University of Southern California Dissertations and Theses

Signaling networks in complex brain disorders

(USC Thesis Other)

Signaling networks in complex brain disorders

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content

SIGNALING NETWORKS IN COMPLEX BRAIN DISORDERS

by

Brent Wilkinson

______________________________________________________________________________

A Dissertation Presented to the
FACULTY OF THE USC GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the Requirements for the Degree
DOCTOR OF PHILOSOPHY
(DEVELOPMENT, STEM CELLS, AND REGENERATIVE MEDICINE)

December 2019
ii

Acknowledgements
First, I would like to thank my mentor, Marcelo Coba, for all the priceless opportunities
and mentorship that I have received while working under his guidance. He has provided guidance
from “A to Z” throughout my graduate career that has fostered growth and learning experiences
that are unapparelled to anything that I would have thought possible. Marcelo has consistently
encouraged me to explore my own ideas while providing just the right amount of support for me
to work independently and learn from my mistakes. I am truly grateful for the foundation he has
provided me with that will be essential through my career as an investigator.
I would like to thank my family for their unwavering support. My wife, Cristina, has been
with me every step of my graduate career. She has been there whenever I have needed advice on
navigating my way through my degree and has always encouraged me to be the best that I can be.
I thank her for always being there, even through the early mornings and late nights that are often
required in science. We have grown with each other over the duration of my PhD and I cannot
wait to see what the future has in store for us. I would like to thank my parents for always being
supportive of the decisions that I have made, both in life and education. This has allowed me to
find my own path and pursue something that I am truly passionate about.
I would like to thank my committee members Justin Ichida and Giorgia Quadrato. Justin
has been a huge influence on how I approach science as his passion for the work we do is
infectious. Without his generosity in both time and resources, much of this would have not been
possible. I thank Giorgia Quadrato for being willing to join my committee on such a short notice
and providing her excellent mentorship. Finally, I would like to thank the countless individuals
that have provided support in learning in new techniques and the other members of the Coba lab.
In particular, Jiazhen Xu, who has been instrumental in driving new and exciting projects forward.
iii

Table of Contents

Acknowledgements ....................................................................................................................................... ii
List of Tables ................................................................................................................................................ v
List of Figures .............................................................................................................................................. vi
Abstract ......................................................................................................................................................... 1
Chapter 1: Introduction ................................................................................................................................. 3
1.1 Complex Brain Disorders .................................................................................................................... 3
1.2 The Use of iPSCs to Model Neural Development and Synaptic Function ......................................... 4
1.3 Synaptic Signaling Mechanisms ......................................................................................................... 6
1.4 Methods Used in the Identification of Protein-Protein Interactions ................................................... 8
Chapter 2: The Developmental Tnik Interactome Informs Tnik Function ................................................. 11
2.1 Introduction ....................................................................................................................................... 11
2.2 Materials and Methods ...................................................................................................................... 13
2.3 Results ............................................................................................................................................... 23
2.3.1 Tnik Developmental Interactome Indicates Differential Functions of Tnik .............................. 23
2.3.2 TNIK hNPC Interactome Indicates Core Functions of TNIK in Early Neural Development ... 30
2.3.3 TNIK Modulates hNPC Proliferation and Wnt Signaling ......................................................... 31
2.4 Discussion ......................................................................................................................................... 38
Chapter 3: Endogenous Cell Type-Specific Disrupted in Schizophrenia 1 Interactomes Reveal Protein
Networks Associated with Neurodevelopmental Disorders ................................................................. 41
3.1 Introduction ....................................................................................................................................... 41
3.2 Materials and Methods ...................................................................................................................... 43
3.3 Results ............................................................................................................................................... 56
3.3.1 Generation of DISC1 FLAG TAG Lines ................................................................................... 56
3.3.2 Incorporation of the 3X-FLAG Tag Does Not Alter Expression or Function of DISC1 ........... 57
3.3.3 Identification of Endogenous DISC1 Interactomes ................................................................... 61
3.3.4 The Endogenous DISC1 Interactome and Previously Reported Protein-Protein Interactions ... 72
3.3.5 Clustering of DISC1 Interactors and Psychiatric Disease ......................................................... 75
3.3.6 Mutations in DISC1 Interactors Regulate Shared Cellular Functions ....................................... 80
3.3 Discussion ......................................................................................................................................... 86
Chapter 4: Synaptic GAP and GEF Complexes Cluster Proteins Essential for GTP Signaling ................. 88
4.1 Introduction ....................................................................................................................................... 88
4.2 Materials and Methods ...................................................................................................................... 90
4.3 Results ............................................................................................................................................... 95
iv

4.3.1 Distribution of GAPs and GEFs Within the Postsynaptic Density ............................................ 95
4.3.2 Interactomes of Agap2, Syngap1, and Kalirin at the Postsynaptic Density ............................... 98
4.3.3 Differential Interactions of Agap2 Between PSD and non-PSD Compartments ..................... 106
4.3.4 Interactions of Agap2, Syngap1, and Kalirin with Risk Factors of Psychiatric Disease ......... 110
4.4 Discussion ....................................................................................................................................... 113
Chapter 5: Summary and Future Directions ............................................................................................. 116
References ................................................................................................................................................. 118

v

List of Tables

Table 2.1. Oligonucleotides Used in TNIK Study ...................................................................................... 22
Table 2.2. Tnik Interactions Identified at P7 and P14 Developmental Stages ............................................ 26
Table 2.3. Interactions Identified Using Recombinant Tnik Kinase Domain ............................................. 27
Table 2.4. TNIK, PDE4DIP, and AKAP9 PPIs in hNPCs .......................................................................... 32
Table 3.1. Oligonucleotides Used in DISC1 Study..................................................................................... 54
Table 3.2. Antibodies Used in DISC1 Study .............................................................................................. 55
Table 3.3. DISC1 gRNA Predicted Off-Target Sequences and Sequencing .............................................. 60
Table 3.4. DISC1 Interactors Identified via HPLC-MS/MS in hNPCs ...................................................... 67
Table 3.5. Gene Ontology Enrichment Within the DISC1 hNPC Interactome........................................... 69
Table 3.6. DISC1 Interactors Identified via HPLC-MS/MS in Astrocytes ................................................. 70
Table 3.7. Gene Ontology Enrichment Within the DISC1 Astrocyte Interactome ..................................... 71
Table 3.8. Endogenous DISC1 Peptides Identified via HPLC-MS/MS ...................................................... 73
Table 3.9. DISC1 Interactors Identified via HPLC-MS/MS Using Recombinant DISC1 in Adult Mouse
Cortex ................................................................................................................................................... 76
Table 3.10. Previously Reported Interactions Detected in DISC1 Interactomes ........................................ 77
Table 3.11. Protein Family Members of Previously Reported DISC1 Interactions Detected in DISC1
Interactomes ......................................................................................................................................... 78
Table 3.12. Clustering of Endogenous DISC1 Interactomes ...................................................................... 82
Table 4.1. Domain Architecture of GAPs and GEFs at the PSD ................................................................ 97
Table 4.2. GAP/GEF Protein Interactions and Functional Annotation ..................................................... 100
Table 4.3. SMART Domain Enrichment Within PSD Complexes ........................................................... 105

vi

List of Figures

Figure 2.1. Characterization of Tnik Protein-Protein Interactions Throughout Mouse Cortical Development
.............................................................................................................................................................. 24
Figure 2.2. Characterization of Protein-Protein Interactions Involving TNIK and TNIK-Associated Proteins
in hNPCs............................................................................................................................................... 31
Figure 2.3. Generation of TNIK K54R and R180X Cell Lines and Characterization of Proliferation Defects
.............................................................................................................................................................. 34
Figure 2.4. Wnt Signaling is increased in hNPCs Lacking TNIK or TNIK kinase activity ...................... 36
Figure 2.5. Wnt Signaling Changes are Cell Type Specific ...................................................................... 37
Figure 3.1. Generation of DISC1
FLAG
human neural progenitor cell lines. ................................................. 58
Figure 3.2. DISC1 Isoforms and Characterization of WT and DISC1
FLAG
iPSC and Differentiated Lines 59
Figure 3.3. Insertion of 3X-FLAG at the C-terminus of DISC1 does not alter expression or function of
DISC1 ................................................................................................................................................... 62
Figure 3.4. qPCR and Upregulation of Endogenous DISC1 ....................................................................... 63
Figure 3.5. Identification of the endogenous DISC1 Interactome in hNPCs and astrocytes ...................... 66
Figure 3.6. Expression of DISC1 Binding Partners in hNPCs and Astrocytes ........................................... 74
Figure 3.7. Recombinant DISC1 Interactions Identified in Adult Mouse Cortex ....................................... 75
Figure 3.8. Clustering and Enrichment of DISC1 Interactome in Psychiatric Disease Risk Factors ......... 81
Figure 3.9. TNIK KO hNPCs Have Decreased Proliferation ..................................................................... 85
Figure 4.1. Distribution of GAPs and GEFs at the PSD ............................................................................. 96
Figure 4.2. Interactomes of the PSD GAP/GEF proteins, Agap2, Syngap1, and Kalirin ........................... 99
Figure 4.3. Distribution of Signaling Molecules within PSD Networks ................................................... 104
Figure 4.4. Differential Interactions and Functions of PSD vs Non-PSD Agap2 ..................................... 108
Figure 4.5. Distribution of Risk Factors for Psychiatric Disease within Protein-Protein Interaction Networks
............................................................................................................................................................ 112

1

Abstract
Thousands of mutations have been identified that may potentially contribute to the genetic
architecture of complex brain disorders such as intellectual disability (ID), autism spectrum
disorder (ASD), developmental delay (DD), and schizophrenia (SCZ). While the affected proteins
have been shown to be enriched in both early neural development and within the postsynaptic
density (PSD) of glutamatergic excitatory neurons, how these risk factors associate in protein-
protein interactions (PPIs) to regulate signaling networks is less clear. Within signaling networks,
proteins with a large number of PPIs (protein hubs) are likely to have a significant impact on the
dysregulation of signaling mechanisms when they harbor damaging mutations and therefore, be
associated with disease. ID causative mutations have previously been identified in the Traf2 and
Nck Interacting Kinase (TNIK) and Tnik has been shown to act as a protein hub in the regulation
of signaling networks. Importantly, these networks are enriched in risk factors for complex brain
disorders including disrupted in schizophrenia 1 (Disc1) and synaptic Ras GTPase activating
protein 1 (Syngap1). Here, PPIs of Tnik and Tnik-interacting proteins were characterized during
specific developmental stages, within precise cellular and subcellular fractions, and under
endogenous conditions relevant to complex brain disorders. We first determined the
spatiotemporal-dependent functions of Tnik by characterizing Tnik interactomes throughout
mouse cortical development and in human pluripotent stem cell (hPSC)-derived neural progenitor
cells (hNPCs). Multiple mutant TNIK hPSC lines were generated to further investigate the
consequences of TNIK dysregulation during early neural development. Second, we determined
cell-type specific PPIs of the TNIK interactor, DISC1, in hNPCs and astrocytes. For this purpose,
we generated a hPSC line containing an endogenous FLAG tag for DISC1 affinity purification.
Lastly, the interactomes of the Tnik interacting proteins, Syngap1, Agap2, and Kalirin, were

2

determined to characterize their involvement in specific G-protein signaling mechanisms within
the PSD. These studies revealed a number of novel interactions that inform cell-type specific
functions of proteins that modulate neural development or neuronal activity and highlight signaling
networks associated with complex brain disorders.

3

Chapter 1: Introduction
1.1 - Complex Brain Disorders
Complex brain disorders are heterogenous neurological disorders that are influenced
through a variety of genetic and environmental factors (1). These include neurodevelopmental
and psychiatric disorders such as autism spectrum disorder (ASD), schizophrenia (SCZ),
intellectual disability (ID), and developmental delay (DD). ASD, ID, and DD may present
themselves from birth up to early childhood while SCZ is frequently diagnosed during young
adulthood (2, 3). These disorders heavily influence the quality of life for affected individuals and
present a heavy financial burden on their families and the medical system as a whole (4). While
scientific progress continues to uncover the genetic underpinnings of complex brain disorders, the
process of therapeutic target discovery lags behind significantly and patients affected by these
disorders have limited treatment options (5).
The genetic contributions of complex brain disorders are polygenic and may be present in
several different forms. These can include single nucleotide polymorphisms (SNPs) identified
through Genome Wide Association Studies (GWAS), inherited or de novo single nucleotide
variants (SNVs), and copy number variations (CNVs). SNPs arise from common variation and
are believed to have less of an effect size when compared to SNVs which are rare and depending
on the class of the mutation, may change the function of a particular protein or completely abolish
its expression. CNVs are genomic rearrangements which can delete or insert large segments of
genetic information (6). Substantial progress has been made in identifying the genetic
underpinnings of complex brain disorders with respect to GWAS (7-11), de novo SNVs (12-22),
and CNVs studies (23-27). Together, these studies have identified thousands of genetic risk factors
that may contribute with varying levels of penetrance.

4

Since several thousand genetic risk factors have been implicated in contributing to complex
brain disorders, it is increasingly important to identify when and where these risk factors may be
expressed and perform functions that affect the disease etiology. Transcriptomic studies mapping
the spatio-temporal expression of genes across human development have found that genes
implicated in contributing to SCZ are enriched in cortical excitatory neurons and fetal neural
progenitor cells. In the same study, risk factors for ASD, ID, and DD were found to be enriched
in both fetal and adult excitatory neurons (28). In addition to mapping spatio-temporal expression,
a common theme amongst studies identifying genetic risk factors is statistically testing the
enrichment of risk factors in particular gene sets that may be relevant to the disease. Common
lists include targets of the fragile X mental retardation 1 (FMR1) protein, protein complexes (such
as NMDA or ARC), and the proteins found to be at the post-synaptic density (PSD) (10, 12, 18).
Although these studies focus on a small number of datasets, they have consistently found
enrichment at the PSD along with several proteins involved in early neural development.
1.2 - The Use of iPSCs to Model Neural Development and Synaptic Function
Functional studies of proteins involved in neural development and the modulation of
synaptic activity have largely been carried out in rodent models. Aside from animal models, the
ability to obtain patient-specific tissue to study the effects of complex brain disorders has
previously been limited to post-mortem brain tissue. This has posed several concerns regarding
the quality and quantity of this tissue source. Furthermore, it represents the end-stage of disease
and may be influenced by therapeutic intervention throughout the lifetime of the individual from
which the tissue is derived from. Studying complex brain disorders in-vitro is further confounded
by the fact that post-mitotic neurons cannot be expanded in culture and common cell lines are not
representative of the primary cell-types affected (29). Human pluripotent stem cells (hPSCs),

5

including embryonic stem cells (hESCs) and induced pluripotent stem cells (iPSCs), offer the
opportunity to study disease-relevant proteins and neurodevelopmental pathways in a cell-type
and/or patient-specific manner. hESCs are derived from human blastocysts while iPSCs are
generated in patient-specific manner through the reprogramming of somatic cells. While both
have the potential to form virtually any cell type in the human body, iPSCs retain the genetic
background of the patient from which they are generated, including all disease-relevant mutations
(30, 31). In addition, the advent of CRISPR/Cas9 genome engineering has enabled the precise
insertion or removal of disease-associated mutations within hPSCs, allowing the study of their
effects within an isogenic background (32, 33).
hPSCs are able to undergo differentiation to specific cell types through the exposure to
small molecules that modulate particular signaling pathways and mimic developmental cues. For
example, by exposing hPSCs to inhibitors of SMAD signaling, a method known as dual-SMAD
inhibition, hPSCs can be directed towards a neuroepithelium/neural progenitor fate to produce
hPSC-derived neural progenitor cells (hNPCs) (34). This provides a platform for the investigation
of the multiple protein-protein interactions (PPIs), signaling pathways, and transcriptional
programs that converge to modulate proliferation and differentiation of hNPCs during early human
neural development. hNPCs have now been used to study numerous disease-related mutations and
signaling pathways (35, 36). For example, iPSCs engineered to contain clinically relevant
mutations in disrupted in schizophrenia 1 (DISC1) were shown to have increased levels of Wnt
signaling and defects in hNPC differentiation (37) which may be due to the previously
characterized PPI between DISC1 and glycogen synthase kinase 3 beta (GSK3B) (38). In addition,
iPSCs derived from ASD patients with damaging mutations in genes that regulate canonical Wnt
signaling, were shown to have decreased levels of Wnt signaling, increased levels of hNPC

6

proliferation, and defects in synaptogenesis upon differentiation to neurons (39). Proliferation and
differentiation of hNPCs are key processes that work together during brain development and the
cellular networks regulating them may constitute a targetable hub for risk factors associated to
neurodevelopmental disorders.
hNPCs can be further exposed to wide variety of signaling cues to instruct them to
differentiate into numerous different types of neurons, including glutamatergic excitatory neurons
(40), GABAergic inhibitory neurons (41), dopaminergic neurons (42), and serotonergic neurons
(43). In addition, through the over-expression of particular transcription factors, hPSCs can be
directly converted to neuronal cell types such as glutamatergic excitatory neurons through the
overexpression of the transcription factor NGN2 (44) and GABAergic neurons through the
overexpression of ASCL1 and DLX2 (45). The direct conversion of hPSCs to neurons has allowed
researchers to rapidly obtain a relatively pure population of cell-type specific neurons. However,
the neural and neuronal cell-types generated from hPSCs are fetal and do not reach maturation
levels compared to cultured primary neurons derived from rodents. Due to this, the investigation
of proteins involved in the modulation of synaptic function in hPSC-derived neurons are currently
limited to the early stages of neurogenesis and synaptogenesis (46).
1.3 – Synaptic Signaling Mechanisms
Synaptic connections can be defined as having a presynaptic terminal which releases
neurotransmitters across the synaptic cleft and then are received by receptors located on the
postsynaptic membrane. The postsynaptic density (PSD) is protein-rich specialization located
along the postsynaptic membrane of excitatory neurons. It was first identified as an electro-dense
region through electron microscopy (47) and its protein composition has been further characterized
through mass spectrometry in a number of studies, identifying 1500 – 2000 proteins (48-50). The

7

PSD is composed of several different classes of proteins including scaffolding molecules, kinases,
phosphatases, cytoskeletal proteins, GAPs, GEFs, and neurotransmitter receptors among others
that work in concert with one another to help relay signaling information from the postsynaptic
site into the cell body of the neuron.
PSD scaffolding proteins are specialized in protein-protein interactions due to their protein
domain composition and are key players in PSD protein organization (51). The most abundant
scaffolding protein families of the PSD can be categorized into three main layers: a top layer
represented by the discs large (DLG) family of proteins which interacts with glutamate and other
neurotransmitter receptors at the PSD membrane; a bottom layer composed of the SH3 and
multiple Ankyrin repeat domains (SHANK) protein family, which connects the scaffolding
machinery to the cytoskeleton; and a middle layer composed of the DLG-associated proteins
(DLGAPs) which connects the top and bottom layers of synaptic scaffolds (52, 53). While the
core-scaffolding molecules of the PSD provide the general architecture to assemble PPIs at the
PSD, this can be disrupted through mutations in risk factors for complex brain disorders. For
example, in mice lacking the scaffolding protein, Dlgap1, there is a decreased association between
proteins representing the upper (Dlg4) and lower layers (Shank3) of the PSD (54). Changes in the
association of PSD scaffolding components have also been observed in mice lacking the ID-
associated Traf2 and Nck interacting kinase (TNIK) (53). Here, decreased association was found
between the middle layers (Dlgap1) and the lower layers (Shank3) of the PSD along with decreased
association of Dlgap1 with the AMPA receptor, Grin2a (53). These structural changes result in a
remodeling of PSD PPIs and may ultimately result in deficits of signaling mechanisms following
synaptic transmission.

8

The ability to strengthen or weaken synaptic transmission in response to stimuli is defined
as synaptic plasticity and is believed to represent the molecular basis of learning and memory (55,
56). Several signaling mechanisms occur at the PSD in response to synaptic activity including the
re-arrangement of PPIs, translocation of synaptic receptors, along the induction of post-
translational modifications such as phosphorylation. Specific examples of these mechanisms can
be observed during the induction of a particular form of synaptic plasticity called long-term
potentiation (LTP). The most well characterized form of LTP is dependent on NMDA and AMPA
receptors and occurs along the Schaffer collateral pathway between CA3 and CA1 pyramidal
neurons of the hippocampus (57). During this process, numerous proteins become phosphorylated
or de-phosphorylated and those proteins that undergo changes in phosphorylation account for the
enrichment of the PSD in proteins implicated in contributing to ASD and SCZ (52). In addition,
the core scaffolding proteins increase their association with receptors at the membrane, while
synaptic Ras GTPase activating protein (Syngap1) moves from the upper to the bottom layers of
scaffolding proteins to enable downstream signaling and LTP (52, 58). Syngap1 is one of the most
abundant proteins at the PSD (59). Syngap1, along with several other proteins that comprise the
PSD core scaffolding machinery such as Dlg4, Shank3, and Dlgap1 play critical roles in the
processes of learning and memory and have all been found to contribute to neurodevelopmental
disorders when damaging mutations are present in them (19, 60-63).
1.4 - Methods Used in the Identification of Protein-Protein Interactions
Although multiple risk factors for complex brain disorders are known to play active roles
in neural development or the modulation of synaptic mechanisms at the PSD, what is less known
is if they interact with one another and how mutations in one protein may affect its protein-binding
partners function and/or localization. PPIs are essential to a number of molecular functions and are

9

dynamic as they can be specific to particular cellular compartments, developmental stages, and
activity conditions. PPIs can be broadly categorized as being either scaffolding, enabling the
presentation of a protein at the correct place and time, or enzymatic, whereby one protein modifies
a binding partner to confer a change of function. Identifying protein-protein interactions within a
spatio-temporal framework can provide insight into the signaling networks where individual
proteins regulate cellular function (64-66). This is in line with protein complexes being modular
units in which interacting proteins act together in order to carry out common cellular processes
(67, 68). Therefore, determining protein interaction signaling networks may aid the development
of novel therapeutics to modify individual signaling hubs and regulate cell function as a whole
(69).
There are several methods used for the identification of PPIs, each with their own set of
advantages and disadvantages. The identification of direct, binary interactions between two
proteins can be determined through the use of yeast-two hybrid assays whereby a bait and prey
protein are co-expressed in host organism and then initiate transcription of a reporter gene when
a physical interaction occurs between them (70). Another method involves the over-expression of
two tagged proteins within cultured mammalian cells followed by immunoprecipitation of one of
the tagged proteins and subsequent identification of the second through immunoblotting. However,
these methods involve co-expressing proteins that may never naturally exist within the same
cellular compartment and inducing their expression at levels much higher than would normally
occur. Therefore, these assays indicate the capacity of proteins to interact with one another as
opposed to whether they are interactors under endogenous conditions. In contrast, through the use
immunoprecipitation followed by tandem mass spectrometry (IP-MS/MS), one can identify PPIs
for a target protein under natural stoichiometric expression levels between binding partners and

10

within a particular cell- or tissue-type of interest (71). Although interactors identified using this
method may include secondary and tertiary interactions in addition to direct interactions, this can
also allow the visualization of how proteins function within protein complexes.
A necessary and limiting requirement of immunoprecipitation is the availability of an
antibody highly specific for your target of interest. One way to circumvent this is through the use
an affinity tag such as a FLAG or HA tag that can be affixed to a terminal region of the protein of
interest (72, 73). This enables the use of specific antibodies as the antibody epitope is unique to
the affinity tag rather than the protein of interest. While these affinity tags have traditionally been
put on through molecular cloning in plasmids followed by the over-expression of the protein of
interest, the advent of CRISPR/Cas9 genome editing has allowed the introduction of the affinity
tag coding sequence into the endogenous genomic region encoding for the protein of interest (74).
The use of CRISPR to insert affinity tags can be particularly useful if carried out in pluripotent
stem cells as these can be differentiated into specific cell types and then the protein of interest can
be analyzed within these specific conditions.

11

Chapter 2 – The Developmental Tnik Interactome Informs Tnik Function
2.1 – Introduction
Protein kinases hold integral roles in the transduction of signaling mechanisms throughout
development. The regulation of proteins via phosphorylation acts as a molecular switch that may
result in the phosphorylated protein undergoing changes in status of activation, conformation, or
subcellular localization among other functional consequences (75). TNIK has been implicated in
the modulation of several signaling pathways including Wnt, Jnk, and Tgf-beta signaling in a
variety of cell types (76-78). TNIK belongs to the Germinal Center Kinases (GCK), a subgroup
of the Ste20 family of kinases, and is composed of a ser/thr kinase domain, a citron homology
(CNH) domain of unknown function, and a linker region of low complexity with no globular
structure (79). Increased expression of TNIK results in increased levels of cellular proliferation
and has been associated with numerous types of cancer (80-82), while a complete lack of TNIK
was recently described as being causative for intellectual disability (83).
In non-neuronal cells, knockdown of TNIK has been shown to decrease cell growth and
proliferation and decrease activation of Wnt signaling (76, 77, 84). In neuronal cells, Tnik is
located within both synaptic and non-synaptic sub-cellular compartments (53) and has been shown
to regulate several synaptic functions (85-88). For example, knockdown of Tnik has been shown
to decrease dendritic complexity (85, 86) and result in abnormal spontaneous activity in cultured
neurons (87). In addition, Tnik knockout mice have been shown to have cognitive impairments
and decreased rates of adult neurogenesis (88). While consequences of both knockdown or a lack
of TNIK have been described in various cell types, how a lack of TNIK or the deficiency of TNIK
kinase activity in human neural progenitor cells (hNPCs) is unknown.

12

We have previously identified Tnik interactomes in mouse cortex in both embryonic and
adult developmental stages. These Tnik interactomes were found to have an enrichment in proteins
found to harbor de-novo mutations in patients affected by complex brain disorders (53). Here, the
Tnik developmental interactome is expanded by identifying Tnik-interacting proteins at
intermediate stages of development in mouse cortex and analyzed with respect to function,
inclusion of protein domains, and enrichment of proteins implicated in contributing to complex
brain disorders. We further characterized potential substrates through the identification of proteins
interacting with a recombinant Tnik kinase domain in embryonic and adult mouse cortical lysates.
We then identified PPIs of TNIK in iPSC-derived hNPCs and generated multiple mutant iPSC
lines via CRISPR/Cas9 genome engineering to model TNIK dysregulation in hNPCs. These
included the introduction of a point mutation to abolish kinase activity (K54R) and a patient-
specific mutation (R180X) which results in a complete lack of TNIK expression and has been
shown to be causative for ID (83). The newly generated cell lines were then used to illustrate how
protein interaction networks can inform the functional consequences of defects in a target protein
in a cell-type specific manner.

