) or PrkG1 mutant
(-/-) littermates were cultured for or 24 hours, starved for 3 hours by removing NGF and
then treated with 8-Br-cGMP (50 μM) or NGF (50 ng/ml) for 30 minutes. Cells were
further lysed and western-blotted for antibody detection. (B) Quantification of relative
band intensity from the blot in (A). Note the reduction of GSK3 phosphorylation in
PrkG1 mutant cells (-/-) treated with cGMP. (C) Same western blot in Fig. 8 A was
further probed with phosphor-specific antibodies against Akt and ERK. Note the
phosphorylation level of either proteins increased in NGF treatment, but stayed same in
cGMP condition.

23
To further verify this connection, we co-expressed PrkG1 and GSK3β in COS cells, and
analyzed GSK3β phosphorylation by western blots (Fig. 8 C). In the presence of the full
length PrkG1α or the S64A mutant, GSK3β phosphorylation increased to 3.4 fold or 4.9
fold respectively over the EGFP control (Fig. 8 D), while the phosphorylation level of
endogenous Akt did not change significantly at all (data not shown). These studies
revealed that GSK3 is a potential downstream target for PrkG1 and suggested a
mechanism in which cGMP signaling leads to GSK3 phosphorylation and thereby
relieves its inhibition of branching in DRG neurons.

24
2.2 PrkG1 Binds and Phosphorylates GSK3 Directly
To test if GSK3 is a direct substrate of PrkG1α indeed, we performed
immunoprecipitation with COS cell lysate. As shown in Fig. 10, when the full length wild
type or the S64A mutant of PrkG1α was co-expressed with HA-tagged GSK3β in COS
cells, antibody pull-down with an HA-specific antibody for GSK3β resulted in
precipitation of PrkG1α proteins. In a reciprocal experiment using a FLAG antibody,
GSK3β could be co-precipitated with either FLAG tagged PrkG1 proteins as well. As
control, when they were co-expressed with EGFP, none of those proteins was precipitated
down by antibodies against GFP, confirming the specificity of the association. This
experiment indicates that GSK3 could physically associate with PrkG1α indeed.
Next, we asked if GSK3β can be phosphorylated in vitro by COS cell extracts containing
overexpressed PrkG1α proteins. In this experiment, native recombinant HA-tagged
GSK3β proteins were obtained with nickel column purification from bacteria, and then
incubated with COS cell lysates containing the S64A mutant of PrkG1α at 30
o
C for
various time. The phosphorylation level of GSK3β, revealed by phosphor-specific
antibodies on western blots, increased by 80% after 5 min incubation, peaked at 15 min
(2.6 fold of the basal level), and plateaued between 30-60 min (Fig. 11 A-B).
Finally, we tested if GSK3β could be phosphorylated directly by purified PrkG1α
proteins. When the recombinant GSK3β protein was incubated with the kinase purified

25
from the bovine lung tissues at 30
o
C for 30 min in a phosphorylation buffer contains no
other proteins (Fig. 11 C), the phosphorylation signal was increased dramatically by ~3
fold over the control condition without the kinase (Fig. 11 D). Moreover, stimulation of
the kinase by including 8-Br-cGMP in the reaction induced an additional 1.8-fold
increase in phosphorylation (Fig. 11 C-D), indicating that PrkG1α phosphorylates
GSK3β in a cGMP-dependence manner.
Take together, these studies in heterologous cells and with purified proteins revealed that
GSK3β is a direct target for PrkG1α phosphorylation in cGMP induced axon branching.

26
Figure 10 Biochemical Interaction of GSK3β β β β with PrkG1α α α α

A) Co-immunoprecipitation of HA-tagged GSK3β and Flag-tagged PrkG1α or
PrkG1α-S64A. COS-7 cells were transfected with a combination of constructs for GSK3β,
PrkG1α, PrkG1α-S64A and EGFP as indicated in the label above the blot. Cell lysates
were immunoprecipitated (IP) with anti-HA or anti-Flag antibodies, and probed with
antibodies against Flag or HA on western blots (WB) as indicated on the left side of the
top four blots. Co-immunoprecipitation of GSK3β and PrkG1α or PrkG1α-S64A is shown
on the fourth and fifth lane. The cell lysates used for immunoprecipitation were also
blotted and probed as input controls (bottom two blots).

27
Figure 11 GSK3β β β β Can Be Phosphorylated by PrkG1α α α α in vitro

(A-B) In vitro phosphorylation of purified GSK3β by COS cell extracts containing
PrkG1α-S64A. Cell extracts were prepared from COS-7 cells transfected a PrkG1α-S64A
expression construct and mixed with recombinant GSK3β proteins (200 ng) purified from
bacteria in the kinase assay buffer at 30
o
C for various time points. Phosphorylation of
GSK3β at serine-9 was detected by western blot using the anti-phospho-GSK3 antibody
and the amount of GSK3β and PrkG1α-S64A was probed with antibodies against HA and
Flag respectively (A). The level of phosphorylation was calculated from the band
intensity that was normalized to that of 0 min (B). (C-D) GSK3β can be directly
phosphorylated by PrkG1α in vitro. Purified bovine PrkG1α proteins (400 ng) and
GSK3β proteins (200 ng) were incubated in the kinase assay buffer with or without 20
μM 8-Br-cGMP at 30
o
C for 30 minutes. GSK3β phosphorylation at serine-9 was
analyzed using the specific antibody on the western blot (C). The relative level of
phosphorylation was calculated from the band intensity (D) normalized to the amount of
GSK3β in the reaction that was determined by the HA antibody (C). Note the huge
signal of GSK3β phosphorylation when the kinase was added to the reaction and another
nearly two-fold increase when 8-Br-cGMP was also included.

28
2.3 A Dominant Active GSK3β β β β Phospho-mutant Inhibits cGMP-induced Branching
in DRG Neurons
With the information from our biochemical studies, we again utilized our in vitro culture
system to further test the role of GSK3 phosphorylation in cGMP-induced axon
branching. We reasoned that a GSK3β mutant (S9A), which can no longer be inactivated
by phosphorylation (Eldar-Finkelman et al., 1996), could serve as a dominant active form
to prevent cGMP signaling from promoting branching. Indeed as expected, DRG neurons
expressing this mutant generated much simpler morphology even in the presence of
8-Br-cGMP (Fig. 12 E-F). At the basal level, the cell distribution as well as the number of
branching points in these neurons (Fig. 12 E, I) was very similar to that of control cells
expressing EGFP (Fig. 12 A, I) from the same culture. However, when these cells were
treated with 8-Br-cGMP (Fig. 12 F), the branched morphology normally exhibited in the
control cells (Fig. 12 B) was completely abolished (Fig. 12 J), and the average number of
branching points dropped back to that of unstimulated neurons (Fig. 12 K, p=0.44,
ANOVA test). In addition, not surprisingly at all, overexpression of the wild type full
length GSK3β also reduced the response to cGMP activation, although to a lesser extent
(Fig. 12 C, D, and I-K). Finally, a control phosphorylated mutant (S9D), had no effect
on the response of DRG neurons to 8-Br-cGMP (Fig. 12 G-H, and I-K), which indicates
the GSK3 regulation on cGMP induced axon branching is phosphorylation dependent.
Thus GSK3 is the mediator that translates cGMP signaling to axon branching.

29
Figure 12 A Dominant Active GSK3β β β β Mutant Blocks Branch Formation
Stimulated by 8-Br-cGMP in DRG Neurons

A-H) Dissociated E14 DRG neurons were electroporated with plasmids expressing EGFP
(A-B), GSK3β full length (FL, C-D), the S9A mutant (E-F) or the S9D mutant (G-H),
cultured in collagen gels, and treated without (A,C,E,G) or with (B,D,F,H) 50 μM
8-Br-cGMP for one day. Cells are visualized with the EGFP fluorescence (A-B) or an
antibody against the HA-tag attached to GSK3β (C-H). Expression of the S9A mutant
blocked cGMP-induced branches (F, arrows), while expression of the wild type or the
S9D mutant had no effect (H). Note, many non-neuronal cells were also stained and
shown in the background and arrows point to the end of axonal branches. Bar, 100 μm.
(I-K) Comparison of the distribution of neurons with different branch numbers (I-J) and
the number of branching points (K) between cells expressing proteins described above.

30
To understand whether GSK3 is the only downstream target shared by both cGMP and
NGF signaling, we also examined other intermediate signaling molecules of the NGF
pathways (Kim et al., 2006; Markus et al., 2002, and Fig 13 A). First, inhibition of PI3-K
by wortmannin (Fig. 13 C) and LY294002 (Fig. 13 D) reduced basal level branching to
different extents but did not block new branch formation stimulated by 8-Br-cGMP;
second, blocking ERK activity by U126 (Fig. 13 E) had no effect on cGMP-stimulated
axon branching at all. As a comparison, the inhibitory effect of these antagonists on
8-Br-cGMP-stimulated branching was <10%, while the inhibition by KT5823 and the
S9A mutant of GSK3β reached 74% and 91% respectively (Fig. 13 B). Therefore, these
results suggest that GSK3 specifically mediates the downstream branching signal of
cGMP via the control of its phosphorylation by PrkG1; NGF-TrkA and cGMP-PrkG1
signaling pathways actually converge on GSK3 in regulating axon branching (Fig. 13 A).

31
Figure 13 Summary of the Common Targets for the cGMP and NGF Pathways
and the Inhibition Studies

A) A working model summarizing the molecular pathways involved in both cGMP and
NGF signaling. The dashed lines include the factors and connections identified in this
study that promote axon branching in DRG neurons. (B) A summary of all inhibition
studies of cGMP-induced axon branching in DRG neurons. They are divided into three
groups: studies targeting PI3-K and ERK with chemical inhibitor wortmannin (wort),
LY294002 (LY) or U0126; studies targeting PrkG1 with KT5823 (KT), by
overexpression of the N-terminal fragment of PrkG1, or using PrkG1 null DRG neurons;
studies targeting GSK3β by overexpressing the full length or the S9A mutant. The
percentage of inhibition is based on the fold change in the number of branching points
comparing between treatment with and without 50 μM 8-Br-cGMP and calculated from
comparing each testing condition and its control [Inhibition% = {1 - (Fold
test
-
1)/(Fold
control
- 1)}*100% ]. (C-E) Quantification of average branching points per neuron
in cultures testing: Wortmannin (C), LY294002 (D) or U126 (E) on cGMP induced axon
branching.

32
Chapter 3:
Natriuretic Peptides Can Regulate Axonal Development
The family of natriuretic peptides contains 3 members, A-type (ANP), B-type (BNP) and
C-type (CNP). ANP and BNP preferably bind to Npr1, CNP binds to Npr2, and all three
can bind to Npr3. ANP and BNP studies were mostly focused on their regulation in
cardiovascular system and maintenance of body water homeostasis. ANP deficient mouse
has been reported to have high blood pressure (John et al. 1995), and BNP deficient
mouse shows abnormal cardiac fibrosis in heart (Tamura et al. 2000). Whereas, CNP
deficient mouse has a problem in bone development which results in a dwarfism
phenotype (Chusho et al. 2001, Tsuji et al. 2005). Even though their expression was
found in various regions of nervous system, little was known about the role of NPs in
neuronal development. However, our studies on cGMP-PrkG signaling provided a hint on
the possible functions of NPs in axonal development, such as promoting branch
formation (Zhao et al. 2009). On the other hand, identical defects in central projection of
DRG sensory axons were found in mouse with a spontaneous mutation in the Npr2 gene
(Tsuji et al. 2005, Schmidt et al. 2007) and PrkG1 null mouse (Schmidt et al. 2007; Zhao
et al. 2009), which suggests that CNP, the ligand that preferentially binds to Npr2, could
be the environmental cue responsible for axon bifurcation, and further implies the
hypothesis that NPs could provide a general extracellular mechanism to regulate axonal
development via cGMP signaling.

33
3. 1The CNP Precursor Is Expressed in the Developing Spinal Cord
To understand the potential in vivo function of NPs, we first used RNA in situ
hybridization to examine the transcripts encoding different NP precursors in the
developing spinal cord. As shown by the previous report (DiCicco-Bloom et al, 2004),
only Nppc, the gene encodes the precursor for CNP, is expressed in the embryonic spinal
cord during early development (Fig.14 E-F), but not the precursor of ANP or BNP (Fig.14
A-D). Nppc transcript appears strongly in the dorsal part of the spinal cord as early as
E10.5 (Fig.14 E), when nascent axons from newborn sensory neurons just start to
bifurcate after reaching the dorsal root entry zone (DREZ); and by E12.5, it becomes
more restricted to the roof plate (Fig.14 F).
We also examined the expression of NP receptors, including Npr1, Npr2 and the
cyclase-deficient receptor Npr3. Consistent with the previous report (Schmidt et al. 2007),
only Npr2, the high affinity receptor for CNP, is expressed in the DRG (Fig.14 K-L). Like
PrkG1 (Fig.14 O-P, and Schmidt et al. 2007), Npr2 expression in the DRG starts at E10.5
and remains at a similar level at E12.5. Thus, CNP is present in the developing spinal
cord at the time when its receptor and downstream signaling target are made in the DRG.

34
Figure 14 Expression of CNP and Npr2 in Mouse Embryos

Transcripts of Nppa, Nppb, Nppc, Npr1, Npr2, Npr3 and PrkG1α were examined by
digoxigenin-labeled RNA in situ hybridization on E10.5 (A, C, E, G, I, K, M, O) and
E12.5 (B, D, F, H, J, L, N, P) embryonic cross sections. Nppa (A-B), Nppb (C-D), Npr1
(I-J) Npr3 (M-N) are not expressed strongly during this developmental window. However,
Nppc (E-F) expression is found to be concentrated in the dorsal region of spinal cord, and
Npr2 (K-L), PrkG1α (O-P) are strongly expressed in DRGs in both stage. A sense probe
for Nppc was used as a control (G-H). Bar: 500 μm.

35
3.2 CNP Promotes Axon Branching of Dissociated DRG Neurons in Culture
Next, we asked if CNP can promote axon branching in the in vitro assay that was
previously used to identify and characterize the cGMP pathway (Zhao et al. 2009). As
previously described, dissociated E14 rat DRG exhibits simple morphology after two
days culture in collagen gels. They grow only one axon and on average having less than
one branch (Fig. 15 A-C). However, addition of 100 nM CNP to the culture on the second
day led to enhanced branch formation: a higher percentage of neurons generated more
than one branch (Fig. 15 J-L and M) and the number of average branching points per
neuron reached 2.3±0.1 (mean±s.e.m.), about a 2.4-fold increase over the control
condition (Fig. 15 J-L and N). This effect is similar to direct activation of PrkG1 by a
cGMP analog, 8-Br-cGMP (Zhao et al. 2009). In addition, CNP activation can be seen at
the concentration as low as 1 nM, suggesting its high sensitivity (Fig. 16 A), and it was
nearly abolished by PrkG1 inhibitor, KT-5823 (Fig. 16 B) indicating the requirement of
PrkG1α. Thus, CNP can indeed regulate axon branching of sensory neurons by activating
cGMP-PrkG1 signaling in culture.

36
Figure 15 NPs Promote Axon Branching of Dissociated DRG Neurons in Culture.

Dissociated E14 DRG neurons were cultured in collagen gels in the presence of NGF (25
ng/ml) for 24 hours and then treated with buffer (ctrl, A-C), or 100 nM ANP (D-F), BNP
(G-I), or CNP (J-L). After cultured for another day, they were fixed and stained with
neurofilament antibodies. Regions of the cultures are shown at low magnification in A, D,
G, and J, while individual cells are shown at high magnification in B, C, E, F, H, I, K, and
L. Neurons treated with all NPs showed a significant increase in branch formation. Bars:
100 μm. (M-N) Distribution of neurons with different number of branches (M) and the
average number of branching points (N, *** p<0.001, ANOVA test) measured from the
above cultures treated with different NPs (100 nM). The results were plotted as mean±s.d.
in (M), and mean±s.e.m. in (N).

37
Figure 16 NPs Promote Axon Branching via PrkG1α α α α

A) Comparison of the average number of branching points in the cultures treated with
different doses of NPs. CNP has effect with concentration as low as 1nM, while ANP or
BNP requires ~100 times more to achieve a similar effect. More than 80 neurons were
analyzed for each condition. (B) Quantification of average branching points per neuron in
dissociated DRG neuron culture, treated with different combination of buffer alone,
10nM of CNP or 5 μM of KT-5823 as indicated. KT-5823 abolished the axon branching
induced by CNP (p=0.18, t-test). (C) Quantification of number of dead cells per every
quarter of DRG explants cultured with buffer or 10nM CNP. More than 5 explants were
analyzed in each condition, and no significant difference in cell death rate was found with
CNP treatment (p=0.47, t-test). The results were plotted as mean±s.e.m. in (A-C).

38
3. 3 CNP Stimulates Axon Outgrowth from DRG Explants
Extracellular cues, such as netrins, neurotrophins and semaphorins, can regulate multiple
aspects of axonal development, including outgrowth, guidance and branching (Dent et al.
2004, Messersmith et al. 1995, Yaron et al. 2005, Zheng et al. 1994). Therefore, we next
asked if CNP, a secreted peptide, has the ability of regulating DRG axon outgrowth. To
avoid the dominant effect of neurotrophins, we cultured DRG explants in the presence of
NGF (5 ng/ml) for a day and then removed it on the second day. Under this condition,
axons grew radically from the explants but did not form dense axonal halos that were
often seen in the culture with the continuous presence of NGF at high dose. Interestingly
however, addition of 100 nM CNP after the removal of NGF allowed the explants to
grow out not only longer but also denser axons around them (Fig.17 D). This can be best
illustrated by the halo size, which increased by nearly two fold in the presence of CNP
(Fig.17 E).
To further test if CNP alone could initiate axonal outgrowth, we performed culture in the
absence of neurotrophins. In this case, only <20% of explants were able to grow axons
longer than 50μm after 48 hrs and none of them had more than five axons (Fig.17 F, J).
However, addition of CNP elicited extensive axonal outgrowth from the explants. At the
lowest CNP concentration tested (1 nM), all of the explants grew axons with half having
1-5 axons (Fig.17 J, gray bar) and half having more than five axons (Fig.17 J, open bar).
At the highest concentration tested (100 nM, Fig.17 I, J), all explants (n=12) had more

39
than five axons and some even formed axonal halos (Fig.17 I), which is never seen in the
control condition.
These two results indicate that CNP can mimic the activity of extracellular factors like
NGF for DRG neurons: not only enhance the growth of existing axons, but also able to
initiate axon outgrowth from explants. In addition, this activity can be replicated by using
Br-cGMP (Fig. 17 J) and blocked by KT-5823 (Fig. 18 A-H), again supporting the
involvement of cGMP signaling. Finally, the activity is not due to the increase in cell
survival that might be promoted by CNP, as TUNEL staining did not reveal any decrease
in the number of apoptotic cells (Fig. 16 C). Thus, CNP can serves as an extracellular
cue to stimulate axon outgrowth.

40
Figure 17 NPs Induce Axon Outgrowth from DRG Explants.

(A-D) E14 DRG explants were cultured in collagen gels for one day in the presence of
NGF (5 ng/ml), and then replaced with the growth medium without NGF and treated with
buffer (A), or 100 nM of ANP (B), BNP (C), or CNP (D). They were cultured for another
day before fluorescent staining using neurofilament antibodies. Explants treated with all
three NPs have increased halo size (B-D) as compared to the buffer control (A). Bar: 100
μm. (E) Quantification of the outgrowth activity by measuring the halo size treated with
NPs at different concentrations. ANP or BNP at 10-100 ng/ml and CNP at 1-100 ng/ml
significantly increases in halo size as compared to the buffer control (***, p < 0.001,
ANOVA test). (F-I) E14 DRG explants were cultured in collagen gels in the absence of
NGF and treated with buffer (F), or 100 nM of ANP (G), BNP (H), or CNP (I) for two
days before immunostaining. Axonal bundles and halos can be often seen associated with
induced outgrowth by NPs. (J) Quantification of the outgrowth activity by counting the
percentage of explants that grew axons in the culture treated with NPs or Br-cGMP at
different concentrations (F-I). Significant numbers of explants grew either 1-5 axons (low,
gray box) or >5 axons (high, open box) in the cultures with ANP or BNP at 10-100 ng/ml
as well as with CNP at 1-100 ng/ml (***, p < 0.001, ANOVA test). Bar: 100 μm.

41
Figure 18 NPs Induce Axon Outgrowth via PrkG1α α α α.

A-H) KT-5823 blocks axon outgrowth from DRG explants induced by CNP. DRG
explants from E14 rat embryos were cultured in collagen gels in the absence of NGF, but
treated with buffer (A and B) or CNP (10 nM, C–H). In some cultures, KT-5823 was
included (E–H). Axon outgrowth induced by CNP (C and D) was reduced by 2 μM
KT-5823 (E and F), but totally blocked by 10 μM KT-5823 (G and H). (Scale bar: 200
μm.) (I) Expression of Npr receptors in cultured DRG neurons: RT-PCR analysis
revealed expression of Npr2, but not Npr1, in the freshly isolated DRG, in cultured
explants, and in cultured dissociated DRG neurons. GAPDH was used as an internal
control, and an Npr1-containing plasmid (DNA) was used as a positive control.