13

2.2. – Materials and Methods
Lysate preparation
Tissue was homogenized in either triton lysis buffer (50 mM Hepes (pH 7.4), 2 mM EGTA, 2 mM
EDTA, 50 mM NaF, 20 mM β-glycerol phosphate, 5 mM sodium orthovanadate, Roche cOmplete,
and 1% Triton X-100) or DOC lysis buffer (50 mM Tris (pH 9), 30 mM NaF, 5 mM sodium
orthovanadate, 20 mM β-glycerol phosphate, 20 µM ZnCl2, Roche cOmplete, and 1% sodium
deoxycholate) dependent on the age of the tissue. hNPCs and cortical tissue derived from E14
mice were lysed in triton lysis buffer while P7, P14, and adult mouse cortical tissue were lysed in
DOC lysis buffer.
Post Synaptic Density Preparation
Postsynaptic density preparations were performed as previously described (88). Briefly, adult
mouse (3-4 months in age) cortex was homogenized in sucrose buffer (0.32 M sucrose, 10 mM
Hepes buffer (pH 7.4), 2 mM EDTA, 30 mM NaF, 20 mM β-glycerol phosphate, 5 mM sodium
orthovanadate, and Roche cOmplete protease inhibitor cocktail) and centrifuged at 500g for 6
minutes. Supernatant was collected and then spun at 10,000g for 10 minutes. The resulting pellet
was solubilized in triton buffer (50 mM Hepes (pH 7.4), 2 mM EGTA, 2 mM EDTA, 50 mM NaF,
20 mM β-glycerol phosphate, 5 mM sodium orthovanadate, Roche cOmplete, and 1% Triton X-
100. The solubilized pellet was centrifuged at 30,000 rpm for 30 min and supernatant was
collected for non-PSD fractions. The resulting pellet was collected and solubilized in DOC buffer
(50 mM Tris (pH 9), 30 mM NaF, 5 mM sodium orthovanadate, 20 mM β-glycerol phosphate, 20
µM ZnCl2, Roche cOmplete, and 1% sodium deoxycholate) and served as the PSD fraction.

14

Immunoprecipitation
Immunoprecipitation experiments were performed as previously described (52). Lysate containing
2 mg of total protein was incubated with 4 µg of the indicated primary antibody at 4 degrees
Celsius overnight with rotation. The following day, IPs were incubated with Dynabeads protein
G (Novex, Thermo Fisher Scientific, Waltham, MA) for 2 hours at 4 degrees Celsius with rotation.
IPs were washed three times with IP wash buffer (25 mM Tris (pH 7.4), 150 mM NaCl, 1 mM
EDTA, and 1% Triton X-100). IPs were re-suspended in 2X LDS sample buffer and incubated at
95 degrees Celsius for 15 minutes to elute protein complexes. The eluate was incubated with DTT
at a final concentration of 1 mM at 56 degrees Celsius for 1 hour followed by incubation with
Iodoacetamide at a final concentration of 20 mM at room temperature for 45 minutes. Primary
antibodies used for immunoprecipitation in this study included TNIK (Bethyl Laboratories, catalog
#A302-695A), Rabbit IgG Isotype Control (Thermo Fisher Scientific, catalog #06-6102), and GST
(NeuroMab, catalog #75-148).
Kinase Domain Immunoprecipitation
The human TNIK kinase domain was purchased from SignalChem (catalog # T27-11G) and
contains a GST tag. 1 µg of TNIK kinase domain was incubated with lysate for 2-3 hours while
rotating at 4 degrees Celsius. E14 experiments were conducted in triton lysis buffer while adult
experiments were carried out in either the triton (extra-synaptic) or DOC (PSD) solubilized lysates.
After the initial incubation, 4 µg of anti-GST primary antibody (NeuroMab, catalog #75-148) was
added, the lysate was placed at 4 degrees Celsius with rotation overnight, and immunoprecipitation
was carried out the next day.

15

Mass Spectrometry Sample Preparation
Samples were loaded onto 4 – 12% Bis-Tris gels and separated at 135V for 1.5 hours. Following
separation, gels were stained using InstantBlue (Expedeon, San Diego, CA) and then destained in
25% ethanol overnight. The next day, gel lanes were cut, individual gel slices were placed into
perforated 96 well plates for destaining, and peptide digestion via trypsin was completed at 37
degrees Celsius overnight. Peptides were then extracted with acetonitrile and dried down. A
Nano/Capillary LC System Ultimate 3000 (Thermo Fisher Scientific, Waltham, MA) was used for
desalting and reverse-phase separation of peptides. The LC system was coupled to a hybrid linear
ion-Fourier transform ion cyclotron resonance LTQ-FT (FTICR) 7 Tesla mass spectrometer
(LC/MS) for data acquisition.
Mass Spectrometry Data Analysis
Proteome Discoverer 1.4 (Thermo Fisher Scientific) was used to process MS data which was
analyzed using both Sequest HT and Mascot (Matrix Science, Boston, MA) against the Uniprot
mouse and human databases combined with its decoy database.
Peptide identification settings: The mass tolerance was set 10 parts per million for precursor ions
and 0.8 daltons for fragment ions, no more than two missed cleavage sites were allowed, static
modification was set as cysteine carboxyamidation, and dynamic modification was set as
methionine oxidation. False discovery rates (FDRs) were automatically calculated by the
Percolator node of Proteome Discoverer and a peptide FDR of 0.01 was used for cut-offs. Peptides
with high confidence were considered as true hits and proteins with at least two different peptides
were accepted. Protein Interactions were considered positive if a minimum of two peptides were
present in at least two assays and absent in anti-GST or IgG controls.

16

Cell Culture and Neural Differentiation
iPSC Culture
The control 03231 iPSC line was generated from a lymphoblastoid cell line derived from a healthy
56 year-old male (NINDS repository, ND03231) as previously described (89). iPSCs were
cultured on Geltrex (Thermo Fisher Scientific) coated plates in mTeSR-1 medium (Stemcell
Technologies, Vancouver, Canada) at 37 degrees Celsius and 5% CO2. Karyotyping was carried
out by the Center for Personalized Medicine at the Children’s Hospital of Los Angeles.
Neural Progenitor Differentiation
Neural progenitors were derived using the embryoid body (EB) differentiation method (90).
Briefly, WT 03231 iPSCs were harvested via accutase (Innovative Cell Technologies, San Diego,
CA) treatment and seeded at a density of 2.0 x 10
6
cells per well into non-adherent 6 well plates
(Corning, Corning, NY) in NPC Differentiation media (1X N2 supplement, 1X B27 supplement
without Vitamin A, DMEM/F12 basal media) supplemented with 10 µM SB431542 (Selleckchem,
Houston TX), 0.1 µM LDN193189 (Selleckchem), and Rock Inhibitor (Y-27632 2HCl,
Selleckchem). EBs were fed every other day with NPC differentiation media supplemented with
10 µM SB431542 and 0.1 µM LDN193189 for 7 days. On day 7, EBs were re-plated onto Geltrex-
coated plates and fed every other day with NPC differentiation media supplemented with 10 µM
SB431542 and 0.1 µM LDN193189 until day 14. On day 14, neural rosettes were harvested via
enzymatic selection using Neural Rosette Selection Reagent (Stemcell Technologies). Rosettes
were re-plated on geltrex coated 6 well plates and fed with NPC media (DMEM/F12, Neurobasal,
50% N2 supplement, 50% B27 supplement without Vitamin A, 1X non-essential amino acids

17

supplement (NEAA), and 1X Glutamax) supplemented with 10 ng/ml EGF and 10 ng/ml bFGF
(Peprotech, Rocky Hill, NJ) every other day until confluent and ready to passage.
Induced Neuron Differentiation
Induced neurons (iNs) were differentiated directly from iPSCs using a modified protocol
derived from Zhang et al. (44). When confluent, iPSCs were split via Accutase and seeded into
Geltrex coated 6 well plates at a density of 3.0 x 10
5
cells per well in mTeSR supplemented with
10 µM rock inhibitor. The next day, the media was changed to mTeSR supplemented with 4 µg/ml
polybrene and iPSCs were infected with hNGN2 (Addgene plasmid # 79049) and rtTA (Addgene
plasmid # 19780) lentiviruses at a concentration to obtain > 90% infection efficiency. Media was
changed daily until iPSCs were ready for passaging. iPSCs were then passaged directly into neural
induction media (DMEM/F12 basal media, 1X N2 supplement, 1X NEAA, 10 ng/ml BDNF
(Shenandoah Biotechnology, Warwick, PA), 10 ng/ml NT-3 (Shenandoah Biotechnology), and 1
µg/ml Doxycycline (Enzo Life Sciences, Farmingdale, NY)) supplemented with 10 µM Rock
Inhibitor. Cells were seeded at a density of 2.0 x 10
5
cells per well in 6 well plates. The next day
media was changed to neural induction media supplemented with 0.7 µg/ml puromycin. 48 hours
later, media was changed to B27 media (Neurobasal, 1X B27 supplement, 1X Glutamax, 10 ng/ml
BDNF, 10 ng/ml NT-3, and 1 µg/ml doxycycline). 48 hours later, media was changed to B27
media supplemented with 2 µM Ara-C (MilliporeSigma, Burlington, MA). Half media changes
were carried out every other day until neurons were harvested at 3 weeks post initiation of
differentiation.

18

Cloning
For gene-editing, the guide RNA sequences listed in Table 2.1 were designed using the CRISPR
design tool (http://crispr.mit.edu). Oligonucleotides were ordered from Integrated DNA
Technologies (San Diego, CA) and cloned into pSpCas9(BB)-2A-Puro (PX459) V2.0 (Addgene
plasmid # 62988) as described by Ran et al., 2013 (32). Insertion of the guide sequence was
confirmed via Sanger sequencing.
For TNIK overexpression, the TNIK ORF was PCR amplified from a full length human TNIK
cDNA ORF clone (Genscript, cat # OHu06777) with addition of a C-terminal FLAG tag and
ligated into pENTR4 (Thermo Fisher Scientific) between KpnI and XhoI restriction enzyme sites.
TNIK was then cloned into pSin-EF2-OCT4-Pur (Addgene plasmid #16579) using the NEBuilder
HiFi DNA Assembly Master Mix (New England Biolabs, Ipswich, MA) between the SpeI and
EcoRI restriction enzyme sites according to the manufacturer protocol. Primers used for
amplification of TNIK are listing in Table 2.1.
Gene Targeting
When confluent, iPSCs were harvested via accutase treatment and 1.5 x 10
6
cells were
nucleofected with 6 µg PX459 (Addgene plasmid # 62988) with a TNIK-specific sgRNA cloned
into the BbsI restriction enzyme sites and 1 µl of a 100 µM custom HDR template ordered as
Ultramer DNA oligos (single stranded oligonucleotide (ssODN)) from Integrated DNA
Technologies using the Amaxa NHDF nucleofector kit (Lonza, Basel, Switzerland) with program
B-016. Nucleofected cells were split in equal numbers between 2-4 wells of a 6 well plate in
mTeSR-1 media supplemented with Rock Inhibitor. The next day, a 48 hour selection period was
started with 0.5 µg/ml puromycin. Cells were then fed daily until colonies became apparent.

19

Individual colonies were transferred into 24 well plates and allowed to grow until genotyping was
subsequently carried out. sgRNA and ssODN sequences are listed in Table 2.1.
Lentiviral Production
HEK293 cells cultured on 0.1% gelatin coated tissue culture dishes were transfected using
polyethylenimine (Polysciences Inc., Warrington PA) at approximately 90% confluency with viral
vectors containing the gene of interest and lentiviral packing plasmids (pPAX2 and VSVG). The
medium was changed 24 hours after transfection and then medium containing lentiviral particles
was harvested at 48 and 72 hours after transfection. Medium containing lentiviral particles was
filtered using 0.45 µM filters and then concentrated using Lenti-X concentrator (Clonetech,
Mountain View, CA).
TNIK Over-Expression
WT 03231 hNPCs were split at a density of 0.75 x 10
6
cells per well of a 6 well plate. The next
day, media was changed to NPC media supplemented with 4 µg/ml polybrene and infected with
either pSIN-TNIK-ORF or pSIN-EF1-alpha-IRES-puro empty vector (Addgene plasmid #90505)
as a control. The media was changed the next day. Cells were harvested and processed for
experiments 48 hours post infection.
Wnt Signaling TCF Luciferase Reporter Assay
hNPCs were passaged to white, clear bottom 96 well plates at a density of 1.0 x 10
4
cells per well.
The next day, hNPCs were infected with 7TFP (7X TCF luciferase reporter) (Addgene plasmid
#24308) and pLenti.PGK.blast-Renilla_Luciferase (Addgene plasmid #74444) or 7TGP (7X TCF
GFP reporter) (Addgene plasmid #24305) alone for visualization in NPC media supplemented with
4 µg/ml polybrene. Media was exchanged with NPC media supplemented with 20 ng/ml

20

recombinant Wnt3a (Time Biosciences, Worcester, MA) after 24 hours. The addition of
recombinant Wnt3a stimulates the recruit of beta catenin to TCF elements and initiates
transcription of the respective reporter. After 48 hours of Wnt3a stimulation, luciferase intensity
was measured using the Dual Glo Luciferase Assay System (Promega Corp, Madison, WI)
according to the manufacturer protocol. Luminescence was measured using a SpectraMax i3x
microplate reader (Molecular Devices, San Jose, CA). An Axio Observer A1 fluorescence
microscope (Zeiss, Oberkochen, Germany) was used to visualize GFP expression in Wnt3a
stimulated hNPCs infected with the 7TGP lentivirus.
Western Blotting
Samples were loaded on 4 – 12% Bis-Tris gels (NuPAGE Novex, Thermo Fisher Scientific) and
separated at 135V for 1.5 hours. Proteins were then transferred to a PVDF membrane using a Bio-
Rad Trans-Blot Turbo Transfer System (Bio-Rad, Hercules, CA). Membranes were blocked for 1
hour at room temperature with 5% bovine serum albumin (BSA) in 0.05% TBS-Tween (TBST)
and then incubated with primary antibody at 1:1000 dilution overnight at 4 degrees Celsius.
Primary antibodies used here include: Phospho-Histone H3 – Ser 10 (Cell Signaling Technologies,
catalog # 3377), AKT1 (Cell Signaling Technologies, catalog #2938), GSK-3β (Cell Signaling
Technologies, catalog #9832), β-catenin (Cell Signaling Technologies, catalog #8480), Non-
phospho β-Catenin – Ser33/37/Thr41 (Cell Signaling Technologies, catalog #8814), TNIK
(Thermo Fisher Scientific, catalog #PA1-20639), TNIK (Bethyl Laboratories, catalog #A302-
695A), Phospho-TNIK – Thr181 (custom, produced by Genscript). Membranes were washed with
0.05% TBST four times, ten minutes each, and then incubated with secondary antibodies for 1
hour at room temperature. Membranes were washed with 0.05% TBST 4 times, 5 minutes each,
and imaged using a 4000MM Pro Image Station (Carestream, Rochester, NY).

21

Immunofluorescence
Cells were fixed in 4% paraformaldehyde at room temperature for 20 min, washed three times
with PBS, and then permeabilized with 0.5% PBS-Tween (PBST) overnight. Cells were then
blocked with blocking solution (10% fetal bovine serum in 0.1% PBST) for 1 hour at room
temperature and then incubated with the primary antibody diluted in blocking buffer overnight at
4 degrees Celsius. Primary antibodies used here include: Phospho-Histone H3 – Ser 10 (Cell
Signaling Technologies, catalog # 3377). The next day, the cells were washed three times with
0.1% PBST and then incubated with the secondary antibody for 1 hour at room temperature
followed by DAPI staining.
Enrichment Analysis
Testing for enrichment within de novo mutations was carried out as previously described (53).
Here, lists of de novo mutations from denovo-db (Version 1.5) (91) for schizophrenia (SCZ),
intellectual disability (ID), autism spectrum disorder (ASD), congenital heart disease (CHD), and
all controls within denovo-db were extracted. These lists were then filtered by removing de novo
mutations present in the non-psychiatric cohort of the Exome Aggregation Consortium (ExAC)
(92) in order to increase the chances of obtaining pathogenic variants. Enrichment was analyzed
using a 2-sided binomial exact test, with the theoretical expectation determined by the aggregate
mutation expectation within each respective Tnik interactome (93). Enrichment tests were carried
out for missense mutations, protein-truncating mutations, nonsynonymous mutations (missense
and protein-truncating mutations combined), and synonymous mutations. All tests were carried
out in the R statistical environment using the binom.test function. In order to correct for multiple
comparisons, we applied the Bonferroni correction taking into account the total number of tests
per interactome (n=20), which set the Bonferroni corrected threshold at P < 2.5 x 10
-3
.

22

Gene ontology enrichment analysis was carried using the Database for Annotation, Visualization,
and Integrated Discovery (DAVID) version 6.8 using default settings (94). Only terms that were
significant following the Bonferroni correction for multiple comparisons are reported in the text.

Table 2.1. Oligonucleotides Used in TNIK Study. Table contains oligonucleotides used in
this study separated by application.

23

2.3 – Results
2.3.1 – Tnik Developmental Interactome Indicates Differential Functions of Tnik
Because protein-protein interactions are dynamic and Tnik is implicated in the modulation
of multiple signaling pathways throughout neuronal development, we wanted to examine the
developmental profile of Tnik interactomes. In order to do so, we determined PPIs of endogenous
Tnik via immunoprecipitation followed by tandem mass spectrometry (IP-MS/MS) in P7 and P14
mouse cortical tissue and integrated these results with our previously published interactions at E14
and adult developmental stages (53). Here, we determined an additional 171 interactions, bringing
the total number of Tnik protein interactions throughout mouse cortical development to 480
(Figure 2.1A). The spatiotemporal profile of Tnik interacting proteins highlights how protein
interactions can gradually change throughout development within the same tissue type. For
example, Tnik interactions at P7 share a greater proportion of interactors with the E14 stage (41
protein interactions, 56%) compared to shared interactions between the P14 and E14 stages (27
protein interactions, 28%). In contrast, P14 interactomes share 32 (33%) and 34 (35%) interactions
with adult non-PSD and adult PSD interactomes, respectively, compared to 20 interactions (27%)
and 18 interactions (25%) shared between P7 and the same adult Tnik interactomes. While
interactors of Tnik change throughout development, the P7 and P14 intermediate developmental
stages share the most interactions with each other, accounting for 41 interactions.
The change in interactors throughout development indicates a shift in function of Tnik and
the proteins that it may regulate. This can be illustrated through enrichment analysis of both gene
ontology (GO) and protein domain composition as protein domains are essential, conserved
segments of protein structure that can indicate the function of a target protein with respect to the

24

Figure 2.1. Characterization of Tnik Protein-Protein Interactions Throughout Mouse
Cortical Development. (A) Tnik developmental interactomes as characterized through
immunoprecipitation followed by mass spectrometry. (B) Enrichment of gene ontology terms
(top) and protein domain composition (bottom) that decrease through development. (C)
Enrichment of gene ontology terms (top) and protein domain composition (bottom) that
increase through development. (D) Enrichment of Tnik developmental interactomes in
proteins found to have de novo mutations in patients affected by intellectual disability (ID),
autism spectrum disorder (ASD), developmental delay (DD), and schizophrenia (SCZ). (E)
Protein interaction network of proteins identified as being associated with the Tnik kinase
domain through immunoprecipitation followed by mass spectrometry. Orange nodes indicate
proteins also found to immunoprecipitate with endogenous Tnik.

25

proteins it may be modulating (95) (Figure 2.1B). In E14 interactomes, there is a significant
enrichment in both “centrosome” and “chromosome” indicating Tnik localization, along with
“Wnt signaling” that decreases through development (Figure 2.1B, top panel). This is
corroborated by an enrichment of interactors that harbor DNA or RNA binding domains such as
DEXDC, SANT, or KH domains at early developmental stages (E14 and P7) which then
drastically decreases or disappears in more advanced developmental stages (P14 and adult) (Figure
2.1B, bottom panel). GO and protein domain enrichment are consistent with Tnik’s role in
proliferation and the modulation of genes involved in Wnt signaling (76, 77, 84), both of which
influence rate of neurogenesis during early development (96, 97). In contrast, the association with
proteins that bind glutamate receptors and modulate synaptic signaling increase through
development with the exception of the non-PSD fraction (Figure 2.1C, top panel). This is expected
as non-PSD fractions are depleted of postsynaptic proteins. Proteins with SH3, PDZ, and GUKC
domains are more abundant in Tnik interactomes throughout development (Figure 2.1C, bottom
panel), which is in part due to Tnik interacting with numerous scaffolding molecules in PSD
fractions. These scaffold molecules, such as Shank3 and Psd95, are enriched in protein domains
whose main function is to interact with protein ligands and are specialized in PPIs (53). Tnik
developmental interactomes illustrate the importance of determining protein interactomes through
development and within different cellular fractions. While there are a few proteins that can interact
with Tnik at all developmental stages assayed, a large majority of Tnik interactors are specific to
only one particular developmental stage. The changes in interactors can highlight changes in
cellular function as shown by the involvement of Tnik in regulating proliferation and transcription
early in development and playing a role in the maintenance and modulation of synaptic activity

26

Table 2.2. Tnik Interactions Identified at P7 and P14 Developmental Stages. Table
contains interactors of Tnik identified by IP-MS/MS in mouse cortex from the specified
developmental stage.

27

Table 2.3. Interactions Identified Using Recombinant Tnik Kinase Domain. Table
contains interactors of the recombinant Tnik kinase domain identified via IP-MS/MS from the
specified developmental stage.

28

at the PSD in mature synapses. This suggests that the involvement of Tnik in regulating signaling
networks related to disease may also change throughout development.
Previously, we had investigated the enrichment of de novo mutations associated with
complex brain disorders in Tnik interactomes at E14 and adult stages (53). With the incorporation
of additional developmental stages in the Tnik interactome, this allowed for a comparison
throughout development (Figure 2.1D). We tested for the enrichment of Tnik interactomes in
missense, protein truncating variants (PTV), nonsynonymous, and synonymous de novo mutations
found in patients affected by intellectual disability (ID), autism spectrum disorder (ASD),
developmental delay (DD), and schizophrenia (SCZ). Here, it is readily apparent that the diseases
associated with neurodevelopmental phenotypes (ID, DD, and ASD) have high enrichment in
damaging mutations (e.g. nonsynonymous and PTV) during early developmental stages which
decreases throughout development to adulthood. As ID, DD, and ASD present themselves early
on in development, Tnik associates to key proteins that aid in the onset of the diseases. In contrast,
enrichment in missense and nonsynonymous mutations in SCZ increase throughout development
with the highest points of enrichment being at the P14 stage and lowest being E14. This may
reflect the switch from high rates of neurogenesis to increased synaptogenesis during these
developmental stages (98) and the modulation of proteins involved in SCZ being later on in
development.
Following the characterization of PPIs of Tnik, we also wanted to characterize which
proteins have the capacity to bind to the Tnik kinase domain and therefore can be potential
substrates or modulators of Tnik protein kinase activity. As a step towards this, we characterized
PPIs of the purified Tnik kinase domain in E14, adult non-PSD, and adult PSD mouse cortical
lysates (Figure 2.1E, Table 2.3). Because the protein kinase domain of Tnik shares a high amino

29

acid homology with other members of the STE family of protein kinases (e.g. Mink1 and Map4k4),
we only considered proteins also identified in endogenous Tnik IPs as potential substrates. We
identified 242 interactions in Tnik kinase domain IPs with E14 kinase domain IPs having the
lowest overlap with endogenous Tnik IPs (approximately 30%) and the adult IPs having the
highest (approximately 40%). There are a limited amount of proteins found to overlap throughout
development including Cttnb1, Gsk3b, Akap9, and Cep170. Cttnb1 and Gsk3b are regulators of
canonical Wnt signaling (76) while Akap9 and Cep170 are centrosomal proteins known to regulate
cellular proliferation (99, 100). While these proteins regulate functions essential for proliferation,
they have also been found to be located at the PSD later in development (53). This suggests that
some core PPI modules may be conserved throughout diverse cellular compartments and serve as
an economic way to regulate multiple cellular functions. As in the endogenous interactomes, the
diversity of proteins found to interact with the Tnik kinase domain also reflects the various
functions of Tnik throughout different developmental stages. For example, endogenous Tnik and
the Tnik kinase domain were both found to bind DNA-binding proteins such as Smarcc2, Zfp462,
Ehmt1 and Ehmt2 in E14 mouse cortical lysate. In contrast, both were found to bind proteins
representing the main layers of the PSD scaffolding machinery such as Syngap1, Dlgap1, Shank3,
and Dlg4 at adult stages (Figure 2.1E). Characterizing the proteins interacting with the Tnik kinase
domain is able to inform which proteins may be regulated by Tnik and illustrates the diverse
cellular functions of Tnik at different developmental stages.

30

2.3.2 – TNIK hNPC Interactome Indicates Core Functions of TNIK in Early Neural
Development
hPSCs and hPSC-derived neural progenitor cells (hNPCs) have been used to represent
some of the earliest stages of fetal brain development (101). This provides an opportunity to
analyze human cell type-specific TNIK interactomes and determine their contribution to
neurodevelopmental disorders such as intellectual disability. To this end, we differentiated a
previously characterized control iPSC line (34, 90) into hNPCs and characterized 109 TNIK PPIs
involving endogenous TNIK via IP-MS/MS (Figure 2.2A, Table 2.4). Compared to the earliest
stage in mouse (E14) we found a 44% overlap as opposed to a 4% and 16% overlap with adult
PSD and non-PSD mouse interactomes, respectively. In addition, DNA/RNA binding domains
such as KH and SANT are among the most prevalent in the hNPC interactome as observed in E14
mouse Tnik interactomes. GO enrichment highlights similar functional characteristics as E14
mouse interactomes such as enrichment in “cell cycle” and “centrosome” (Figure 2.2B). This is
further corroborated through some of the most abundant interactors of TNIK being CEP170,
AKAP9, and PDE4DIP, which have all been associated with centrosome function (99, 100) and
are present both in E14 mouse and hNPC interactomes. Therefore, integrating results from Tnik
kinase domain IPs in mouse cortical tissue is also able to highlight several potential substrates of
TNIK in hNPCs (Figure 2.2A).
Protein interactomes within PPI networks may reflect the association of a target protein to
multiple protein complexes. Therefore, by increasing the coverage of PPI networks through the
determination of protein interactomes from multiple nodes within the network, the composition of
specific protein complexes can be more accurately determined. In order to expand the TNIK
interactome in hNPCs, we further characterized the protein interaction networks of the abundant

31

TNIK interactors PDE4DIP and AKAP9. We determined 86 and 117 interactors of AKAP9 and
PDE4DIP, respectively (Figure 2.2C, Table 2.4). Within these interactomes, we determined a core
regulatory module of cell proliferation and neurogenesis in the TNIK-PDE4DIP-AKAP9 PPI
network. This core module shares approximately 11% (AKAP9) and 16% (PDE4DIP) of their total
PPIs with respect to hNPC-TNIK interactions. This shows that TNIK, AKAP9, and PDE4DIP are
involved in specific network modules and share similar functions with respect to the proteins they
regulate.
2.3.3 – TNIK modulates hNPC Proliferation and Wnt Signaling
After analyzing the composition of Tnik interactomes throughout mouse cortical development,
along with TNIK and the TNIK-interacting proteins AKAP9 and PDE4DIP in hNPCs, we then
wanted to generate cellular models to investigate the consequences of either a lack of TNIK kinase
activity or a lack of TNIK protein. Here, we used CRISPR/Cas9 genome engineering within a
Figure 2.2. Characterization of Protein-Protein Interactions Involving TNIK and TNIK-
Associated Proteins in hNPCs. (A) Characterization of the endogenous TNIK interactome in
hNPCs. Yellow nodes indicate proteins found to associate with the purified Tnik kinase domain
in mouse cortical tissues. (B) Gene ontology enrichment analysis of the TNIK hNPC
Interactome. (C) Combined protein interaction network of proteins identified as being
associated with TNIK, AKAP9, and PDE4DIP in hNPCs.

32

Table 2.4. TNIK, PDE4DIP, and AKAP9 PPIs in hNPCs. Table contains interactors of the
specified target protein identified via IP-MS/MS in hNPCs.