42
3.4 A Point Source of CNP Attracts the Growth Cones of DRG Neurons
cGMP has been shown to regulate the growth cone turning response to Sema3A (Song et
al. 2001). However, it was still unclear whether cGMP signaling itself can modulate
growth cone behavior, and none of the NPs has actually been tested in any growth cone
turning experiment yet. Since Npr2 is strongly expressed in the DRG sensory neurons,
we hypothesize that the presence of Npr2 in the growth cone would enable the neuron to
sense CNP, and asymmetric activation of Npr2 by CNP may able to modify axon
pathfinding.
To test this, we utilized a well characterized growth cone turning assay (Song et al. 2001)
with modification for rat DRG neurons. In brief, E14 rat DRG neurons were dissociated
and cultured on cover glass coated with laminin, and after one day of initial growth
period, they were transfer to a stage with heating on an inverted microscope; CNP
gradient was produced closely to the growth cone (~50μm) by pressure pulse at 2Hz
generated by picospritzer; the trajectory of the growth cone was recorded by CCD camera.
As shown in Fig. 19 A, growth cones of those DRG neurons normally followed a linear
trajectory. Occasionally, they changed growth directions (Fig. 19 C), but the overall
turning angle is around 0.7
o
±1.9
o
(n=27, Fig. 19 F). However, significant turning
responses were observed for the growth cones that were presented with a CNP gradient,
as illustrated by the reorientation of the growth cones (Fig. 19 B). At the end of 30-min
recording, the majority of the growth cones migrated toward the tip of the pipette (Fig.19

43
D,E) with an average turning angle of 18.9
o
±4.5
o
(n=26, Fig. 19 F). On the other hand, no
significant change in overall growth rates was detected (Fig.19 G). This attractive
response is best illustrated by the time-lapse movie (http://branching.usc.edu/), in which
the growth cone first reoriented itself toward the pipette, but after the pipette was
repositioned to a new location, the growth cone was able to follow it and adjust its growth
direction. Thus, CNP can serve as a guidance cue to positively control the growth
direction of sensory axons.

44
Figure 19 Attractive Turning Responses of DRG Growth Cones in CNP Gradient

(A-B) Growth cone turning assays were carried out in the overnight culture of dissociated
E14 DRG neurons. Representative phase-contrast images with 5-minute interval
demonstrate the nearly linear growth of a control axon (A) and the turning response of a
growth cone (B) in the CNP gradient created from a pipette tip (indicated by the arrows at
the upper left corner). Bar: 50 μm. (C-D) Composite drawings of the growth cone
trajectory for a period of 30 min of a population of neurons in the control condition (C) or
in the CNP gradient (D). Bar: 20 μm. (E) A cumulative plot of the turning angles of all
growth cones at the end of 30 min in the absence (○) or presence (●) of CNP gradients.
(F-G) The average turning angles (F, mean±s.e.m, *** p < 0.005, Mann-Whitney test)
and the average distance of axonal extension (G, mean±s.e.m, p = 0.57, Mann-Whitney
test) are compared during the 30-min period calculated for the above growth cones in the
control condition (Ctrl, n=27) or in the CNP gradient (n=26).

45
3.5 The Specificity of NPs in Axonal Development
CNP is structurally related to ANP and BNP, which contain a disulfide-linked ring
structure with conserved 17-amino acid residues (Potter et al. 2006). Since NPs are all
found in the brain and the circulation, we asked whether the positive activities described
above could be also elicited by the other two NPs. With the same in vitro assays, we
found that addition of either ANP or BNP to dissociated DRG neurons promoted branch
formation (Fig.15 D-I) and the presence of either NP stimulated axonal outgrowth in the
explant assay (Fig.17 G, H). Interestingly, however, the doses needed for both NPs were
much higher (10-100 fold) than that needed for CNP. In the branching assay, CNP was
active as low as 1 nM whereas ANP and BNP required at least 100 nM to show a
comparable activity (Fig.15 O). The same trend holds for axon outgrowth, as CNP
promoted outgrowth at 1 nM, while ANP or BNP was not active at the same
concentration (Fig.17 J), and the effect of 100 nM ANP or BNP is only somewhat
comparable to that of CNP at 1-10 nM (Fig.17 J). These results are consistent with Npr2
being the only NP receptor expressed in the neurons used in these assays (Fig. 14, and
Schmidt et al. 2007), and correlate well with the binding affinity of NPs to their cognate
receptors, as ANP and BNP preferably bind to Npr1 (Barr et al. 1996), but still able to
bind Np2 with a much higher EC50. Thus, this family of natriuretic peptides hormones is
capable of regulating axonal development in various aspects, although the sensitivity
depends on the presence of specific Npr receptors.

46
Chapter 4:
cGMP Pathway Controls Bifurcation of Sensory Axons
Even though, cGMP signaling pathway is well studied for decades, the physiological
significance in neuronal development, especially in terms of axon branching, wasn’t
revealed until recently. The first piece of evidence came from an in vivo phenotype study
on PrkG1 knockout mice, in which DRG central afferents labeled by DiI failed to form
rectangular pattern in the DREZ, and was found more in one direction than the other
although without a clear rostral or caudal preference (Fig. 20 and Schmidt et al., 2002).
Based on the expression of PrkG1 in DRG, and our in vitro data about cGMP induced
axon branching, we hypothesized that loss of PrkG1 could cause abnormal axonal
development, particularly loss of axonal branches, which results in this abnormal
phenotype.

47
4.1 Developmental Requirement of PrkG1 in DRG Axon Bifurcation in Mice
To examine this hypothesis, we used DiI iontophoresis to label individual E13.5 mouse
DRG sensory axons, and trace their projection in spinal cord (Fig. 21 B-E). As described
previously (Ma and Tessier-Lavigne, 2007; Ozaki and Snider, 1997), nearly all axons
formed symmetric T- or Y-shaped bifurcation fork in wild type or heterozygous
littermates, and very rarely did they turn to the longitudinal tract without bifurcation
(1.4% in wild type and 5.5% in the heterozygous) (Fig. 21 B-C, Table1). However in the
mutants, the majority of DiI-labeled axons (98.6%) did not have the bifurcation and lost
either the rostral or the caudal branch in a random fashion (Fig. 21 D-E, Table1). The
remaining branch turned correctly into the longitudinal track, and some already sprouted
nascent collateral branches, which suggesting that the defect is not likely due to a delay in
bifurcation (Fig. 20 D-E). This defect is consistent with a recent report, and confirmed in
adult mutant mice as well (Schmidt et al., 2007). Thus, these in vivo analyses indicate
that cGMP signaling indeed plays an important role in sensory axon bifurcation, which is
consistent with its activity in promoting branching in culture.

48
Figure 20 Misprojection of Sensory Afferents in PrkG1 Null Mice

(A-D) Visualization of sensory afferent with DiI bulk labeling in E13 mouse embryos.
Images were taken from the dorsal side of E13.5 spinal cords in an open book preparation.
In the wild type spinal cord, sensory axons form rectangular DREZ where sensory
afferents extend evenly along the rostrocaudal axis (A-B). However, in the PrkG1-/-
mutant spinal cords (C-D), axons form irregular DREZ, with fewer axons on one side
than the other. (E) Quantification of width of DREZ in wild-type (n=2) and PrkG1-/-
mutant (n=2) embryos, after normalized to the size of the spinal cord.

49
Figure 21 Requirement of PrkG1 in DRG Axon Bifurcation during Early
Embryonic Development

A) A cartoon showing the projections of sensory neurons in the E13.5 spinal cord. (B-E)
Visualization of DiI labeled DRG axon projections at the single cell resolution. Images
were taken from the lateral side of E13.5 spinal cords in an open book preparation.
Normally in the wild type spinal cord, sensory axons bifurcate (arrows) at the DREZ,
resulting two daughter branches that extend in opposite directions along the rostrocaudal
axis (B-C). However, in the PrkG1-/- mutant spinal cords (D-E), one of the branches is
missing, while the other appears to grow normally but turn to either rostral or caudal
direction (arrows point to the turning point). Note that the collaterals (arrow heads) are
still formed from the remaining axons. Bar: 500 μm. (F-O) Expression of three different
PrkG isoforms is analyzed in E10.5 (F), E12.5 (G,J,M), E15.5 (H,K,N), or E18.5 (I,L,O)
mouse DRGs by RNA in situ hybridization using digoxigenin-labeled RNA probes for
PrkG1α (F-I), PrkG1β (J-L), and PrkG2 (M-O). Bar: 100 μm.

50
Table 1 Quantification of Sensory Axons with or without Bifurcation in PrkG1
Transgenic Mouse Embryos

Genotype
Number of
embryos
Total neurites
analyzed
Bifurcation
No
Bifurcation
PrkG1 +/+ 2 481 98.6% 1.4%
PrkG1 +/- 4 115 94.5% 5.5%
PrkG1 -/- 3 528
1.4%
(***)
98.6%
(***)

Quantification of the DiI labeled sensory axons from Fig. 21 B-E. The number of
embryos and the total neurites analyzed in different genotype categories, the percentage
of axons with correct bifurcation and loss of bifurcation are listed for comparison. Almost
none of the afferents bifurcated in the PrkG -/- mutants as compared to the wild type or
heterozygote littermates (***, p < 0.001, t test).

51
4.2 CNP Is the Environmental Cue for the Sensory Neurons to Bifurcate
Next, we asked what could be the environmental signal that stimulates the production of
cGMP in the DRG sensory neurons when they reach the DREZ, and initiates the
bifurcation process. As shown earlier, YC-1, an activator of sGC, can mimic the cGMP
activity on promoting axon branching (Zhao et al. 2009); furthermore sGC and one of its
endogenous activator, neuronal nitric oxidase synthase (nNOS), have been shown to be
present in DRG as well (Mungrue et al. 2004, Schmidt et al. 2007, 2009). So we
examined the central projections of embryos lacking nNOS (Packer et al. 2004). However,
DiI labeling of single sensory axons showed no change of bifurcation in DREZ (Fig. 22
D-E). And more recently, sGC knock-out mouse has also been shown with normal
sensory axon bifurcations (Schmidt et al. 2009). However on the other hand, our RNA
in situ hybridization data proved the existence of another possible upstream activator of
cGMP production, which mediated by CNP-Npr2 (Fig.14). More importantly, Npr2 was
just shown to be same essential as PrkG1α in regulating sensory axon bifurcation in
spinal cord, because mice lacking Npr2 expression exhibited identical lose-of-bifurcation
phenotype (Fig. 24 F-G, Schmidt et al. 2007). Thus, the expression of Nppc in the dorsal
spinal cord is very likely to be the extracellular signal for sensory axon to bifurcate.
To establish the developmental function for CNP in axon branching, we examined
sensory afferents in the long bone abnormality (lbab) mouse, a spontaneous autosomal
recessive mutant that was isolated in the Jackson laboratory based on its reduced

52
postnatal body size and long bone morphology (Chusho et al. 2001). These phenotypes
were attributed to abnormal bone development in the absence of CNP-Npr2 signaling, as
similar defects were found in mice with targeted deletion of CNP or Npr2 (Tamura et al.
2004), as well as in the spontaneous Npr2 mutant (cn) (Tsuji et al. 2001). Furthermore, a
missense mutation was recently mapped to the Nppc gene that results Arg to Gly
substitution in the CNP functional domain and hence the loss of CNP activity (Jiao et al.
2007). Since the general embryonic development is not affected in this mutant, the lbab
mouse provides a useful loss-of-function model to investigate CNP functions in the
developing spinal cord.
As shown in table 3, almost all of the sensory axons in the wild type embryo (100%,
n=63) or in the three heterozygous embryos (97.1%, n=175) bifurcated at the DREZ and
have two branches extending in opposite directions along the rostrocaudal axis (Fig.22
B-C). However, 98% of the afferents analyzed in the six lbab/lbab mutant embryos
(n=387) turned randomly into the longitudinal track without making the second branch
(Fig.22 H-I). In contrast, collateral branches from the proprioceptive neurons that
normally just start to sprout at this age can still be found, and have normal length and
projection inside the spinal cord as revealed by bulk labeling (L-Collateral, Fig. 24 A-E).
This phenotype is identical to the bifurcation defect found in the Npr2 and PrkG1 mutant
embryos (Schmidt et al 2007, Zhao et al 2009), which demonstrates the requirement of
CNP for sensory axon bifurcation in the developing spinal cord.

53
Figure 22 CNP Is Required for Sensory Axon Bifurcation.

A) A cartoon showing the projections of sensory neurons in the E13.5 spinal cord. (B-I)
Visualization of DiI labeled DRG axon projections at the single cell resolution from E13
wild-type, nNOS -/-, cn/cn and lbab/lbab mouse embryos. Images were taken from the
lateral side of spinal cords in an open book preparation. In wild-type (B-C) and nNOS -/-
(D-E) spinal cord, sensory axons bifurcate (arrows) at the DREZ, resulting two daughter
branches that extend in opposite directions along the rostrocaudal axis. However, in the
cn/cn (F-G) and lbab/lbab (H-I) mutant spinal cords, one of the branches is missing,
while the other appears to grow normally but turn to either rostral or caudal direction
(arrows point to the turning point). Note that the collaterals (arrow heads) are still formed
from the remaining axons. Bar: 500 μm.

54
Table 2 Analysis of Sensory Axon Phenotypes from Individual Nppc Mutant
Embryos

Litter Embryo Genotype
Total neurites
analyzed
Bifurcation
No
Bifurcation
1
#2 lbab/lbab 60 1 59
#7 +/+ 63 63 0
2
#7 lbab/lbab 41 0 41
#8 lbab/lbab 50 2 48
3
#1 lbab/+ 95 93 2
#2 lbab/lbab 91 1 90
#5 lbab/lbab 116 3 113
#9 lbab/+ 80 77 3
#10 lbab/lbab 70 3 67

List of detailed analysis of the total number of neurites, the number of axons with correct
bifurcation or missing bifurcation in all the embryos analyzed from three E13.5 litters.

Table 3 Quantification of Sensory Axons with or without Bifurcation in Nppc
Mutant Mouse Embryos

Genotype
Number of
embryos
Total neurites
analyzed
Bifurcation
No
Bifurcation
+/+ 1 63 100% 0%
lbab/+ 2 175 97.1% 2.9%
lbab/lbab 6 387 2.6% 97.4%

List of the number of embryos and total neurites analyzed in different genotype
categories, the percentage of axons with correct bifurcation and loss of bifurcation are
listed for comparison. Almost none of the afferents bifurcated in the lbab/lbab mutants as
compared to the wild type or heterozygote littermates (***, p < 0.001, t test).

55
4.3 CNP Is Expressed in a Gradient Pattern in Dorsal Part of Spinal Cord.
Molecules like Wnt and Eph, regulate axonal development by forming a gradient
expression pattern (Lyuksyutova et al. 2003). Since the expression of Nppc is restricted in
the dorsal region of embryonic spinal cord, we wonder if there is a possible CNP gradient
in the spinal cord which attracts sensory axons and contributes to the formation of DREZ.
Thus, we carefully examined the expression of Nppc with RNA in situ hybridization in
both sagittal (Fig. 23 B) and coronal (Fig. 23 D) sections of the E10.5 spinal cord. The
expression pattern shown in sagittal sections (Fig. 23 B) strengthened our idea of a
possible dorsal to ventral gradient. Line scans of the labeling intensity along the
dorsal-ventral axis on both sagittal (Fig. 23 E) and coronal (Fig. 23 F) sections further
confirmed the gradient expression of Nppc within the dorsal half of the spinal cord.
To further address the possible role of the CNP gradient on the formation of DREZ in the
spinal cord, we performed DiI bulk labeling on E13.5 embryos, which revealed that in the
lbab/lbab mutants, neither the width of the DREZ (L-DREZ, Fig. 24 E), nor the distance
between the left and right DREZ (W-mid, Fig. 24 E) was significantly different from that
in the wild-type littermates. This indicates that the position of the DREZ is independent of
Nppc expression. Thus, even though Nppc expression has a dorsal to ventral gradient, it
only provides a necessary cue for the sensory axons to bifurcate, but has little influence
on the formation of DREZ.

56
Figure 23 Gradient Expression of Nppc along the Dorsal-ventral Axis in the Spinal
Cord

(A-D) The expression pattern of Npr2 and Nppc in E10.5 mouse embryos showing by
RNA in situ hybridization on segital (A, B) and coronal sections (C, D). Bar: 500 μm.
(E-F) line scans of intensity of RNA in situ hybridization signal on 100 pixels (dorsal half
of the spinal cord) alone the dorsal ventral axis. Individual traces in (E) represent
different locations in (B) (colored arrow heads), and individual traces in (F) represent the
spinal cord region from different coronal sections (SC1-8). Note the expression level is
high around roof plate, and gradually decline towards ventral side.

57
Figure 24 CNP Is Not Required for the Formation of DREZ or Collateral
Branches

(A-D) Coronal sections after DiI bulk labeling in E 13 DRG reveals no difference in the
size and location of DREZ in the lbab/lbab mutants (C, D), compared to wild-type
littermates (A, B). (E) Quantification of the average width of DREZ (L-DREZ), the
length of the collateral branches from the proprioceptive afferents (L-Collateral) and the
distance between left and right DREZ (W-mid) showed no significant difference between
lbab/lbab mutants (n=3) and wild-type littermates (n=2) (t-test).

58
4.4 Peripheral Sensory Projections in the Absence of CNP
RNA in situ hybridization confirmed both PrkG1 and Npr2 are expressed in dorsal root
(Fig. 14) and trigeminal ganglia (data not shown), and in vivo phenotype studies showed
sensory afferents missing bifurcation in central projection in spinal cord, so we further
asked if there could be any defect in peripheral projection of sensory axon during
development as well.
Our approach was visualizing peripheral axon projections with florescence based
wholemount staining method using antibodies against neurofilament on E11.5 embryos
(Wickramasinghe et al, 2008). As shown in Fig 25 A-B, the facial projections from
trigeminal ganglion, and the somatosensory projections alone the body wall originated
from dorsal root ganglia were clearly labeled by florescent secondary antibodies.
However, no defect was found in those axon tracts in lbab/lbab mutants (Fig 25 E-F)
compared to their wild-type littermates (Fig 25 C-D), indicating peripheral axon
projections might not have been affected by loss of CNP signaling, at least in early
developmental stage, which is consistent with the Nppc expression around this stage, as it
is mainly expressed strongly in the dorsal part of spinal cord, revealed by Lacz reporter
mice (Schmidt H, et al. 2009).

59
Figure 25 Peripheral Sensory Projections in lbab/lbab Mutant Embryos

(A-B) low magnification of florescence based wholemount staining using antibodies
against neurofilament on E11.5 embryos. The red window indicates facial projections
from trigeminal ganglion (A) and somatosensory projections (B). (C-D) high
magnification of those areas revealed there is no gross defect in early sensory projections
in peripheral in lbab/lbab mutant embryo (E-F), compared with wild-type littermates
(C-D).

60
4.5 Bifurcation Pattern in GSK3α/β 21A/21A/9A/9A Double Knock-in Mice
As we have shown, CNP-Npr2-cGMP-PrkG pathway signaling regulated both axon
branching in vitro and sensory bifurcation in vivo. However, their downstream target,
GSK3, has only been validated in culture. Since GSK3α and GSK3β are believed to be
functionally redundant, information regarding mouse phenotype with manipulation in
both genes would provide the final missing piece to our study. However, considering the
essential role of GSK3 in general cellular function and the double inhibition mechanism
from cGMP to GSK3 to axon branching, knock-out of both GSK3 α and β in mice is not
the best approach for this purpose. On the other hand, our data also showed the mutant
GSK3β S9A can sever as a phosphorylation resistant form to block cGMP signaling on
axon branching in culture, so we hypothesis that if the endogenous GSK3 α and β can be
replaced with mutant forms, it would be more likely to phenocopy the loss of Nppc, Npr2,
or PrkG1. Interestingly, GSK3α/β 21A/21A/9A/9A double knock-in mice were generated
by Dario R Alessi's group in Scotland. Those mice develop and grow normally. Wnt
signaling still able to phosphorylate b-catenin and stimulate Wnt-dependent transcription
in neurons isolated from those mice, and hippocampal neurons from those mice exhibited
normal development of polarity in vitro (McManus et al. 2005). These data suggests
phosphorylation on GSK3α Ser21 and GSK3β Ser9 is not required for axon formation or
outgrowth, which is consistent with our finding that the phosphorylation modifies axon
branching.

61
Figure 26 No Bifurcation Defect in GSK3 Double Knock-in Mice

A-D) Visualization of DiI labeled DRG axon projections at the single cell resolution from
E13 GSK3α/β 21A/21A/9A/9A double knock-in embryos. Example pictures of axons
with loss of bifurcation are shown in (A-B), and examples of correct bifurcations are
shown in (C-D). Images were taken from the lateral side of spinal cords in an open book
preparation. Bar: 100 μm. (E-F) P21 adult Thy-1 YFP-H reporter mice showed no defect
in sensory bifurcation at mature stages in GSK3α/β 21A/21A/9A/9A double knock-in
mutants (E), compared to GSK3α/β +/21A/+/9A double heterozygous mice (E). Images
were taken from the dorsal side of spinal cords. Bar: 100 μm.

62
Table 4 Quantification of Sensory Axons with or without Bifurcation in GSK3α/β
21A/21A/9A/9A Double Knock-in Mouse Embryos

Litter Embryo Genotype
Total neurites
analyzed
Bifurcation
No
Bifurcation
% no
bifurcatio
n
1 #4
GSK3α/β
21A/21A/9
A/9A
double
knock-in
139 124 15 10.8%
2 #6 161 141 20 12.4%
3 #2 135 116 19 14.1%
4 #10 69 60 9 13.0%
Total 4 504 441 63 12.5%

Quantification of the DiI labeled sensory axons in GSK3α/β21A/21A/9A/9A double
knock-in embryos. The number of embryos and total neurites analyzed in different
genotype categories, the percentage of axons with correct bifurcation and loss of
bifurcation are listed for comparison. In all 4 double knock-in embryos, there are more
than 10% of axons failed to bifurcate after entering DREZ in spinal cord.