33

control iPSC line (89) to generate a kinase dead cell model by mutating a conserved lysine residue
in the ATP-binding pocket of the kinase domain to arginine (K54R) (79) and a TNIK knockout
cell line through the insertion of a truncating mutation found in a patient affected by ID (R180X)
(83) (Figure 2.3A). Individual iPSC colonies were screened following genome engineering via
restriction enzymes specific for the intended mutations and sanger sequencing which was able to
confirm the successful generation of the TNIK K54R and R180X cell lines (Figure 2.3B). We
then differentiated WT, TNIK K54R, and TNIK R180X iPSCs to hNPCs. We first verified that
TNIK remained expressed in the TNIK K54R hNPCs (Figure 2.3C, top panel). In order to ensure
that TNIK was not active in the K54R cell line, we generated a phosphorylation-site specific
antibody against an autophosphorylation residue, T181, to probe the activity of TNIK. Because
this site is conserved among TNIK and the TNIK family members, MINK1 and MAP4K4, we first
immunoprecipitated TNIK from WT and K54R cell lines and then immunoblotted with the
phosphorylation site specific T181 antibody. This was able to show a complete lack of
phosphorylation in the K54R cell line, ensuring that TNIK was not active (Figure 2.3C, bottom
panel) and that TNIK T181 phosphorylation cannot be compensated by other protein kinases in
the hNPC K54R cell line. We then verified that TNIK expression was completely ablated in TNIK
R180X hNPCs (Figure 2.3D).
In colorectal cancer cells, knockdown of Tnik reduces proliferation (84) and in Tnik
knockout mice, a marked decrease in ki67 positive cells in the dentate gyrus has been observed,
indicating decreased rates of adult neurogenesis (88). TNIK K54R hNPCs also showed a reduction
in proliferation as measured by phospho-histone H3 staining (Figure 2.3E). Immunoblotting was
able to verify the downregulation of phospho-histone H3 staining in K54R hNPCs (Figure 2.3F)
and show the same phenotype in TNIK R180X hNPCs (Figure 2.3G). In order to investigate the

34

Figure 2.3. Generation of TNIK K54R and R180X Cell Lines and Characterization of
Proliferation Defects. (A) Schematic illustrating TNIK mutant lines generated by
CRISPR/Cas9 genome engineering in WT iPSCs. (B) Agarose gels showing polymerase
chain reaction followed by restriction enzyme digest screening for TNIK K54R and R180X
cell lines (top panel). Sanger sequencing conformation of cell lines homozygous for TNIK
K54R and R180X (bottom panel). (C) Immunoblot showing TNIK expression is maintained
in K54R hNPCs (top panel). Immunoprecipitation of TNIK followed by immunoblot using
an antibody directed against a TNIK autophosphorylation site (T181) is able to show a lack
of kinase activity in K54R hNPCs (bottom panel). (D) Immunoprecipitation followed by
immunoblot of TNIK in R180X hNPCs is able to confirm a lack of TNIK. (E)
Immunofluorescence of phospho-histone H3 expression is able to show TNIK K54R hNPCs
have decreased proliferation compared to isogenic controls (n = 3 biological replicates). (F)
Immunoblot and quantification of phospho-histone H3 expression in TNIK K54R hNPCs
compared to isogenic controls (n = 3 biological replicates). (G) Immunoblots and
quantification of phospho-histone H3 expression in TNIK R180X compared to isogenic
controls (n = 4 biological replicates). (H) Immunoblot and quantification of phospho-histone
H3 expression in WT hNPCs following transient expression of empty vector or TNIK
lentiviral vectors 48 hours post infection (n = 3 biological replicates). (I) Immunoblot and
quantification of TNIK expression 48 hours post infection with TNIK or empty vector
lentiviral vectors (n = 3 biological replicates). Error bars are mean ± SEM. *=p<0.05,
Student’s t test was used for immunofluorescence quantification and two tailed unpaired t test
was used for immunoblot quantification.

35

direct involvement of TNIK in hNPC cell proliferation, we over-expressed TNIK in WT hNPC
cells and measured phospho-histone H3 levels via immunoblotting after 48 hours (Figure 2.3H-I).
TNIK overexpression showed an increase in phospho histone H3 expression relative to the empty
vector control transduction, indicating TNIK expression has a direct effect on proliferation of
hNPCs (Figure 2.3H).
TNIK has been shown to associate with TCF4 and beta catenin in order to regulate Wnt
signaling in non-neuronal cells. In colorectal crypt cells, siRNA-mediated knockdown of TNIK
or over-expression of TNIK K54R results in a decrease in Wnt signaling as shown by a decrease
in TCF-reporter activity which is driven by the activity of beta catenin (76). Surprisingly, we
observed the opposite effect, with an increase in the expression of Wnt-related proteins such as
beta catenin, the active (non-phosphorylated) form of beta catenin, GSK3β, and AKT1 in both the
K54R and R180X hNPC cell lines relative to isogenic controls (Figure 2.4A). In order to
investigate whether the upregulation of Wnt-related proteins translated to an increase in Wnt
signaling, we performed a luciferase TCF-reporter assay (102) in TNIK K54R hNPCs which was
able to show a significant increase in TCF-reporter activity under non-stimulated conditions along
with an increased responsiveness to Wnt3a stimulation compared to isogenic control hNPCs
(Figure 2.4B). There was no significant difference in TCF reporter activity between non-
stimulated TNIK K54R hNPCs and WT Wnt3a stimulated hNPCs indicating that TNIK K54R
hNPCs are similar to a stimulated level of Wnt signaling under normal culture conditions. The
same result was obtained using a GFP-TCF reporter construct (Figure 2.4C). Unlike the regulation
of proliferation, transient overexpression of TNIK did not result in changes in levels of the Wnt-
related proteins AKT1, GSK3β, beta catenin, or the active (non-phosphorylated) form of beta
catenin (Figure 2.5A). We also examined the expression of these proteins in induced neurons (iNs)

36

Figure 2.4. Wnt Signaling is increased in hNPCs Lacking TNIK or TNIK kinase activity.
(A) Immunoblot and quantification of Wnt signaling-associated proteins AKT1, GSK3β,
CTNNB1, and non-phospho CTNNB1 (Ser33/37/Thr41) in TNIK K54R (n = 3 biological
replicates) and TNIK R180X (n = 4 biological replicates) hNPCs. Error bars are mean ± SEM,
* = p<0.05, two-tailed unpaired t-test. (B) Quantification of luciferase intensity in WT and
K54R hNPCs either untreated or stimulated with Wnt3a (n = 3 biological replicates). Error
bars are mean ± SEM. * = p<0.05, *** = p<0.001, one-way ANOVA with Bonferroni post hoc
test. (C) Immunofluorescence illustrating increased of WT and K54R hNPCs infected with a
7X-TCF GFP reporter, illustrating increased response to Wnt3a stimulation in K54R hNPCs.

37

generated from TNIK K54R iPSCs and did not observe any changes in total protein or protein
phosphorylation levels (Figure 2.5B). Collectively, these results indicate that the up-regulation of
Wnt signaling is cell-type specific to hNPCs and may occur during the generation of hNPCs from
hPSCs rather than being an immediate consequence of modulating TNIK expression levels. While
the transient expression of TNIK directly affects proliferation of hNPCs, this may do so
independently of Wnt signaling as it does not result in any significant changes in either total or
active beta catenin levels.

Figure 2.5. Wnt Signaling Changes are Cell Type Specific. (A) Immunoblot and
quantification of Wnt signaling-associated proteins in WT hNPCs following transient
expression of empty vector or TNIK lentiviral vectors 48 hours post infection (n = 3 biological
replicates). (B) Immunoblot and quantification of Wnt signaling-associated proteins in WT or
TNIK K54R induced neurons (n = 4 biological replicates). Error bars are mean ± SEM. No
significant difference was found, two-tailed unpaired t-test.

38

2.4 - Discussion
Here, we have expanded the mouse cortical Tnik developmental interactome by
incorporating 171 novel interactions at intermediate developmental stages between embryonic and
adult stages. This was able to illustrate how the interactome of any one particular protein can
change in a spatiotemporal manner. A target protein may have differential interactions not only
within a tissue or cell type specific setting, but also within a subcellular manner, as illustrated
previously using PSD and non-PSD cellular fractionations (53, 103). By analyzing Tnik
interactions throughout development, we were able to visualize the involvement of Tnik in
transcription/translation and cellular division in early development and the regulation of synaptic
function during adult stages. While these are in line with previously reported functions of Tnik
(76, 84, 88), the use of IP-MS/MS was able to assign binding partners of Tnik that may participate
in the dysregulation of these processes when TNIK function or expression is ablated.
Because TNIK is a protein kinase, many of the identified proteins in the developmental
interactome may be potential substrates. A number of small-scale studies have also identified
substrates of Tnik. For example, Tnik has been shown to phosphorylate TCF4 in-vitro, which
modulates Wnt signaling (76). In another study, a phosho-antibody against a Tnik consensus site
was generated to immunoprecipitate proteins containing this site in hippocampal neurons treated
with either okadaic acid to stimulate Tnik activity or okadaic acid in combination with a Tnik
kinase family inhibitor. Following phospho-peptide enrichment, they were able to identify 16
differentially phosphorylated peptides with half of them belonging to p120-catenin, δ-catenin, and
ARVCF. However, these may not all be specific to Tnik as the Tnik inhibitor used here can inhibit
Tnik, Map4k4, and Mink1 (104). In the current study, the use of the Tnik kinase domain was able
to provide candidates for phosphorylation by Tnik. Overlaying the identified interactions with

39

those that interact with endogenous Tnik is able to increase the confidence with respect to the
specificity of the potential Tnik substrates. Moreover, some of the identified proteins may solely
bind to the kinase domain rather than undergo phosphorylation. These may represent Tnik
regulatory proteins and warrant further investigation in the future. The hNPC K54R line generated
in this study may also be useful in the future to compare differences in phosphorylation status
between different cell types and/or modulation via stimuli to identify individual phosphorylation
sites that are regulated via TNIK.
The identification of TNIK PPIs in hNPCs was able to show the similarities between
human and mouse interactors at early embryonic stages. The hNPC interactome, in combination
with the TNIK K54R cell line, was able to further illustrate how protein interactions can inform
protein function. With an enrichment in proteins found to regulate proliferation, we observed that
a lack of TNIK kinase activity resulted in decreased proliferation of hNPCs. This is in line with
decreased proliferation of progenitor cells in the dentate gyrus in the Tnik knockout mouse model
which was further associated with decreased adult neurogenesis and cognitive defects (88).
In contrast to the previously described role of Tnik in the regulation of Wnt signaling (84),
we uncovered an upregulation of the expression of Wnt-related proteins, active (non-
phosphorylated) beta catenin, and beta catenin-driven TCF reporter activity in hNPCs with a lack
of TNIK kinase activity or TNIK expression. Collectively, this shows an upregulation of Wnt
signaling in hNPCs when TNIK is dysregulated and illustrates how deficiencies in TNIK can result
in cell-type specific phenotypes. This is re-enforced by the observation that iNs derived from
TNIK K54R iPSCs do not show the same upregulation of active beta catenin as observed in
hNPCs. This may be due to iNs being differentiated directly from hPSCs and bypassing the
intermediate stage of hNPC generation or cell-type specific functions as hNPCs proliferate and

40

differentiate while iNs are post-mitotic. However, the differences between hNPCs, iNs, and the
previously described observations in cancer cells illustrate how Wnt signaling may be up-
regulated, down-regulated, or not changed in cells with defects in TNIK expression or kinase
activity depending on the specific cell type in question.
Both proliferation and defects in Wnt signaling have been associated with
neurodevelopmental phenotypes. For example, hNPCs derived from patients with fragile X
syndrome, have a lack of FMR1 (a TNIK interactor and putative target (105)) and dysregulation
of Wnt signaling transcriptional network (106). hNPCs derived from ASD patients with truncating
mutations in CTNNB1 have a lack of CTNNB1, decreased Wnt signaling, and exhibit increased
rates of proliferation (39). This, in combination with the results observed in mutant TNIK hNPCs,
suggests that altering the critical balance of CTNNB1 protein concentration or activity can result
in decreased rates of proliferation in hNPCs.
The Tnik interactome is also enriched in proteins that may contribute to several
neurodevelopmental disorders such ID, ASD, and DD. Proteins that influence disease are thought
to be closer to one another in protein interaction networks as opposed to two proteins that may not
contribute to disease (107). This may implicate the TNIK PPI network as a signaling hub that
integrates several pathways and modulates risk factors for neurodevelopmental disorders, making
the Tnik signaling network a potential therapeutic target. However, as outlined above, the
modulation of Tnik with respect to therapeutic potential should be carried out in a cell-type specific
manner.

41

Chapter 3 - Endogenous Cell Type-Specific Disrupted in Schizophrenia 1 Interactomes
Reveal Protein Networks Associated with Neurodevelopmental Disorders.
3.1 - Introduction
Since its discovery, Disrupted in schizophrenia 1 (DISC1) has been implicated in
contributing to multiple psychiatric disorders (108-118) and functional studies have shown DISC1
to play critical roles in several cell-types throughout brain development (38, 119). Induced
pluripotent stem cells (iPSCs) with clinically-relevant mutations in DISC1 have been shown to
exhibit neurodevelopmental phenotypes following differentiation into neural progenitor cells
(hNPCs) such as abnormalities in proliferation and altered cell fate (37, 120, 121). The
proliferation of primary NPCs has also been shown to be affected by astrocytes over-expressing a
dominant negative form of DISC1 (122). DISC1 does not have an enzymatic function and has been
considered a scaffold protein (123), with its function usually characterized in association with
modulation of activity and/or localization of its binding partners (124-126). However, the
identification of endogenous DISC1 protein interactions has not been straightforward.
The large-scale identification of DISC1 protein interactions has been largely composed of
in-vitro systems such as yeast two hybrid based assays (127-133). These studies have identified
dozens of proteins with the potential to directly interact with DISC1 in in-vivo interactions. In
addition, a relatively low number of interactions have also been observed in low-throughput
studies using recombinant expression systems (114, 124, 125). However, these assays usually do
not account for tissue or cell specificity and the relative stoichiometry between interacting partners.
The investigation of endogenous DISC1 protein-protein interactions has been limited to low-
throughput assays (134, 135).

42

In order to determine DISC1 protein complexes at endogenous expression levels in cell
types representing early brain development and to eliminate non-specific interactions, we inserted
a 3X-FLAG (136) coding sequence at the C-terminal end of the endogenous DISC1 gene using
CRISPR/Cas9 genome engineering in human iPSCs. DISC1
FLAG
iPSCs, were differentiated into
hNPCs and astrocytes and endogenous DISC1 binding partners were determined by
immunoisolation of DISC1-FLAG followed by HPLC-MS/MS in cell type-specific settings. We
identified 165 non-redundant proteins in DISC1 protein complexes associated to different
subcellular compartments. We confirm the involvement of DISC1 in centrosomal dynamics (137)
and mRNA transport (138) along with novel cell-type specific functions such as regulation of
transcription in hNPCs and cytoskeleton processes in astrocytes. We also show that DISC1 is
expressed at sub-stoichiometric levels when compared to its binding partners. This indicates
DISC1 might associate with a discrete subset of its interacting molecules involved in specific
molecular functions. Finally, with DISC1 being implicated in a variety of psychiatric disorders,
we also show a novel set of protein interactors contributing to complex brain disorders such as
schizophrenia and intellectual disability in protein interaction networks identified in hNPCs but
not astrocytes or recombinant immunopurification experiments.

43

3.2 – Materials and Methods
Cell culture and neural differentiation
The control 03231 iPSC line was generated from a lymphoblastoid cell line derived from a healthy
56 year-old male (NINDS repository, ND03231) as previously described (89). iPSCs were
cultured on Geltrex (Thermo Fisher Scientific, Waltham, MA) coated plates in mTeSR-1 medium
(Stemcell Technologies, Vancouver, Canada) at 37 degrees Celsius and 5% CO2.
Neural progenitors were derived using a modified protocol from Chambers et al. and Qi et al. (34,
139). Here, DISC1
FLAG
and WT iPSCs cultured in mTeSR-1 media were split via accutase
treatment and re-plated at a density of .80 x 10
4
cells per cm
2
in mTeSR-1 supplemented with Rock
Inhibitor (Y-27632 2HCl, Selleckchem, Houston, TX). The next day, cells were half-fed with
KSOR media (DMEMF12/15% KSOR/NEAA/Glutamax) supplemented with 10 µM SB431542
(Selleckchem, Houston, TX) and 0.1uM LDN193189 (Selleckchem, Houston, TX). Cells were
transitioned from KSOR to N2 media (DMEMF12/N2/NEAA/Glutamax) supplemented with 10
µM SB431542 and 0.1 µM LDN193189 being fed with KSOR on days 3 -4 and N2 on days 5 – 8.
On the 8
th
day, cells were passaged using accutase onto geltrex coated plates at a density of 2x10
5

cells per cm
2
in NPC media (DMEMF12/Neurobasal/50% N2/ 50% B27 without Vitamin
A/NEAA/Glutamax) supplemented with Rock Inhibitor, 10 ng/ml EGF (Peprotech, Rocky Hill,
NJ), and 10 ng/ml bFGF (Peprotech, Rocky Hill, NJ). Following the first passage, hNPCs were
fed every other day with NPC media supplemented with 10 ng/ml EGF and 10 ng/ml bFGF. Cells
were further propagated until the third passage when experiments were carried out.
WT hNPCs used for gene expression profiling via RNA-seq were obtained using the embryoid
body (EB) differentiation method (90). Briefly, WT 03231 iPSCs were harvested via accutase

44

treatment and seeded at a density of 2.0 x 10
6
cells per well into non-adherent 6 well plates
(Corning, Corning, NY) in NPC Differentiation media (1X N2/1X B27/DMEM/F12/10 µM
SB431542/0.1 µM LDN193189) supplemented with Rock Inhibitor. EBs were fed every other day
with NPC differentiation media for 7 days. On day 7, EBs were re-plated onto geltrex coated
plates and fed every other day with NPC differentiation media until day 14. On day 14, neural
rosettes were harvested via enzymatic selection using Neural Rosette Selection Reagent (Stemcell
Technologies, Vancouver, Canada). Rosettes were re-plated on geltrex coated 6 well plates and
fed with NPC media supplemented with 10 ng/ml EGF and 10 ng/ml bFGF every other day until
confluent and ready to passage. Cells were passaged three times prior to harvesting for RNA-seq
library preparation.
iPSC-derived astrocytes were generated as previously described (140). Briefly, hNPCs were
dissociated with accutase and re-plated in complete astrocyte media (ScienCell Research
Laboratories, Carlsbad, CA) and fed every other day. Differentiating cells were split
approximately once a week and re-plated at the initial seeding density. After 30 days of
differentiation, astrocytes were expanded for two passages before experiments were carried out.
To differentiate NPCs to neurons, hNPCs were re-plated and cultured in Neurobasal media
supplemented with 1X B27, 10 ng/ml BDNF, 10ng/ml GDNF, 10 µM cAMP. Half of the media
was changed every other day until 4 weeks later when the neurons were stained and imaged.
Cloning
For gene-editing, the guide RNA sequences listed in Table 3.1 was designed using the CRISPR
design tool (http://crispr.mit.edu). Oligonucleotides were ordered from Integrated DNA
Technologies and cloned into pSpCas9(BB)-2A-Puro (PX459) V2.0 (Addgene plasmid # 62988)

45

as described by Ran et al., 2013 (32). Insertion of the guide sequence was confirmed via Sanger
sequencing.
The homologous recombination donor vector was constructed using the pFETCH donor plasmid
(Addgene plasmid # 63934) according to Savic et al. (74). Here, the pFETCH donor plasmid was
digested with BsaI and BbsI restriction enzymes (New England Biolabs, Ipswich, MA) and
fragments corresponding to 4.5 kb and 1 kb were subsequently gel purified. Homology arms with
attached Gibson assembly tails were ordered as gBlocks from Integrated DNA Technologies (San
Diego, CA) (sequences are listed in Table 3.1. Homology Arms and plasmid backbone fragments
were then assembled using NEBuilder HiFi DNA Assembly Master Mix (New England Biolabs,
Ipswich, MA) according to the manufacturer’s protocol.
Off-target predictions were carried out using CRISPOR (141). The top ten off predicted off-target
sites for the DISC1
FLAG
iPSC line were PCR amplified and the resulting PCR products were cloned
into pMiniT 2.0 using the NEB PCR Cloning Kit (New England Biolabs, Ipswich, MA).
Individual clones then underwent Sanger Sequencing. All primers used for PCR amplification of
predicted off-target sites are listed in Table 3.1.
For CRISPR gene activation, sgRNAs were designed using the CRISPR-ERA design tool (142)
(Table 3.1). Oligonucleotides were ordered from Integrated DNA Technologies, annealed, and
cloned into the BsaI sites of pGL3-U6-sgRNA-PGK-puromycin (Addgene plasmid # 51133).
Insertion of sgRNAs were confirmed via Sanger sequencing.
Gene Targeting
When confluent, iPSCs were harvested via accutase treatment and 1.5 x 10
6
cells were
nucleofected with 5 µg PX459 (Addgene plasmid # 62988) with the DISC1 sgRNA cloned into it

46

along with 5 µg DISC1-FLAG-pFETCH-Donor vector using the Amaxa NHDF nucleofector kit
(Lonza, Basel, Switzerland) with program B-016. Nucleofected cells were split in equal numbers
between 6 wells of a 6 well plate in mTeSR-1 media supplemented with Rock Inhibitor. The next
day, a 48 hour selection period was started with 0.5 µg/ml puromycin. Cells were then fed daily
until colonies became apparent. Individual colonies were transferred into 24 well plates and
allowed to grow until genotyping was subsequently carried out. Targeting the CNH domain in
TNIK was completed above with the exception of the TNIK sgRNA being used and a custom HDR
template ordered as Ultramer DNA oligos (single stranded oligonucleotide (ssODN)) from
Integrated DNA Technologies (San Diego, CA). For nucleofection, 6 µg of PX459 with the TNIK
sgRNA cloned into it along with 1 µl of 100 µM ssODN was used. gRNA and ssODN sequences
are listed in Table 3.1.
Gene Activation
sgRNAs targeting the promoter region of DISC1 (Table 3.1) were cloned into pGL3-U6-sgRNA-
PGK-puromycin as described above. DISC1-FLAG and WT hNPCs were then nucleofected using
the Amaxa Nucleofector II, program A-33, with 4 µg SP-dCas9-VPR (Addgene plasmid # 63798)
and 1 µg of each cloned sgRNA plasmid. Cells were seeded onto Geltrex-coated coverslips in
NPC growth media. After 48 hours, cells were either harvested or processed for
immunofluorescence.
Immunofluorescence
Cells were fixed in 4% paraformaldehyde at room temperature for 20 min, washed three times
with PBS, and then permeabilized with 0.5% PBS-Tween (PBST) overnight. Cells were then
blocked with blocking solution (10% fetal bovine serum in 0.1% PBST) for 1 hour at room

47

temperature and then incubated with the primary antibody diluted in blocking buffer overnight at
4 degrees Celsius. The next day, the cells were washed three times with 0.1% PBST and then
incubated with the secondary antibody for 1 hour at room temperature followed by DAPI staining.
All primary and secondary antibodies used, along with their dilutions, can be found in Table 3.2.
BrdU Experiments
DISC1
FLAG
iPSCs were seeded into Geltrex-coated 24 well plates at equal densities. After 72
hours, NPC growth media was supplemented with 10 µM BrdU for one hour. Cells were then
fixed with 4% Paraformaldehyde at room temperature for 20 min and then permeabilized with
0.2% triton for 20 min at room temperature. Cells were then incubated with 2M HCl for 20 min
to facilitate DNA denaturation and the reaction was neutralized with 0.1 M sodium borate for 15
min. Immunofluorescence was then carried out as described above. For each sample, 5 random
fields were counted using CellProfiler cell image analysis software (http://www.cellprofiler.org)
(143) in order to determine the percentage of BrdU positive cells.
Immunoprecipitation
iPSC-derived hNPCs or astrocytes cultured in geltrex coated 10 cm
2
tissue culture dishes with
harvested via accutase treatment and lysed in cell lysis buffer (50 mM Tris pH 7.4, 2 mM EDTA,
10 mM NaVO4, 30 mM NaF, 20 mM β-glycerophosphate, 1% n-Dodecyl-β-Maltopyranoside
(Anatrace, Maumee, OH), supplemented with cOmplete Protease Inhibitor Cocktail Tablets
(Sigma, St. Louis, MO)) using mechanical homogenization followed by incubation at 4 degrees
Celsius while rotating for 40 minutes. The cell lysate was then centrifuged at 35,000 RPM for 30
minutes at 4 degrees Celsius and the protein concentration of the supernatant was subsequently
determined using the BCA assay (Thermo Fisher Scientific, Waltham, MA). For hNPCs, 2 mg of

48

total protein was mixed with 50 µl of Anti-FLAG M2 Magnetic Beads (Sigma) and incubated at 4
degrees Celsius overnight with gentle rotation. Beads were washed 3 times in IP wash buffer (25
mM Tris pH 7.4, 150 mM NaCl, 1 mM EDTA, 0.5% n-Dodecyl-β-Maltopyranoside) and then
incubated with 100 µl of 150 ng/µl 3X-FLAG peptide (Apex Bio, Houston, TX) for 2 hours at 4
degrees Celsius. Incubation was repeated twice and individual fractions were collected. Fractions
were dried down to less than 50 µl in volume using a Thermo Scientific Savant SPD 1010
Speedvac Concentrator and combined with LDS buffer. In addition, the Anti-FLAG M2 magnetic
beads were boiled at 95 degrees Celsius in 2X LDS buffer when measuring elution efficiency. The
first elution was for mass spectrometry-based analyses.
For recombinant DISC1 experiments, adult mouse (3-4 months in age) cortex was homogenized
in triton buffer (50 mM Hepes (pH 7.4), 2 mM EDTA, 50 mM NaF, 20 mM β-glycerophosphate,
5 mM sodium orthovanadate, and Roche cOmplete protease inhibitor cocktail). The cell lysate
was then centrifuged at 35,000 RPM for 30 minutes at 4 degrees Celsius and 2 mg of lysate was
incubated with or without 0.5 µg of recombinant DISC1 (TP312015, Origene Technologies,
Rockville, MD) for 2 hours at 4 degrees Celsius with rotation. Following this, 2 µg of anti-FLAG
antibody (Sigma) was added to the lysate and was incubated with rotation for overnight. The next
day, IPs were incubated with Dynabeads protein G (Novex) for 2 hours at 4 degrees Celsius with
rotation. IPs were washed three times with IP wash buffer (25 mM Tris (pH 7.4), 150 mM NaCl,
1 mM EDTA, and 1% Triton X-100). IPs were re-suspended in 2X LDS sample buffer and
incubated at 95 degrees Celsius for 15 minutes to elute protein complexes. Immunoprecipitation
of DISC1 from Astrocytes followed the same protocol as above with the exception of recombinant
DISC1 being used and cells being lysed in DDM lysis buffer instead of triton buffer.

49

Western Blotting
Following immunoprecipitation using either DISC1
FLAG
or WT hNPC cell lysate, samples were
combined with LDS sample buffer and 10 mM DTT and then incubated at 95 degrees Celsius for
15 minutes. Samples were then loaded on 4 – 12% Bis-Tris gels (NuPAGE Novex, Thermo Fisher
Scientific, Waltham, MA) and separated at 135V for 1.5 hours. Proteins were then transferred to
a PVDF membrane using a Bio-Rad Trans-Blot Turbo Transfer System (Bio-Rad, Hercules, CA).
Membranes were blocked for 1 hour at room temperature with 5% bovine serum albumin (BSA)
in 0.05% TBS-Tween (TBST) and then incubated with primary antibody overnight at 4 degrees
Celsius. Membranes were washed with 0.05% TBST four times, ten minutes each, and then
incubated with secondary antibodies for 1 hour at room temperature. Membranes were washed
with 0.05% TBST 4 times, 5 minutes each, and imaged using a 4000MM Pro Image Station
(Carestream, Rochester, NY). All primary and secondary antibody antibodies used along with their
dilutions can be found in Table 3.2.
Mass Spectrometry
Samples were combined with LDS sample buffer and 10 mM DTT and incubated at 56 degrees
Celsius for 1 hour. Iodoacetamide was added to the samples to a final concentration of 20 mM
followed by a 45 minute incubation in the dark at room temperature. Samples were loaded onto 4
– 12% Bis-Tris gels and separated at 135V for 1.5 hours. Following separation, gels were stained
using InstantBlue (Expedeon, San Diego, CA) and then destained in 25% ethanol overnight. The
next day, gel lanes were cut, individual gel slices were placed into perforated 96 well plates for
destaining, and peptide digestion via trypsin was completed at 37 degrees Celsius overnight.
Peptides were then extracted with acetonitrile and dried down. A Nano/Capillary LC System
Ultimate 3000 (Thermo Fisher Scientific, Waltham, MA) was used for desalting and reverse-phase

50

separation of peptides. The LC system was coupled to a hybrid linear ion-Fourier transform ion
cyclotron resonance LTQ-FT (FTICR) 7 Tesla mass spectrometer (LC/MS) for data acquisition.
Mass Spectrometry Data Analysis
Proteome Discoverer 1.4 (Thermo Fisher Scientific, Waltham, MA) was used to process MS data
which was analyzed using Sequest HT against the Uniprot human database combined with its
decoy database. With respect to analysis settings, the mass tolerance was set 10 parts per million
for precursor ions and 0.8 daltons for fragment ions, no more than two missed cleavage sites were
allowed, static modification was set as cysteine carboxyamidation, and dynamic modification was
set as methionine oxidation. False discovery rates (FDRs) were automatically calculated by the
Percolator node of Proteome Discoverer and a peptide FDR of 0.01 was used for cut-offs. Peptides
with high confidence were considered as true hits and proteins interactions were divided into
categories as stated in the results: DISC1 interactors that were identified with multiple unique
peptides in duplicate assays and zero peptides in control elutions, proteins identified with 1 unique
peptide in duplicate and zero peptides in control elutions, and interactors identified with at least 3
times more peptides, and 5 times more peptide-spectral matches (PSMs) in DISC1-FLAG elutions
than in controls. The datasets generated and/or analyzed during the current study have been
deposited to the ProteomeXchange Consortium via the Proteomics IDEntifications (PRIDE)
partner repository under the dataset identifier PXD007103 (144).
RNA-sequencing and DISC1 coding region analysis
RNA was extracted from WT hNPCs using the Direct-zol RNA MiniPrep Plus kit (Zymo
Research, Irvine, CA). Quality of the RNA was assessed using an Agilent Technologies 2200
TapeStation Instrument in order to obtain RNA Integrity Numbers and quantification was
performed using the NanoDrop ND-1000 Spectrophotometer (Thermo Fisher Scientific, Waltham,

51

MA). Starting with 1 µg of total RNA per sample, polyA(+) RNA was purified and cDNA libraries
were prepared using the Truseq Stranded mRNA LT Kit (Illumina, San Diego, CA). Prepared
libraries were sequenced on an Illumina HiSeq2000 sequencer in order to generate 101 bp single
end reads. An average of 5.6 million reads were obtained per replicate (range = 5.4 – 5.8 million
reads). Reads were mapped to the UCSC Human Reference Genome (hg19) using TopHat version
2.1.1 (145) and relative gene expression was quantified using Cufflinks version 2.2.1 (146). Reads
were visualized with Integrated Genome Viewer (IGV, Version 2.3.92) and manual inspection was
carried out in order to observe the coding sequence of DISC1.
Quantitative PCR
hNPCs were harvested and cell pellets were processed for RNA isolation using the Direct-zol RNA
Miniprep Plus kit (Zymo research, Irvine, CA). RNA concentration and 260/280 absorbance ratios
were measured using the NanoDrop ND-1000 Spectrophotometer (Thermo Fisher Scientific,
Waltham, MA). Reverse transcription was carried out using Protoscript II First Strand cDNA
Synthesis Kit (New England Biolabs, Ipswich, MA) according to the manufacturer’s instructions.
cDNA was used for quantitative PCR using Taqman Gene Expression Master Mix (Thermo Fisher
Scientific, Waltham, MA) on a Roche LightCycler 96 system. Three technical triplicates were ran
for each biological replicate. Taqman gene expression assays used for quantitative PCR include
DISC1 (Hs00257791_s1) and GAPDH (Hs03929097_g1). Additional custom qPCR probe-based
assays were ordered from Integrative DNA Technologies (San Diego, CA). Both primer and probe
sequences for custom qPCR assays are located in Table 3.1. Relative quantity of the assayed
transcript was determined by the 2
-ΔΔCt
method with GAPDH as a reference (147).

52

Enrichment Analyses
Testing for enrichment within de novo mutations was carried out as previously described (53).
Here, lists of de novo mutations from denovo-db (Version 1.5) (91) for schizophrenia (SCZ),
intellectual disability (ID), autism spectrum disorder (ASD), congenital heart disease (CHD), and
all controls within denovo-db were extracted. These lists were then filtered by removing de novo
mutations present in the non-psychiatric cohort of the Exome Aggregation Consortium (ExAC)
(92) in order to increase the chances of obtaining pathogenic variants. Enrichment was analyzed
using a 2-sided binomial exact test, with the theoretical expectation determined by the aggregate
mutation expectation within the DISC1 hNPC interactome (93). Enrichment tests were carried out
for missense mutations, protein-truncating mutations, nonsynonymous mutations (missense and
protein-truncating mutations combined), and synonymous mutations. All tests were carried out in
the R statistical environment using the binom.test function. In order to correct for multiple
comparisons, we applied the Bonferroni correction taking into account the total number of tests
per interactome (n=20), which set the Bonferroni corrected threshold at P < 2.5 x 10
-3
.
Fisher exact tests were carried out in order to measure the enrichment of DISC1 interactomes or
all previously reported genes in differentially expressed genes. The background was set to the
total number of protein-coding genes according to Ensembl Human Genome GrCH38.p10 (n =
20,338) in order to account for all previously reported DISC1 interactors. The Bonferroni
correction was applied to each dataset being tested (n = 4).
Gene ontology enrichment analysis was carried using the Database for Annotation, Visualization,
and Integrated Discovery (DAVID) version 6.8 using default settings (94). Only terms that were
significant following the Bonferroni correction for multiple comparisons are reported in the text.

53

Analysis of previously reported protein-protein interactions
For analysis of previously reported DISC1 protein-protein interactions, a list was curated through
the use public databases including BioGRID (148), Mentha (149), and The Center for Cancer
Systems Biology (CCSB) at the Dana-Farber Cancer Institute HI-III dataset accessed via HuRI:
the Human Reference Protein Interactome Mapping Project on February 20, 2018 DISC1
interactions within BioGRID and Mentha databases included the species: H. sapiens, M. musculus,
and R. norvegicus. Additional DISC1 interactions were obtained through manual curation of
literature using PubMed. Additional references used are supplied in Supplemental Table S8. For
the clustering of the DISC1 hNPC and astrocyte interactomes based on previously reported
interactions, interactions were obtained using the GeneMANIA app for Cytoscape (150) and
visualized using Cytoscape version 3.5.0.

54

Table 3.1. Oligonucleotides Used in DISC1 Study. Table contains all oligonucleotides used
in this study separated by application.

55

Table 3.2. Antibodies Used in DISC1 Study. Table contains all antibodies used in this study
including each respective application and dilution at which the antibody was used.

56

3.3 - Results
3.3.1 - Generation of DISC1 FLAG TAG Lines
In order to detect DISC1 interactions and avoid non-specific interactions, we utilized the
CRISPR-Cas9 genome engineering system to insert a 3X-FLAG sequence (136) at the C-terminal
region of the canonical DISC1 sequence in a previously characterized control iPSC line derived
from a healthy 56-year-old male with no known psychiatric disorders (89) (Figure 3.1A, Figure
3.2A). It should be noted that the particular iPSC line used in this study carries the common
missense mutation of R264Q (rs3738401). However, this is the most common missense mutation
in DISC1 with an allele frequency of 29.23% (92) and recent studies have shown no link between
this mutation and schizophrenia (151, 152). Thus, it is highly unlikely that this common missense
mutation will have an effect in the protein interactions reported. PCR-based genotyping of
individual iPSC colonies was able to show the successful homozygous insertion of the 3X-FLAG
tag (designated DISC1
FLAG
) which was confirmed via Sanger sequencing (Figure 3.1B and 3.1C).
The top predicted off-target sequences of the guide-RNA used for gene editing were also
sequenced to confirm no undesired mutagenesis occurred within these regions (Table 3.3). The
newly derived DISC1
FLAG
iPSC line maintained expression of pluripotency markers such as OCT4
and SSEA4 (Figure 3.1D, Figure 3.2B and 3.2C).
We then differentiated the DISC1
FLAG
iPSCs along with the WT parental iPSCs into hNPCs
using a dual SMAD inhibition-based differentiation scheme (34, 139). The iPSC-derived NPCs
were positive for the NPC markers PAX6, SOX2, and NESTIN in both the DISC1
FLAG
and WT
cell lines (Figure 3.1E, Figure 3.2D and 3.2E). Further, due to the ability of DISC1 to influence
gliogenesis and proliferation of NPCs through astrocytes (122, 153), hNPCs were then

57

differentiated into S100β/GFAP+ astrocytes over a 30 day period (140) and were confirmed to be
multipotent. (Figure 3.1F, Figure 3.2F-H).
In order to visualize DISC1 hNPCs and astrocytes, we performed immunoisolation of
DISC1-FLAG followed by immunoblotting with anti-FLAG antibodies (Figure 3.1G-H). Previous
studies have shown over 50 different mRNA transcript splice variants of DISC1 in the human brain
(154) and numerous molecular weights of DISC1 have been observed in western blot assays (155).
Western blot analysis shows both cell types express a FLAG-DISC1 isoform corresponding to the
commonly reported long isoform of 98 kDa (isoform L, NCBI reference sequence
NP001158009.1) (155). We also observed additional bands between 100 and 75 kDa which may
correspond to the smaller predicted molecular weight isoform (isoform a, NCBI reference
sequence NP_001158012.1). Thus, the insertion of the 3X-FLAG tag labeled multiple isoforms
that encompass the canonical C-terminal region of DISC1.
3.3.2 - Incorporation of the 3X-FLAG Tag Does Not Alter Expression or Function of DISC1
Decreased expression of DISC1 has been observed in whole blood from patients affected by
schizophrenia (126, 156) and has been reported to modulate neural progenitor proliferation (38,
119). Therefore, we wanted to verify that the insertion of the 3X-FLAG tag did not alter DISC1
expression or hNPC proliferation. Quantification of DISC1 mRNA levels via qPCR using primers
that were specific to either a large majority of the reported DISC1 isoforms (exons 2-4) or only
those isoforms containing the conical C-terminal sequence (exons 12-13) (Figure 3.4) was able to
show no significant difference in DISC1 expression between WT and DISC1
FLAG
hNPC lines
regardless of the primer set used for qPCR analysis (Figure 3.3A). In addition, quantification of
BrdU incorporation was able to show no significant difference in proliferation between DISC1
FLAG

and parental hNPCs (Figure 3.3B-C).

58

Figure 3.1. Generation of DISC1
FLAG
human neural progenitor cell lines. (A) Gene
targeting scheme illustrating the insertion of a 3X-FLAG sequence at the C-terminal region
of DISC in induced pluripotent stem cells using the CRISPR/Cas9 genome engineering
System. The site of double strand break induction is annotated by the red triangles. (B)
Agarose gel showing PCR-based screening for the insertion of the 3X-FLAG tag sequence.
(C) Sanger sequencing conformation of the 3X-FLAG tag sequence. (D) DISC1
FLAG
iPSCs
express markers of pluripotency such as OCT4 and SSEA4. (E) DISC1
FLAG
iPSC-derived
neural progenitors express the neural progenitor markers SOX2 and NESTIN. (F)
DISC1
FLAG
Astrocytes express the astrocyte marker S100β. (G) Immunoprecipitation
followed by immunoblot of DISC1 using anti-FLAG antibodies in WT and DISC1
FLAG

hNPCs. (H) Immunoprecipitation followed by immunoblot of DISC1 using anti-FLAG
antibodies in WT and DISC1
FLAG
Astrocytes. All bands corresponding to DISC1 isoforms
were absent in WT hNPC controls and include a shift of 3.92 kDa in their molecular weight
due to the insertion of the 3X-FLAG sequence. Scale bars in (D), (E), and (F) = 100 µm.

59

Figure 3.2. DISC1 Isoforms and Characterization of WT and DISC1
FLAG
iPSC and
Differentiated Lines. (A) DISC1 isoforms annotated according to the NCBI Reference
Sequence Database (Refseq, release 81) as visualized in Integrative Genome Viewer (IGV,
Version 2.3.92). The isoforms with the canonical stop codon that have the in-frame integration
of the 3X-FLAG tag are annotated. (B and C) Immunostaining of WT and DISC1
FLAG
iPSCs
for the pluripotency markers OCT4, SSEA4, SOX2, and TRA-1-60. (D and E)
Immunostaining of WT and DISC1
FLAG
hNPCs for PAX6, SOX2, and NESTIN. (F and G)
Immunostaining of WT and DISC1
FLAG
astrocytes for the S100β and GFAP. (H)
Immunostaining of WT and DISC1
FLAG
neurons for MAP2. Scale bars = 100 µm.

60

Table 3.3. DISC1 gRNA Predicted Off-Target Sequences and Sequencing. Table contains
the top predicted off-target regions of the guide RNA used to generate DISC1
FLAG
iPSCs along
with sequencing results of each target that was PCR amplified and then verified via Sanger
sequencing.

61

In addition to effects on expression and proliferation, we wanted to ensure the 3X-FLAG tag did
not alter the localization of DISC1. Multiple studies have determined that DISC1 is localized to
the centrosome (128, 137, 157, 158) and the perinuclear region (158). These studies have indicated
that the C-terminal region of DISC1 is essential for its subcellular localization as the lack of the
C-terminal region can disrupt localization and cause DISC1 to be diffusely spread throughout the
cytoplasm (157, 158). However, we found that traditional immunofluorescence methods using an
anti-FLAG antibody were unable to visualize the endogenous tagged-DISC1 (Figure 3.4B). This
may be a result of the relatively low expression of DISC1 as it has been previously reported that
DISC1 is a low-abundance protein in mouse brain (159) and we were unable to detect endogenous
DISC1 in total cell lysate without the use of immunoisolation enrichment assays (data not shown).
Due to its low abundance, we utilized the CRISPR-VPR activation system to upregulate
the expression of DISC1 in hNPCs for the sole purpose of visualizing DISC1-FLAG (160). Here,
we employed four single-guide RNAs (sgRNAs) spread throughout the first 250 BP of the DISC1
promoter region (Figure 3.3D). With an approximate 34% transfection efficiency, this resulted in
an approximate 2-fold increase in gene expression and allowed us to detect FLAG-DISC1
localization in hNPCs (Figure 3.3E-F, Figure 3.4C-D). Co-staining with the centrosomal marker
and DISC1 interactor, AKAP9 (127), along with DAPI, was able to show DISC1 localization at
both the centrosome and peri-nuclear region in hNPCs coinciding with previous reports (128, 137,
161) (Figure 3.3). This suggests that the 3X-FLAG tag does not impair DISC1 localization.
3.3.3 - Identification of Endogenous DISC1 Interactomes
We co-immunopurified endogenous DISC1 and its binding partners in hNPCs using an
anti-FLAG-M2 antibody (73) attached to magnetic beads and performed sequential elutions using

62

Figure 3.3. Insertion of 3X-FLAG at the C-terminus of DISC1 does not alter expression
or function of DISC1. (A) DISC1 expression levels showed no significant difference between
DISC1
FLAG
hNPCs and the WT hNPC line as measured by qPCR (n = 3 biological replicates).
(B) Representative immunofluorescence images of BrdU incorporation in WT and DISC1
FLAG

hNPCs. (C) Quantification of BrdU incorporation in both WT and DISC1
FLAG
hNPC cells lines
shows no significant difference in proliferation (n = 3 biological replicates with 5 random fields
imaged and quantified for each). (D) Schematic illustrating the position of guide RNAs used
to upregulate expression of DISC1 relative to the transcription start site of DISC1. (E) qPCR
demonstrating up-regulation of DISC1 following CRISPR activation. (F) Immunofluorescence
following upregulation of DISC1 using the CRISPR activation system in WT and DISC1
FLAG
hNPCs. Images display co-staining of DAPI and the centrosomal marker, AKAP9, along with
FLAG (DISC1). Data in panels (A) (C), and (E) are plotted as mean ± SD. Scale bars in (B) =
100 µm. Scale bars in (F) = 20 µm.

63

Figure 3.4. qPCR and Upregulation of Endogenous DISC1. (A) Diagram displays location
of primers used for qPCR assays to determine levels of endogenous DISC1 in WT and
DISC1
FLAG
hNPCs. (B) Traditional immunofluorescence in DISC1
FLAG
and WT hNPCs using
the anti-FLAG antibody is unable to detect DISC1 above background levels. (C) GFP
expression in hNPCs following nucleofection with Cas9-GFP plasmid. (D) Quantification of
GFP expression in hNPCs following nucleofection. Graph displays results of two
independent nucleofections. Error bar is ± SD. Scale bars = 100 µm.

64

an anti-3X-FLAG peptide followed by a final release of all bound protein via boiling of the
magnetic beads. Each step was subject to immunoblotting in order to measure the efficiency, yield,
and specificity of DISC1 protein elution (Figure 3.5A, right panel). We observed that more than
90% of DISC1 could be released within a single elution. This was followed by identification of
DISC1 interactors via high performance liquid chromatography (HPLC) coupled to tandem mass
spectrometry (HPLC-MS/MS).
DISC1 hNPC-interactors were considered if they were identified with multiple unique
peptides in duplicate assays and no peptides in controls (36 interactors). However, because of the
low expression level of DISC1, this stringent criteria might impair the identification of true DISC1
interactors. Thus, we also considered interactors with different levels of confidence bringing the
number of interactors identified via mass spectrometry to a total of 107 interactors (Figure 3.5A
left panel, Table 3.4). In order to confirm DISC1 interactions and the validity of our approach, we
performed DISC1 immunoisolation followed by western blot analysis from DISC1 interactors that
we identified with different levels of confidence (Figure 3.5B). Moreover, we were also able to
confirm several previously reported DISC1 protein interactors including TNIK, PDE4DIP,
CEP170, and AKAP9 bringing the total number of interactors to 112 (Figure 3.5A-B).
Inference of hNPC DISC1 localization through its interactome indicates DISC1 associates
with proteins in different subcellular structures such as the nucleus, centrosome, endoplasmic
reticulum (ER), and the cytoplasm. Gene ontology (GO) enrichment analysis of DISC1 interactors
provides further evidence for this as terms such as “cytosol”, “ER”, and “nucleoplasm” are
significantly enriched (P < 0.05 for all, Bonferroni corrected) (Figure 3.5E, Table 3.5). This
corresponds with previous reports of DISC at the ER and nucleus in neuronal systems (162, 163).

65

We then identified DISC1-interacting partners in astrocytes differentiated from hNPCs.
Here, we were able to identify 51 interactors that were categorized according the various
confidence levels as was completed for hNPCs (Figure 3.5C, Supplemental Table 3.6). Western
blot analysis of DISC1 interactors was able to confirm a number of the identified interactors
including NDE1, GRIPAP1, and GSN. In addition, we were again able to confirm reported DISC1
interactors including centrosomal proteins LIS1 and NDEL1, bringing the total number of
interactors to 56 (Figure 3.5D). This Indicates that a core DISC1 function, such as modulation of
centrosomal dynamics, is conserved in astrocytes. The astrocyte interactome also shows an
increased number of cytoskeletal proteins which is corroborated through GO enrichment analysis
of DISC1 astrocyte interactors. Here, terms such as “actin filament binding,” “focal adhesion,”
and “actin cytoskeleton” are significantly enriched (P < 0.05 for all, Bonferroni corrected) (Figure
3.5F, Table 3.7), indicating that DISC1 may also play a role in astrocyte cytoskeletal organization
and therefore, astrocyte vesicle mobility and trafficking.
While the enrichment of cytoskeletal proteins may indicate a cell type-specific function or
localization of DISC1 in astrocytes, “Poly(A) RNA binding” is significantly enriched in both cell
types (P < 0.05 for all, Bonferroni corrected) (Figure 3.5E-F, Table 3.5, Table 3.7). A recent report
characterized DISC1 as an RNA binding protein and showed its involvement in RNA transport in
adult hippocampal dendrites (138), suggesting a common regulatory function in hNPCs,
astrocytes, and neurons. This is exemplified by the association of DISC1 with RNA binding
proteins, including members of the nuclear pore complex (e.g. NUP210, NUP98, and RANGAP1)
in hNPCs, along with ALYREF which is involved in nuclear mRNA transport and is a common
DISC1 interactor in both hNPCs and astrocytes.

66

Figure 3.5. Identification of the endogenous DISC1 Interactome in hNPCs and astrocytes.
(A) Right Panel - Western blot following immunoprecipitation and elution of DISC1 from
DISC1
FLAG
and WT hNPCs using an anti-3X-FLAG peptide. Comparing elutions shows that a
majority of DISC1 is eluted within the first elution. Lane 1 = first elution, Lane 2 = second
elution, Lane 3 = boiling of magnetic beads post elutions. Left Panel - DISC1 Interactome in
hNPCs as characterized via immunoprecipitation followed by mass spectrometry or western
blotting. Criteria for inclusion is indicated by colored border of nodes. (B) Western blots in
hNPCs showing conformation of proteins identified via immunoprecipitation followed by mass
spectrometry or previously identified protein-protein interactions involving DISC1. (C) DISC1
Interactome in astrocytes as characterized via immunoprecipitation followed by mass
spectrometry or western blotting. (D) Western blots in astrocytes showing conformation of
proteins identified via immunoprecipitation followed by mass spectrometry or previously
identified protein-protein interactions involving DISC1. (E) Gene ontology analysis of the
identified DISC1 interactome in hNPCs. (F) Gene ontology analysis of the identified DISC1
interactome in Astrocytes.

67

Table 3.4. DISC1 Interactors Identified via HPLC-MS/MS in hNPCs. Table contains
interactors of DISC1 identified via HPLC-MS/MS in hNPCs. Identified interactors contain the
confidence group in which the interactor was assigned to, the number of unique peptides in each
run and if applicable, the number of unique peptides identified in controls.

68

Table 3.4. Continued

69

Table 3.5. Gene Ontology Enrichment Within the DISC1 hNPC Interactome. Table
contains results of gene ontology enrichment analysis within the endogenous DISC1
interactome in hNPCs. Results were obtained using the DAVID database.

70

Table 3.6. DISC1 Interactors Identified via HPLC-MS/MS in Astrocytes. Table contains
interactors of DISC1 identified via HPLC-MS/MS in astrocytes. Identified interactors
contain the confidence group in which the interactor was assigned to, the number of unique
peptides in each run and if applicable, the number of unique peptides identified in controls.

71

Table 3.7. Gene Ontology Enrichment Within the DISC1 Astrocyte Interactome. Table
contains results of gene ontology enrichment analysis within the endogenous DISC1
interactome in astrocytes. Results were obtained using the DAVID database.

72

We identified multiple unique DISC1 peptides accounting for 21.66% and 34.13%
coverage of the DISC1 canonical (L) isoform in hNPCs and astrocytes, respectively (Table 3.8).
This relatively low coverage, along with the inability to detect DISC1 above background with
traditional immunofluorescence or immunoblotting assays, is in agreement with low expression
levels as previously reported (159). Characterization of the expression of DISC1 in hNPCs and
astrocytes relative to the identified DISC1 interactors using publicly available data that directly
compares the expression of genes in both iPSC-derived hNPCs and astrocytes (140) was able to
show over 90% of DISC1-interacting proteins have higher expression levels than DISC1 in both
cell types (Figure 3.6). While transcript levels may not always be directly correlated to protein
levels, our MS results are in agreement with DISC1 low expression levels in hNPCs and astrocytes.
This suggests that for a particular interactor, only a small fraction of the overall protein levels
might be interacting with DISC1 at any given time, indicating DISC1 may interact at sub-
stoichiometric ratios.
3.3.4 - The Endogenous DISC1 Interactome and Previously Reported Protein-Protein
Interactions
With this being the first report of endogenous DISC1 protein-protein interactions in iPSC-
derived hNPCs and astrocytes, we wanted to compare our data with what has been previously
reported. Here, we curated a list of previously reported DISC1 interactions (Methods) and were
able to identify a total of 9 proteins in hNPCs that we confirmed either by mass spectrometry or
via immunoprecipitation followed by western blot and 9 proteins in astrocytes (Table 3.10). We
also observed an additional 7 and 3 proteins derived from yeast two-hybrid assays present in at
least one sample in hNPCs and astrocytes, respectively (Table 3.10).

73

Table 3.8. Endogenous DISC1 Peptides Identified via HPLC-MS/MS. Table contains high
confidence peptides of DISC1 identified via HPLC-MS/MS in hNPCs and astrocytes.
Identified peptides are denoted by the run in which they were identified, the number of peptide
spectrum matches (PSMs), number of miss cleavages, along with the amino acid start and
finish position relative to the conical isoform of DISC1.

74

Since a large number of protein interactors in hNPCs and astrocytes were novel, we
hypothesized that this may be due to cell-type specific interactions and DISC1 stoichiometry
ratios. To test this, we performed immunoprecipitation of the full-length recombinant DISC1 in
adult mouse cortex. This is comparable to previous studies that have characterized DISC1
interactions with respect to both non-stoichiometric conditions and developmental context. As
opposed to the 20 – 30% coverage obtained under endogenous conditions, we obtained close to
100% using the recombinant protein. Under these conditions, we identified 148 proteins in DISC1
immunoprecipitation experiments and absent in corresponding controls, several of which were
verified via western blot (Figure 3.7A-B, Table 3.9). As opposed to 9 proteins identified in NPCs
or astrocytes, we identified a total of 42 previously reported interactions using recombinant DISC1
in a tissue similar to what has been used previously (Table 3.10).
We also found a number of related proteins to previously reported interactions in both structure
and function (Table 3.11). For example, while the nuclear pore protein, NUP160, was previously
Figure 3.6. Expression of DISC1 Binding Partners in hNPCs and Astrocytes. (A)
Expression of DISC1 relative to the DISC1-hNPC interactome in iPSC-derived hNPCs. (B)
Expression of DISC1 relative to the DISC1-astrocyte interactome in iPSC-derived astrocytes.
All data is derived from Julia et al., 2017 (133) and was obtained from processed data under
GEO accession: GSE97904. Blue circles indicate DISC1 while red circles indicate DISC1
binding partners.