63
To test this hypothesis, we again labeled sensory axons with DiI iontophoresis on E13.5
GSK3α/β 21A/21A/9A/9A double knock-in embryos. As shown in Fig. 26 A-D and table
4, we found, although the majority of sensory axons (87.5%) still bifurcated after
reaching DREZ, there was an increase in frequency in which axons failed to do so (12.5%
in Table 4 ), as compared to that in wild-type (1.4% in Table 1, 0% in Table 3).
In order to examine sensory axon bifurcation at adult stages, we crossed GSK3α/β
+/21A/+/9A double heterozygous with Thy-1 YFP-H reporter mouse, which has been
shown to label a small subset of sensory afferents in the adult spinal cord (Feng et al
2000). Thus, we were able to generate GSK3α/β 21A/21A/9A/9A double knock-in mice
in the Thy-1-YFP background. However, the sensory axons visualized in those mice at
age of P21 showed no defect in bifurcation (Fig. 26 E).
Since pyramidal neurons in mouse cortex are also labeled by Thy-1-YFP, we were also
able to look at cortical neuron development in GSK3α/β 21A/21A/9A/9A double
knock-in mouse with the Thy-1-YFP allele. As shown in Fig. 27, there is no gross
morphological change of pyramidal neurons labeled by Thy-1-YFP in the P21 double
knock-in mouse (Fig. 27, B) as compared to their heterozygous littermate (Fig. 27, A). In
addition, the thickness of the cortex or layer 5 (Fig. 27, A-B bracket) also remains the
same (Fig. 27, C), which indicates there is no migration defect in those pyramidal
neurons either, even though it was found in PrkG1 mutant mouse ((Demyanenko et al.
2005)

64
Figure 27 No Cortical Development Defect in GSK3 Double Knock-in Mice

(A-B) the morphology of pyramidal neurons in P21 GSK3α/β 21A/21A/9A/9A double
knock-in mouse was visualized by crossing to Thy-1 YFP-H reporter mouse, and
compared with GSK3α/β +/21A/+/9A double heterozygous littermates. (C) The thickness
of cortex or layer 5 (labeled by green cell bodies) in GSK3α/β 21A/21A/9A/9A double
knock-in mouse is not significant from that in GSK3α/β +/21A/+/9A double heterozygous
littermates (t-test).

65
Conclusion
From establishing neuronal polarity to axon elongation, from branch formation to
synaptic contact, formation of elaborated axonal morphology is a critical part in
assembling functional neuronal circuits. The diversity and complexity in axon branching
pattern was documented more than a century ago by Santiago Ramón y Cajal, however
the underlining mechanisms are still not well defined. In this study, we provided strong
evidence showing that cGMP signaling pathway is indeed one of such mechanisms. Our
model is: embryonic sensory neuron from DRG can sense the extracellular CNP signal
with membrane bound receptor Npr2, which locally catalyzes the production of cGMP;
cGMP activates PrkG1α, which results in the phosphorylation of GSK3 and further
removal of inhibition on axon branching (Fig. 28 A); in culture, the cGMP or CNP
treatment cause a global activation of PrkG1α, which results in an excessively branched
morphology of dissociated DRG neurons (Fig. 28 B); in the embryonic spinal cord, after
sensory axon reaches DREZ, it start to contact with CNP signal, which leads to the
creation of an addition branch, and further establishment of bifurcation (Fig. 28 C),
although this step may also need other local signals, like Slit.

66
Figure 28 A Working Model of cGMP Signaling in Axon Branching

A) A diagram showing the axon branching mechanism mediated by cGMP signaling
pathway (inside dash line). (B) A cartoon showing the global activation of cGMP
signaling in culture. (C) A cartoon showing the mechanism which controls the sensory
axon bifurcation in the spinal cord: CNP is produced in the dorsal region of the spinal
cord, which is sensed by the Npr2 receptors on the sensory axon and activates PrkG1 in
those DRG neurons, thus bifurcation process is initiated.

67
Discussion
cGMP Signaling Pathway Regulates Axon Branching in DRG Neurons
The finding of cGMP as a regulatory factor on axon guidance in response to semaphorin
in vitro (Song and Poo, 1999) pointed out a potential role of cGMP signaling axonal
development. However, no direct evidence was documented until recently. By in vitro
culture, we demonstrated that cGMP has a potent effect on promoting axon branching in
dissociated DRG neurons (Fig. 1-2); and PrkG1α activation is both necessary and
sufficient for axon branching induced by cGMP (Fig. 4-6) in culture. Together with
another group (Schmidt et al 2007 and 2009), our studies on PrkG1 null mice identified
that protein kinase PrkG1α is required for the bifurcation of sensory afferents in spinal
cord (Fig. 21, Table 1). Thus the compelling evidences from both in vitro and in vivo data
revealed a novel role of cGMP signaling in regulating axon branching.
However, two questions were brought up to us. First, how does PrkG1α mediate the
cGMP signal to the remodeling of cytoskeleton in order to generate additional branches?
Fortunately with a candidate approach, we identified GSK3 as a new substrate for PrkG1
in this particular regulation (Fig. 8-11): GSK3 and PrkG1 are biochemically associated
with each other if expressed together; more importantly GSK3 can be directly
phosphorylated by PrkG1 in vitro; furthermore, a GSK3 mutant (GSK3β S9A)
successfully blocked cGMP signaling on stimulating branching. GSK3 has been shown to

68
regulate microtubule assembly via CRMP2 and APC (Yoshimura et al., 2005; Zhou et al.,
2004), and is also downstream of NGF-Akt signaling in DRG neurons. Pharmacological
studies with GSK3 inhibitors indicated that complete inactivation of GSK3 in the growth
cone mediated NGF-dependent axon growth, while partial inactivation along the axons
led to non-polarized activation of primed substrates such as CRMP2 and APC along the
axon and hence increase in branch formation (Kim et al., 2006). Our data from the
phosphorylation study in neurons showed cGMP induced a lesser increase in GSK3
phosphorylation, compared to NGF treatment, which suggested that cGMP signaling may
just partially inactive GSK3, thus release the inhibition on CRMP2 and APC, which
further affects microtubule assembly. Thus, GSK3 is the missing link between cGMP
synthesis and branch formation.
The other question is how neurons achieve local activation of cGMP signaling to generate
precise branching patterns. This wasn’t easily answered until we mapped CNP as the
upstream activator. As shown by RNA in situ hybridization, only Nppc, the gene encodes
the precursor of CNP, is expressed in the dorsal part of spinal cord during early
development, while the receptor Npr2 and kinase PrkG1α are synthesized in DRG
neurons. So, CNP-Npr2 signaling provides a likely mechanism for local activation.
Interestingly, this ligand-to-receptor activation is extremely sensitive: as shown by in
vitro studies, CNP at 1 nM concentration is sufficient and reliable to induced significant
axon branching (Fig. 15). In addition, growth cones of dissociated DRG neurons are able

69
to sense CNP gradient and migrate towards the source (Fig. 19), which suggests that the
local machinery is assembled in the growth cone, and ready to respond to extracellular
CNP signals.
CNP-Npr2-PrkG1α α α α Signaling Controls Sensory Axon Bifurcation in the spinal Cord
In order to determine the physiological significance of cGMP signaling in neuronal
development, we examined several mouse lines with genetic modification in the pathway,
including PrkG1 targeted knockout, Npr2 spontaneous mutant (cn/cn), CNP spontaneous
mutant (lbab/lbab) and nNOS targeted knockout. Interestingly, PrkG1 null, cn/cn, and
lbab/lbab mutant mice all share an identical phenotype, which is missing bifurcation in
the central projection of sensory axons from DRG; however this defect was not found in
nNOS targeted knockout mouse, or sGC targeted knockout mouse as reported (Schmidt et
al., 2009). Reduction in sensory input from DRG to spinal cord, including less capsaicin
responding cell and decrease in mEPSC frequency, has been reported in those mice with
bifurcation defect as well (Schmidt et al., 2007 and 2009). These studies established the
requirement of the CNP-Npr2-PrkG1α signaling pathway in controlling sensory afferent
bifurcation, hence building precise peripheral sensory circuits.
However, bifurcation is only a specific type of axon branching, as mentioned earlier, at
least formation of collateral branches in those sensory axons are not affected by
deficiency in cGMP signaling. So, the in vivo defect we have reported does not

70
completely reflect the effect of activating cGMP signaling in vitro culture. However, the
consistency here is cGMP signaling is not the only pathway that regulates axon branching.
Maybe it indeed only controls axon bifurcation, and different signaling pathways are
deployed in other type of branching, which need to be discovered. Another explanation is
every type of axon branching is a result of a combination of environmental cues, and
CNP maybe only one of many extracellular signals that tells sensory axon to bifurcate.
This kind of hypothesis has been proposed with the notion that Slit-Robo signaling is also
important to establish the correct bifurcation in the same location (Ma et al. 2007, Zhao et
al., 2009). No matter what, in order to fully understand this simple yet faithful bifurcation
process, establishing an in vitro culture system with reliable reproduction of bifurcation
event in those neurons will offer great help to us.
GSK3 Is a Downstream Target Shared by NGF, Wnt and NP Signaling
Recent independent studies on axonal morphogenesis in DRG neurons have pointed out
GSK3 as a downstream target in 3 different signaling pathways: NGF, Wnt and CNP
(Kim et al., 2006, Purro et al., 2008, Zhao et al., 2009). This not only strengthened the
significance of GSK3 in neuronal development, but also initiated a proposal that GSK3 is
a convergent point for regulatory pathways to control axonal morphology. Although NGF,
Wnt and CNP are three unrelated proteins with no similarity, their signaling mechanisms
have a lot in common: they bind to transmembrane receptors, activate protein kinases. On
the other hand, GSK3 provides a perfect target for receiving and conducting multiple

71
regulations: it can be phosphorylated by Ser/Thr kinases, and itself is also a Ser/Thr
kinase capable of phosphorylating various proteins, such as CRMP2, APC and β-catenin.
However, most studies of GSK3 on axonal development were from in vitro studies with
various inhibitors or RNA interference. GSK3β knockout was reported as embryonic
lethality after E16 caused by severe liver degeneration (Hoeflich et al. 2000).
Heterozygous mice have been reported with memory reconsolidation problem (Kimura et
al. 2008). However, no obvious sensory axonal projection defect was found at E13.5 in
GSK3β null embryos; E16 hippocampal neuron developed normal polarity in culture also
(Kim et al. 2006). GSK3α targeted knockout (MacAulay et al. 2007) was available more
recently, but not much is known for its role in neuronal development so far yet. Since
GSK3α and GSK3β are structurally and functionally closely related, and found to
co-exist in many regions of the nervous system (Kim et al. 2006, Kimura et al. 2008), so
they might be able to compensate for the loss of the other. This may be the reason genetic
deletion of single isoform didn’t result in any clear defect in neuronal development.
On the other hand, we focused our study on an alternative approach with a hypothesis that
GSK3α/β 21A/21A/9A/9A double knockin mouse could phenocopy the bifurcation defect.
However, only a mild increase in failure rate of axon bifurcation (12.5%, Table 4) was
detected as compared to almost complete loss in PrkG null or lbab/lbab embryos.
Possible explanations are: (1) the Ser-to-Ala mutation doesn’t totally mimic the
phosphorylation and activation of GSK3, thus PrkG1 activity is not fully bypassed in the

72
double knock-in mice; (2) in the knock-in condition, there are only two copies of
GSK3α/β 21A/21A/9A/9A compared to the GSK3β-S9A over-expression condition; (3)
GSK3 is not downstream of cGMP signaling in this bifurcation process. Regardless of
this, further study on GSK3α/β 21A/21A/9A/9A double knockin mouse will eventually
provide us valuable information of GSK3 in regulating axonal development.
Natriuretic Peptides are a New Class of Extracellular Growth Factor like Molecules
NPs are known for its unique function in regulating vascular physiology, but in this study,
several novel activities of these hormones in axonal development have been revealed.
First, biochemical and genetic evidence confirmed the positive regulation of CNP in axon
branching: in culture, all three NPs stimulated the formation of axon branches (Fig. 15);
whereas in the CNP-deficient mouse embryos, the sensory afferents failed to generated
bifurcated branches in spinal cord (Fig. 22). In addition, explants culture illustrated that
NPs not only can stimulate the growth of existing axons, but also initiate axon outgrowth
from explants in culture (Fig. 17). The growth cone turning assay with a CNP gradient
demonstrated the role of CNP as an axon guidance Furthermore, growth molecule, as
growth cones from dissociated DRG neurons were able to sense the CNP cue and tended
to migrate towards it (Fig. 19). Thus, the family of NPs is a new class of extracellular cue
that can stimulate axon outgrowth, promote branch formation and guide growth cone
pathfinding.

73
Although we extended our knowledge of natriuretic peptides in regulating axonal
development by in vitro study, our approach towards understanding the physiological
significance of those NPs in neuron development in vivo succeeded only in central
projection of sensory axon, where the bifurcation in spinal cord is precisely controlled by
CNP-Npr2-PrkG1α signaling pathway. The search for defects in peripheral projection of
sensory neurons in CNP deficient embryos by florescence based wholemount staining
returned without positive results. This can be explained as in the early developmental
stage (E10-E12), CNP can only be found in the dorsal region of spinal cord. Thus,
detailed study on the expression pattern of CNP in later stages, or even adulthood, may
provide us useful information to further examine the in vivo relevance of NPs in nervous
system.
On the other hand, the idea of those peptide hormones can regulate axonal development is
consistent with a recent discovery, as endothelins, a family of 21-amino acid peptides
produced primarily in the endothelium with an essential role in vascular homeostasis, also
have been found to guide sympathetic axonal development (Makita et al. 2008). Thus,
both families of peptide hormones and their receptors can potentially mediate the
crosstalk between cardiovascular and nervous systems. The fluctuation of those peptides
in local tissue or even blood flow can be sensed by nearby neurons, which lead to
possible changes in function and morphology. However, this hypothesis is still need to be
tested.

74
References
Ayoob, J. C., Yu, H. H., Terman, J. R., and Kolodkin, A. L. (2004). The Drosophila
receptor guanylyl cyclase Gyc76C is required for semaphorin-1a-plexin A-mediated
axonal repulsion. J Neurosci 24, 6639-6649.
Barnstable, C. J., Wei, J. Y ., and Han, M. H. (2004). Modulation of synaptic function by
cGMP and cGMP-gated cation channels. Neurochem Int 45, 875-884.
Barr, C. S., Rhodes, P. & Struthers, A. D. (1996) C-type natriuretic peptide. Peptides 17,
1243-51.
Browning, D. D., Mc Shane, M., Marty, C., and Ye, R. D. (2001). Functional analysis of
type 1alpha cGMP-dependent protein kinase using green fluorescent fusion proteins. J
Biol Chem 276, 13039-13048.
Busch, J. L., Bessay, E. P., Francis, S. H., and Corbin, J. D. (2002). A conserved serine
juxtaposed to the pseudosubstrate site of type I cGMP-dependent protein kinase
contributes strongly to autoinhibition and lower cGMP affinity. J Biol Chem 277,
34048-34054.
Cao, L. H. & Yang, X. L. (2008) Natriuretic peptides and their receptors in the central
nervous system. Prog Neurobiol 84, 234-48.
Chusho, H., et al. (2001) Dwarfism and early death in mice lacking C-type natriuretic
peptide. Proc Natl Acad Sci U S A 98, 4016-21.
Cogen, J., and Cohen-Cory, S. (2000). Nitric oxide modulates retinal ganglion cell axon
arbor remodeling in vivo. J Neurobiol 45, 120-133.
Demyanenko GP; Halberstadt AI; Pryzwansky KB; Werner C; Hofmann F; Maness PF
(2005). Abnormal neocortical development in mice lacking cGMP-dependent protein
kinase I., Brain Res Dev Brain Res. 160(1):1-8
Dent, E. W., Tang, F., and Kalil, K. (2003). Axon guidance by growth cones and branches:
common cytoskeletal and signaling mechanisms. Neuroscientist 9, 343-353.
Dent, E. W., Barnes, A. M., Tang, F. & Kalil, K. (2004) Netrin-1 and Semaphorin 3A
Promote or Inhibit Cortical Axon Branching, Respectively, by Reorganization of the
Cytoskeleton. J. Neurosci. 24, 3002-3012.

75
DiCicco-Bloom, E., et al. (2004) Embryonic expression and multifunctional actions of
the natriuretic peptides and receptors in the developing nervous system. Dev Biol 271,
161-75.
Dundore, R. L., Clas, D. M., Wheeler, L. T., Habeeb, P. G., Bode, D. C., Buchholz, R. A.,
Silver, P. J., and Pagani, E. D. (1993). Zaprinast increases cyclic GMP levels in plasma
and in aortic tissue of rats. Eur J Pharmacol 249, 293-297.
Eldar-Finkelman, H., Argast, G. M., Foord, O., Fischer, E. H., and Krebs, E. G. (1996).
Expression and characterization of glycogen synthase kinase-3 mutants and their effect
on glycogen synthase activity in intact cells. Proc Natl Acad Sci U S A 93, 10228-10233.
Feil, R., Hartmann, J., Luo, C., Wolfsgruber, W., Schilling, K., Feil, S., Barski, J. J.,
Meyer, M., Konnerth, A., De Zeeuw, C. I., and Hofmann, F. (2003). Impairment of LTD
and cerebellar learning by Purkinje cell-specific ablation of cGMP-dependent protein
kinase I. J Cell Biol 163, 295-302.
Feil, R., Hofmann, F., and Kleppisch, T. (2005). Function of cGMP-dependent protein
kinases in the nervous system. Rev Neurosci 16, 23-41.
Feng G, Mellor RH, Bernstein M, Keller-Peck C, Nguyen QT, Wallace M, Nerbonne JM,
Lichtman JW, Sanes JR. (2000). Imaging neuronal subsets in transgenic mice expressing
multiple spectral variants of GFP. Neuron. 28(1):41-51.
Fiscus, R. R. (2002). Involvement of cyclic GMP and protein kinase G in the regulation
of apoptosis and survival in neural cells. Neurosignals 11, 175-190.
Friebe, A., and Koesling, D. (2003). Regulation of nitric oxide-sensitive guanylyl cyclase.
Circ Res 93, 96-105.
Galle, J., Zabel, U., Hubner, U., Hatzelmann, A., Wagner, B., Wanner, C., and Schmidt, H.
H. (1999). Effects of the soluble guanylyl cyclase activator, YC-1, on vascular tone,
cyclic GMP levels and phosphodiesterase activity. Br J Pharmacol 127, 195-203.
Gibbs, S. M., Becker, A., Hardy, R. W., and Truman, J. W. (2001). Soluble guanylate
cyclase is required during development for visual system function in Drosophila. J
Neurosci 21, 7705-7714.
Gibbs, S. M., and Truman, J. W. (1998). Nitric oxide and cyclic GMP regulate retinal
patterning in the optic lobe of Drosophila. Neuron 20, 83-93.
Hidaka, H., and Kobayashi, R. (1992). Pharmacology of protein kinase inhibitors. Annu
Rev Pharmacol Toxicol 32, 377-397.

76
Hoeflich KP, Luo J, Rubie EA, Tsao MS, Jin O, Woodgett JR. (2000). Requirement for
glycogen synthase kinase-3beta in cell survival and NF-kappaB activation. Nature.
406(6791):86-90.
Hofmann, F., Ammendola, A., and Schlossmann, J. (2000) Rising behind NO:
cGMP-dependent protein kinases. J Cell Sci 113 ( Pt 10), 1671-1676.
Jiao, Y ., et al. (2007) A single nucleotide mutation in Nppc is associated with a long bone
abnormality in lbab mice. BMC Genet 8, 16.
Kawasaki, Y ., Kugimiya, F., Chikuda, H., Kamekura, S., Ikeda, T., Kawamura, N., Saito,
T., Shinoda, Y ., Higashikawa, A., Yano, F., et al. (2008). Phosphorylation of GSK-3beta
by cGMP-dependent protein kinase II promotes hypertrophic differentiation of murine
chondrocytes. J Clin Invest 118, 2506-2515.
Kennedy, T. E. & Tessier-Lavigne, M. (1995) Guidance and induction of branch
formation in developing axons by target-derived diffusible factors. Curr Opin Neurobiol 5,
83-90.
Kim, W. Y ., Zhou, F. Q., Zhou, J., Yokota, Y ., Wang, Y . M., Yoshimura, T., Kaibuchi, K.,
Woodgett, J. R., Anton, E. S., and Snider, W. D. (2006). Essential roles for GSK-3s and
GSK-3-primed substrates in neurotrophin-induced and hippocampal axon growth. Neuron
52, 981-996.
Kimura T, Yamashita S, Nakao S, Park JM, Murayama M, Mizoroki T, Yoshiike Y , Sahara
N, Takashima A. (2008). GSK-3beta is required for memory reconsolidation in adult
brain. PLoS ONE. 3(10):e3540.
Kishimoto, I., et al. (2008) C-type natriuretic peptide is a Schwann cell-derived factor for
development and function of sensory neurones. J Neuroendocrinol 20, 1213-23.
Kleppisch, T., Wolfsgruber, W., Feil, S., Allmann, R., Wotjak, C. T., Goebbels, S., Nave,
K. A., Hofmann, F., and Feil, R. (2003). Hippocampal cGMP-dependent protein kinase I
supports an age- and protein synthesis-dependent component of long-term potentiation
but is not essential for spatial reference and contextual memory. J Neurosci 23,
6005-6012.
Lein, E. S., Hawrylycz, M. J., Ao, N., Ayres, M., Bensinger, A., Bernard, A., Boe, A. F.,
Boguski, M. S., Brockway, K. S., Byrnes, E. J. (2007) Genome-wide atlas of gene
expression in the adult mouse brain. Nature 445, 168-176.
Lentz, S. I., Knudson, C. M., and Snider, W. D. (1999). Neurotrophins support the
development of diverse sensory axon morphologies. J Neurosci 19, 1038-1048.