75

characterized via yeast-two hybrid and detected in the DISC1 recombinant IP, we identified
NUP210 and NUP98 as interactors of DISC1 in hNPCs. This highlights the role cellular context
plays in determining protein interactions.
3.3.5 - Clustering of DISC1 Interactors and Psychiatric Disease
Next, we asked if the endogenous DISC1 interactomes could cluster in subsets of protein
interactions that can indicate both protein functions and complexes that DISC1 may associate with.
Here, we extracted reported protein interactions for each DISC1 interactor identified in our screens
and clustered them in a hNPC- or astrocyte-DISC1 protein interaction network. We identified a
number of molecular functions and cellular locations where DISC1 interactors associated in
interconnected clusters of protein interactions. Both hNPCs and astrocyte interactomes contain
Figure 3.7. Recombinant DISC1 Interactions Identified in Adult Mouse Cortex. (A)
DISC1 Interactome in adult mouse cortex as characterized via immunoprecipitation followed
by mass spectrometry. Previously reported DISC1 interactors along with protein family
members of previously reported interactors are noted. (B) Western blots showing
conformation of proteins identified via immunoprecipitation followed by mass spectrometry
involving recombinant DISC1 in adult mouse cortex.

76

Table 3.9. DISC1 Interactors Identified via HPLC-MS/MS Using Recombinant DISC1 in
Adult Mouse Cortex. Table contains interactors of recombinant DISC1 identified via HPLC-
MS/MS in adult mouse cortex. Identified interactors contain the number of unique peptides in
each run and if applicable, the number of unique peptides identified in controls.

77

Table 3.10. Previously Reported Interactions Detected in DISC1 Interactomes. Table
contains interactors identified in DISC1-hNPC, -astrocyte, or -recombinant interactomes in
common with previously reported DISC1 interactions.

78

Table 3.11. Protein Family Members of Previously Reported DISC1 Interactions
Detected in DISC1 Interactomes. Table contains interactors identified in DISC1-hNPC, -
astrocyte, or -recombinant interactomes within the same protein families as previously
reported DISC1 interactors.

79

the centrosomal interactome including TNIK, PDE4DIP, and CEP170 along with proteins that
clustered into functions such as RNA binding and transport (Figure 3.8A, Table 3.12). hNPCs
interactions also clustered in the peri-nuclear region including the nuclear pore and nuclear lamina
complexes. In contrast, DISC1 interactors identified in astrocytes clustered into complexes
involving cytoskeletal processes such as actin-related proteins or the spectrin complex (Figure
3.8A, Table 3.12). While clusters of proteins involved in centrosome dynamics have been
previously described, a majority of the clusters uncovered are novel and include proteins
characterized as contributing to psychiatric disorders such as GATAD2B and FOXP1 which were
identified in hNPCs (164, 165).
DISC1 has been associated with numerous psychiatric disorders (116, 118, 166, 167).
However, the lack of association to schizophrenia through both genome wide association (GWAS)
and exome sequencing studies has brought the applicability of DISC1 to psychiatric disorders into
question (168, 169). Irrespective of this, proteins may be able to regulate disease-relevant
pathways through the modulation of protein interactions and function in protein networks where
disease risk factors are associated, as the case has been made for DISC1 (125). Therefore, we
tested the hNPC, astrocyte, and recombinant DISC1 interactomes for enrichment in proteins
affected by de novo mutations, separated by mutation class, in Schizophrenia (SCZ), Autism
Spectrum Disorder (ASD), Intellectual Disability (ID), Congenital Heart Disease (CHD), and an
aggregate of all controls in denovo-db (91) (Figure 3.8B). For both the astrocyte and recombinant
DISC1 interactomes, there was no statistically significant enrichment in any category tested.
However, in the hNPC DISC1 interactome, we did find enrichment of nonsynonymous mutations
in SCZ following correcting for multiple comparisons (nominal p-value = 1.84 x 10
-3
, Bonferroni
threshold p < 2.40 x 10
-3
). ID also presented enrichment in nonsynonymous mutations, but did not

80

pass the Bonferroni threshold (nominal p-value = 3.34 x 10
-3
) (Figure 3.8B). While the SCZ
enrichment was driven by a diverse set of proteins, the non-significant enrichment in ID was driven
by proteins that were found to have recurrent mutations including FOXP1, GATAD2B, and USP7
(Figure 3.8B). This Suggests that DISC1 function within hNPCs may contribute more to the risk
of disease when compared to DISC1 function in astrocytes or in the adult cortex.
In addition to proteins that have been implicated through exome-sequencing studies, we
also focused on proteins known to be syndromic for neurodevelopmental disorders as classified
by the Online Mendelian Inheritance in Man (OMIM) database (Figure 3.8C). Here, DISC1
interactors included the previously described DISC1 interacting kinases CIT (microcephaly) and
TNIK (ID) along with NDE1 (Microhydranencephaly) and PAFAH1B1 (Lissencephaly). Novel
interactors that are also known to by syndromic include the transcription factors FOXP1 (ID and
ASD) and GATAD2B (ID), USP7 (ID and ASD), PGAP1 (ID), SGSH (Mucopolysaccharidosis
type IIIA), and EMC1 (Developmental Delay). This is in accordance with our previous findings,
showing that the interactomes of molecules involved in developmental disorders, such as TNIK
(83), show an enrichment in de novo mutations in a variety of complex brain disorders (53).
3.3.6 - Mutations in DISC1 Interactors Regulate Shared Cellular Functions
Proteins associated in protein interaction networks are usually involved in the regulation
of shared cellular processes and mutations in genes encoding for these proteins are more likely to
be associated to the same disease (170-172). With several novel DISC1 interactors identified in
iPSC-derived hNPCs and astrocytes, we also wanted to determine if disruptions in DISC1 affect
the expression levels of the newly identified interactors. We compiled a list of differentially
expressed genes or proteins from multiple studies (37, 120, 173, 174) and compared these against
the newly identified interactors. While these studies examined a variety of DISC1 mutations or

81

Figure 3.8. Clustering and Enrichment of DISC1 Interactome in Psychiatric Disease Risk
Factors. (A) Clustering of the endogenous DISC1 hNPC and astrocyte interactomes based on
protein-protein interactions derived from public databases. (B) Enrichment of the DISC1
hNPC interactome in proteins found to have de novo mutations in patients affected by
schizophrenia (SCZ), intellectual disability (ID), autism spectrum disorder (ASD), congenital
heart disease (CHD), along with all controls listed in denovo-db. Variant classes are split
between missense mutations, protein truncating variants (PTV), nonsynonymous mutations
(includes missense and protein truncating variants), and synonymous mutations. Error bars
represent the 95% confidence interval and nominal p-values are displayed in bold for each test.
(C) Heatmap representing proteins implicated in contributing to neurological disorders via
OMIM along with proteins containing nonsynonymous de novo mutations in SCZ, ASD, or
ID when found in two or more gene lists.

82

Table 3.12. Clustering of Endogenous DISC1 Interactomes. Table contains proteins and
clusters presented in Figure 3.8A.

83

manipulations of DISC1 protein levels, several of the novel DISC1 binding proteins described here
were also found to be differentially expressed. For example, Wen et al., 2014 (120) examined
iPSC-derived forebrain neurons derived from patients with a 4 bp frameshift mutation in exon 12
of DISC1 that affects the isoforms of DISC1 that are also selectively tagged in this study.
Differentially expressed genes found under these conditions contain the
neurodevelopmental proteins GATAD2B and FOXP1 among others. We asked whether or not the
endogenous DISC1 interactomes were enriched in these genes more than the recombinant
interactome in adult mouse cortex or all previously reported interactions in DISC1. While we
found no statistically significant enrichment in previously reported interactions or the recombinant
interactome, the hNPC interactome was enriched in all differentially expressed genes and the
astrocyte interactome was enriched in down-regulated genes (P = 0.021 and 0.010, respectively,
Fishers Exact Test, Bonferroni corrected).
In the only study to examine changes in protein expression in primary astrocytes expressing
a dominant negative form of DISC1, several cytoskeletal proteins were found to undergo changes,
including the newly identified DISC1-interacting protein CKAP4 (174). Therefore, a number of
alterations in DISC1 regulate expression levels of DISC1 interactors and are associated in the
regulation of common cellular processes.
In addition, with DISC1 interacting with several centrosomal proteins in hNPCs, we
hypothesized that similarly to DISC1, they might also affect hNPC proliferation. One of the
proteins, the TRAF and Nck interacting Kinase (TNIK), has also been involved in a number of
psychiatric disorders and intellectual disability (83, 175-177). This suggests that TNIK mutations
might also be associated with alterations in cellular proliferation in hNPCs. Thus, we introduced a
truncating mutation at the Citron homology domain (CNH) domain of TNIK, which resulted in

84

TNIK knockout cell line (Figure 3.9A). Compared to WT cell lines, we found a marked decrease
of proliferation in TNIK KO hNPCs, as measured by a 40% decrease in phospho-histone H3
positive cells (p < 0.001, Student’s t Test) (Figure 3.9B-C). Interestingly, it has been reported that
DISC1 may inhibit TNIK function in mature neurons in vitro (178). However, our results indicate
a synergistic interaction of DISC1 and TNIK where both molecules are needed to regulate hNPC
proliferation, highlighting the importance of the identification of cell- specific protein networks.

85

Figure 3.9. TNIK KO hNPCs Have Decreased Proliferation. (A) Immunoprecipitation
followed by western blot of TNIK in TNIK
-/-
and WT hNPCs. (B) Immunofluorescence images
showing staining for the proliferation marker phopsho-histone H3 (pHH3) in WT and TNIK
-/-

hNPCs. Scale bars = 100 µm. (C) Quantification of pHH3+ hNPCs. Data are derived from
three independent differentiations. Error bars are ± SD. *** = p<0.001, Student’s t test.

86

3.4 - Discussion
Through the combination of genome-engineering and mass-spectrometry, we are able to
report the first endogenous interactome of DISC1 in human iPSC-derived NPCs and astrocytes.
While we ensured that there were no transcriptional changes at the mRNA level due to the insertion
of the endogenous tag, we should note the unlikely possibility of an endogenous tag interfering
with protein expression levels, distribution, or protein interactions. Nonetheless, the novel
interactions identified here can be categorized into numerous functional modules that are in line
with previously reported functions of DISC1 and expand upon them. For example, DISC1 is
associated with a number of proteins that are known to reside at the centrosome including AKAP9,
CEP170, NDEL1, LIS1. Our results are able to include molecules such as PDE4DIP into the
DISC1/centrosome protein interaction network in both hNPCs and astrocytes.
Using the DISC1 interactions determined in hNPCs, astrocytes, the recombinant protein,
or all previously reported interactions, we were able to show that only the DISC1 hNPC
interactome is enriched in nonsynonymous mutations as a whole identified in sporadic cases of
SCZ. Determining DISC1 interactors in hNPCs is also able to provide context to
neurodevelopmental phenotypes observed in previous studies. For example, clinically relevant
mutations in DISC1 created via genome engineering in iPSCs were shown to result in differential
sup-populations of neural progenitors following differentiation (37). Here, we show that DISC1
binds to nuclear pore proteins such as NUP210 and NUP98, which have both been shown to
influence neural differentiation (179, 180). While relatively little is known concerning the function
of DISC1 in astrocytes, it has been shown to influence gliogenesis of primary NPCs (153),
indicating a common function of influencing differentiation potential. Here, the common

87

functional module of regulatory RNA-binding proteins between hNPCs and astrocytes may play a
role in this process.
While DISC1 has been regularly described as a scaffold protein, the expression of DISC1
appears to be extremely low which is corroborated through both transcriptomic and proteomic
data. Mass spectrometry data derived from public repositories such as PeptideAtlas do not contain
any DISC1 peptides observed in any neural cell types or brain tissue as a whole (181). Thus, the
sub-stoichiometric characteristics of DISC1 protein interactions suggests that DISC1 binds and
affects only a specific subset of its interactors at any given time.
Our results are also able to provide the first confirmation via immunoprecipitation followed
by western for DISC1 interactors identified in Y2H screens including AKAP9, CEP170, and
GRIPAP1. However, comparison of the endogenous interactome of DISC1 in iPSC-derived
hNPCs and astrocytes to the multiple DISC1 yeast-two hybrid screens is able to show that a vast
majority of the interactions detected here were not apparent in those studies. While protein
interactions derived from affinity purification may include secondary or tertiary interactions
compared to binary interactions identified via yeast-two hybrid assays, these assays do not take
context or stoichiometric ratios into account. In contrast to the endogenous interactions, when we
performed immunopurification of recombinant DISC1 in a non-stoichiometric fashion and in an
experimental setting resembling what has been previously published, we found an abundance of
previously reported interactions. This underscores the importance of defining endogenous protein-
protein interactions within a particular cellular context of interest with the correct stoichiometric
ratios between interacting proteins.

88

Chapter 4: Synaptic GAP and GEF Complexes Cluster Proteins Essential for GTP
Signaling
4.1 - Introduction
Small GTPases are molecular switches that can rapidly interconvert between two conformational
states, depending on association to guanosine triphosphate (GTP) or guanosine diphosphate (GDP)
(182). The cellular actions of GTPases are frequently initiated by GTP binding, enabled by guanine
nucleotide exchange factors (GEFs), and finalized by GTP hydrolysis facilitated by GTPase-
activating proteins (GAPs) (182, 183). The large number of existing GTPases require a multitude
of GEFs and GAPs that utilize a variety of combinatorial mechanisms to achieve functional
specificity (184, 185). These proteins not only need to respond to different inputs, but also direct
their output to specific small GTPases. Proteins containing GAP and GEF domains are usually
distributed in families containing a variety of protein domains (185, 186). These domains are
arranged in multi-domain architectures including a family specific GAP/GEF domain together
with domains that have the capacity to respond to different chemical messages and associate in
protein complexes through protein-protein interactions (185, 186).
Small GTPases can regulate a variety of cellular functions. In neurons, they have been
shown to be involved in different aspects of synaptic function such as regulation of the actin
cytoskeleton, spine remodeling, and synaptic plasticity (187-190). Proteomics analyses have
localized a number of GAPs and GEFs at the postsynaptic specialization of glutamatergic
excitatory neurons known as the postsynaptic density (PSD) (50, 52). The PSD contains a
collection of more than 1500 proteins arranged in macromolecular complexes and many of these
proteins have been linked to a variety of brain disorders (18, 48, 51, 52, 191, 192). PSD protein
complexes link glutamate receptors to downstream signaling pathways through a number of

89

scaffold proteins with multi-modular protein domain architectures (51, 52). However, it is not
known how synaptic GAPs and GEFs are organized within this signaling machinery, if they have
overlapping binding partners, and if they also associate to brain disease risk factors.
Here, using immunoisolation and high performance liquid chromatography tandem mass
spectrometry (HPLC-MS/MS), we determined the interactomes of the RasGAP Syngap1, the
RhoGEF Kalirin, and the ArfGAP Agap2 in PSD fractions of adult mouse cortex. These proteins
are able to interact with each other at the PSD and have also been described to be involved in
intellectual disability and psychiatric disease (193-197). We determined 281 protein interactions
within the PSD involving these proteins. We described the functional properties and organization
of these complexes, along with their role in the organization of GAP and GEF families within the
PSD. Because synaptic proteins may function within multiple subcellular locations, we utilized
Agap2 as an example of this scenario and determined the interactome of Agap2 in non-PSD
fractions, which involved an additional 110 interactions. By comparing domain architecture and
binding interactions in different subcellular fractions, we illustrate both common and unique
functions of Agap2 based on its cellular localization. With Agap2, Kalirin, and Syngap1 being
implicated in contributing to the risk of psychiatric disease, we also show that they interact with
and cluster proteins involved in autism spectrum disorder (ASD), schizophrenia (SCZ), and
intellectual disability (ID).

90

4.2 - Materials and Methods
Postsynaptic Density Preparations
Postsynaptic density preparations were performed as described (88, 198). Briefly, adult mouse (3-
4 months in age) cortex was homogenized in sucrose buffer (0.32 M sucrose, 10 mM Hepes buffer
(pH 7.4), 2 mM EDTA, 30 mM NaF, 20 mM β-glycerol phosphate, 5 mM sodium orthovanadate,
and Roche cOmplete protease inhibitor cocktail) and centrifuged at 500g for 6 minutes.
Supernatant was collected and then spun at 10,000g for 10 minutes. The resulting pellet was
solubilized in triton buffer (50 mM Hepes (pH 7.4), 2 mM EGTA, 2 mM EDTA, 50 mM NaF, 20
mM β-glycerol phosphate, 5 mM sodium orthovanadate, Roche cOmplete, and 1% Triton X-100.
The solubilized pellet was centrifuged at 30,000 rpm for 30 min and supernatant was collected for
non-PSD fractions. The resulting pellet was collected and solubilized in DOC buffer (50 mM Tris
(pH 9), 30 mM NaF, 5 mM sodium orthovanadate, 20 mM β-glycerol phosphate, 20 µM ZnCl2,
Roche cOmplete, and 1% sodium deoxycholate) and served as the PSD fraction.
Immunoprecipitation
Immunoprecipitation experiments were performed as previously described (52, 107). Lysate
containing 2 mg of total protein was incubated with the indicated primary antibody at a
concentration of 1 µg/µl at 4 degrees Celsius overnight with rotation. The following day, IPs were
incubated with Dynabeads protein G (Novex, Thermo Fisher Scientific, Waltham, MA) for 2 hours
at 4 degrees Celsius with rotation. IPs were washed three times with IP wash buffer (25 mM Tris
(pH 7.4), 150 mM NaCl, 1 mM EDTA, and 1% Triton X-100 or sodium deoxycholate where
appropriate). IPs were re-suspended in 2X LDS sample buffer and incubated at 95 degrees Celsius
for 15 minutes to elute protein complexes. The elutant was incubated with DTT at a final

91

concentration of 1 mM at 56 degrees Celsius for 1 hour followed by incubation with Iodoacetamide
at a final concentration of 20 mM at room temperature for 45 minutes. Primary antibodies used
for immunoprecipitation in this study included PIKE (Agap2) (Bethyl Laboratories, catalog
#A304-262A), Kalirin (Millipore, catalog #07-0122), Syngap1 (Cell Signaling Technologies,
catalog #5539), Rabbit IgG Isotype Control (ThermoFisher, catalog #06-6102) and GST
(NeuroMab, catalog #75-148). All antibodies were used as obtained from the manufacturer.
Mass spectrometry sample preparation
Samples were prepared for mass spectrometry as described in Li et al.(52). Briefly, samples were
separated on 4-12% Bis-Tris gels (NuPAGE Novex, Thermo Fisher Scientific) followed by
staining with InstantBlue (Expedeon, San Diego, CA) for protein visualization. Following
destaining, lanes were cut and placed in 96-well perforated plates for destaining and peptide
digestion via trypsin at 37 degrees Celsius overnight. Peptides were then extracted with
acetonitrile, dried down using a Savant SPD 1010 SpeedVac Concentrator (Thermo Fisher
Scientific), and then suspended in 3% ACN/0.1% FA. A Nano/Capillary LC System Ultimate
3000 (Thermo Fisher Scientific/Dionex) was used for desalting and reverse-phase separation of
peptides. The LC system was coupled to a hybrid linear ion-Fourier transform ion cyclotron
resonance LTQ-FT (FTICR) 7 Tesla mass spectrometer (LC/MS) for data acquisition.
Data analysis
Proteome Discoverer 1.4 (Thermo Fisher Scientific) was used to process MS data which was
analyzed using both the Sequest and Mascot V2.4 (Matrix Science, Boston, MA) against a
modified mouse database from Uniprot combined with its decoy database. With respect to analysis
settings, the mass tolerance was set 10 parts per million for precursor ions and 0.8 daltons for

92

fragment ions, no more than two missed cleavage sites were allowed, static modification was set
as cysteine carboxyamidation, and dynamic modification was set as methionine oxidation. False
discovery rates (FDRs) were automatically calculated by the Percolator node of Proteome
Discoverer and a peptide FDR of 0.01 was used for cut-offs. Peptides with high confidence were
considered as true hits and proteins with at least two different peptides were accepted. Protein
Interactions were considered positive if a minimum of two peptides were present in at least two
assays and absent in anti-GST controls. The datasets generated and/or analyzed during the current
study have been deposited to the ProteomeXchange Consortium via the Proteomics
IDEntifications (PRIDE) partner repository under the dataset identifier PXD006326 (144).
Western blotting
Immunoprecipitation was carried out as described above in non-PSD (triton soluble) for Agap2
and PSD (DOC soluble) fractions for Agap2, Kalirin. and Syngap1. For conformation of protein
interactions, total lysate (20 µg) and a control rabbit IgG immunoprecipitation were ran alongside
Agap2, Kalirin, or Syngap1 immunoprecipitation experiments. Protein was loaded onto 4 – 12%
Bis-Tris gels and separated at 135V for 1.5 hours. Proteins were transferred to a PVDF membrane
using a BIO-RAD Trans Blot Turbo System. Membranes were then blocked using 5% bovine
serum albumin (BSA) in 0.05% TBST (TBS-Tween 20) for 1 hour at room temperature and then
incubated with the primary antibody at a 1:1000 dilution overnight at four degrees Celsius.
Primary antibodies include PIKE (Agap2) (Bethyl Laboratories, catalog #A304-262A), Syngap1
(Cell Signaling Technologies, catalog #5539), Kalirin (Millipore, catalog #07-0122), Camk2a
(ThermoFisher, catalog #13-7300), Cit (Bethyl Laboratories, catalog #A302-303A), Dlg4
(ThermoFisher, catalog #MA1-045), Fmr1 (Abcam, catalog #ab17722), Git1 (NeuroMab, catalog
#75-094), NR2B (NeuroMab, Catalog #75-097), Gsk3β (Cell Signaling Technologies, catalog

93

#9832), mGluR2 (Cell Signaling Technologies, catalog #76012), mGluR5 (Cell Signaling
Technologies, catalog #55920), Nckap1 (Proteintech, catalog #12140-1-AP), and Tnik (Bethyl
Laboratories, catalog #A302-695A). All antibodies were used as obtained from the manufacturer.
The following day, membranes were washed with 0.05% TBST four times for 10 minutes each,
incubated with the respective secondary antibody for 1 hour, and subsequently washed 4 time for
5 minutes each. Western blots were incubated with Pierce ECL Western Blotting Substrate
(Thermo Fisher Scientific) for 5 minutes and then imaged using a Carestream Image Station 4000
MM Pro.
Over-representation analyses
Enrichment of GAP and GEF domain containing proteins within PSD protein complexes was
carried out using the one-tailed Fisher’s exact test with the option “greater,” followed by the
Bonferroni correction for multiple comparisons using the (fisher.test) and (p.adjust) packages in
R, respectively. The total number of proteins within the PSD served as the background.
SMART protein domain enrichment was completed using the Database for Annotation,
Visualization, and Integrated Discovery (DAVID) version 6.8 (94, 199) using the default settings
and the number of protein coding genes expressed in purified mouse cortical neurons derived from
Zhang et al. 2014 (200) as the background for each gene list analyzed. To obtain the background,
we accessed processed transcriptomic data from NCBI Gene Expression Omnibus (GEO)
accession number GSE52564 and compiled a list of genes that were consistently expressed at a
statistically significant level as defined by Zhang et al. 2014 (200) (FPKM > 0.1) between
biological replicates. After filtering out non-coding genes, this resulted in a list of 13,761 protein
coding genes. Only domains that reached statistical significance following Bonferroni correction
of multiple comparisons are reported in the text.

94

For disease gene set enrichment, we incorporated several different lists encompassing genes that
are likely to contribute to autism spectrum disorder (ASD), schizophrenia (SCZ), and intellectual
disability (ID). These included: 1) Supplementary table 15 from Turner et al. 2016 (193) which
encompasses a list of 57 genes found to have recurrent de novo likely gene disrupting (LGD)
mutations in probands with autism from a collection of exome sequencing studies (12, 13), 2)
Supplementary table 7 from Iossifov et al. (13) which contains 353 genes harboring validated de
novo LGD mutations in probands with autism, 3) Supplementary table 3 from De Rubeis et al.
(12) which included 107 genes with an FDR < 0.3 implicated in contributing to autism through the
TADA statistical model, 4) SFARI gene database (201) as of February 2016 which was restricted
genes containing the SFARI gene rankings 1 – 4 (strong evidence to minimal evidence) and S
(syndromic) 5) Supplementary table 5 from Fromer et al. (18) which contains de novo SNVs
identified in probands with SCZ excluding silent mutations, 6) Supplementary table 3 from Ripke
et al. (10) restricted to loci reaching genome wide significance in patients with SCZ and
implicating a single gene, 7) Supplementary table 4 from Leliveld et al. (19) which contains a list
of 1537 genes known to contribute to ID. For all tests, protein coding genes expressed in purified
mouse cortical neurons derived from Zhang et al (200)

served as the background measured by one-
tailed Fisher’s exact test with the option “greater,” followed by the Bonferroni correction for
multiple comparisons using the (fisher.test) and (p.adjust) packages in R, respectively.

95

4.3 - Results
4.3.1 - Distribution of GAPs and GEFs Within the Postsynaptic Density
In order to explore the PSD signaling landscape of postsynaptic GEF and GAP proteins, we first
determined the distribution of GAP and GEF protein domains within PSD proteins. We extracted
protein domains from SMART and Pfam databases (202, 203) and mapped them to 1524 PSD
proteins isolated from mouse cortex and identified by HPLC-MS/MS. From this, we determined
59 PSD proteins containing GAP and GEF domains including 29 GAPs and 30 GEFs (Table 4.1).
GAP proteins were distributed among the subfamilies of GTPases, including 10 RhoGAPs, 8
ArfGAPs, 7 RasGAPs, 3 RapGAPs and 1 RabGAP (Figure 4.1A). Although RabGAP domain
containing proteins are the most abundant within the family of mammalian GTPases (185), the
fact that the Arf, Ras, and Rho GAPs comprise 86% of the total GAPs within the PSD suggests
that PSD signaling mechanisms may preferentially use these three sub-families (Figure 4.1A).
GEF domain containing proteins had a more even distribution with RhoGEF domains (37%) being
the most abundant GEF family members at the PSD, followed by SEC7 (30%), DOCK (20%) and
RasGEF (13%) (Figure 4.1B).
One characteristic of the mammalian PSD is the organization of signaling proteins in
macromolecular protein complexes (48, 51, 52). These protein complexes use scaffold proteins as
master-organizers of protein interactions, based on their multi-modular protein domain
composition (51, 204). Several proteins containing GAP and GEF domains also contain a variety
of other protein domains, forming multi-domain architectures (186). Therefore, we determined the
protein domain composition of GAP and GEF proteins in the PSD (Table 4.1). While PH domains
were the most abundant domain co-occurring with GAPs and GEFs, we also determined the
presence of a variety of protein domains whose main function is to organize protein-protein

96

Figure 4.1. Distribution of GAPs and GEFs at the PSD. (A) Distribution of GTPase activating proteins
(GAPs) at the postsynaptic density (PSD). (B) Distribution of guanine nucleotide exchange factors (GEFs)
at the PSD. (C) Distribution of protein domain architecture of GAPs at the PSD. (D) Distribution of protein
domain architecture of GEFs at the PSD. (E) Domain architecture of the GAP and GEF proteins used in
this study.

97

Table 4.1. Domain Architecture of GAPs and GEFs at the PSD. Table contains GAP and
GEF proteins identified in PSD fractions of adult mouse cortex along with their associated
protein domain structure. Each protein contains their associated gene symbol, MGI ID,
associated gene name, the GAP or GEF domain present within the respective protein, followed
by other protein domains co-occurring with the GAP or GEF domains.

98

interaction modules (Figure 4.1C-D). Therefore, the co-occurrence of GAPs and GEFs with
protein domains such as SH3, WW, ANK and PDZ suggests that PSD GTPases and exchange
factors can also be associated with and incorporated into PSD protein complexes.
4.3.2 - Interactomes of Agap2, Syngap1, and Kalirin at the Postsynaptic Density
We have previously reported that the RasGAP family member and intellectual disability
gene, Syngap1, associates with both Kalirin (RhoGEF) and Agap2 (ArfGAP) in the CA1 area of
mouse hippocampus

(52). These three proteins have also been linked to a variety of psychiatric
conditions (193-197) suggesting that they might share a set of common protein interactions within
the PSD. Therefore, we immunoprecipitated and isolated the PSD interactomes of Agap2,
Syngap1, and Kalirin from adult mouse cortex (Figure 4.1E) and determined protein interactions
by HPLC-MS/MS as previously described (52). Here, we identified 104, 124, and 53 interactions
in Agap2, Syngap1 and Kalirin PSD protein complexes, respectively (Figure 4.2, Table 4.2).
Protein interactions for each interactome where then validated by immunoisolation and western
blotting assays (Figure 4.2A).
We first focused on analyzing different functional groups associated to each interactome
individually. Clustering the interactors of the PSD GAPs and GEFs by function shows that Agap2
and Syngap1 were mainly associated to PSD scaffolds, protein kinases, and proteins involved in
GTP signaling (Figure 4.2B-E, Table 4.2). Although the interactors of Kalirin were also proteins
involved in GTP signaling, kinases, and scaffolds, they also included a large number of
cytoskeletal proteins in accordance with the role of Kalirin in regulating dendritic spine
morphology through RhoGEF activity (Figure 4.2F-G) (205, 206). Thus, Agap2, Syngap1 and
Kalirin share three main functional associations through: a) interactions to the core-scaffold

99

Figure 4.2. Interactomes of the PSD GAP/GEF proteins, Agap2, Syngap1, and Kalirin.
(A) Cropped images of immunoprecipitation followed by western blot for interactors of
Agap2 (left), Syngap1 (middle), and Kalirin (right) that were identified via HPLC-MS/MS.
(B-G) Interactome of each respective protein is clustered by protein function which is
quantified to right. (B-C) Agap2, (D-E) Syngap1, (F-G) Kalirin.

100

Table 4.2. GAP/GEF Protein Interactions and Functional Annotation. Table contains
interactions of Agap2, Syngap1, and Kalirin within the PSD fractions along with Agap2 in non-
PSD fractions determined via HPLC-MS/MS.

101

Table 4.2. Continued

102

machinery of the PSD, b) association with PSD kinases, and c) clustering of proteins involved in
GTP signaling mechanisms.
Similarly to Syngap1, Agap2 and Kalirin also associate with the core-scaffold machinery
of the PSD, composed of MAGUKs (Dlg1-4), DLGAPs (Dlgap1-4) and SHANKs (Shank1-3)
(Figure 4.2, Figure 4.3A, Table 4.2). This association thus links PSD GAPs and GEFs to NMDA
and AMPA ionotropic glutamate receptors through the PSD core-scaffolds, while also binding to
metabotropic glutamate receptors (Table 4.2).
While all three PSD protein complexes were found to associate with kinases belonging to
an array of protein families, the three complexes shared protein interactions with members of the
CAMK family of protein kinases such as the Camk2a and Camk2b kinases (Figure 4.3B, Table
4.2). This in accordance with the modulation of Syngap1, Kalirin and Agap2 phosphorylation by
Camk2a (52, 206-208). We recently reported that the induction of LTP increases Agap2 and
Syngap1 phosphorylation at multiple CamkIIa phosphosites (52). Moreover, induction of LTP also
induced the increase of Akt1 phosphorylation sites in Agap2 (52), which is reflected in the diverse
set of kinases associated to Agap2 (Figure 4.3B, Table 4.2). Some kinases, including MAPK and
PKC kinase families, were shared solely between Syngap1 and Agap2 in accordance with the role
of Syngap1 role in the modulation of MAPK families (190, 207). However, Kalirin was found to
associate with a decreased number of kinases, suggesting a more restricted role, not as a modulator
of kinases but as a substrate.
In addition to the scaffolds and kinases associated with the PSD GAP/GEF proteins, Syngap1,
Agap2 and Kalirin complexes are highly associated with proteins involved in signaling cascades
such as GTP signaling (Figure 4.2). These three complexes are associated with a multitude of
GAP/GEF proteins (Figure 4.3C-D) and we hypothesized that they might associate with more

103

GAPs/GEFs than would be expected by chance. Within the PSD there are a total of 29 GAPs and
30 GEFs (Table 4.1). Within the Agap2 interactions we found a total of 9 GAPs and 7 GEFs which
are highly enriched relative to the PSD (P = 8.55x10
-5
and P = 4.24x10
-3
, respectively, Bonferroni
corrected). The same was true for Kalirin complexes (P = 0.029 and P = 5.31x10
-3
for GAPs and
GEFs, respectively). While Syngap1 complexes were highly enriched in GAP proteins (P =
6.08x10
-6
), they lacked enrichment within GEF proteins (P = 0.79) (Figure 4.3E). Overall, this
shows that GAPs and GEFs are preferentially organized in multi-GAP and GEF family complexes
and not isolated in individual protein interactions at the PSD.
In addition to analyzing Agap2, Kalirin, and Syngap1 complexes by protein function, we
also performed over-representation analysis to test for enrichment of protein domains using the
SMART database annotation (202) (Table 4.3). The strong association with the PSD core-
scaffolding machinery is reflected through significant enrightment in SH3 domains, characteristic
of PSD protein-protein interactions (50), while the association with NMDA and AMPA glutamate
receptors is reflected through enrichment of PBPe (glutamate Rc) domains (P < 0.05 for SH3 and
PBPe domain enrichment within all complexes, Bonferroni corrected). Within the Kalirin
complex, the SPEC (spectrin) domain has the highest enrichment, corrosponding to the increased
number of cytoskeletal interactors (P = 5.02x10
-4
). The enrichment of S_TKc (serine/threonine
kinase) and RasGAP domains in Agap2 PSD complexes (P = 1.63x10
-6
and 3.75x10
-4
for S_TKc
and RasGAP, respectively) and RasGAP domains within Syngap1 (P = 0.03) confirms the
clustering of multiple kinases families and RasGAP proteins in these complexes. Moreover,
Agap2 is also eniched in the G_alpha domain (P = 1.64x10
-5
), corresponding to the association of
Agap2 to 6 out of the 10 G-protein alpha subunits localized within the PSD (Table 4.3).

104

Figure 4.3. Distribution of Signaling Molecules within PSD Networks. Figure shows the
distribution of several different classes of signaling molecules within the three PSD interactomes
including: (A) Scaffolding proteins and glutamate receptors, (B) kinases, (C) GAPs, and (E)
GEFs. (F) Enrichment of GAPs and GEFs within the PSD interactomes. P-values represent the
results of one-tailed fisher exact tests followed by the Bonferroni correction for multiple
comparisons.

105

Table 4.3. SMART Domain Enrichment Within PSD Complexes. Table contains results of
enrichment analyses of protein domains within protein complexes determined in this study
according to the SMART protein domain database. Results were obtained using the DAVID
database.

106

4.3.3 - Differential Interactions of Agap2 Between PSD and Non-PSD Compartments
GAPs and GEFs can also localize in neuronal compartments apart from the PSD (209-212).
Contrary to Syngap1, Agap2 can also localize in non-PSD compartments. Biochemical
fractionation shows the presence of Agap2 within both triton soluble fractions (210) and PSD
preparations (Figure 4.4A). Thus, we selected Agap2 as an example to compare protein
interactions from GAP/GEF complexes between PSD and non-PSD fractions. Therefore, we
isolated Agap2 complexes from triton-soluble fractions in addition to PSD preparations (198) and
determined protein interactions by HPLC-MS/MS. Here, we were able to confidently identify 110
Agap2 interacting proteins in the non-PSD fractions (Table 4.2). We also validated a number of
novel interactions via immunoprecipitation followed by western blot (Figure 4.4B).
Similar to the Agap2 PSD interactome, Agap2 in non-PSD fractions interacted with a large
number of kinases and proteins involved in GTP signaling (Figure 4C-D). However, only 30
interactions were shared with Agap2 interacting proteins within the PSD fraction, corresponding
to 75% unique interactions in the non-PSD fraction. As with the PSD interactomes, we chose to
compare the Agap2 PSD versus non-PSD interactomes with respect to scaffolding proteins and
glutamate receptors, kinases, followed by the associated GAPs and GEFs (Figure 4E-H).
Functional analysis of interacting partners within Agap2 non-PSD complexes shows that the high
association with kinases and GTP signaling proteins remains, but scaffolding protein associations
decrease in accordance with a lack of the PSD core scaffolding machinery in non-PSD
compartments (Figure 4.4E). Although NMDA and AMPA glutamate receptors can also localize
at non-PSD compartments, Agap2 in unable to connect to ionotropic glutamate receptors here.
While some scaffolding molecules that are not solely localized to the PSD remain, there is an
absence of the core-scaffolding molecules of the PSD. This suggests that the link to PSD core-

107

scaffolds is essential to connect Agap2 to AMPA and NMDAR receptors. However, there is a
common association of several metabotropic glutamate receptors with Agap2 in both fractions as
metabotropic glutamate receptors can localize to peri-PSD sites and they can directly bind to
Agap2 in binary protein interactions (213) (Figure 4.4E).
In contrast to Agap2 within the PSD, Agap2 residing in the non-PSD fraction is associated
with a higher number of cytoskeletal and adhesion-related proteins (Figure 4.4D). This is
consistent with the ArfGAP family being involved in the remodeling of the actin cytoskeleton
(214). In particular, Agap2 has been implicated in the regulation of focal adhesion and neurite
outgrowth (188), suggesting that these functions correspond to Agap2 molecules localized in non-
PSD fractions.
With respect to kinases, a large majority of those that are shared between the two different
fractions include kinases involved in the CAMK family (Figure 4.4F). There are also specific
groups of kinases that reside in each particular fraction that can allow the inference of Agap2
function depending on its localization. For example, while the non-PSD fractions contain a
number of kinases that can regulate cytoskeletal function (e.g. neurite outgrowth) such as Pkn2
and Dclk1 (215, 216), the PSD fraction has a larger number of PKA and PKC kinases. These
families of kinases have been previously observed to associate mainly to upper and middle layers
of PSD scaffolds, which are also associated to Agap2 in PSD complexes (Figure 4.4F) (52). Thus,
while Agap2 has an overall capacity to interact with numerous proteins families, the spatial
localization within neuronal compartments, restricts the interactions of Agap2 to a different set of
protein-protein interactions and therefore its function within the PSD.

108

Figure 4.4. Differential Interactions and Functions of PSD vs Non-PSD Agap2. (A)
Cropped images of immunoprecipitation followed by western blot of Syngap1 and Agap2
post PSD enrichment in adult mouse cortex. Syngap1 is restricted to the PSD fraction while
Agap2 is present in both PSD and non-PSD fractions. (B) Cropped images of
immunoprecipitation followed by western blot for interactors of Agap2 in non-PSD fractions
that were identified via HPLC-MS/MS. (C) Non-PSD Agap2 interactome clustered by
protein function. (D) Quantitation of Agap2 interactome by protein function. (E - H)
Comparison of Agap2 PSD interactome versus Agap2 non-PSD interactome illustrating the
distributions of (E) scaffolding proteins and glutamate receptors, (F) Kinases, (G) GAPs, and
(H) GEFs.

109

Because the main functional groups associated to GAPs and GEFs were those involved in
GTP signaling, we also compared the clustering of Agap2 associated GAPs and GEFs in PSD and
non-PSD complexes (Figure 4.4G-H). Both Agap2 PSD and non-PSD complexes have their own
respective RhoGAP and ArfGAP proteins. However, Agap2 PSD and non-PSD complexes shared
the RasGAP, Dab2ip, which has been implicated in neurite outgrowth and cytoskeletal processes
such as the Pkn2 and Dclk1 kinases in the non-PSD fraction (217)

(Figure 4.4G). These processes
can occur in PSD and non-PSD fractions and therefore the correspondent RasGAP effectors are
localized in both fractions. Those that are not shared between the two cellular compartments
include Syngap1, which is PSD-specific, along with Iqgap1, both of which are involved in the
modulation of NMDA receptor signaling and synaptic plasticity (190, 218-220) and therefore
represent specific functions within PSD compartments. In contrast to PSD Agap2 complexes,
which associate with a total of 7 GEF proteins, we were only able to identify a single GEFs in
Agap2 non-PSD complexes (Figure 4.4H). Thus, Agap2 present a differential clustering of GAPs
and GEFs in PSD and non-PSD compartments, which might indicate a differential protein domain
composition in their protein interactors.
Protein domain enrichment of Agap2 non-PSD interactors shows that they are also
enriched in G_alpha and S_TKc domains (P = 4.76x10
-4
and 1.75x10
-3
for G_alpha and S_TKc,
respectively, Bonferroni corrected), but also contained significant enrichment in SPEC domains
(P = 7.30x10
-4
) which correspond to an increased quantity of cytoskeletal interactors outside of
the PSD (Table 4.3). Comparison of protein domain composition within Agap2 interactors located
solely in the PSD again reflect the association of the PSD core-scaffolding machinery through the
enrichment of SH3, PDZ, and PBPe domains (P = 5.62x10
-7
, 1.75x10
-5
, and 1.01x10
-3
for SH3,
PDZ, and PBPe, respectively). Surprisingly, a majority of the G-protein alpha subunit proteins

110

found in Agap2 PSD complexes are found to be in common with Agap2 non-PSD complexes,
leading to a significant enrichment of G_alpha domains among proteins in common (P = 1.21x10
-
4
) (Table 4.3).
4.3.4 - Interactions of Agap2, Syngap1, and Kalirin with Risk Factors of Psychiatric Disease
With Agap2, Syngap1, and Kalirin all being implicated in contributing to the risk of
psychiatric disease (13, 177, 193-197), we also mapped the distribution of SCZ and ASD risk
factors (Figure 4.5A) along with ID risk factors (Figure 4.5B) within their protein-protein
interaction networks. Over-representation analysis was performed in order to test for enrichment
within different datasets for each psychiatric disease (Figure 4.5C). Because of the heterogeneity
of these disorders we tested risk factors implicated in Genome Wide Association Studies (GWAS),
de novo mutations, single nucleotide variants, and recurrent variants.
We tested for the enrichment of genes implicated in ASD using the SFARI database (201),
single nucleotide variants (13), recurring single nucleotide variants (193), and a subset of 107
genes implicated in contributing to ASD through the TADA statistical model (12) (Figure 4.5C).
All three PSD complexes were enriched in recurrent mutations (P < 0.05 for all, Bonferroni
corrected) as these include several PSD-specific proteins such as Grin2b, Shank3, and Syngap1.
Within the SFARI database, Syngap1 and Agap2 PSD complexes were highly enriched (P =
8.89x10
-7
and 9.10x10
-9
for Syngap1 and Agap2 respectively) followed by a decreased enrichment
in Kalirin complexes (P = 0.02). In the single nucleotide variants gene list, Agap2 PSD and non-
PSD complexes gained significance through the interaction of proteins with very little overlap
between the two cellular compartments (P < 0.05 for all) (Figure 4.5C). This shows that Agap2
interacts with different sets of risk factors contributing to ASD based on its subcellular localization

111

and therefore its role in the pathophysiology of ASD cannot be circumscribed to only one synaptic
compartment.
With respect to SCZ, we analyzed each protein complex for enrichment in single nucleotide
variants and GWAS hits implicating a single loci. Contrary to ASD, SCZ risk factors showed a
higher PSD component. In all cases, Syngap1 and Agap2 PSD complexes are highly enriched for
SCZ risk factors (P < 4.0x10
-3
for all). However, Kalirin and Agap2 non-PSD complexes, failed
to reach significance (Figure 4.5C). As the PSD complexes analyzed in this study interact with
the core-scaffolding machinery, the results obtained for the ASD and SCZ gene lists is consistent
with the hypothesis that the risk for these disorders is not spread throughout the entire PSD, but
localized to this core machinery (52). This is especially true for the Agap2 and Syngap1
complexes. However, the Kalirin complex which interacts with less members of the core-
scaffolding complex, and therefore has a decreased association, is less significant in ASD and not
significant in SCZ. In contrast to ASD and SCZ enrichment, all 4 complexes analyzed have similar
levels of enrichment in known ID genes (P < 0.05 for all) (Figure 4.5C). As seen for the ASD
SNVs, Agap2 PSD and non-PSD complexes both have a very distinct set of interacting proteins
that account for their significance of enrichment.

112

Figure 4.5. Distribution of Risk Factors for Psychiatric Disease within Protein-Protein
Interaction Networks. (A) Distribution of risk factors implicated in contributing to autism
spectrum disorder (ASD) and schizophrenia (SCZ) within the protein-protein interaction
network of Syngap1, Kalirin, Agap2 within the PSD fraction, and Agap2 outside of the PSD.
(B) Distribution of risk factors implicated in contributing to intellectual disability (ID) within
the same protein-protein interaction network as in (A). (C) Enrichment of risk factors
implicated in contributing to SCZ, ASD, and ID within Syngap1, Kalirin, Agap2 PSD, and
Agap2 non-PSD complexes.

113

4.4 - Discussion
Through immunoisolation followed by HPLC-MS/MS we were able to determine protein
complexes involving the interacting GAP/GEF proteins Agap2, Syngap1, and Kalirin which
comprised a total of 281 interactions. In addition, we analyzed non-PSD Agap2 complexes (110
interactions), which was able to show functions of Agap2 based on differential localization. In
analyzing the composition of GAP/GEF proteins complexes, we show for the first time that PSD
GAPs and GEFs are clustered in subsets of protein complexes highly enriched in other GAPs and
GEFs and that these protein complexes are associated to the PSD through its core-scaffold
machinery.
While we and others have shown Syngap1 to interact with the PSD core-scaffolding
machinery (52, 221, 222) and Kalirin has been shown to interact with Dlg4 (189), here we are able
to show that all three of the GAP/GEF proteins associate with multiple layers of the core-
scaffolding machinery at the PSD. All of the complexes analyzed here also associate with
glutamate receptors, reflecting their ability to relay signaling information from the receptors while
relying on the scaffolding machinery for positioning.
In addition to the widespread interactions with the core-scaffolding machinery, the three
PSD complexes were also enriched in the quantities of GAP and GEF proteins contained within
each complex. This suggests that, GAP and GEF proteins are highly associated with and can
cluster other GAPs and GEFs within protein interaction networks. In addition, Agap2 and
Syngap1 complexes were also highly associated with G-protein subunit alpha proteins of the
heterotrimeric G protein complexes with Agap2 complexes containing all four classes of G-protein
alpha subunits and Syngap1 complexes mainly associated to the G-protein alpha subunit group I
signature. This suggests a wide regulation of small G-protein signaling mechanisms by Agap2 and

114

more specific modulation by Syngap1 at the PSD. In addition, all of the G protein alpha subunits
within Agap2 PSD complexes were also shared with Agap2 non-PSD complexes. Thus, this
represents a core functional unit within Agap2 complexes, while different interactors within each
subcellular compartment provide the differential modulation of Agap2 function in a location-
specific manner.
With Syngap1, Agap2, and Kalirin being individually implicated in psychiatric disease,
here we show that their protein interactors are also highly enriched in recurrent mutations found
in ASD. However, the enrichment for PSD complexes is not widespread as Kalirin PSD
complexes are not enriched and any SCZ gene set analyzed. While Agap2 non-PSD interactors are
also not enriched in SCZ risk factors, Agap2 PSD and non-PSD complexes have similar levels of
enrichment with respect to all ASD SNVs and genes implicated in contributing to ID. This reflects
the diverse sets of Agap2 interactors implicated in contributing to psychiatric disease as the
majority of interactors that contribute to the similar levels of enrichment between PSD and non-
PSD complexes are unique to Agap2 based on its subcellular location. Furthermore, protein
interactions contributing to disease relevant aspects of Agap2 have been verified within this study.
While we have shown that the interactomes of baits implicated in contributing to psychiatric
disease are enriched with other psychiatric disease risk factors, this concept has also been shown
to be true with respect to neurodegenerative diseases (223). For example, the interactomes of the
Alzheimer’s Disease (AD) risk factors, APP and PSEN1, have been shown to be significantly
associated to AD risk factors identified through GWAS data (223). This provides evidence for the
concept that proteins linked to a particular disease may be closer together within protein-protein
interaction networks than by chance (107).

115

Recently, increased Agap2 expression have been associated to the impaired signaling,
synaptic function, and behavior in observed in fragile X syndrome (FXS) (224). It was proposed
that Agap2 contributed in the dysregulation of metabotropic glutamate receptors 1/5 (Grm1/5)
signaling observed in FXS. Our results show that Agap2 can associate not only to Grm1 and 5, but
also to type II Grm2 and 3, suggesting a wide regulation of metabotropic glutamate signaling by
Agap2. Moreover, Agap2 was also able to interact with Fmr1 binding partners Cyfip1 and Cyfip2,
indicating that Agap2 might modulate Fmr1 function not only at the level of glutamate receptor
inputs but also through protein-protein interactions within the Fmr1 protein interaction network.
A number of these protein interactions were also identified in non-PSD fractions, including Grm1,
Grm2 and Cyfip2. In addition, we were also able to identify Fmr1 as an Agap2 protein interactor
in non-PSD fractions, confirming a role of Agap2 in the Fmr1 protein interaction network.
Interestingly, recent studies have also suggested that altered signaling pathways observed in
Syngap1 mutation mice, might converge with FXS (224). In line with this, we also identified Grm5
and Cyfip2 as Syngap1 interactors, suggesting a convergence in a protein interaction network
within glutamate metabotropic receptors, synaptic Agap2, and Fmr1 protein interactors.
We have identified numerous interactors of PSD GAPs and GEFs and provide an example
of how interacting proteins can change based on cellular compartmentalization. This shows that
GAPs and GEFs may play roles in multiple cellular processes and future studies will need to
address the role of GAPs and GEFs not in isolation but within protein interaction networks. This
will allow the investigation of how disease-relevant mutations affect not only their activity, but
also the composition of protein interactions within the synapse.

116

Chapter 5 – Summary and Future Directions

Here, PPIs of Tnik and Tnik interacting-proteins were determined throughout development
and within specific cellular and sub-cellular fractions. The determination of PPIs for a target
protein within different contexts is able to highlight several overarching principles throughout each
study. The first being that endogenous PPIs can elucidate the function and signaling pathway
involvement of a target protein in question. As protein complexes act as modular units, several
interacting proteins can work together to carry out common cellular processes (67, 68). In line
with this, a target protein may have differential functions and therefore PPIs dependent on the
cellular or developmental context in which the target protein is expressed. Because PPIs can
change based on cellular context, it is essential to determine endogenous PPIs in settings that
closely resemble biological conditions pertinent to the question being investigated. Without taking
this into consideration, the determined PPIs may contain a large number of interactions that are
not naturally occurring and not contain interactions essential for determination of disease
mechanisms. Here, determination of PPIs within specific experimental conditions is particularly
relevant as proteins involved in contributing to disease have been shown to close to one another in
PPI networks (107). Tnik interactomes are able to illustrate this concept as they are highly
enriched in proteins contributing to neurodevelopmental disorders early in development.
The characterization of Tnik interactomes was able to assign protein interactions to Tnik
functions in both early development and at the mature synapse. This was further corroborated
with mutant TNIK hNPCs displaying decreased proliferation and increased levels of Wnt
signaling. While the generated mutant TNIK cells lines were used to investigate phenotypes in
hNPCs, these studies can be extended to several other specific cell types in the future. In addition,
with TNIK kinase activity being abolished, they could be used to identify endogenous substrates

117

of TNIK through the use of phospho-proteomics. Performing these experiments in iPSC-derived
neurons would allow for biochemical fractionation and analysis of both PSD and non-PSD regions.
Like the newly generated TNIK mutant lines, the DISC1
FLAG
iPSC cell line could be differentiated
to numerous other cell types in order to determine PPIs involving DISC1 within additional disease-
relevant contexts.
In line with the expansion of cell types, the conditions in which PPIs are determined can
also be manipulated. In the present studies, all PPIs were determined under basal conditions. PPIs
are dynamic and can change in response to stimuli, as has been illustrated for scaffolding proteins
at the PSD upon the induction of LTP (52). Due to this, all of the proteins investigated here could
also be investigated following stimulation of particular signaling pathways. For example, with
both DISC1 and TNIK being involved in the regulation of Wnt signaling in hNPCs (225), changes
in PPIs or phosphorylation could be investigated following stimulation of Wnt. In addition,
Agap2, Syngap1, and Kalirin were all shown to bind to an assortment of different types of
glutamate receptors such AMPA, NMDA, and mGlur receptors. The PPIs of each of these proteins
could be determined following the stimulation or blocking of the different types of glutamate
receptors which may be able to inform how interactions change when there are deficits in these
signaling pathways in neurodevelopmental disorders.
The determination of PPIs via IP-MS/MS is able to provide a wealth of information
concerning each target protein assayed. This is important to identify different functions and
pathways a target protein may be involved in, along with potential therapeutic targets. In addition
to the determination of PPIs using proteomics-based methods, it will be important to further
investigate individual PPIs in order to decipher their function and prioritize their importance with
respect to disease.