77
Lohof, A. M., Quillan, M., Dan, Y . & Poo, M. M. (1992) Asymmetric modulation of
cytosolic cAMP activity induces growth cone turning. J Neurosci 12, 1253-61.
Lyuksyutova AI, Lu CC, Milanesio N, King LA, Guo N, Wang Y , Nathans J,
Tessier-Lavigne M, Zou Y . (2003). Anterior-posterior guidance of commissural axons by
Wnt-frizzled signaling. Science. 302(5652):1984-8.
Ma, L., and Tessier-Lavigne, M. (2007). Dual branch-promoting and branch-repelling
actions of Slit/Robo signaling on peripheral and central branches of developing sensory
axons. J Neurosci 27, 6843-6851.
MacAulay K, Doble BW, Patel S, Hansotia T, Sinclair EM, Drucker DJ, Nagy A,
Woodgett JR. (2007). Glycogen synthase kinase 3alpha-specific regulation of murine
hepatic glycogen metabolism. Cell Metab. 6(4):329-37.
Makita T, Sucov HM, Gariepy CE, Yanagisawa M, Ginty DD. (2008). Endothelins are
vascular-derived axonal guidance cues for developing sympathetic neurons. Nature 452,
759-63.
Markus, A., Zhong, J., and Snider, W. D. (2002). Raf and akt mediate distinct aspects of
sensory axon growth. Neuron 35, 65-76.
McManus EJ, Sakamoto K, Armit LJ, Ronaldson L, Shpiro N, Marquez R, Alessi
DR.(2005). Role that phosphorylation of GSK3 plays in insulin and Wnt signalling
defined by knockin analysis. EMBO J. 24(8):1571-83.
Messersmith, E. K., et al. (1995) Semaphorin III can function as a selective
chemorepellent to pattern sensory projections in the spinal cord. Neuron 14, 949-59.
Mungrue, I. N., and Bredt, D. S. (2004). nNOS at a glance: implications for brain and
brawn. J Cell Sci 117, 2627-2629.
Nishiyama, M., Hoshino, A., Tsai, L., Henley, J. R., Goshima, Y ., Tessier-Lavigne, M.,
Poo, M. M., and Hong, K. (2003). Cyclic AMP/GMP-dependent modulation of Ca2+
channels sets the polarity of nerve growth-cone turning. Nature 423, 990-995.
O'Leary, D. D., and Terashima, T. (1988). Cortical axons branch to multiple subcortical
targets by interstitial axon budding: implications for target recognition and "waiting
periods". Neuron 1, 901-910.
Ozaki, S., and Snider, W. D. (1997). Initial trajectories of sensory axons toward laminar
targets in the developing mouse spinal cord. J Comp Neurol 380, 215-229.

78
Packer MA, Stasiv Y , Benraiss A, Chmielnicki E, Grinberg A, Westphal H, Goldman SA,
Enikolopov G.(2003). Nitric oxide negatively regulates mammalian adult neurogenesis.
Proc Natl Acad Sci U S A. 100(16):9566-71.
Polleux, F., Morrow, T., and Ghosh, A. (2000). Semaphorin 3A is a chemoattractant for
cortical apical dendrites. Nature 404, 567-573.
Portera-Cailliau, C., Weimer, R. M., De Paola, V., Caroni, P., and Svoboda, K. (2005).
Diverse modes of axon elaboration in the developing neocortex. PLoS Biol 3, e272.
Potter, L. R., Abbey-Hosch, S. & Dickey, D. M. (2006) Natriuretic peptides, their
receptors, and cyclic guanosine monophosphate-dependent signaling functions. Endocr
Rev 27, 47-72
Purro, S. A., Ciani, L., Hoyos-Flight, M., Stamatakou, E., Siomou, E., and Salinas, P. C.
(2008). Wnt regulates axon behavior through changes in microtubule growth
directionality: a new role for adenomatous polyposis coli. J Neurosci 28, 8644-8654.
Qian, Y ., Chao, D. S., Santillano, D. R., Cornwell, T. L., Nairn, A. C., Greengard, P.,
Lincoln, T. M., and Bredt, D. S. (1996). cGMP-dependent protein kinase in dorsal root
ganglion: relationship with nitric oxide synthase and nociceptive neurons. J Neurosci 16,
3130-3138.
Ryan, M. C. & Gundlach, A. L. (1998) Ontogenic expression of natriuretic peptide
mRNAs in postnatal rat brain: implications for development? Brain Res Dev Brain Res
105, 251-68.
Salonen, J. T. & Heinonen, O. P. (1984) Mental retardation and mother's hypertension
during pregnancy. J Ment Defic Res 28 (Pt 1), 53-6.
Serulle Y , Zhang S, Ninan I, Puzzo D, McCarthy M, Khatri L, Arancio O, Ziff EB. A.
(2007) GluR1-cGKII interaction regulates AMPA receptor trafficking. Neuron.
56(4):670-88.
Schmidt, H., Stonkute, A., Juttner, R., Schaffer, S., Buttgereit, J., Feil, R., Hofmann, F.,
and Rathjen, F. G. (2007). The receptor guanylyl cyclase Npr2 is essential for sensory
axon bifurcation within the spinal cord. J Cell Biol 179, 331-340.
Schmidt H, Stonkute A, Jüttner R, Koesling D, Friebe A, Rathjen FG. (2009). C-type
natriuretic peptide (CNP) is a bifurcation factor for sensory neurons. Proc Natl Acad Sci
U S A. 29;106(39):16847-52.

79
Schmidt, H., Werner, M., Heppenstall, P. A., Henning, M., More, M. I., Kuhbandner, S.,
Lewin, G. R., Hofmann, F., Feil, R., and Rathjen, F. G. (2002). cGMP-mediated signaling
via cGKIalpha is required for the guidance and connectivity of sensory axons. J Cell Biol
159, 489-498.
Schwarzschild, M. A., and Zigmond, R. E. (1991). Effects of peptides of the
secretin-glucagon family and cyclic nucleotides on tyrosine hydroxylase activity in
sympathetic nerve endings. J Neurochem 56, 400-406.
Seidel, C., and Bicker, G. (2000). Nitric oxide and cGMP influence axonogenesis of
antennal pioneer neurons. Development 127, 4541-4549.
Song, H., and Poo, M. (2001). The cell biology of neuronal navigation. Nat Cell Biol 3,
E81-88.
Song, H. J., and Poo, M. M. (1999). Signal transduction underlying growth cone
guidance by diffusible factors. Curr Opin Neurobiol 9, 355-363.
Starr, J. M. (1999) Blood pressure and cognitive decline in the elderly. Curr Opin Nephrol
Hypertens 8, 347-51.
Tamura, N., et al. (2004) Critical roles of the guanylyl cyclase B receptor in endochondral
ossification and development of female reproductive organs. Proc Natl Acad Sci U S A
101, 17300-5.
Thippeswamy, T., and Morris, R. (1997).Cyclic guanosine 3',5'-monophosphate-mediated
neuroprotection by nitric oxide in dissociated cultures of rat dorsal root ganglion
neurones. Brain Res 774, 116-122.
Tsuji, T. & Kunieda, T. (2005) A loss-of-function mutation in natriuretic peptide receptor
2 (Npr2) gene is responsible for disproportionate dwarfism in cn/cn mouse. J Biol Chem
280, 14288-92.
Wang, K. H., Brose, K., Arnott, D., Kidd, T., Goodman, C. S., Henzel, W., and
Tessier-Lavigne, M. (1999). Biochemical purification of a mammalian slit protein as a
positive regulator of sensory axon elongation and branching. Cell 96, 771-784.
Wedel, B., and Garbers, D. (2001). The guanylyl cyclase family at Y2K. Annu Rev
Physiol 63, 215-233.
Wegener, J. W., Nawrath, H., Wolfsgruber, W., Kuhbandner, S., Werner, C., Hofmann, F.,
and Feil, R. (2002). cGMP-dependent protein kinase I mediates the negative inotropic
effect of cGMP in the murine myocardium. Circ Res 90, 18-20.

80
Wickramasinghe SR, Alvania RS, Ramanan N, Wood JN, Mandai K, Ginty DD.(2008)
Serum response factor mediates NGF-dependent target innervation by embryonic DRG
sensory neurons. Neuron. 58(4):532-45
Xiong, G., Mojsilovic-Petrovic, J., Perez, C. A., and Kalb, R. G. (2007). Embryonic motor
neuron dendrite growth is stunted by inhibition of nitric oxide-dependent activation of
soluble guanylyl cyclase and protein kinase G. Eur J Neurosci 25, 1987-1997.
Yaron, A., Huang, P. H., Cheng, H. J. & Tessier-Lavigne, M. (2005) Differential
requirement for Plexin-A3 and -A4 in mediating responses of sensory and sympathetic
neurons to distinct class 3 Semaphorins. Neuron 45, 513-23.
Yoshimura, T., Kawano, Y ., Arimura, N., Kawabata, S., Kikuchi, A., and Kaibuchi, K.
(2005). GSK-3beta regulates phosphorylation of CRMP-2 and neuronal polarity. Cell 120,
137-149.
Zhao, X., Zhuang, S., Chen, Y ., Boss, G. R., and Pilz, R. B. (2005). Cyclic GMP
dependent protein kinase regulates CCAAT enhancer-binding protein beta functions
through inhibition of glycogen synthase kinase-3. J Biol Chem 280, 32683-32692.
Zhao Z, Ma L. (2009). Regulation of axonal development by natriuretic peptide
hormones. Proc Natl Acad Sci U S A. in print
Zhao, Z., et al. (2009) Regulate axon branching by the cyclic GMP pathway via inhibition
of glycogen synthase kinase 3 in dorsal root ganglion sensory neurons. J Neurosci 29,
1350-60.
Zheng, J. Q., Felder, M., Connor, J. A. & Poo, M. M. (1994) Turning of nerve growth
cones induced by neurotransmitters. Nature 368, 140-4.
Zhong, J., Li, X., McNamee, C., Chen, A. P., Baccarini, M., and Snider, W. D. (2007).
Raf kinase signaling functions in sensory neuron differentiation and axon growth in vivo.
Nat Neurosci 10, 598-607.
Zhou, F. Q., Zhou, J., Dedhar, S., Wu, Y . H., and Snider, W. D. (2004). NGF-induced
axon growth is mediated by localized inactivation of GSK-3beta and functions of the
microtubule plus end binding protein APC. Neuron 42, 897-912.

81
Appendix:
Materials and Methods
Chemicals, Animals, and DNA Constructs
8-Br-cGMP and 8-Br-cAMP, purified ANP, BNP, and CNP were obtained from Sigma
and KT5823, YC-1, Wortmannin, LY294002, U0126 were obtained from Axxora (San
Diego). All animal works were done according to the protocols approved by the
Institutional Animal Care and Use Committees at the University of Southern California
following the NIH regulations. Rat embryos were collected from pregnant Sprague
Dawley females (purchased from Charles River Laboratories) with the plug day
designated as E0. PrkG1, nNOS and GSK3 double Knock-in mutant and Thy-1 YFP-H
reporter mouse embryos in a mixed CD-1/C57B6 background were generated by crossing
heterozygous animals with the plug day as E0.5. Genotyping was done by PCR as
previously described (Wegener et al., 2002, Packer et al., 2003, McManus et al., 20005,
Feng et al., 2000). The lbab mouse was obtained from the Jackson Laboratory, and
mutant animals were generated by crossing heterozygous animals with the plug day as
E0.5. The genotype was determined by AvaII digestion of the PCR product using primers
5’-CTCTTGGGTGCAGAGCTAGG-3’ and 5’-AGCTGGTGGCAATCAGAAAA-3’
(Jiao et al., 2007). The mouse PrkG1α full length clone was obtained from RIKEN
(DB0073L22), and then cloned into the pCAGGS expression vector by PCR with the
FLAG sequence incorporated into the primers. The full length HA-tagged GSK3β (gift

82
from Xi He) was cloned into the pCAGGS vector by restriction digestions. All single
amino acid mutations were generated using QuickChange Mutagenesis kit (Stratagene)
and truncation mutants were generated by PCR. For bacterial expression, His-tagged
GSK3β was generated by subcloning the full length HA-tagged GSK3β to carboxyl
terminal of the 6xHis sequence in the pQE30 vector (Qiagen).
In vitro Primary DRG Neuron Culture
All cultures were done in an F12 medium (Gibco BRL) with the N3 supplement, 40 mM
Glucose, 0.5% fetal calf serum (FCS) plus NGF (25 ng/ml, 7s, Sigma). For the in vitro
branching assay, rat DRG neurons from E14 embryos were dissociated and cultured at
8000 cells per 20 μl collagen gel (Wang et al., 1999). After one day in culture, they were
treated with 8-Br-cGMP and other pharmacological reagents. Cells were fixed after
another 24 hrs and stained for visualization. For PrkG1 mutant mouse neurons, they were
collected and pooled together from E12.5 embryos with the same genotype and then
dissociated and cultured as described above.
For expression of various constructs of PrkG1α or GSK3β in rat E14 DRG neurons,
~1x10
6
dissociated cells were electroporated with 2-3 μg plasmid DNA using the
nucleofection reagent for rat DRG (Amaxa). After electroporation, cells were diluted with
400 μl growth medium, incubated at 37°C for 10 minutes, and then plated together with
EGFP-transfected cells (1:1 or 1:2 ratio) at 3-5x10
4
in 20 μl collagen gels.

83
Immunohistochemistry and Analysis of Axon Length and Branching Points
Cultured neurons were fixed with 4% paraformaldehyde in PBS (pH 7.4) for 1 hour, and
permeabilized and blocked with PBS containing 0.1% triton X-100 and 1% Goat Serum.
For untransfected culture, neurons were stained with a monoclonal antibody against
neurofilament (RMO270, 1:1000, gift from Dr. Virginia Lee), and HRP conjugated
secondary antibody (Jackson ImmunoResearch, 1:1000), then developed with
diaminobenzadine (Sigma). Images were taken on a dissection scope with AxioCamHRc
(Zeiss). For TUNEL staining, cells were fixed as above and stained following the
manufacturer’s protocol (Roche). The transfected cells were stained with mouse
anti-FLAG (M2, 1:2000, Sigma) or rabbit anti-HA (1:500, Santa Cruz), and then Alexa-
(Molecular Probes) or Cy3- (Jackson ImmunoResearch) conjugated secondary antibodies
to detected PrkG1α or GSK3β expression. Images were taken on an Axiovert 200
inverted microscope with AxioCamMRm (Zeiss).
Images were analyzed using the AxioVision software (Zeiss) or ImageJ (NIH) for the
following parameters: number of neurons with neurites; neuritic length, number of
branching points. Only neurons with axons longer than 50 μm were counted. Neurites
longer than 20 μm (about the diameter of a soma) were traced and counted as branches.
Each condition was repeated at least twice, and in a single experiment of each condition,
more than 50 neurons were analyzed. Statistical significance was determined with
independent t-test and one way ANOV A. For distribution, the data were presented as

84
mean ± STD, and for the rest of the analysis, the data were presented as mean ± SEM.
Outgrowth Assay
DRGs were cut into small pieces and cultured in collagen gels in the presence or absence
of NGF (5 ng/ml), as well as with different doses of NPs or 8-Br-cGMP. They were fixed
and stained using neurofilament antibodies and Cy3 or HRP-labeled secondary antibodies.
To quantify the size of axonal halos in the presence of NGF, the radius of the entire halo
was measured at every 45
o
and then subtracted the radius of the explants. To quantify
outgrowth in the absence of NGF, explants with axons longer than 50 μm were classified
into two groups: low for those making 1-5 axons; and high for those growing more than
five axons or axonal halos. All experiments were repeated twice and 5-6 explants from
each condition were analyzed.
Growth Cone Turning Assay
The turning assay was performed following the method described previously (19, 27)
with modifications. Dissociated DRG neurons were grown on PDL/Laminin coated glass
coverslips. After cultured overnight, they were transferred to a heat-controlled (37
o
C)
open chamber on an Axiovert microscope (Zeiss). Microscopic CNP gradients were
produced from a capillary glass pipette (1-μm opening) loaded with a 10 μM CNP
solution by repetitive application (2 Hz and 20 ms) of a positive pressure of 3 psi (19).
The pipette tip was positioned 50 μm away from the growth cone and ~45° from the

85
initial direction of extension (defined by the 50 μm segment from the growth cone tip).
Images were captured by an Axiocam (Zeiss) and analyzed. The turning angle was
defined by the angle between the initial direction of neurite extension and a straight line
connecting the positions of the growth cone at the onset and the end of the 30-minute
period. Only growth cones with a net extension >10 μm were included in the analysis.
In Situ Hybridization
Mouse embryos from different stages were cut on a cryostat (Microm HM560).
Sections (16 μm) were processed for in situ analysis following a standard procedure
using digoxigenin-labeled RNA probes. PrkG1α (1-267 bp), PrkG1β (1-322bp), and
PrkG2 (1062-1740bp) were used as probes for PrkGs, the entire coding region was used
for labeling mouse Nppa, Nppb, and Nppc. Probes for Npr1 (1891-2769bp), Npr2
(493-1342bp), and Npr3 (707-1305bp) were adapted from the Allen Brain Atlas (22).
Images were obtained on a dissection microscope with AxioCamHMc (Zeiss).
Analysis of Sensory Axonal Projections in Mouse Embryos by DiI Labeling
To label DRG axons inside the spinal cord, E13.5 mouse embryos were fixed and
implanted with large DiI crystals or injected with small amount of DiI by iontophoresis as
previously described (Ma and Tessier-Lavigne, 2007). The dye was allowed to diffuse
at 25
o
C overnight before visualization on an AxioImager microscope with AxioCamMRm
(Zeiss). The labeled axons were visualized from the lateral side of the dorsal spinal cord

86
in an open book preparation.
Analysis of Sensory Axonal Projections in Adult Mouse with Florescence Label
For analyzing adult sensory axon projections, mutant mice with heterozygous genetic
background were crossbreed with Thy-1 YFP-H reporter mouse (Feng et al., 2000), from
which adult homozygous mutant mice carrying Thy-1-YFP were generated by
heterozygous inbreedings. Spinal cords were collected from P21 mice and fixed in 4%
paraformaldehyde. The labeled axons were visualized from dorsal side of the spinal cord
under a fluorescence microscope with AxioCamMRm (Zeiss).
Analysis of Protein Phosphorylation by Western Blots
Proteins were separated by SDS-PAGE and transferred onto nitrocellulose membranes,
which were probed with the following antibodies: rabbit anti-HA (Santa Cruz); rabbit
anti-phospho-GSK3 (Ser9), rabbit anti-phospho-Akt (Ser473), rabbit anti-phospho-ERK
(Thr202) (Cell Signaling); and mouse anti-tubulin (DM1α). Alexa680- (Molecular probes)
and IRDye800-labeled secondary antibodies were used for detection on an infrared
imaging system (Odyssey, LI-COR). Experiments were done multiple times and
quantification was done for the blots represented in the figures.
GSK3 Phosphorylation Assay in Neurons or COS Cells
Neuronal cell lysates were prepared in ice-cold RIPA buffer with protease inhibitors

87
cocktail and phosphotase inhibitors from E14 rat DRG neurons (4x10
5
cell/well) that
were cultured overnight, starved for 3 hours without NGF and stimulated for different
times with 50 μM 8-Br-cGMP or NGF. COS cells transfected with different combinations
of GSK3β and PrkG1α were grown in serum-free medium for one day before lysed in the
above buffer. Proteins from the lysates were then subjected to western blotting
described above.
Co-immunoprecipitation
COS cells were cultured in 6-well plates and transfected by FuGene 6 (Roche) with
different combinations of HA-tagged GSK3β, Flag-tagged PrkG1α, Flag-tagged
PrkG1α-S64A and EGFP construct. Two days after transfection, cells from each well
were collected by scraping, pelleted by centrifugation, and then lysed in 100 μl buffer
containing 50 mM Tris-HCl (pH7.5), 50 mM NaCl,1 mM EDTA, 1.25% Triton X-100
and a protease inhibitor cocktail (Roche). Each 50 μl lysate was diluted to a final
volume of 1 ml with PBS, incubated overnight with anti-HA (Santa Cruz) or anti-Flag
antibody (Sigma) at 4
o
C, then precipitated with pre-cleaned protein A Sapharose beads
(GE Healthcare Life Sciences). The beads were washed 3 times with PBS before elution
with 50 μl gel loading buffer for western blot.
In vitro Phosphorylation Assays
In vitro phosphorylation of GSK3β using COS cell lysates or the purified kinase was

88
performed in a total volume of 40 μl kinase assay buffer containing 20 mM Tris-HCl (pH
7.5), 20 mM magnesium acetate, 100 μM ATP, 2 mM dithiothreitol, with or without 20
μM 8-Br-cGMP. The reaction was incubated at 30°C for different time points, and
stopped by heating at 95°C for 5 min. Protein samples were separated by SDS-PAGE and
blotted as described above. His-tagged GSK3β was expressed in E.coli (strain M15) and
purified from the soluble lysate using a Hi-Trap Ni column (Pharmacia) under the native
condition. Purified GSK3β proteins were dialyzed against a buffer containing 20 mM
Tris-HCl (pH7.5) and 20 mM magnesium acetate and 200 ng was used in the kinase assay.
The PrkG1α-S64A lysate was prepared from transfected COS cells (100 μl per 10-cm
culture dish) and 5 μl of the lysate was used in the assay. Purified PrkG1α protein from
bovine lung was purchased from Calbiochem and ~400ng/1500 units of proteins were
used per reaction.
Florescence Based Whole Mount Staining in E11.5 Embryos
Florescence based whole mount staining with an antibody against neurofilament was used
to examine the axon projection in embryos of early developmental stages
(Wickramasinghe et al., 2007) . In brief, E11.5 mouse embryos were collected, fixed in
4% paraformaldehyde, followed by dehydration in a methanol series of 10%, 30%, 50%
in PBS and 80% in H
2
O; they were further bleached in an H
2
O
2
-DMSO-methanol (1: 2: 8)
mixture overnight at 4
o
C, and re-fix in DMSO-methanol (1:4), followed by re-hydration
in a methanol series of and 80% in H
2
O, and 50%, 30%, 10% in PBS; primary antibodies

89
(2H3, the Developmental Studies Hybridoma Bank) were diluted in PBS containing 20%
DMSO and 5% normal goat serum, and incubated for 3 days at room temperature,
followed by 6 washes at one hour each in PBS with 20% DMSO; Cy3- (Jackson
ImmunoResearch) conjugated secondary antibodies were incubated overnight in the same
block solution at room temperature; embryos were further cleared with BA/BB (1:2
benzyl alcohol:benzyl benzoate) after excessive washes and visualized under a
fluorescence microscope with AxioCamMRm (Zeiss).

Abstract (if available)

Abstract The involvement of cyclic guanosine-3’, 5’-monophosphate (cGMP) signaling in axonal development has been proposed. However, the underlying molecular and cellular mechanism is not well understood yet. In this study, we used rodent embryonic sensory neurons from the dorsal root ganglion (DRG) as a model system, provided pharmacological, genetic and biochemical evidences, demonstrated the role of cGMP signaling in axon branching, further identified GSK3 as the downstream kinase, and CNP as upstream activator, thus established a functional signaling pathway mediating this regulation.

Linked assets