118

References
1. Gandal MJ, Leppa V, Won H, Parikshak NN, Geschwind DH (2016): The road to
precision psychiatry: translating genetics into disease mechanisms. Nat Neurosci. 19:1397-1407.
2. Investigators AaDDMNSYP, Prevention CfDCa (2012): Prevalence of autism spectrum
disorders--Autism and Developmental Disabilities Monitoring Network, 14 sites, United States,
2008. MMWR Surveill Summ. 61:1-19.
3. Millan MJ, Andrieux A, Bartzokis G, Cadenhead K, Dazzan P, Fusar-Poli P, et al.
(2016): Altering the course of schizophrenia: progress and perspectives. Nat Rev Drug Discov.
15:485-515.
4. Lavelle TA, Weinstein MC, Newhouse JP, Munir K, Kuhlthau KA, Prosser LA (2014):
Economic burden of childhood autism spectrum disorders. Pediatrics. 133:e520-529.
5. McPheeters ML, Warren Z, Sathe N, Bruzek JL, Krishnaswami S, Jerome RN, et al.
(2011): A systematic review of medical treatments for children with autism spectrum disorders.
Pediatrics. 127:e1312-1321.
6. McCarthy MI, Abecasis GR, Cardon LR, Goldstein DB, Little J, Ioannidis JP, et al.
(2008): Genome-wide association studies for complex traits: consensus, uncertainty and
challenges. Nat Rev Genet. 9:356-369.
7. Consortium ASDWGoTPG (2017): Meta-analysis of GWAS of over 16,000 individuals
with autism spectrum disorder highlights a novel locus at 10q24.32 and a significant overlap
with schizophrenia. Mol Autism. 8:21.
8. Wang K, Zhang H, Ma D, Bucan M, Glessner JT, Abrahams BS, et al. (2009): Common
genetic variants on 5p14.1 associate with autism spectrum disorders. Nature. 459:528-533.
9. Pardiñas AF, Holmans P, Pocklington AJ, Escott-Price V, Ripke S, Carrera N, et al.
(2018): Common schizophrenia alleles are enriched in mutation-intolerant genes and in regions
under strong background selection. Nat Genet. 50:381-389.
10. Consortium SWGotPG (2014): Biological insights from 108 schizophrenia-associated
genetic loci. Nature. 511:421-427.
11. Niemi MEK, Martin HC, Rice DL, Gallone G, Gordon S, Kelemen M, et al. (2018):
Common genetic variants contribute to risk of rare severe neurodevelopmental disorders. Nature.
562:268-271.
12. De Rubeis S, He X, Goldberg AP, Poultney CS, Samocha K, Cicek AE, et al. (2014):
Synaptic, transcriptional and chromatin genes disrupted in autism. Nature. 515:209-215.
13. Iossifov I, O'Roak BJ, Sanders SJ, Ronemus M, Krumm N, Levy D, et al. (2014): The
contribution of de novo coding mutations to autism spectrum disorder. Nature. 515:216-221.
14. Iossifov I, Ronemus M, Levy D, Wang Z, Hakker I, Rosenbaum J, et al. (2012): De novo
gene disruptions in children on the autistic spectrum. Neuron. 74:285-299.
15. O'Roak BJ, Vives L, Girirajan S, Karakoc E, Krumm N, Coe BP, et al. (2012): Sporadic
autism exomes reveal a highly interconnected protein network of de novo mutations. Nature.
485:246-250.
16. Turner TN, Coe BP, Dickel DE, Hoekzema K, Nelson BJ, Zody MC, et al. (2017):
Genomic Patterns of De Novo Mutation in Simplex Autism. Cell. 171:710-722.e712.
17. Sanders SJ, Murtha MT, Gupta AR, Murdoch JD, Raubeson MJ, Willsey AJ, et al.
(2012): De novo mutations revealed by whole-exome sequencing are strongly associated with
autism. Nature. 485:237-241.

119

18. Fromer M, Pocklington AJ, Kavanagh DH, Williams HJ, Dwyer S, Gormley P, et al.
(2014): De novo mutations in schizophrenia implicate synaptic networks. Nature. 506:179-184.
19. Lelieveld SH, Reijnders MR, Pfundt R, Yntema HG, Kamsteeg EJ, de Vries P, et al.
(2016): Meta-analysis of 2,104 trios provides support for 10 new genes for intellectual disability.
Nat Neurosci. 19:1194-1196.
20. de Ligt J, Willemsen MH, van Bon BW, Kleefstra T, Yntema HG, Kroes T, et al. (2012):
Diagnostic exome sequencing in persons with severe intellectual disability. N Engl J Med.
367:1921-1929.
21. Rauch A, Wieczorek D, Graf E, Wieland T, Endele S, Schwarzmayr T, et al. (2012):
Range of genetic mutations associated with severe non-syndromic sporadic intellectual
disability: an exome sequencing study. Lancet. 380:1674-1682.
22. Study DDD (2017): Prevalence and architecture of de novo mutations in developmental
disorders. Nature. 542:433-438.
23. Pinto D, Delaby E, Merico D, Barbosa M, Merikangas A, Klei L, et al. (2014):
Convergence of genes and cellular pathways dysregulated in autism spectrum disorders. Am J
Hum Genet. 94:677-694.
24. Pinto D, Pagnamenta AT, Klei L, Anney R, Merico D, Regan R, et al. (2010): Functional
impact of global rare copy number variation in autism spectrum disorders. Nature. 466:368-372.
25. Sanders SJ, Ercan-Sencicek AG, Hus V, Luo R, Murtha MT, Moreno-De-Luca D, et al.
(2011): Multiple recurrent de novo CNVs, including duplications of the 7q11.23 Williams
syndrome region, are strongly associated with autism. Neuron. 70:863-885.
26. Kirov G, Pocklington AJ, Holmans P, Ivanov D, Ikeda M, Ruderfer D, et al. (2012): De
novo CNV analysis implicates specific abnormalities of postsynaptic signalling complexes in the
pathogenesis of schizophrenia. Mol Psychiatry. 17:142-153.
27. Marshall CR, Howrigan DP, Merico D, Thiruvahindrapuram B, Wu W, Greer DS, et al.
(2017): Contribution of copy number variants to schizophrenia from a genome-wide study of
41,321 subjects. Nat Genet. 49:27-35.
28. Li M, Santpere G, Imamura Kawasawa Y, Evgrafov OV, Gulden FO, Pochareddy S, et
al. (2018): Integrative functional genomic analysis of human brain development and
neuropsychiatric risks. Science. 362.
29. Soliman MA, Aboharb F, Zeltner N, Studer L (2017): Pluripotent stem cells in
neuropsychiatric disorders. Mol Psychiatry. 22:1241-1249.
30. Takahashi K, Yamanaka S (2006): Induction of pluripotent stem cells from mouse
embryonic and adult fibroblast cultures by defined factors. Cell. 126:663-676.
31. Thomson JA, Itskovitz-Eldor J, Shapiro SS, Waknitz MA, Swiergiel JJ, Marshall VS, et
al. (1998): Embryonic stem cell lines derived from human blastocysts. Science. 282:1145-1147.
32. Ran FA, Hsu PD, Wright J, Agarwala V, Scott DA, Zhang F (2013): Genome
engineering using the CRISPR-Cas9 system. Nat Protoc. 8:2281-2308.
33. Cong L, Ran FA, Cox D, Lin S, Barretto R, Habib N, et al. (2013): Multiplex genome
engineering using CRISPR/Cas systems. Science. 339:819-823.
34. Chambers SM, Fasano CA, Papapetrou EP, Tomishima M, Sadelain M, Studer L (2009):
Highly efficient neural conversion of human ES and iPS cells by dual inhibition of SMAD
signaling. Nat Biotechnol. 27:275-280.
35. Ardhanareeswaran K, Mariani J, Coppola G, Abyzov A, Vaccarino FM (2017): Human
induced pluripotent stem cells for modelling neurodevelopmental disorders. Nat Rev Neurol.
13:265-278.

120

36. Mulligan KA, Cheyette BN (2017): Neurodevelopmental Perspectives on Wnt Signaling
in Psychiatry. Mol Neuropsychiatry. 2:219-246.
37. Srikanth P, Han K, Callahan DG, Makovkina E, Muratore CR, Lalli MA, et al. (2015):
Genomic DISC1 Disruption in hiPSCs Alters Wnt Signaling and Neural Cell Fate. Cell Rep.
12:1414-1429.
38. Mao Y, Ge X, Frank CL, Madison JM, Koehler AN, Doud MK, et al. (2009): Disrupted
in schizophrenia 1 regulates neuronal progenitor proliferation via modulation of GSK3beta/beta-
catenin signaling. Cell. 136:1017-1031.
39. Marchetto MC, Belinson H, Tian Y, Freitas BC, Fu C, Vadodaria K, et al. (2017):
Altered proliferation and networks in neural cells derived from idiopathic autistic individuals.
Mol Psychiatry. 22:820-835.
40. Shi Y, Kirwan P, Smith J, Robinson HP, Livesey FJ (2012): Human cerebral cortex
development from pluripotent stem cells to functional excitatory synapses. Nat Neurosci.
15:477-486, S471.
41. Maroof AM, Keros S, Tyson JA, Ying SW, Ganat YM, Merkle FT, et al. (2013):
Directed differentiation and functional maturation of cortical interneurons from human
embryonic stem cells. Cell Stem Cell. 12:559-572.
42. Kriks S, Shim JW, Piao J, Ganat YM, Wakeman DR, Xie Z, et al. (2011): Dopamine
neurons derived from human ES cells efficiently engraft in animal models of Parkinson's disease.
Nature. 480:547-551.
43. Lu J, Zhong X, Liu H, Hao L, Huang CT, Sherafat MA, et al. (2016): Generation of
serotonin neurons from human pluripotent stem cells. Nat Biotechnol. 34:89-94.
44. Zhang Y, Pak C, Han Y, Ahlenius H, Zhang Z, Chanda S, et al. (2013): Rapid single-step
induction of functional neurons from human pluripotent stem cells. Neuron. 78:785-798.
45. Yang N, Chanda S, Marro S, Ng YH, Janas JA, Haag D, et al. (2017): Generation of pure
GABAergic neurons by transcription factor programming. Nat Methods. 14:621-628.
46. Mariani J, Simonini MV, Palejev D, Tomasini L, Coppola G, Szekely AM, et al. (2012):
Modeling human cortical development in vitro using induced pluripotent stem cells. Proc Natl
Acad Sci U S A. 109:12770-12775.
47. PALAY SL (1956): Synapses in the central nervous system. J Biophys Biochem Cytol.
2:193-202.
48. Bayés A, van de Lagemaat LN, Collins MO, Croning MD, Whittle IR, Choudhary JS, et
al. (2011): Characterization of the proteome, diseases and evolution of the human postsynaptic
density. Nat Neurosci. 14:19-21.
49. Bayés A, Collins MO, Croning MD, van de Lagemaat LN, Choudhary JS, Grant SG
(2012): Comparative study of human and mouse postsynaptic proteomes finds high
compositional conservation and abundance differences for key synaptic proteins. PLoS One.
7:e46683.
50. Collins MO, Husi H, Yu L, Brandon JM, Anderson CN, Blackstock WP, et al. (2006):
Molecular characterization and comparison of the components and multiprotein complexes in the
postsynaptic proteome. J Neurochem. 97 Suppl 1:16-23.
51. Sheng M, Kim E (2011): The postsynaptic organization of synapses. Cold Spring Harb
Perspect Biol. 3.
52. Li J, Wilkinson B, Clementel VA, Hou J, O'Dell TJ, Coba MP (2016): Long-term
potentiation modulates synaptic phosphorylation networks and reshapes the structure of the
postsynaptic interactome. Sci Signal. 9:rs8.

121

53. Li J, Zhang W, Yang H, Howrigan DP, Wilkinson B, Souaiaia T, et al. (2017):
Spatiotemporal profile of postsynaptic interactomes integrates components of complex brain
disorders. Nat Neurosci. 20:1150-1161.
54. Coba MP, Ramaker MJ, Ho EV, Thompson SL, Komiyama NH, Grant SGN, et al.
(2018): Dlgap1 knockout mice exhibit alterations of the postsynaptic density and selective
reductions in sociability. Sci Rep. 8:2281.
55. Whitlock JR, Heynen AJ, Shuler MG, Bear MF (2006): Learning induces long-term
potentiation in the hippocampus. Science. 313:1093-1097.
56. Kennedy MB (2013): Synaptic Signaling in Learning and Memory. Cold Spring Harb
Perspect Biol. 8:a016824.
57. Lüscher C, Malenka RC (2012): NMDA receptor-dependent long-term potentiation and
long-term depression (LTP/LTD). Cold Spring Harb Perspect Biol. 4.
58. Araki Y, Zeng M, Zhang M, Huganir RL (2015): Rapid dispersion of SynGAP from
synaptic spines triggers AMPA receptor insertion and spine enlargement during LTP. Neuron.
85:173-189.
59. Cheng D, Hoogenraad CC, Rush J, Ramm E, Schlager MA, Duong DM, et al. (2006):
Relative and absolute quantification of postsynaptic density proteome isolated from rat forebrain
and cerebellum. Mol Cell Proteomics. 5:1158-1170.
60. Stessman HA, Xiong B, Coe BP, Wang T, Hoekzema K, Fenckova M, et al. (2017):
Targeted sequencing identifies 91 neurodevelopmental-disorder risk genes with autism and
developmental-disability biases. Nat Genet. 49:515-526.
61. Du X, Gao X, Liu X, Shen L, Wang K, Fan Y, et al. (2018): Genetic Diagnostic
Evaluation of Trio-Based Whole Exome Sequencing Among Children With Diagnosed or
Suspected Autism Spectrum Disorder. Front Genet. 9:594.
62. Cochoy DM, Kolevzon A, Kajiwara Y, Schoen M, Pascual-Lucas M, Lurie S, et al.
(2015): Phenotypic and functional analysis of SHANK3 stop mutations identified in individuals
with ASD and/or ID. Mol Autism. 6:23.
63. Hamdan FF, Daoud H, Piton A, Gauthier J, Dobrzeniecka S, Krebs MO, et al. (2011): De
novo SYNGAP1 mutations in nonsyndromic intellectual disability and autism. Biol Psychiatry.
69:898-901.
64. Mitra K, Carvunis AR, Ramesh SK, Ideker T (2013): Integrative approaches for finding
modular structure in biological networks. Nat Rev Genet. 14:719-732.
65. Pawson T, Scott JD (1997): Signaling through scaffold, anchoring, and adaptor proteins.
Science. 278:2075-2080.
66. Pawson T (1995): Protein modules and signalling networks. Nature. 373:573-580.
67. Hartwell LH, Hopfield JJ, Leibler S, Murray AW (1999): From molecular to modular cell
biology. Nature. 402:C47-52.
68. Spirin V, Mirny LA (2003): Protein complexes and functional modules in molecular
networks. Proc Natl Acad Sci U S A. 100:12123-12128.
69. Scott DE, Bayly AR, Abell C, Skidmore J (2016): Small molecules, big targets: drug
discovery faces the protein-protein interaction challenge. Nat Rev Drug Discov. 15:533-550.
70. Suter B, Kittanakom S, Stagljar I (2008): Two-hybrid technologies in proteomics
research. Curr Opin Biotechnol. 19:316-323.
71. De Las Rivas J, Fontanillo C (2010): Protein-protein interactions essentials: key concepts
to building and analyzing interactome networks. PLoS Comput Biol. 6:e1000807.

122

72. Field J, Nikawa J, Broek D, MacDonald B, Rodgers L, Wilson IA, et al. (1988):
Purification of a RAS-responsive adenylyl cyclase complex from Saccharomyces cerevisiae by
use of an epitope addition method. Mol Cell Biol. 8:2159-2165.
73. Brizzard BL, Chubet RG, Vizard DL (1994): Immunoaffinity purification of FLAG
epitope-tagged bacterial alkaline phosphatase using a novel monoclonal antibody and peptide
elution. Biotechniques. 16:730-735.
74. Savic D, Partridge EC, Newberry KM, Smith SB, Meadows SK, Roberts BS, et al.
(2015): CETCh-seq: CRISPR epitope tagging ChIP-seq of DNA-binding proteins. Genome Res.
25:1581-1589.
75. Ardito F, Giuliani M, Perrone D, Troiano G, Lo Muzio L (2017): The crucial role of
protein phosphorylation in cell signaling and its use as targeted therapy (Review). Int J Mol Med.
40:271-280.
76. Mahmoudi T, Li VS, Ng SS, Taouatas N, Vries RG, Mohammed S, et al. (2009): The
kinase TNIK is an essential activator of Wnt target genes. EMBO J. 28:3329-3340.
77. Shkoda A, Town JA, Griese J, Romio M, Sarioglu H, Knöfel T, et al. (2012): The
germinal center kinase TNIK is required for canonical NF-κB and JNK signaling in B-cells by
the EBV oncoprotein LMP1 and the CD40 receptor. PLoS Biol. 10:e1001376.
78. Kim J, Moon SH, Kim BT, Chae CH, Lee JY, Kim SH (2014): A novel aminothiazole
KY-05009 with potential to inhibit Traf2- and Nck-interacting kinase (TNIK) attenuates TGF-
β1-mediated epithelial-to-mesenchymal transition in human lung adenocarcinoma A549 cells.
PLoS One. 9:e110180.
79. Fu CA, Shen M, Huang BC, Lasaga J, Payan DG, Luo Y (1999): TNIK, a novel member
of the germinal center kinase family that activates the c-Jun N-terminal kinase pathway and
regulates the cytoskeleton. J Biol Chem. 274:30729-30737.
80. Yu DH, Zhang X, Wang H, Zhang L, Chen H, Hu M, et al. (2014): The essential role of
TNIK gene amplification in gastric cancer growth. Oncogenesis. 2:e89.
81. Li D, Jiao J, Zhou YM, Wang XX (2015): Epigenetic regulation of traf2- and Nck-
interacting kinase (TNIK) in polycystic ovary syndrome. Am J Transl Res. 7:1152-1160.
82. Takahashi H, Ishikawa T, Ishiguro M, Okazaki S, Mogushi K, Kobayashi H, et al.
(2015): Prognostic significance of Traf2- and Nck- interacting kinase (TNIK) in colorectal
cancer. BMC Cancer. 15:794.
83. Anazi S, Shamseldin HE, AlNaqeb D, Abouelhoda M, Monies D, Salih MA, et al.
(2016): A null mutation in TNIK defines a novel locus for intellectual disability. Hum Genet.
135:773-778.
84. Shitashige M, Satow R, Jigami T, Aoki K, Honda K, Shibata T, et al. (2010): Traf2- and
Nck-interacting kinase is essential for Wnt signaling and colorectal cancer growth. Cancer Res.
70:5024-5033.
85. Kawabe H, Neeb A, Dimova K, Young SM, Takeda M, Katsurabayashi S, et al. (2010):
Regulation of Rap2A by the ubiquitin ligase Nedd4-1 controls neurite development. Neuron.
65:358-372.
86. Hussain NK, Hsin H, Huganir RL, Sheng M (2010): MINK and TNIK differentially act
on Rap2-mediated signal transduction to regulate neuronal structure and AMPA receptor
function. J Neurosci. 30:14786-14794.
87. MacLaren EJ, Charlesworth P, Coba MP, Grant SG (2011): Knockdown of mental
disorder susceptibility genes disrupts neuronal network physiology in vitro. Mol Cell Neurosci.
47:93-99.

123

88. Coba MP, Komiyama NH, Nithianantharajah J, Kopanitsa MV, Indersmitten T, Skene
NG, et al. (2012): TNiK is required for postsynaptic and nuclear signaling pathways and
cognitive function. The Journal of neuroscience : the official journal of the Society for
Neuroscience. 32:13987-13999.
89. Pepper JP, Wang TV, Hennes V, Sun SY, Ichida JK (2016): Human Induced Pluripotent
Stem Cell-Derived Motor Neuron Transplant for Neuromuscular Atrophy in a Mouse Model of
Sciatic Nerve Injury. JAMA Facial Plast Surg.
90. Topol A, Tran NN, Brennand KJ (2015): A guide to generating and using hiPSC derived
NPCs for the study of neurological diseases. J Vis Exp.e52495.
91. Turner TN, Yi Q, Krumm N, Huddleston J, Hoekzema K, F Stessman HA, et al. (2017):
denovo-db: a compendium of human de novo variants. Nucleic Acids Res. 45:D804-D811.
92. Lek M, Karczewski KJ, Minikel EV, Samocha KE, Banks E, Fennell T, et al. (2016):
Analysis of protein-coding genetic variation in 60,706 humans. Nature. 536:285-291.
93. Samocha KE, Robinson EB, Sanders SJ, Stevens C, Sabo A, McGrath LM, et al. (2014):
A framework for the interpretation of de novo mutation in human disease. Nat Genet. 46:944-
950.
94. Huang dW, Sherman BT, Lempicki RA (2009): Systematic and integrative analysis of
large gene lists using DAVID bioinformatics resources. Nat Protoc. 4:44-57.
95. Jin J, Xie X, Chen C, Park JG, Stark C, James DA, et al. (2009): Eukaryotic protein
domains as functional units of cellular evolution. Sci Signal. 2:ra76.
96. Lie DC, Colamarino SA, Song HJ, Désiré L, Mira H, Consiglio A, et al. (2005): Wnt
signalling regulates adult hippocampal neurogenesis. Nature. 437:1370-1375.
97. Varela-Nallar L, Inestrosa NC (2013): Wnt signaling in the regulation of adult
hippocampal neurogenesis. Front Cell Neurosci. 7:100.
98. Rice D, Barone S (2000): Critical periods of vulnerability for the developing nervous
system: evidence from humans and animal models. Environ Health Perspect. 108 Suppl 3:511-
533.
99. Bouguenina H, Salaun D, Mangon A, Muller L, Baudelet E, Camoin L, et al. (2017):
EB1-binding-myomegalin protein complex promotes centrosomal microtubules functions. Proc
Natl Acad Sci U S A. 114:E10687-E10696.
100. Zhang W, Yang SL, Yang M, Herrlinger S, Shao Q, Collar JL, et al. (2019): Modeling
microcephaly with cerebral organoids reveals a WDR62-CEP170-KIF2A pathway promoting
cilium disassembly in neural progenitors. Nat Commun. 10:2612.
101. Hoffman GE, Hartley BJ, Flaherty E, Ladran I, Gochman P, Ruderfer DM, et al. (2017):
Transcriptional signatures of schizophrenia in hiPSC-derived NPCs and neurons are concordant
with post-mortem adult brains. Nat Commun. 8:2225.
102. Fuerer C, Nusse R (2010): Lentiviral vectors to probe and manipulate the Wnt signaling
pathway. PLoS One. 5:e9370.
103. Wilkinson B, Li J, Coba MP (2017): Synaptic GAP and GEF Complexes Cluster Proteins
Essential for GTP Signaling. Sci Rep. 7:5272.
104. Wang Q, Amato SP, Rubitski DM, Hayward MM, Kormos BL, Verhoest PR, et al.
(2016): Identification of Phosphorylation Consensus Sequences and Endogenous Neuronal
Substrates of the Psychiatric Risk Kinase TNIK. J Pharmacol Exp Ther. 356:410-423.
105. Darnell JC, Van Driesche SJ, Zhang C, Hung KY, Mele A, Fraser CE, et al. (2011):
FMRP stalls ribosomal translocation on mRNAs linked to synaptic function and autism. Cell.
146:247-261.

124

106. Boland MJ, Nazor KL, Tran HT, Szücs A, Lynch CL, Paredes R, et al. (2017): Molecular
analyses of neurogenic defects in a human pluripotent stem cell model of fragile X syndrome.
Brain. 140:582-598.
107. Ideker T, Sharan R (2008): Protein networks in disease. Genome Res. 18:644-652.
108. Hennah W, Varilo T, Kestila M, Paunio T, Arajarvi R, Haukka J, et al. (2003): Haplotype
transmission analysis provides evidence of association for DISC1 to schizophrenia and suggests
sex-dependent effects. Human molecular genetics. 12:3151-3159.
109. Hashimoto R, Numakawa T, Ohnishi T, Kumamaru E, Yagasaki Y, Ishimoto T, et al.
(2006): Impact of the DISC1 Ser704Cys polymorphism on risk for major depression, brain
morphology and ERK signaling. Human molecular genetics. 15:3024-3033.
110. Hennah W, Thomson P, Peltonen L, Porteous D (2006): Genes and schizophrenia:
beyond schizophrenia: the role of DISC1 in major mental illness. Schizophrenia bulletin. 32:409-
416.
111. Porteous DJ, Thomson P, Brandon NJ, Millar JK (2006): The genetics and biology of
DISC1--an emerging role in psychosis and cognition. Biological psychiatry. 60:123-131.
112. Mackie S, Millar JK, Porteous DJ (2007): Role of DISC1 in neural development and
schizophrenia. Current opinion in neurobiology. 17:95-102.
113. Wang Q, Charych EI, Pulito VL, Lee JB, Graziane NM, Crozier RA, et al. (2010): The
psychiatric disease risk factors DISC1 and TNIK interact to regulate synapse composition and
function. Molecular psychiatry.1-18.
114. Brandon NJ, Sawa A (2011): Linking neurodevelopmental and synaptic theories of
mental illness through DISC1. Nature reviews Neuroscience. 12:707-722.
115. Porteous DJ, Millar JK, Brandon NJ, Sawa A (2011): DISC1 at 10: connecting
psychiatric genetics and neuroscience. Trends Mol Med. 17:699-706.
116. Kilpinen H, Ylisaukko-Oja T, Hennah W, Palo OM, Varilo T, Vanhala R, et al. (2008):
Association of DISC1 with autism and Asperger syndrome. Mol Psychiatry. 13:187-196.
117. Millar JK, Christie S, Semple CA, Porteous DJ (2000): Chromosomal location and
genomic structure of the human translin-associated factor X gene (TRAX; TSNAX) revealed by
intergenic splicing to DISC1, a gene disrupted by a translocation segregating with schizophrenia.
Genomics. 67:69-77.
118. Millar JK, Wilson-Annan JC, Anderson S, Christie S, Taylor MS, Semple CA, et al.
(2000): Disruption of two novel genes by a translocation co-segregating with schizophrenia.
Human molecular genetics. 9:1415-1423.
119. Ishizuka K, Kamiya A, Oh EC, Kanki H, Seshadri S, Robinson JF, et al. (2011): DISC1-
dependent switch from progenitor proliferation to migration in the developing cortex. Nature.
120. Wen Z, Nguyen HN, Guo Z, Lalli MA, Wang X, Su Y, et al. (2014): Synaptic
dysregulation in a human iPS cell model of mental disorders. Nature. 515:414-418.
121. Murai K, Sun G, Ye P, Tian E, Yang S, Cui Q, et al. (2016): The TLX-miR-219 cascade
regulates neural stem cell proliferation in neurodevelopment and schizophrenia iPSC model. Nat
Commun. 7:10965.
122. Terrillion CE, Abazyan B, Yang Z, Crawford J, Shevelkin AV, Jouroukhin Y, et al.
(2017): DISC1 in Astrocytes Influences Adult Neurogenesis and Hippocampus-Dependent
Behaviors in Mice. Neuropsychopharmacology. 42:2242-2251.
123. Yerabham AS, Weiergraber OH, Bradshaw NJ, Korth C (2013): Revisiting disrupted-in-
schizophrenia 1 as a scaffold protein. Biol Chem. 394:1425-1437.

125

124. Tomppo L, Hennah W, Lahermo P, Loukola A, Tuulio-Henriksson A, Suvisaari J, et al.
(2009): Association between genes of Disrupted in schizophrenia 1 (DISC1) interactors and
schizophrenia supports the role of the DISC1 pathway in the etiology of major mental illnesses.
Biological psychiatry. 65:1055-1062.
125. Bradshaw NJ, Porteous DJ (2012): DISC1-binding proteins in neural development,
signalling and schizophrenia. Neuropharmacology. 62:1230-1241.
126. Rampino A, Walker RM, Torrance HS, Anderson SM, Fazio L, Di Giorgio A, et al.
(2014): Expression of DISC1-interactome members correlates with cognitive phenotypes related
to schizophrenia. PloS one. 9:e99892.
127. Camargo LM, Collura V, Rain JC, Mizuguchi K, Hermjakob H, Kerrien S, et al. (2007):
Disrupted in Schizophrenia 1 Interactome: evidence for the close connectivity of risk genes and a
potential synaptic basis for schizophrenia. Molecular psychiatry. 12:74-86.
128. Morris JA, Kandpal G, Ma L, Austin CP (2003): DISC1 (Disrupted-In-Schizophrenia 1)
is a centrosome-associated protein that interacts with MAP1A, MIPT3, ATF4/5 and NUDEL:
regulation and loss of interaction with mutation. Human molecular genetics. 12:1591-1608.
129. Ozeki Y, Tomoda T, Kleiderlein J, Kamiya A, Bord L, Fujii K, et al. (2003): Disrupted-
in-Schizophrenia-1 (DISC-1): mutant truncation prevents binding to NudE-like (NUDEL) and
inhibits neurite outgrowth. Proc Natl Acad Sci U S A. 100:289-294.
130. Park YU, Jeong J, Lee H, Mun JY, Kim JH, Lee JS, et al. (2010): Disrupted-in-
schizophrenia 1 (DISC1) plays essential roles in mitochondria in collaboration with Mitofilin.
Proc Natl Acad Sci U S A. 107:17785-17790.
131. Miyoshi K, Honda A, Baba K, Taniguchi M, Oono K, Fujita T, et al. (2003): Disrupted-
In-Schizophrenia 1, a candidate gene for schizophrenia, participates in neurite outgrowth. Mol
Psychiatry. 8:685-694.
132. Corominas R, Yang X, Lin GN, Kang S, Shen Y, Ghamsari L, et al. (2014): Protein
interaction network of alternatively spliced isoforms from brain links genetic risk factors for
autism. Nat Commun. 5:3650.
133. Millar JK, Christie S, Porteous DJ (2003): Yeast two-hybrid screens implicate DISC1 in
brain development and function. Biochemical and biophysical research communications.
311:1019-1025.
134. Ishizuka K, Chen J, Taya S, Li W, Millar JK, Xu Y, et al. (2007): Evidence that many of
the DISC1 isoforms in C57BL/6J mice are also expressed in 129S6/SvEv mice. Molecular
psychiatry. 12:897-899.
135. Kvajo M, McKellar H, Arguello PA, Drew LJ, Moore H, MacDermott AB, et al. (2008):
A mutation in mouse Disc1 that models a schizophrenia risk allele leads to specific alterations in
neuronal architecture and cognition. Proc Natl Acad Sci U S A. 105:7076-7081.
136. Terpe K (2003): Overview of tag protein fusions: from molecular and biochemical
fundamentals to commercial systems. Appl Microbiol Biotechnol. 60:523-533.
137. Bradshaw NJ, Ogawa F, Antolin-Fontes B, Chubb JE, Carlyle BC, Christie S, et al.
(2008): DISC1, PDE4B, and NDE1 at the centrosome and synapse. Biochemical and biophysical
research communications. 377:1091-1096.
138. Tsuboi D, Kuroda K, Tanaka M, Namba T, Iizuka Y, Taya S, et al. (2015): Disrupted-in-
schizophrenia 1 regulates transport of ITPR1 mRNA for synaptic plasticity. Nat Neurosci.
18:698-707.

126

139. Qi Y, Zhang XJ, Renier N, Wu Z, Atkin T, Sun Z, et al. (2017): Combined small-
molecule inhibition accelerates the derivation of functional cortical neurons from human
pluripotent stem cells. Nat Biotechnol. 35:154-163.
140. Tcw J, Wang M, Pimenova AA, Bowles KR, Hartley BJ, Lacin E, et al. (2017): An
Efficient Platform for Astrocyte Differentiation from Human Induced Pluripotent Stem Cells.
Stem Cell Reports. 9:600-614.
141. Haeussler M, Schönig K, Eckert H, Eschstruth A, Mianné J, Renaud JB, et al. (2016):
Evaluation of off-target and on-target scoring algorithms and integration into the guide RNA
selection tool CRISPOR. Genome Biol. 17:148.
142. Liu H, Wei Z, Dominguez A, Li Y, Wang X, Qi LS (2015): CRISPR-ERA: a
comprehensive design tool for CRISPR-mediated gene editing, repression and activation.
Bioinformatics. 31:3676-3678.
143. Kamentsky L, Jones TR, Fraser A, Bray MA, Logan DJ, Madden KL, et al. (2011):
Improved structure, function and compatibility for CellProfiler: modular high-throughput image
analysis software. Bioinformatics. 27:1179-1180.
144. Vizcaíno JA, Csordas A, del-Toro N, Dianes JA, Griss J, Lavidas I, et al. (2016): 2016
update of the PRIDE database and its related tools. Nucleic Acids Res. 44:D447-456.
145. Kim D, Pertea G, Trapnell C, Pimentel H, Kelley R, Salzberg SL (2013): TopHat2:
accurate alignment of transcriptomes in the presence of insertions, deletions and gene fusions.
Genome Biol. 14:R36.
146. Trapnell C, Roberts A, Goff L, Pertea G, Kim D, Kelley DR, et al. (2012): Differential
gene and transcript expression analysis of RNA-seq experiments with TopHat and Cufflinks. Nat
Protoc. 7:562-578.
147. Schmittgen TD, Livak KJ (2008): Analyzing real-time PCR data by the comparative C(T)
method. Nat Protoc. 3:1101-1108.
148. Stark C, Breitkreutz BJ, Reguly T, Boucher L, Breitkreutz A, Tyers M (2006): BioGRID:
a general repository for interaction datasets. Nucleic Acids Res. 34:D535-539.
149. Calderone A, Castagnoli L, Cesareni G (2013): mentha: a resource for browsing
integrated protein-interaction networks. Nat Methods. 10:690-691.
150. Montojo J, Zuberi K, Rodriguez H, Bader GD, Morris Q (2014): GeneMANIA: Fast gene
network construction and function prediction for Cytoscape. F1000Res. 3:153.
151. Wang HY, Liu Y, Yan JW, Hu XL, Zhu DM, Xu XT, et al. (2018): Gene polymorphisms
of DISC1 is associated with schizophrenia: Evidence from a meta-analysis. Prog
Neuropsychopharmacol Biol Psychiatry. 81:64-73.
152. Xu Y, Ren J, Ye H (2018): Association between variations in the disrupted in
schizophrenia 1 gene and schizophrenia: A meta-analysis. Gene. 651:94-99.
153. Wang S, Liang Q, Qiao H, Li H, Shen T, Ji F, et al. (2016): DISC1 regulates astrogenesis
in the embryonic brain via modulation of RAS/MEK/ERK signaling through RASSF7.
Development. 143:2732-2740.
154. Nakata K, Lipska BK, Hyde TM, Ye T, Newburn EN, Morita Y, et al. (2009): DISC1
splice variants are upregulated in schizophrenia and associated with risk polymorphisms. Proc
Natl Acad Sci U S A. 106:15873-15878.
155. Soares DC, Carlyle BC, Bradshaw NJ, Porteous DJ (2011): DISC1: Structure, Function,
and Therapeutic Potential for Major Mental Illness. ACS Chem Neurosci. 2:609-632.

127

156. Millar JK, Pickard BS, Mackie S, James R, Christie S, Buchanan SR, et al. (2005):
DISC1 and PDE4B are interacting genetic factors in schizophrenia that regulate cAMP signaling.
Science. 310:1187-1191.
157. Millar JK, James R, Christie S, Porteous DJ (2005): Disrupted in schizophrenia 1
(DISC1): subcellular targeting and induction of ring mitochondria. Mol Cell Neurosci. 30:477-
484.
158. Kamiya A, Kubo K, Tomoda T, Takaki M, Youn R, Ozeki Y, et al. (2005): A
schizophrenia-associated mutation of DISC1 perturbs cerebral cortex development. Nat Cell
Biol. 7:1167-1178.
159. Kuroda K, Yamada S, Tanaka M, Iizuka M, Yano H, Mori D, et al. (2011): Behavioral
alterations associated with targeted disruption of exons 2 and 3 of the Disc1 gene in the mouse.
Hum Mol Genet. 20:4666-4683.
160. Chavez A, Scheiman J, Vora S, Pruitt BW, Tuttle M, P R Iyer E, et al. (2015): Highly
efficient Cas9-mediated transcriptional programming. Nat Methods. 12:326-328.
161. Brandon NJ, Schurov I, Camargo LM, Handford EJ, Duran-Jimeniz B, Hunt P, et al.
(2005): Subcellular targeting of DISC1 is dependent on a domain independent from the Nudel
binding site. Molecular and cellular neurosciences. 28:613-624.
162. Norkett R, Modi S, Birsa N, Atkin TA, Ivankovic D, Pathania M, et al. (2016): DISC1-
dependent Regulation of Mitochondrial Dynamics Controls the Morphogenesis of Complex
Neuronal Dendrites. J Biol Chem. 291:613-629.
163. Soda T, Frank C, Ishizuka K, Baccarella A, Park YU, Flood Z, et al. (2013): DISC1-
ATF4 transcriptional repression complex: dual regulation of the cAMP-PDE4 cascade by
DISC1. Mol Psychiatry. 18:898-908.
164. Willemsen MH, Nijhof B, Fenckova M, Nillesen WM, Bongers EM, Castells-Nobau A,
et al. (2013): GATAD2B loss-of-function mutations cause a recognisable syndrome with
intellectual disability and are associated with learning deficits and synaptic undergrowth in
Drosophila. J Med Genet. 50:507-514.
165. Hamdan FF, Daoud H, Rochefort D, Piton A, Gauthier J, Langlois M, et al. (2010): De
novo mutations in FOXP1 in cases with intellectual disability, autism, and language impairment.
Am J Hum Genet. 87:671-678.
166. Crepel A, Breckpot J, Fryns JP, De la Marche W, Steyaert J, Devriendt K, et al. (2010):
DISC1 duplication in two brothers with autism and mild mental retardation. Clin Genet. 77:389-
394.
167. Thomson PA, Parla JS, McRae AF, Kramer M, Ramakrishnan K, Yao J, et al. (2014):
708 Common and 2010 rare DISC1 locus variants identified in 1542 subjects: analysis for
association with psychiatric disorder and cognitive traits. Mol Psychiatry. 19:668-675.
168. Sullivan PF (2013): Questions about DISC1 as a genetic risk factor for schizophrenia.
Mol Psychiatry. 18:1050-1052.
169. Porteous DJ, Thomson PA, Millar JK, Evans KL, Hennah W, Soares DC, et al. (2014):
DISC1 as a genetic risk factor for schizophrenia and related major mental illness: response to
Sullivan. Mol Psychiatry. 19:141-143.
170. Carter H, Hofree M, Ideker T (2013): Genotype to phenotype via network analysis. Curr
Opin Genet Dev. 23:611-621.
171. Goh KI, Cusick ME, Valle D, Childs B, Vidal M, Barabási AL (2007): The human
disease network. Proc Natl Acad Sci U S A. 104:8685-8690.

128

172. Menche J, Sharma A, Kitsak M, Ghiassian SD, Vidal M, Loscalzo J, et al. (2015):
Disease networks. Uncovering disease-disease relationships through the incomplete interactome.
Science. 347:1257601.
173. Sialana FJ, Wang AL, Fazari B, Kristofova M, Smidak R, Trossbach SV, et al. (2018):
Quantitative Proteomics of Synaptosomal Fractions in a Rat Overexpressing Human DISC1
Gene Indicates Profound Synaptic Dysregulation in the Dorsal Striatum. Front Mol Neurosci.
11:26.
174. Xia M, Broek JA, Jouroukhin Y, Schoenfelder J, Abazyan S, Jaaro-Peled H, et al. (2016):
Cell Type-Specific Effects of Mutant DISC1: A Proteomics Study. Mol Neuropsychiatry. 2:28-
36.
175. Potkin SG, Turner JA, Guffanti G, Lakatos A, Fallon JH, Nguyen DD, et al. (2009): A
genome-wide association study of schizophrenia using brain activation as a quantitative
phenotype. Schizophr Bull. 35:96-108.
176. Shi J, Levinson DF, Duan J, Sanders AR, Zheng Y, Pe'er I, et al. (2009): Common
variants on chromosome 6p22.1 are associated with schizophrenia. Nature. 460:753-757.
177. Genovese G, Fromer M, Stahl EA, Ruderfer DM, Chambert K, Landén M, et al. (2016):
Increased burden of ultra-rare protein-altering variants among 4,877 individuals with
schizophrenia. Nat Neurosci. 19:1433-1441.
178. Wang Q, Charych EI, Pulito VL, Lee JB, Graziane NM, Crozier RA, et al. (2011): The
psychiatric disease risk factors DISC1 and TNIK interact to regulate synapse composition and
function. Mol Psychiatry. 16:1006-1023.
179. D'Angelo MA, Gomez-Cavazos JS, Mei A, Lackner DH, Hetzer MW (2012): A change
in nuclear pore complex composition regulates cell differentiation. Dev Cell. 22:446-458.
180. Liang Y, Franks TM, Marchetto MC, Gage FH, Hetzer MW (2013): Dynamic association
of NUP98 with the human genome. PLoS Genet. 9:e1003308.
181. Desiere F, Deutsch EW, King NL, Nesvizhskii AI, Mallick P, Eng J, et al. (2006): The
PeptideAtlas project. Nucleic Acids Res. 34:D655-658.
182. Bourne HR, Sanders DA, McCormick F (1990): The GTPase superfamily: a conserved
switch for diverse cell functions. Nature. 348:125-132.
183. Scheffzek K, Ahmadian MR, Wittinghofer A (1998): GTPase-activating proteins: helping
hands to complement an active site. Trends Biochem Sci. 23:257-262.
184. Raaijmakers JH, Bos JL (2009): Specificity in Ras and Rap signaling. J Biol Chem.
284:10995-10999.
185. Cherfils J, Zeghouf M (2013): Regulation of small GTPases by GEFs, GAPs, and GDIs.
Physiol Rev. 93:269-309.
186. Bos JL, Rehmann H, Wittinghofer A (2007): GEFs and GAPs: critical elements in the
control of small G proteins. Cell. 129:865-877.
187. Naisbitt S, Kim E, Weinberg RJ, Rao A, Yang FC, Craig AM, et al. (1997):
Characterization of guanylate kinase-associated protein, a postsynaptic density protein at
excitatory synapses that interacts directly with postsynaptic density-95/synapse-associated
protein 90. J Neurosci. 17:5687-5696.
188. Dwane S, Durack E, O'Connor R, Kiely PA (2014): RACK1 promotes neurite outgrowth
by scaffolding AGAP2 to FAK. Cell Signal. 26:9-18.
189. Penzes P, Johnson RC, Sattler R, Zhang X, Huganir RL, Kambampati V, et al. (2001):
The neuronal Rho-GEF Kalirin-7 interacts with PDZ domain-containing proteins and regulates
dendritic morphogenesis. Neuron. 29:229-242.

129

190. Komiyama NH, Watabe AM, Carlisle HJ, Porter K, Charlesworth P, Monti J, et al.
(2002): SynGAP regulates ERK/MAPK signaling, synaptic plasticity, and learning in the
complex with postsynaptic density 95 and NMDA receptor. J Neurosci. 22:9721-9732.
191. Grant SG (2012): Synaptopathies: diseases of the synaptome. Curr Opin Neurobiol.
22:522-529.
192. Ting JT, Peça J, Feng G (2012): Functional consequences of mutations in postsynaptic
scaffolding proteins and relevance to psychiatric disorders. Annu Rev Neurosci. 35:49-71.
193. Turner TN, Hormozdiari F, Duyzend MH, McClymont SA, Hook PW, Iossifov I, et al.
(2016): Genome Sequencing of Autism-Affected Families Reveals Disruption of Putative
Noncoding Regulatory DNA. Am J Hum Genet. 98:58-74.
194. Hamdan FF, Gauthier J, Spiegelman D, Noreau A, Yang Y, Pellerin S, et al. (2009):
Mutations in SYNGAP1 in autosomal nonsyndromic mental retardation. N Engl J Med. 360:599-
605.
195. Mignot C, von Stülpnagel C, Nava C, Ville D, Sanlaville D, Lesca G, et al. (2016):
Genetic and neurodevelopmental spectrum of SYNGAP1-associated intellectual disability and
epilepsy. J Med Genet. 53:511-522.
196. Kushima I, Nakamura Y, Aleksic B, Ikeda M, Ito Y, Shiino T, et al. (2012):
Resequencing and association analysis of the KALRN and EPHB1 genes and their contribution
to schizophrenia susceptibility. Schizophr Bull. 38:552-560.
197. Cahill ME, Xie Z, Day M, Photowala H, Barbolina MV, Miller CA, et al. (2009): Kalirin
regulates cortical spine morphogenesis and disease-related behavioral phenotypes. Proc Natl
Acad Sci U S A. 106:13058-13063.
198. Coba MP, Pocklington AJ, Collins MO, Kopanitsa MV, Uren RT, Swamy S, et al.
(2009): Neurotransmitters drive combinatorial multistate postsynaptic density networks. Sci
Signal. 2:ra19.
199. Huang DW, Sherman BT, Tan Q, Collins JR, Alvord WG, Roayaei J, et al. (2007): The
DAVID Gene Functional Classification Tool: a novel biological module-centric algorithm to
functionally analyze large gene lists. Genome Biol. 8:R183.
200. Zhang Y, Chen K, Sloan SA, Bennett ML, Scholze AR, O'Keeffe S, et al. (2014): An
RNA-sequencing transcriptome and splicing database of glia, neurons, and vascular cells of the
cerebral cortex. J Neurosci. 34:11929-11947.
201. Abrahams BS, Arking DE, Campbell DB, Mefford HC, Morrow EM, Weiss LA, et al.
(2013): SFARI Gene 2.0: a community-driven knowledgebase for the autism spectrum disorders
(ASDs). Mol Autism. 4:36.
202. Letunic I, Doerks T, Bork P (2015): SMART: recent updates, new developments and
status in 2015. Nucleic Acids Res. 43:D257-260.
203. Bateman A, Coin L, Durbin R, Finn RD, Hollich V, Griffiths-Jones S, et al. (2004): The
Pfam protein families database. Nucleic Acids Res. 32:D138-141.
204. Kim E, Sheng M (2009): The postsynaptic density. Curr Biol. 19:R723-724.
205. Yan Y, Eipper BA, Mains RE (2015): Kalirin-9 and Kalirin-12 Play Essential Roles in
Dendritic Outgrowth and Branching. Cereb Cortex. 25:3487-3501.
206. Xie Z, Srivastava DP, Photowala H, Kai L, Cahill ME, Woolfrey KM, et al. (2007):
Kalirin-7 controls activity-dependent structural and functional plasticity of dendritic spines.
Neuron. 56:640-656.

130

207. Krapivinsky G, Medina I, Krapivinsky L, Gapon S, Clapham DE (2004): SynGAP-
MUPP1-CaMKII synaptic complexes regulate p38 MAP kinase activity and NMDA receptor-
dependent synaptic AMPA receptor potentiation. Neuron. 43:563-574.
208. Walkup WG, Washburn L, Sweredoski MJ, Carlisle HJ, Graham RL, Hess S, et al.
(2015): Phosphorylation of synaptic GTPase-activating protein (synGAP) by Ca2+/calmodulin-
dependent protein kinase II (CaMKII) and cyclin-dependent kinase 5 (CDK5) alters the ratio of
its GAP activity toward Ras and Rap GTPases. J Biol Chem. 290:4908-4927.
209. Shiba Y, Römer W, Mardones GA, Burgos PV, Lamaze C, Johannes L (2010): AGAP2
regulates retrograde transport between early endosomes and the TGN. J Cell Sci. 123:2381-2390.
210. Wu Y, Zhao Y, Ma X, Zhu Y, Patel J, Nie Z (2013): The Arf GAP AGAP2 interacts with
β-arrestin2 and regulates β2-adrenergic receptor recycling and ERK activation. Biochem J.
452:411-421.
211. Patrakitkomjorn S, Kobayashi D, Morikawa T, Wilson MM, Tsubota N, Irie A, et al.
(2008): Neurofibromatosis type 1 (NF1) tumor suppressor, neurofibromin, regulates the neuronal
differentiation of PC12 cells via its associating protein, CRMP-2. J Biol Chem. 283:9399-9413.
212. Sot B, Kötting C, Deaconescu D, Suveyzdis Y, Gerwert K, Wittinghofer A (2010):
Unravelling the mechanism of dual-specificity GAPs. EMBO J. 29:1205-1214.
213. Gross C, Chang CW, Kelly SM, Bhattacharya A, McBride SM, Danielson SW, et al.
(2015): Increased expression of the PI3K enhancer PIKE mediates deficits in synaptic plasticity
and behavior in fragile X syndrome. Cell Rep. 11:727-736.
214. Randazzo PA, Inoue H, Bharti S (2007): Arf GAPs as regulators of the actin
cytoskeleton. Biol Cell. 99:583-600.
215. Buchser WJ, Slepak TI, Gutierrez-Arenas O, Bixby JL, Lemmon VP (2010):
Kinase/phosphatase overexpression reveals pathways regulating hippocampal neuron
morphology. Mol Syst Biol. 6:391.
216. Shin E, Kashiwagi Y, Kuriu T, Iwasaki H, Tanaka T, Koizumi H, et al. (2013):
Doublecortin-like kinase enhances dendritic remodelling and negatively regulates synapse
maturation. Nat Commun. 4:1440.
217. Lee GH, Kim SH, Homayouni R, D'Arcangelo G (2012): Dab2ip regulates neuronal
migration and neurite outgrowth in the developing neocortex. PLoS One. 7:e46592.
218. Gao C, Frausto SF, Guedea AL, Tronson NC, Jovasevic V, Leaderbrand K, et al. (2011):
IQGAP1 regulates NR2A signaling, spine density, and cognitive processes. J Neurosci. 31:8533-
8542.
219. Gao C, Liang C, Nie Z, Liu Y, Wang J, Zhang D (2015): Alkannin inhibits growth and
invasion of glioma cells C6 through IQGAP/mTOR signal pathway. Int J Clin Exp Med. 8:5287-
5294.
220. Oh JS, Manzerra P, Kennedy MB (2004): Regulation of the neuron-specific Ras GTPase-
activating protein, synGAP, by Ca2+/calmodulin-dependent protein kinase II. J Biol Chem.
279:17980-17988.
221. Chen HJ, Rojas-Soto M, Oguni A, Kennedy MB (1998): A synaptic Ras-GTPase
activating protein (p135 SynGAP) inhibited by CaM kinase II. Neuron. 20:895-904.
222. Fernández E, Collins MO, Uren RT, Kopanitsa MV, Komiyama NH, Croning MD, et al.
(2009): Targeted tandem affinity purification of PSD-95 recovers core postsynaptic complexes
and schizophrenia susceptibility proteins. Mol Syst Biol. 5:269.

131

223. Hosp F, Vossfeldt H, Heinig M, Vasiljevic D, Arumughan A, Wyler E, et al. (2015):
Quantitative interaction proteomics of neurodegenerative disease proteins. Cell Rep. 11:1134-
1146.
224. Barnes SA, Wijetunge LS, Jackson AD, Katsanevaki D, Osterweil EK, Komiyama NH, et
al. (2015): Convergence of Hippocampal Pathophysiology in Syngap+/- and Fmr1-/y Mice. J
Neurosci. 35:15073-15081.
225. Ishizuka K, Kamiya A, Oh EC, Kanki H, Seshadri S, Robinson JF, et al. (2011): DISC1-
dependent switch from progenitor proliferation to migration in the developing cortex. Nature.
473:92-96.

Abstract (if available)

Abstract Thousands of mutations have been identified that may potentially contribute to the genetic architecture of complex brain disorders such as intellectual disability (ID), autism spectrum disorder (ASD), developmental delay (DD), and schizophrenia (SCZ). While the affected proteins have been shown to be enriched in both early neural development and within the postsynaptic density (PSD) of glutamatergic excitatory neurons, how these risk factors associate in protein-protein interactions (PPIs) to regulate signaling networks is less clear. Within signaling networks, proteins with a large number of PPIs (protein hubs) are likely to have a significant impact on the dysregulation of signaling mechanisms when they harbor damaging mutations and therefore, be associated with disease. ID causative mutations have previously been identified in the Traf2 and Nck Interacting Kinase (TNIK) and Tnik has been shown to act as a protein hub in the regulation of signaling networks. Importantly, these networks are enriched in risk factors for complex brain disorders including disrupted in schizophrenia 1 (Disc1) and synaptic Ras GTPase activating protein 1 (Syngap1). Here, PPIs of Tnik and Tnik-interacting proteins were characterized during specific developmental stages, within precise cellular and subcellular fractions, and under endogenous conditions relevant to complex brain disorders. We first determined the spatiotemporal-dependent functions of Tnik by characterizing Tnik interactomes throughout mouse cortical development and in human pluripotent stem cell (hPSC)-derived neural progenitor cells (hNPCs). Multiple mutant TNIK hPSC lines were generated to further investigate the consequences of TNIK dysregulation during early neural development. Second, we determined cell-type specific PPIs of the TNIK interactor, DISC1, in hNPCs and astrocytes. For this purpose, we generated a hPSC line containing an endogenous FLAG tag for DISC1 affinity purification. Lastly, the interactomes of the Tnik interacting proteins, Syngap1, Agap2, and Kalirin, were determined to characterize their involvement in specific G-protein signaling mechanisms within the PSD. These studies revealed a number of novel interactions that inform cell-type specific functions of proteins that modulate neural development or neuronal activity and highlight signaling networks associated with complex brain disorders.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

The autism-associated gene SYNGAP1 regulates human cortical neurogenesis

PDF

Modeling SynGAP1 truncating mutations in neurodevelopmental disease using iPSC-derived neurons

PDF

Neural crest-derived cranial pericytes and their dysfunction in disease

PDF

Identification of therapeutic targets in human cerebral brain organoid models of neurodegeneration

PDF

Generation and long-term culture of human cerebellar organoids to study development and disease.

PDF

Functional compensation between hematopoietic stem cell clones in vivo

PDF

Molecular programs in epithelial morphogenesis in mammalian organ systems

PDF

Exploring stem cell pluripotency through long range chromosome interactions

PDF

Transcriptional regulation in nephron progenitor cells

PDF

The role of ERK1/2 in mouse embryonic stem cell fate control

PDF

Genetics and the environment: evaluating the role of noncoding RNA in autism spectrum disorder

PDF

Transcriptional and epigenetic mechanisms underlying sensory hair cell differentiation and regeneration

PDF

Investigating molecular roadblocks to enhance direct cellular reprogramming

PDF

Transcriptomic maturation of developing human cone precursors in fetal and 3D hESC-derived tissues

PDF

Elucidating the functional role of CHD7 associated nuclear PDH complex and other associated proteins on neural crest development

PDF

Neuronal master regulator SRRM4 in breast cancer cells facilitates CNS-acclimation and colonization leading to brain metastasis

PDF

Cross-species comparison of non-canonical Wnt signaling in the developing retina

PDF

Skeletal cell fate plasticity in zebrafish bone development and regeneration

PDF

Identification of therapeutic targets for neurons and microglia in amyotrophic lateral sclerosis

PDF

Functional study of C9ORF72 and its implication in the pathogenesis of amyotrophic lateral sclerosis

Asset Metadata

Creator Wilkinson, Brent (author)

Core Title Signaling networks in complex brain disorders

School Keck School of Medicine

Degree Doctor of Philosophy

Degree Program Development, Stem Cells and Regenerative Medicine

Publication Date 12/06/2020

Defense Date 10/04/2019

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

Tag neural development,OAI-PMH Harvest,postsynaptic density,protein-protein interactions,synaptic

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Ichida, Justin (committee chair), Coba, Marcelo (committee member), Quadrato, Giorgia (committee member)

Creator Email brentwil@usc.edu,brentwilkinson89@gmail.com

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c89-245661

Unique identifier UC11673360

Identifier etd-WilkinsonB-8017.pdf (filename),usctheses-c89-245661 (legacy record id)

Legacy Identifier etd-WilkinsonB-8017.pdf

Dmrecord 245661

Document Type Dissertation

Rights Wilkinson, Brent

Type texts

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

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA