Functions of Drosophila Pak (P21-Activated Kinase) in Morphogenesis: a Mechanistic Model Based on Cellular, Molecular, and Genetic Studies

Functions of Drosophila Pak (p21-activated kinase) in Morphogenesis: A Mechanistic Model based on Cellular, Molecular, and Genetic Studies

Item Type text; Electronic Dissertation

Authors Lewis, Sara Ann

Publisher The University of Arizona.

Rights Copyright © is held by the author. Digital access to this material is made possible by the University Libraries, University of Arizona. Further transmission, reproduction or presentation (such as public display or performance) of protected items is prohibited except with permission of the author.

Download date 07/10/2021 10:07:41

Link to Item http://hdl.handle.net/10150/594389 FUNCTIONS OF DROSOPHILA PAK (P21-ACTIVATED KINASE) IN MORPHOGENESIS: A MECHANISTIC MODEL BASED ON CELLULAR, MOLECULAR, AND GENETIC STUDIES

by Sara Lewis

A Dissertation Submitted to the Faculty of the

GRADUATE INTERDISCIPLINARY PROGRAM

in Partial Fulfillment of the Requirements

for the Degree of

DOCTOR OF PHILOSOPHY IN NEUROSCIENCE

In the Graduate College

THE UNIVERSITY OF ARIZONA

2015

THE UNIVERSITY OF ARIZONA GRADUATE COLLEGE

As members of the Dissertation Committee, we certify that we have read the dissertation prepared by Sara Lewis, titled “Functions of Drosophila Pak (p21-activated kinase) in morphogenesis: a mechanistic model based on cellular, molecular and genetic studies” and recommend that it be accepted as fulfilling the dissertation requirement for the Degree of Doctor of Philosophy.

______Date: November 18, 2015

Dr. Linda Restifo

______Date: November 18, 2015

Dr. Jean Wilson

______Date: November 18, 2015

Dr. Konrad Zinsmaier

______Date: November 18, 2015

Dr. Lynne Oland

Final approval and acceptance of this dissertation is contingent upon the candidate’s submission of the final copies of the dissertation to the Graduate College.

I hereby certify that I have read this dissertation prepared under my direction and recommend that it be accepted as fulfilling the dissertation requirement.

______Date: November 18, 2015 Dissertation Director: Dr. Alan Nighorn

Statement by Author

This dissertation has been submitted in partial fulfillment of the requirements for an advanced degree at the University of Arizona and is deposited in the University Library to be made available to borrowers under rules of the Library.

Brief quotations from this dissertation are allowable without special permission, provided that an accurate acknowledgement of the source is made. Requests for permission for extended quotation from or reproduction of this manuscript in whole or in part may be granted by the head of the major department or the Dean of the Graduate College when in his or her judgment the proposed use of the material is in the interests of scholarship. In all other instances, however, permission must be obtained from the author.

______

SIGNED: Sara Lewis

Acknowledgements

I am grateful to numerous undergraduates and other students who assisted with this research, both directly and indirectly, over the years: Monica Chaung, Adrian Ostertag, Aubrey Downey, Allegra Cohen, Cori Schwartzlow, Cecilia Brown, Rachel Doser, and Grace Samtani. I especially appreciate the hard work of Lauren O’Neill and Josh Paree in immunostaining, morphological analysis, NeuronMetrics, and fly rearing. Many thanks to Judith Tello for years of support as my peer and friend.

Thank you to Jennifer Tifft and William Lewis for editing help on this thesis.

I would like to express my appreciation to Anna Burns for tissue culture training and Patty Jansma for outstanding microscopy assistance and knowledge.

I am forever indebted to the expertise and mentoring of Dave Andrew and Robert Kraft. Thank you for all of your training, feedback, and support to help me develop as a scientist.

I would like to thank the members of my committee: Lynne Oland, Jean Wilson, Konrad Zinsmaier, and Alan Nighorn. I appreciate your advice and the numerous ways you supported my research over the years. I am also grateful to the GIDP in Neuroscience for funding and support and to all of the students and faculty who made graduate school such a rich learning opportunity.

Finally, I would like to thank my mentor, Linda Restifo, for constantly challenging me. I am grateful to my husband, William Lewis, for constant support and encouragement.

Table of Contents

List of Tables and Figures ...... 7

Abstract ...... 8

List of abbreviations ...... 10

Chapter I: Introduction ...... 11

Morphogenesis is the genetically-regulated process of changing tissue shape to form

body structures ...... 11

Intellectual disability, microcephaly, and PAK3 ...... 13

PAK3 regulates proteins that stabilize the actin cytoskeleton ...... 16

Rodent and Drosophila models of PAK3 ...... 19

Drosophila mushroom bodies are higher-order brain structures whose morphology is

genetically regulated ...... 23

Non-muscle myosins have diverse, critical roles in morphology and nervous system

physiology ...... 25

Statement of problem addressed in dissertation ...... 27

Chapter II: Selective regulation of appendage and mushroom body morphogenesis by Drosophila Pak

(p21-activated kinase) ...... 31

Abstract ...... 31

Introduction ...... 32

Results ...... 35

Discussion ...... 42

Materials and Methods ...... 48

Figures and Legends ...... 54

Chapter III: Pak (p21-activated kinase) regulates neurite arbor size and brain structure morphology with apparent roles in branch stability and inhibition of myosin regulatory light chain phosphorylation ...... 68

Abstract ...... 68

Introduction ...... 69

Results ...... 72

Discussion ...... 79

Materials and Methods ...... 83

Figures and Legends ...... 89

Chapter IV: Summary and Future directions ...... 101

Appendix A: Cofilin (twinstar) and Pak interact to regulate neurite branches, but not total neurite length in vitro ...... 106

References ...... 110

List of Tables and Figures

Table 1: Human mutations in PAK3, protein products and function ...... 29

Table 2: Antibodies used for immunolabeling ...... 49

Fig. 1: Pak molecular signaling pathway with other genetic causes of ID ...... 30

Fig. 2: Pak proteins: evolutionary relationships, domain organization, mutations, and accumulation in Drosophila tissues ...... 54

Fig. 3: Effects of Pak mutations on survival to adulthood, longevity, and leg morphology...... 56

Fig. 4: Pak regulates leg morphogenesis...... 58

Fig. 5: The crumpled-wing phenotype of Pak mutants...... 60

Fig. 6: Persistent cells in the wing blades of Pak-mutant adults ...... 62

Fig. 7: Mushroom body α- and β-lobe defects in Pak-mutant brains ...... 64

Fig. 8: Progression of Pak-mutant α- and β-lobe defects during metamorphosis ...... 66

Fig. 9: Pak regulates neurite arbor size and complexity ...... 89

Fig. 10: Truncated Pak protein retains partial function for neurite length, but not higher-order branch number ...... 91

Fig. 11: Regulation of the actin cytoskeleton by Pak ...... 93

Fig. 12: Pak regulates neurite branch stability and defects in the growth cone F-actin and microtubule organization precede branch loss ...... 95

Fig. 13: Pak and myosin interact to regulate neurite arbor size in vitro ...... 97

Fig. 14: Pak and myosin interact to regulate MB morphology in vivo ...... 99

Fig. A1: Interaction between tsr and Pak ...... 108

Abstract

Intellectual disability (ID) is a common phenotype of brain-development disorders and is heterogeneous in etiology with numerous genetic causes. PAK3 is one gene with multiple mutations causing ID.

Affected individuals have microcephaly, and other brain-structure defects have been reported.

Additionally, PAK3 is in a genetic network with eighteen other genes whose mutations cause ID, suggesting the molecular mechanisms by which PAK3 regulates of cognitive function may be shared by other genetic ID disorders.

Studies in rodent models have shown that the orthologs of PAK3 are important for regulating dendrite spine morphology and postnatal brain size. In Drosophila melanogaster, the morphological processes of oogenesis, dorsal closure during embryogenesis, and salivary gland-lumen formation require Pak, the

Drosophila ortholog of PAK3. Additionally, Pak is important for development of the subsynaptic reticulum of the neuromuscular junction, sensory axon pathfinding and terminal arborization in the

Drosophila central nervous system (CNS). However, the role of Pak in mushroom body (MB) structure and intrinsic neurite arbor morphogenesis, as well as details of the underlying cellular and molecular mechanisms are unknown. To address this gap, I used Drosophila models of PAK3 gene mutations, Pak, and a combination of immunostaining, primary cell culture, and genetic interaction studies to elucidate these mechanisms.

I performed a detailed characterization of the previously reported adult Pak phenotypes of decreased survival as well as leg and wing morphology. I found that decreased survival is a low-penetrance phenotype that is enhanced by chromosomes from the same mutagenesis. Defects of the adult wing include folding and misalignment between the layers, blisters, and missing or partial cross veins. The

Pak-mutant legs are short and often misdirected in the pupal case with morphological defects in the shape of the leg segments themselves.

The mushroom bodies are important insect learning and memory brain structures whose lobes are composed of axon bundles with individual axons bifurcating to form the α and β lobes. Mutations in Pak cause defects in the length, thickness, and direction of the MB α and β lobes. These defects increase in severity during metamorphosis, when neurogenesis and differentiation of these structures occur, suggesting that Pak stabilizes the branches of the α/β mushroom body neurons.

Pak-mutant cultured neurons have reduced neurite arbor size with defects in neurite caliber. Initial outgrowth was normal, followed by a decrease in neurite branch number, again supporting the role of Pak in neurite-branch stability. There are defects in the cytoskeleton in growth cones at six hours post-plating as well as in neurons after three days in vitro. The Pak-mutant phenotype severity depends on the phosphorylation status of myosin regulatory light chain, supporting the mechanistic hypothesis that Pak regulates neurite-branch stability by inhibiting myosin light chain kinase. The neuronal phenotype of decreased branch stability suggests a mechanism of excessive retraction as the cellular pathogenesis underlying PAK3 mutation-associated brain disorders.

I used western blotting to characterize the protein products of four nonsense mutations in Drosophila Pak to interpret genotype-phenotype relationships. Each allele has molecularly unique consequences: Pak11, stop-codon read through and truncated protein; Pak16, no read through, but truncated protein; Pak6, read through with no truncated protein; Pak14, neither readthrough nor truncated protein. Truncated proteins produced by Pak11 and Pak16 alleles retained partial function for survival, wing blistering, leg morphology, and neurite length. Conversely, truncated protein increased the severity of the mushroom body defects. Truncated proteins have no effect on neuron branch number, wing folding, or vein defects.

Together, these results demonstrate a role of Pak in regulating epithelial morphology, brain structure, and neurite arbor size and complexity. These closely resemble features of the human disorder, providing evidence that this is a good genetic model for this cause of ID.

List of abbreviations

AID auto-inhibitory domain

ASD autism spectrum disorder

CRIB Cdc42/Rac-interactive binding domain div days in vitro

Dpp decapentaplegic

ECM extracellular matrix

GAP GTPase-activating protein

GEF guanine exchange factor

HE head eversion

ID intellectual disability

LIMK LIM kinase

LTD long term depression

LTP long term potentiation

MB mushroom bodies)

MLCK myosin light chain kinase

MR mental retardation

MRLC myosin regulatory light chain

N.A. numerical aperture

PAK p21-activated kinase

PIX PAK-interacting exchange factor

PP1 protein phosphatase 1

Chapter I: Introduction

Morphogenesis is the genetically-regulated process of changing tissue shape to form body structures

Tissue changes shape to form body structures and organs through a developmental process called morphogenesis; this term can also describe shape changes of individual cells to achieve their final morphology. In Drosophila, the adult legs and wings are formed by the evagination of imaginal discs during metamorphosis in response to the hormone 20-hydroxyecdysone (Fristrom et al., 1973). Imaginal discs undergo massive shape changes by cellular rearrangements, cell division, and cell shape changes

(Condic et al., 1991; Taylor and Adler, 2008). The cells that will become leg segments are arranged in concentric rings within the imaginal disc. The outer ring develops into the part of the thorax and the proximal-most leg segment, the coxa, whereas the central part becomes the distal-most segments, the tarsi. During metamorphosis, the entire disc is drawn out in a telescopic fashion (Nöthiger and Schubiger,

1966; Poodry and Schneiderman, 1970). Legs first elongate to their maximal length as large sacs, then constrict, and finally differentiate the adult bristles and cuticle (Fristrom and Fristrom, 1993). Like the leg disc, the wing disc is also arranged with future distal structures in the center and proximal structures at the periphery of the disc, with the wing blade deriving from the wing disc central pouch (Bryant, 1975).

Like the legs, the entire wing disc first lengthens and folds (Fristrom and Fristrom, 1975), then contracts

(Milner et al., 1984), and undergoes apolysis, where cuticle is deposited on the outer layers (Milner,

1977). During this time, active extracellular transport of decapentaplegic (Dpp) from the putative longitudinal vein cells pattern the cells that will become the cross veins (Blair et al., 2007). After eclosion, wings are unfolded by a sudden influx of hemolymph (Peabody and White, 2013). The final step is wing maturation, during which the epithelial cells undergo programmed cell death (Link et al., 2007), delaminate from the cuticle, and are washed out of the wing and into the thorax by hemolymph (Kiger et al., 2007).

Rac, Rho and Cdc42 GTPase networks regulate morphogenesis

Wing and leg development defects arise from mutations in genes controlling evagination which encode ecdysone-inducible transcription factors, Rho/Rac/Cdc42 GTPase signaling, and non-muscle myosin (von

Kalm, 1995). Rho-family GTPases regulate a wide variety of cellular processes requiring reorganization of the cytoskeleton and cell-shape change. Rac and Cdc42 are required for recruiting F-actin for macrophage phagocytosis (Leverrier and Ridley, 2001) and tissue invagination (Simões et al., 2006).

Additionally, GTPases are crucial for regulating epithelial morphogenesis (Van Aelst and Symons, 2002), such as dorsal closure of the Drosophila embryo which requires Rac, Cdc42 and Rho (Harden et al.,

1999). The processes by which GTPases mediate morphogenesis include cell-cell cadherin-dependent adhesion (Braga et al., 1997), cell-substrate integrin-based focal adhesions (Clark et al., 1998), actin- cytoskeleton remodeling (Eaton et al., 1995), and acto-myosin contractility (Halsell et al., 2000). GTPases also mediate aspects of neuron morphology including Rac-mediated dendrite arborization (Iyer et al,

2012) and increased filopodia formation (Anderson et al., 2005), which is spatially controlled by the localization of an activator, such as Trio Rac GEF (Shivalkar and Giniger, 2012). GTPases are important for axon guidance with the response of the growth cone to attractive cues mediated by Rac/Cdc42 and response to repulsive cues mediated by Rho (Kozma et al., 1997).

Cytoskeleton dynamics change cell shape and drive morphogenesis in the nervous system

Regulated modification of the actin cytoskeleton drives the cellular shape changes, rearrangements, and migration essential for tissue morphogenesis (Lecuit et al., 2011; Davies, 2013). In the central nervous system (CNS), actin cytoskeleton reorganization is a fundamental mechanism of neuronal differentiation and synaptic plasticity (Penzes and Ravolovich, 2012), as well as of large-scale movements such as neural tube closure and cortical morphogenesis (Suzuki et al., 2012; Azzarelli et al., 2015). Mutations in human genes whose products are involved, directly or indirectly, in reorganization of the actin cytoskeleton can cause developmental brain disorders such as intellectual disability (Moon and Wynshaw-Boris, 2013;

Verpelli et al., 2013; Restifo et al., manuscript in preparation). Some of these CNS disorders are

12 syndromes whose defining features also include structural and functional defects of non-neural tissues, further supporting the notion that mechanisms for morphogenesis can be shared between tissues.

Intellectual disability, microcephaly, and PAK3

Etiology of intellectual disability and microcephaly

Intellectual disability (ID), a common phenotype of developmental brain disorders, is defined in DSM-5 as significant limitations in intellectual functioning and adaptive behaviors with onset before age 18

(American Psychiatric Association, 2013). "ID" has replaced the former term "mental retardation” (MR)

(Schalock et al., 2007). The estimated percentage of affected individuals in the US is ~1%, with wide- ranging economic and educational impacts (Boyle et al., 2011; Maulik et al., 2011). The etiology of ID is heterogeneous with causes including chromosomal aberrations, copy number variants, single-gene mutations, and environmental toxins (Kaufman et al., 2010). To date, there are over 700 single-gene causes of ID in the Online Mendelian Inheritance in Man (OMIM) database maintained by NCBI

(National Center for Biotechnology Information) at the NIH (Inlow and Restifo, 2004; Restifo et al., in preparation). A number of these genes have been well-studied, but the roles of most of them in brain development and cognitive function are unknown. These genes are members of interconnected molecular pathways essential for neuronal development and function.

ID can be roughly divided into two major categories: syndromic ID, with a suite of characteristic additional features that accompany the phenotype of reduced cognitive ability, and non-syndromic ID, where the decrease in IQ is the only apparent feature (Kaufman et al., 2010). One of the co-morbid features of ID in some syndromic disorders is microcephaly, defined as a head circumference more than two standard deviations below the mean for age and gender (Roche et al., 1987). Microcephaly is also heterogeneous in etiology, with several genetic and environmental causes (Abuelo, 2007, Kaufman et al.,

2010), and is strongly correlated with reduced cognitive ability (Ashwal et al., 2009). Microcephaly can be either congenital or postnatal, depending on when it appears during development. In a number of well-

13 studied disorders of the nervous system, including developmental disorders (Rett syndrome, schizophrenia, ASD, Fragile X syndrome, and Down syndrome) and neurodegenerative diseases

(Alzheimer's disease), abnormalities of dendritic arbors have been detected in autopsy brain material.

These abnormalities include changes in branch morphology, neurite-arbor length, and in the density and shape of dendritic spines (Kulkarni and Firestein, 2012). These findings suggest that there could be microscopic anatomical defects in many more ID disorders caused by single genes mutations, but more studies are needed to understand the underlying pathogenesis of developmental brain disorders.

PAK3 mutations cause X-linked ID with microcephaly, craniofacial and neuropsychiatric features

PAK3 is a gene on the X chromosome encoding a p21-activated kinase. Mutations in PAK3 cause intellectual disability (Donnelly et al. 1996; Allen et al., 1998; Bienvenu et al., 2000; Gedeon et al., 2003;

Peippo et al., 2007; Rejeb et al., 2008; Magini et al., 2014). Some carrier females have cognitive deficits, but without reaching the threshold for a diagnosis of intellectual disability (Peippo et al., 2007). In another family in which none of the carrier females were affected, they had skewed X-inactivation with inactivation of more than 80% of the chromosomes carrying the PAK3 mutation (Rajeb et al., 2008).

ID caused by PAK3 mutations was initially classified as MRX30 (MIM #300142); the phenotype designation of “MRX” means nonspecific, or non-syndromic, X-linked ID (Donnelly et al., 1996; Allen et al., 1998). However, further clinical investigation revealed additional phenotypes. These include craniofacial features such as long ears, high palate, and high-bridged nose. Affected individuals have microcephaly (Donnelly et al., 1996; Allen et al., 1998; Gedeon et al., 2003; Peippo et al., 2007; Rejeb et al., 2008; Magini et al., 2014) and one patient whose brain was imaged by MRI has corpus callosum agenesis and cerebellar hypoplasia (Magini et al., 2014). Other features include oral hypotonia (Bienvenu et al., 2000; Gedeon et al., 2003; Peippo et al., 2007; Rejeb et al., 2008; Magini et al., 2014) and epilepsy

(Gedeon et al., 2003, Peippo et al., 2007; Rejeb et al., 2008; Magini et al., 2014). Some individuals have neuropsychiatric features such as anxiety, restlessness, aggression, paranoia, and even psychosis (Allen et al., 1998; Gedeon et al., 2003; Peippo et al. 2007; Rejeb et al., 2008, Magini et al., 2014). In one family,

14 the intellectually disabled boys with PAK3 mutations had ichthyosis, a severe skin condition (Magini et al., 2014). Recently, whole-exome DNA sequencing revealed two novel PAK3 mutations in boys with cerebral palsy (McMichael et al., 2015), suggesting that PAK3 function is required for much more than merely cognitive performance as the original description of MRX had suggested.

Functional and molecular consequences of mutations in PAK3

PAK3 is expressed in the brain (Allen et al., 1998) from early embryonic stages throughout adulthood

(Brain Span Consortium of the Allen Human Brain Atlas, Hawrylycz et al., 2012; Human Brain

Transcriptome Project, Kang et al., 2011). Disease-causing mutations are described in Table 1 and Fig. 2.

Three missense and one nonsense mutation are located in the kinase domain (Allen et al., 1998, Gedeon et al., 2003, Peippo et al., 2007, Magini et al., 2014) and decrease the function of the kinase domain as assessed by a cellular assay (Magini et al., 2014). One missense mutation is in the CRIB domain which decreases binding of Cdc42/Rac and activation of PAK3 (Bienvenu et al., 2000). One mutation is a splice variant causing early truncation with no protein accumulation (Rejeb et al., 2008). These mutations suggested a model of loss-of-function causality of PAK3-mutant phenotypes.

However, in the family studied by Magini and colleagues (2014), the PAK3 mutation (K389N) allowed accumulation of the kinase-inactive protein that activated MAPK signaling, and these individuals have additional brain structure and skin defects. Expressing kinase-dead mutant PAK3 protein (R419X and

A365E) in rodent hippocampal neurons resulted in slightly decreased spine number, but caused elongated, immature-appearing spines, while expressing mutant protein with decreased GTPase binding (R67C) dramatically decreased spine density (Kreis et al., 2007). The range of PAK3 clinical and experimental phenotypes suggests the possibility that a kinase-inactive mutant protein may dysregulate multiple signaling pathways and cause developmental abnormalities that are more severe than simple loss-of- function phenotypes (Allen et al., 1998; Magini et al., 2014).

PAK3 regulates proteins that stabilize the actin cytoskeleton

PAK is in a GTPase signaling pathway with eighteen other ID genes

PAKs are activated by small GTPases in the Rho family, Rac and Cdc42 (Edwards et al., 1999). The greatest activation comes from Cdc42 (Kreis et al., 2007). Rac and Cdc42 activity is regulated by

GTP/GDP: when bound to GTP, the protein can interact with effectors, but when bound to GDP, the

GTPase is inactive. The activity of GTPases is controlled by regulatory proteins of the GTPase cycle.

Guanine nucleotide exchange factors (GEFs) activate Rho GTPases by catalyzing the exchange of bound

GDP for GTP. GTPase activating proteins (GAPs) which stimulate the endogenous GTPase activity of the

GTPase, hydrolyzing GTP to GDP. Guanine dissociation inhibitors (GDIs) inhibit the removal of GDP and keep the GTPase in an inactive form (Boguski and McCormick, 1993). There are over 70 GEFs and

GAPs that have been identified in mammals (Bai et al., 2015). Despite evidence that these proteins regulate localized GTPase activity in different tissues, respond to different cues, and activate different downstream targets, the function of very few specific GEFs and GAPs are known (Bai et al., 2015;

Rossman et al., 2005). However, some have known roles in disease including neurodegenerative disease, schizophrenia, and intellectual disability (Bai et al., 2015), suggesting that they have distinct, physiologically relevant roles.

According to Kaufman (2010), a number of the single-gene causes of ID, including PAK3, fall into conserved molecular pathways with other genes with mutations causative for ID. Therefore, there are conserved pathways required for cognitive abilities (Inlow and Restifo, 2004). PAK3 is part of a genetic network with eighteen other genes with mutations causing ID (Fig. 1, pg. 29). These genes include the

GEFs (shown in orange) ARHGEF6 (encoding alpha PIX, a Rac/Cdc42 GEF; Ramakers, 2002),

ARHGEF9 (Rac GEFl; Ba et al., 2013), FGD1 (Cdc42 GEF; Ramakers, 2002). GAPs include OPHN1

(Rho GAP; Ramakers, 2002), SRGAP3 (Rac GAP; Ba et al., 2013).

PAKs decrease the activity of myosin light chain kinase (Sanders et al., 1999; Goeckeler et al., 2000;

Wirth et al., 2003) and increase the activity of LIMK (Edwards et al., 1999). In addition to the upstream

16 activators of PAK3 activity, possible downstream targets of PAK3 are also known ID genes, including two isoforms of non-muscle myosin (MYH10; Tuzovic et al., 2013; MYOVa; Pastural et al., 1999) and two actin genes (ACTB, ACTG1; Rivière et al., 2012) and LIMK (Ramakers, 2002).

PAK3 also regulates activity of MAPK which is part of a signal-transduction cascade including the products of the genes HRAS, KRAS, BRAF, MAPK1, MAPK2, PTPN11, and RAF1 (Aoki et al., 2008). A target transcription factor of MAPK signaling is CREB, whose activity is regulated by the products of two other ID genes, CREBBP and P300 (Roelfsema et al., 2005). Together, these findings suggest that PAK3 activity may be a pivotal mechanism for regulating several signaling pathways that are essential for normal cognitive function. Studies of this pathway may elucidate mechanisms of cognitive function and brain development as well as provide potential targets for therapeutics.

Pak phosphorylates target kinases which regulate the cytoskeleton

The small GTPases Rho and Rac/Cdc42 are known to have important, but differing, roles in regulating cytoskeleton assembly (Moorman et al., 1999). Constitutively active RhoA causes stress fiber formation and focal adhesion assembly (Ridley and Hall, 1992; Moorman et al., 1999). Constitutively active Cdc42 causes increased growth cone size and number of filopodia (Brown et al., 2000). Constitutively active

Rac1 results in membrane ruffling and formation of lamellipodia (Ridley et al., 1992). Rac and Rho have antagonistic functions in dendrite arborization (Lee et al., 2000), epithelial to mesenchymal transition

(Royal et al., 2000), and F-actin regulation in Drosophila oogenesis (Vlachos and Harden, 2011).

One of the known targets of Rac and Cdc42 GTPases is PAK (Manser et al., 1994). PAK phosphorylates downstream effectors including LIMK (Edwards et al., 1999), myosin light chain kinase (MLCK) (Zeng et al., 2000, Chew et al., 1998; Wirth et al., 2003) and MEK1 (Park et al., 2007). LIMK phosphorylates and decreases the activity of cofilin (Arber et al., 1998; Yang et al., 1998), a protein that cleaves filamentous actin (Bamburg, 1999). PAK phosphorylation and increases activity of LIMK which reduces cofilin activity (Edwards et al., 1999). Without PAK, cofilin can be overactive, which prevents the

17 accumulation of F-actin (Arber et al., 1998), increases the rate of growth cone motility and neurite extension (Endo et al., 2003), and increases the size of axonal terminals at the neuromuscular junction

(Ang et al., 2006). Therefore, PAK stabilizes the F-actin cytoskeleton.

MLCK phosphorylates myosin regulatory light chains (MRLC) and allows the myosin hexamer to change from a folded to an extended conformation (Ikebe et al., 1988), such that it can bind actin and initiate contraction (Adelstein and Conti, 1975). Myosin classes are defined by the type of heavy chains which associate with diverse regulatory and essential light chains (Weiss and Leinwand, 1996). Myosin has various roles in regulating the cytoskeleton including neurite retraction in response to repulsive cues, short-distance transport of essential proteins, and retrograde actin flow to recycle actin during growth cone extension (Kneussel and Wagner, 2013). PAK phosphorylation of MLCK decreases its activity and subsequently decreases myosin activity. Therefore, PAK activity inhibits myosin contraction.

PAK can also phosphorylate MAPK/ERK kinase 1 (MEK1), activating the MAP kinase cascade (Menges et al., 2012; Roberts and Der, 2007). MAPK has a host of targets including CREB binding protein (Liu et al., 1999), a transcription co-factor of CREB (Meng et al., 2005) which is important for learning (Impey et al., 1998) and long-term memory (Yin et al., 1994). Together, these pathways show that PAK3 is important for cytoskeleton stability. Loss-of-function mutations in PAK3 could cause cofilin and myosin overactivity or reduced activity in MAPK which could result in dysregulation of the actin cytoskeleton in neurons.

Rodent and Drosophila models of PAK3

Rodent PAK3 regulates dendrite spine morphology

In the vertebrate nervous system, the majority of excitatory synapses are made onto dendritic spines of the post-synaptic neuron (Alvarez and Sabatini, 2007). High-frequency or chemical stimulation results in long term changes in synaptic strength (long term potentiation, or LTP) which can lead to increased spine size (Bosch et al., 2014) and filopodia extension to form new spines (Malenka and Nicoll, 1999). These spine changes require changes in the actin cytoskeleton: stimuli that induce LTP increased actin polymerization and stimuli that cause long term depression (LTD) increased actin depolymerization at the same time that spine shape changes (Okamoto et al., 2004). LTP can be blocked by applying drugs that inhibit actin polymerization or depolymerization (Kim and Linsman, 1999). Even without changes in synaptic strength, the actin cytoskeleton in mature spines is undergoing constant turnover. Around 85% of the actin in the spine is dynamic (Star et al., 2002).

Boda and colleagues used hippocampal slice cultures to express PAK3 RNAi or the dominant-negative

R419X mutation to reduce endogenous PAK3 function. Both treatments resulted in fewer mature dendritic spines and more elongated, filopodial-like spines. Adding R419X mutant PAK3 reduced synaptic function within the spines with reductions in levels of AMPA receptors, reduced spontaneous activity, and defects in LTP (Boda et al., 2004). Transgenic mice expressing the PAK AID by a forebrain driver have reduced density of dendritic spines, increased spine length, enhanced LTP and reduced LTD in cortical neurons (Hayashi et al., 2004). These findings demonstrate that mutations in a protein important for both cognitive function and the regulation of actin cause bidirectional synaptic defects and changes in synaptic morphology.

PAK3 and PAK1 together regulate postnatal brain growth

Mice with the kinase region of PAK3 knocked out have normal brain size, dendritic arbor size, spine morphology, spine density, and F-actin distribution as assessed by hippocampal cell culture and Golgi impregnations. These PAK3 kinase domain knockout mice have defects in LTP. Assessments of the

19 phosphorylation status of known molecular targets of PAK showed that there was no change in basal cofilin or MAPK phosphorylation, but levels of phosphorylated/activated CREB were reduced (Meng et al., 2005). Mice with the PAK1 locus deleted also have normal neuronal and brain morphology, but have decreased F-actin in the dendritic spines and LTP defects. Knockout of PAK1 caused decreases in activated cofilin in response to NMDA receptor-activation, although basal levels of cofilin and basal and activated levels of MAPK and MLCK activity were all normal (Asrar et al., 2009).

However, double-mutant mice with knockout of PAK1 and the kinase domain of PAK3, although born with normal brain size and normal neuron number, develop reduced brain size during the postnatal period, with reduced dendritic arbors and synaptic density in the hippocampus (Huang et al., 2011). Double- mutant mice have reduced synapse density with enlarged spines and a higher proportion of spines with immature morphology in both hippocampal and cortical neurons. In dendrites of wild-type hippocampal neurons cultured for 21 div, F-actin is upregulated in actively growing processes and spines of neurons compared to the dendritic shaft. In the PAK1 and PAK3 double-mutant neurons; there is no difference in intensity of F-actin between spine and shaft, suggesting that the levels of F-actin are reduced in dendritic spines. The brains of these animals had increased levels of basal cofilin phosphorylation with no change seen in levels of phosphorylated MLCK or MAPK (Huang et al., 2011), suggesting that cofilin is sensitive to levels of PAK. MLCK or MAPK are dynamically regulated and lack of a basal change does not necessarily rule out PAKs as an important regulator of MLCK or MAPK activity. Together, these findings suggest that PAK1 and PAK3 could have partially redundant roles in rodent brains, possibly related to the role of both PAK1 homodimers and PAK3/PAK1 heterodimers in signaling (Combeau et al., 2012; Parrini et al., 2002).

Drosophila Pak is the ortholog of human PAK3 with conserved protein domains

Drosophila melanogaster is an excellent model system for studying hereditary developmental brain disorders due to conservation of genes important for brain development and complex behaviors.

Additionally, experimental methods, such as genetic labeling and primary neuron culture, allow for study

20 of structure and function at the single-gene level (Restifo, 2005). Of 282 ID genes examined by Inlow and

Restifo in 2004, 87% had a fruit fly homolog and 76% had at least one functional ortholog; this shows that genes important for cognition are surprisingly well-conserved between flies and humans. Drosophila

Pak is the closest ortholog to human group-1 PAK genes which are PAK1, PAK2, and PAK3 (Fig. 2A).

Gene orthology was determined by Mentzel and Raabe in a phylogenetic analysis of the DNA encoding the PAK kinase domain (2005). There is an insect-specific Pak, indicated as Pakins. Group-2 PAK genes are PAK4, PAK5, and PAK6 in vertebrates (Jaffer and Chernoff, 2002; Wells and Jones, 2010). The

Drosophila group 2 ortholog is known as mushroom body tiny (mbt; Mentzel and Raabe, 2005), which also acts to regulate the actin cytoskeleton of photoreceptors (Mentzel et al., 2007) and CNS neurogenesis

(Melzig et al., 1998).

Protein domains of PAKs are well-conserved between both human PAK3 and Drosophila Pak (Fig. 2B).

The amino-terminal PxxP domain interacts with the SH3 domain of Nck/Dock and is important for recruiting Pak to the membrane for proper function (Hing et al., 1999). Activation by Cdc42/Rac occurs by disrupting the physical interaction between PAKs autoinhibitory domain (AID) and the kinase domain

(Manser et al., 1994), allowing the protein to autophosphorylate the activation loop between the AID and the kinase domain (Zenke et al., 1999; Pirruccello et al., 2006; Chong; 2001; Parrini et al., 2002). This allows the carboxyl-terminal kinase domain to adopt an active conformation (Zhao et al., 1997) and phosphorylate substrates including LIMK (Edwards et al., 1999), MLCK (Zeng et al., 2000, Chew et al.,

1998; Wirth et al., 2003) and MEK1 (Park et al., 2007). As revealed by X-ray crystallography and in vitro biophysical studies, the majority of PAK1 in vivo is present in either homodimers (Lei et al., 2000) or heterodimers with PAK3, and dimerization is important for regulating PAK3 activity (Combeau et al.,

2012). Therefore, the AID domain of a kinase-dead Pak protein could have negative function by binding to other Pak or dPak* proteins and inhibiting their function (Conder et al., 2004).

Functions of Drosophila Pak

Drosophila Pak is required for regulating cytoskeletal changes that drive oogenesis, dorsal closure during early embryogenesis, and morphogenesis of specialized tissues. Elongation of the egg chamber requires

Pak to bundle F-actin in follicle cells (Conder et al., 2007). Dorsal closure relies on Pak for formation of adherens junctions and epithelial cell constriction (Conder et al., 2004). Salivary gland lumen formation requires Pak for endocytosis of E-cadherin from adherens junctions and relocation to the basolateral membrane (Pirraglia et al., 2010). Pak is a mediator of programmed cell death in the larval salivary glands during metamorphosis (Ihry and Bashirullah, 2014) with constitutively active Pak causing apoptosis of salivary gland cells (Pirraglia et al., 2010).

In vivo Drosophila phenotypes of Pak mutations in sensory neurons include ectopic targeting of olfactory neuron axons within the antennal lobe (Ang et al., 2003) and disorganization of photoreceptors entering the optic lobe (Hing et al., 1999). These findings identify a role for Pak in axon guidance and terminal arborization. Pak is required for post-synaptic glutamate receptor localization and subsynaptic reticulum ultrastructure at the neuromuscular junction (Albin and Davis, 2004), demonstrating a role in synaptic maturation. Together, these findings demonstrate that Pak is important for morphogenesis and has an important role in the development of the nervous system. Pak may play an important role in survival, leg extension, and wing development (Hing et al., 1999; Ihry et al., 2014), although these functions have not been studied in detail.

Drosophila mushroom bodies are higher-order brain structures whose morphology is genetically regulated

Mushroom bodies are important for learning, memory, and complex behaviors

Mushroom bodies (MBs) are paired neuropil structures in the Drosophila central brain that are essential for certain types of learning and memory (Pascual and Préat T, 2001; Heisenberg, 2003) with well- characterized developmental processes (Lee et al., 1999). Drosophila adults and larvae with mutations in genes that disrupt MB development are impaired in olfactory associative learning and memory

(Heisenberg et al., 1985; Pauls et al., 2010). Genes with mutations that cause learning and/or memory defects in Drosophila are expressed in and required for function of the mushroom bodies, such as rutabaga (Han et al., 1992; Zars et al., 2000; Mao et al., 2004), dunce (Nighorn et al., 1991; Gervasi et al., 2010), and dfmr1 (Pan et al., 2004; Bolduc et al., 2008). Feeding hydroxyurea to newly-hatched larvae kills the proliferating neuroblasts, chemically ablates the MBs, and also creates defects in olfactory learning and memory (de Belle and Heisenberg, 1994). Together, these studies illustrate the critical role of the MBs as a learning and memory structure.

Other complex behaviors determined to be dependent on the MBs. For instance, MB function suppresses locomotor activity. Flies with disrupted MB function via genetic mutation or chemical ablation, showed an increase in time spent walking. This was due to increased time of individual but no change in the initiation of walking bouts (Martin et al., 1998). MBs are also necessary for courtship memory (McBride et al., 1999; Sakai and Kitamoto, 2006), aggressive behavior (Edwards et al., 2009), and sleep (Pitman et al., 2006).

Mushroom body structure and development

Each intrinsic MB neuron extends a primary neurite hat bifurcates near the cell body to send a dendritic arbor into the calyx and a single axon projecting anteriorly and ventrally through the peduncle and into the lobes. The peduncle and lobes are tightly-packed bundles of fibers. Axons of the α/β and α′/β′ neurons

23 bifurcate and project into the medial and dorsal lobes. Adult γ neurons do not bifurcate and project only into a medial lobe (Lee et al., 1999; Crittenden et al., 1998).

The three sets of neurons (γ, α′/β′, and α/β) that form the MBs differentiate sequentially during development from common neuroblast precursors (Lee et al., 1999) that continue to proliferate until 90 hours after puparium formation, which is shortly before eclosion (Ito and Hotta, 1992). Gamma mushroom body neurons are born before the 3rd instar stage. The α′/β′ neurons are born between mid-3rd instar larval stage and puparium formation. Finally, α/β neurons are born after puparium formation (Lee et al., 1999). Larval γ neurons bifurcate into dorsal and medial lobes. During metamorphosis, γ MB neurons are induced to retract by exposure to ecdysone. The axons then regrow into a single, unbifurcated axon that projects medially (Lee et al., 1999).

Genetic mutations can cause defects in MB structure

Many genetic mutations that cause MB structural defects have been identified. dfmr1 encodes the fragile

X mental retardation protein which is a RNA-binding protein, localization factor, and translational repressor. dfmr1 mutations cause β-lobe fusion, due to extension of β lobes across the midline of the brain (Michel et al., 2004). The gene ciboulot encodes a β-thymosin ortholog that binds to G-actin and enhances polymerization. Overexpression of ciboulot also results in β-lobe fusion (Boquet et al., 2000).

Twinstar encodes cofilin, an actin depolymerization factor that severs filamentous actin (F-actin)

(Bamburg, 1999). Loss-of-function mutations in twinstar cause short α/β axons. MARCM clones with twinstar mutations have neuronal processes that do not reach the end of the lobe (Ng and Luo, 2004).

Cofilin is inhibited via phosphorylation by LIMK (Arber et al., 1998). Overexpressing LIMK or LIMK- activating factors, including Rac, Rho, Cdc42, Rho-activating GEF, Pak and Rok (Maekawa et al., 1999), in MB neurons also causes short lobes (Ng and Luo, 2004).

The Rho-Rok-myosin pathway regulates MB lobe stability (Billuart et al., 2001). Inhibiting p190

RhoGAP, a negative regulator of Rho signaling, results in short α and β lobes with defects that become progressively worse during development. This was shown to require Rok and MRLC function, indicating that the short lobes were due to overactive Rho-mediated myosin retraction. Conversely, overexpression of p190 RhoGAP caused overextension, where some axon processes of the MB continue past the rest of the lobe and may cross the midline (Billuart et al., 2001). Taken together, this suggests that Rac, Cdc42 signaling and their downstream effectors are important for MB axon extension and Rho GTPase signaling and their effectors are important for stabilization.

Non-muscle myosins have diverse, critical roles in morphology and nervous system physiology

Myosin plays an important role in growth cone traction and dynamics. According to the ‘clutch hypothesis’, myosin can either mediate retraction of filopodia or move the growth cone forward (Suter and Forscher, 2000). In this model, filopodia protrusion is driven by actin polymerization when the base of the actin filament is anchored to the membrane. Additionally, myosin moves membranes or other material, such as organelles, forward in extending filopodia (Jay, 2000). Myosin stabilization of the actin cytoskeleton with the N-cadherin complex is crucial for stabilization of dendritic spines in hippocampal neurons. Adhesion reduces F-actin flow and is required for myosin-mediated tension (Chazeau et al.,

2015).

Multiple forms of myosin have been demonstrated to have important roles in neuronal physiology and morphology. Myosin I (Ruppert et al., 1993), II (Rochlin et al., 1995), V (Evans et al., 1997), and VI

(Suter and Forscher, 2000) are localized to neuronal growth cones. Myosin Ic is required for retaining lamellipodia and retrograde actin flow in growth cones (Diefenbach et al., 2002). Myosin IIA causes neurite retraction in response to Rho kinase-dependent thrombin treatment (Wylie and Chantler, 2003), neurite retraction in response to the repulsive cure semaphorin 3A (Gallo, 2006), and F-actin turnover

25 after treatment with jasplakinolide, an inhibitor of F-actin recycling (Gallo et al., 2002). Myosin IIB is important for spreading of lamellipodia in neuronal growth cones, filopodial traction, and growth cone motility (Bridgman et al., 2001). Myosin II also drives the cell shape changes necessary for Drosophila leg disc eversion (Edwards and Kiehart, 1996). Myosin V is an organelle motor for short-range transport, including into synaptic spines (Hammer and Wagner, 2009) and myosin Va is responsible for moving

FMRP-bound granules of mRNAs and ribosomes to sites of mRNA translation (Lindsay and McCaffrey,

2014). Myosin Va also participates in targeting transmembrane proteins, such as GluR1, to the dendrites

(Lewis et al., 2009). Myosin Vb is responsible for transporting recycling endosomes into synaptic spines via Rab11 (Wang et al., 2008), such as in rhodopsin secretory transport (Li et al., 2007). Myosin VI traffics proteins to the axon (Lewis et al., 2011), and localizes and tethers synaptic vesicles at the

Drosophila neuromuscular junction (Kisiel et al., 2014). Myosin VI is also part of a complex that regulates F-actin, total dendrite length, and primary dendrite number (Lv et al., 2015)

Myosin is a hexamer of three pairs of subunits: the heavy chain which distinguishes the different forms of myosin, regulatory light chain (MRLC), and essential light chain. These genes are well conserved between humans and Drosophila. In Drosophila, spaghetti squash encodes the regulatory light chain

(Karess et al., 1991). Myosin ATPase and motor activities are greatly increased with phosphorylation of

MRLC on residues 20 and 21 (Jordan and Karess, 1997). Phosphorylation of MRLC is known to be mediated by myosin light chain kinase (MLCK) (reviewed in Tan et al., 1992), or Rho kinase (Kosako et al., 2000; Ueda et al., 2002) and negatively regulated by protein phosphatase 1 (Vereshchagina et al.,

2004). The protein conformation change caused by phosphorylation can be mimicked by replacing serine with glutamic acid (Kamisoyama et al., 1994); in the case of sqh, the glutamate substitution mimics phosphorylation at S21 by MLCK. Constitutively active MLCK and excessive MRLC phosphorylation caused midline crossing defects of pCC interneurons in the ventral nerve cord of Drosophila embryos

(Kim et al., 2001). Together, this shows that MRLC phosphorylation is probably tightly regulated during development to achieve the right balance of neurite extension and retraction.

Statement of problem addressed in dissertation

ID is a common phenotype of developmental brain disorders that affects 1-3% of the population.

Developmental brain disorders can be caused by mutations of genes, many of them in shared pathways.

PAK3 has mutations known to cause ID and is centrally positioned in a pathway with eighteen other ID genes, indicating that its function is crucial for cognitive function. Numerous lines of evidence suggest that PAK3 may regulate neuron morphology. For instance, agenesis of the corpus callosum indicates that

PAK3 is important for axon tract development. PAK3 mutations cause microcephaly and there is one reported instance of cerebellar hypoplasia, indicating that PAK3 is required for brain size. Affected individuals also have cranial-facial malformations which indicate a role of PAK3 in morphogenesis of other tissues. PAK activators Cdc42 and Rac are known regulators of the actin cytoskeleton. These phenotypes imply PAK3 is important for neuronal and brain morphology. However, a role of PAK3 in regulating neuronal morphogenesis has not been demonstrated, except for excitatory spines.

Based on phenotypes present in humans, rodents, and flies, I predicted that Pak, the Drosophila PAK3 ortholog, is important for neuronal morphology, MB morphology, and cytoskeleton organization. I utilized published evidence from biochemical assays to infer molecular interactions between Pak and non-muscle myosin. This had not been previously studied in neurons. The aim of these studies was to characterize the role of Pak on external structures and MB morphogenesis, identify genotype-phenotype relationships, and explore cellular and molecular mechanisms of neurite arbor morphogenesis.

In Chapter 2, I address the following questions:

 Is Pak required for adult survival and morphogenesis of the leg and wing? What features are

present?

 Does the accumulation of truncated Pak have partial, negative, or no effect on the severity or

penetrance of Pak-mutant phenotypes?

 Does Pak regulate MB morphology and how do these defects develop?

In Chapter 3, I examine the cellular and molecular mechanisms of Pak in regulating nervous system morphology. Questions addressed include:

 What is the cell-intrinsic phenotype of the Pak-mutant neurite arbor?

 How does this phenotype develop?

 Does Pak regulate the actin cytoskeleton? Is important during or post-outgrowth?

 What are the downstream molecular targets of Pak in neurons?

Table 1: Human PAK3 mutations, resulting protein products, and effects on function

Mutation Protein present? Change in function References R419X truncated, accumulates* no kinase function*, ** Allen et al., 1998 R67C accumulates* no Cdc42/Rac GTPase binding** Bienvenu et al., 2000 A365E accumulates* no kinase function*, ** Gedeon et al., 2003 W446S accumulates* decreased kinase function* Peippo et al., 2007 Frameshift none predicted no GTPase binding or kinase function Rejeb et al., 2008 128X predicted K389N accumulates* no kinase function* Magini et al., 2014 R439C not tested predict reduced kinase function McMichael et al., 2015 M100I not tested predict reduced Cdc42/Rac GTPase McMichael et al., binding 2015 *=tested by Magini et al. in cellular assay, **=tested by Kreis et al., 2007 in pull-down assay

Fig. 1: Pak molecular signaling pathway. Numbers in parenthesis indicate number of genes with mutations known to cause ID.

Chapter II: Selective regulation of appendage and mushroom body morphogenesis by Drosophila Pak (p21-activated kinase)

Abstract

Mutations in p21-activated kinase 3 (PAK3; MIM#300142) cause X-linked intellectual disability with microcephaly and craniofacial defects. PAKs are activated by Cdc42/Rac GTPases, and regulate the actin cytoskeleton in diverse tissues. Drosophila Pak is the sole ortholog of the group-1 PAK-gene family, encoding PAK1, PAK2, and PAK3. We studied the function of Drosophila Pak in morphogenesis as a model for understanding PAK3-mutant phenotypes.

Pak mutations disrupt morphogenesis of thoracic appendages and mushroom-body lobes. Mutant wings are small with numerous tiny folds involving the distal blade. Blisters and missing cross veins are also common. Legs are short, malformed, with both anterior>posterior and proximal>distal gradients. During metamorphic remodeling, the adult-specific mushroom-body α/β lobes show abnormal size and position, with these defects worsening over time. Of four allegedly-null nonsense mutations, some produce truncated protein with no kinase domain and/or full-length proteins via stop-codon read-through. The genotype-phenotype relationships are consistent with partial retained function, and in some cases, with additional negative effects.

We found that Pak is required for morphogenesis of structures that undergo dramatic elongation and adhesion changes during metamorphosis, including axon tracts required for cognition. The functionality of truncated and read-through PAK proteins deserves further investigation. These findings parallel the clinical phenotypes caused by PAK3 mutations.

Introduction

Regulated modification of the actin cytoskeleton drives the cellular shape changes, rearrangements, and migration that are essential for normal tissue morphogenesis (Lecuit et al., 2011; Davies, 2013). In the central nervous system (CNS), actin cytoskeleton reorganization is a fundamental mechanism of neuronal differentiation and synaptic plasticity (Penzes and Ravolovich, 2012), as well as of large-scale movements such as neural tube closure and cortical morphogenesis (Azzarelli et al., 2015; Suzuki et al.,

2012). Therefore, it is not surprising that developmental brain disorders can result from mutations in human genes whose products are involved, directly or indirectly, in reorganization of the actin cytoskeleton (Inlow and Restifo, 2004; Verpelli et al., 2013; Wynshaw-Boris, 2013). Some of these CNS disorders are syndromes whose defining features also include structural and functional defects of non- neural tissues.

Among the key regulators of the actin cytoskeleton are PAKs, members of a small family of highly conserved p21-activated serine/threonine kinases (Fig. 2A-B; Rane and Minden, 2014). In this report, we focus on the group-1 PAK subfamily, which includes mammalian PAK1, PAK2, and PAK3, as well as

Drosophila melanogaster Pak (Fig. 2A; Mentzel and Raabe, 2005). The amino-terminal PxxP domain

(Fig. 2B) binds the SH3 domain of Nck/Dock, which is important for recruiting PAK to the membrane for proper function. The highly-regulated carboxy-terminal kinase domain has multiple phosphorylation substrates, including other kinases with important signaling roles. As their name indicates, PAKs are activated by small GTPases, specifically by Rac and Cdc42, in response to extracellular signals (Edwards et al., 1999; Van Aelst and Symons, 2002). Binding of GTP-bound Cdc42 or Rac to the CRIB domain of

PAK disrupts the physical interaction between the autoinhibitory (AID) and kinase domains that blocks the active site. Once relieved of this blockade, PAK autophosphorylates the activation loop between AID and the kinase domain and, thereby, activates the PAK kinase (Pirruccello et al., 2006).

The phosphorylation substrates of group-1 PAK kinase activity depend on the specific system studied and include myosin light chain kinase (MLCK; Goeckeler et al., 2000; Sanders et al., 1999; Wirth et al., 2003) and LIM kinase (LIMK; Edwards et al., 1999). Phosphorylation of these targets results in inhibition of actomyosin contraction and inhibition of cofilin-based cleavage of filamentous actin (F-actin), respectively. Thus, PAK function tends to stabilize the actin cytoskeleton. In addition, PAK can phosphorylate MAPK/ERK kinase 1 (MEK1), activating the MAP kinase (MAPK) cascade (Menges et al., 2012; Roberts and Der, 2007). MAPK plays important roles in morphogenesis of diverse tissues

(Bromberg-White et al., 2012) and, through CREB-mediated transcription, in mechanisms of synaptic plasticity (Kreis and Barnier, 2009).

Mutations in the X-linked human gene PAK3 (MIM #300142) cause intellectual disability (Allen et al.,

1998). Subsequent evaluation of patients’ clinical records revealed more complex phenotypes. Many of the affected boys also have microcephaly (indicating small brain size), hypotonia of the mouth and throat musculature, distinct craniofacial malformations, and behavioral abnormalities including psychosis and aggression (Magini et al., 2014; Peippo et al., 2007; Rejeb et al., 2008). In one family, the intellectually disabled boys with PAK3 mutations had a severe skin condition (ichthyosis) and gross morphological defects of brain development – agenesis of the corpus callosum and cerebellar hypoplasia (Magini et al.,

2014). Recently, whole-exome DNA sequencing revealed two novel PAK3 mutations in boys with cerebral palsy (McMichael et al., 2015). Clearly, PAK3 has more diverse roles in regulating development than was suggested by the original phenotype designation of “MRX,” meaning nonspecific, or non- syndromic, X-linked intellectual disability (ID). (MR refers to “mental retardation,” a term that has been replaced by “intellectual disability”; Schalock et al., 2007.)

Owing in part to the rarity of identified PAK3 mutations, efforts to understand genotype–phenotype relationships are at a relatively early stage. Several disease-causing PAK3 mutations (Fig. 2B; Allen et al.,

1998; Gedeon et al., 2003; Peippo et al., 2007) encode proteins with reduced or absent kinase activity in a cell-based assay (Magini et al., 2014). Other alterations in PAK3 sequence prevent binding of Cdc42/Rac

33 to the CRIB domain (Bienvenu et al., 2000), which would prevent PAK activation, or encode truncated proteins (Rejeb et al., 2008). These data suggested a model of loss-of-function causality of PAK3-mutant phenotypes. However, in the family with the additional skin and brain phenotypes (Magini et al., 2014), the PAK3 mutation allowed accumulation of the kinase-dead protein. The range of PAK3 clinical phenotypes suggests the possibility that a kinase-inactive mutant protein can dysregulate signaling pathways and cause developmental abnormalities that are more severe than simple loss-of-function phenotypes (Allen et al., 1998; Magini et al., 2014).

Pak is the sole Drosophila ortholog of the group-1 PAK gene subfamily (Fig. 2A; Harden et al., 1996;

Mentzel and Raabe, 2005). The availability of Pak-mutant alleles and the already published phenotype data suggest that Drosophila genetics could provide a whole-animal model with which to explore the range of mutant-tissue phenotypes and their relationships to different mutant alleles.

Pak for bundling F-actin in follicle cells (Conder et al., 2007). Dorsal closure relies on Pak for formation of adherens junctions and epithelial cell constriction (Conder et al., 2004). Salivary gland lumen formation requires Pak for endocytosis of E-cadherin from adherens junctions and relocation to the basolateral membrane (Pirraglia et al., 2010). In the Drosophila nervous system, Pak is essential for the normal pattern of sensory axon projections into the primary visual (Hing et al., 1999) and olfactory (Ang et al., 2003) areas of the CNS. At the neuromuscular junction, Pak regulates the clustering of postsynaptic glutamate receptors and formation of the subsynaptic reticulum (Albin and Davis, 2004).

In this study, we focused first on external adult structures, wings and legs, defects of which had been briefly noted in Pak mutants by previous investigators (Hing et al., 1999; Ihry and Bashirullah, 2014).

Because of the fully penetrant ID phenotype of human PAK3 mutations, we then looked for evidence of

Pak function in the mushroom bodies, the brain region best established as essential for learning and

34 memory in insects (Heisenberg, 2003; Menzel, 2014). Interpretation of our findings required re- evaluation of fundamental features of Pak developmental genetics, including the functional status of extant mutant alleles and the question of whether the gene is essential for adult viability. Together, our data provide novel insights into Pak function during Drosophila development and establish the value of

Drosophila genetics for future investigation of human PAK3 mutations.

Results

Characterization of Pak-encoded proteins in wild-type and mutant tissues

The four Pak alleles used for this study were generated in two separate EMS mutageneses, with each resulting from a different nonsense mutation (Fig. 2B, pg. 54): Pak6 and Pak11 (Hing et al., 1999), Pak14 and Pak16 (Newsome et al., 2000). Chromosomes bearing each of these mutant alleles are homozygous lethal early in development. Based on the predicted absence of a kinase domain and lack of detectable full-length Pak protein by immunoblotting, these alleles were reported to be null mutations (Hing et al.,

1999; Newsome et al., 2000). We analyzed Pak protein accumulation in larval tissue homogenates of hemizygous animals (Pak allele/deficiency), using more sensitive detection methodology (Fig. 2C-G), in order to better understand the similarities and differences among the Pak-mutant alleles. As in the previous immunoblotting studies by other groups, we used the polyclonal antibody raised against the amino-terminal end of Pak (Harden et al., 1996).

Based on migration distance, the most abundant Pak proteins in wild-type homozygous (+/+) and hemizygous (+/Df) larval tissue have mean estimated sizes of 91 kDa and 113 kDa (Fig. 2C-D), based on five independent experiments using either whole wandering 3rd instar larvae or the CNS only. These represent the previously reported small and large Pak isoforms, which have predicted molecular weights of 76 and 90 kDa, respectively, based on amino acid sequence (NP_731074 and NP_001262310; NCBI

Protein; Amino Acid Calculator). Cross-reacting material of average apparent molecular weight ~77 kDa was seen at the same low-abundance level in all genotypes (Fig. 2D, asterisk). This may represent insect-

35 specific Pak (Fig. 2A; Mentzel and Raabe, 2005), which shares conserved protein features and antigenicity (S. Wang and N. Harden, personal communication). An additional low-abundance band migrating just ahead of the small Pak isoform with an apparent size of 82 kDa appears only in the presence of the Df(3R)BSC745 (Fig. 2D, pg. 54) or Df(3R)BSC744 (not shown) chromosomes. The origin of this deficiency-specific Pak-immunoreactive protein is uncertain.

In the hemizygous Pak-mutant samples from CNS (Fig. 2D box) or whole larvae (Fig. 2G), tiny Pak proteins were detected in Pak11 (~29 kDa vs. 24 kDa predicted) and in Pak16 (~30 kDa vs. 24.5 kDa predicted) tissue. This confirmed the accumulation of low-abundance truncated Pak due to premature

STOP codons in these two mutations just downstream from the PIX-coding domain (Fig. 2B). The truncated Pak proteins were not detectable until the experimental methods allowed high sensitivity

(compare Figs. 2C and 2D). Accumulation was estimated at ~4% of the small Pak isoform in +/Df CNS and in whole-larval tissue.. In contrast, there was no truncated protein detected in Pak6 or Pak14 hemizygous larval CNS samples (Fig. 2D, dashed box). When gel electrophoresis was optimized for small proteins, a very faint band was detected in Pak14/Df CNS samples migrating at ~17 kDa (Fig. 2F, dashed box), which is considerably larger than the predicted molecular weight of 11.5 kDa.

An unexpected finding was the detection of wild-type-sized small and large Pak isoforms in hemizygous mutant tissues, suggesting STOP-codon read-through. This was most evident with Pak11 and Pak6 alleles in both the CNS (Fig. 2D) and whole-larval samples (data not shown). When compared with +/Df control samples (Fig. 2E), the estimated abundance of the full-length Pak proteins in Pak11 or Pak6 hemizygous tissue was 5-6% (if only the small and large Pak isoforms were included) or 19-34% (if the Df-specific band was included as well). Also notable in the Pak-mutant tissues is the marked decrease in the relative abundance of the small full-length Pak isoform (Fig. 2E).

Taking all of the larval protein data into consideration, it appears that each of the four Pak mutations has unique molecular consequences (Fig. 2E). Of the four alleles, Pak14 is the closest to a protein-null.

Conversely, Pak11 gives rise to both a truncated product and full-length read-through proteins. The other two alleles express either a truncated protein (Pak16) or full-length read-through proteins (Pak6).

Adult survival and longevity are differentially affected by Pak-mutant alleles

Contrary to previous reports that Pak is required for survival to adulthood (Hing et al., 1999; Ihry and

Bashirullah, 2014), we estimated lethality in Pak-mutants to be 0-33% for compound heterozygotes of alleles from independent mutageneses (Fig. 3A, pg. 56). This incompletely penetrant lethality is uncovered by two deficiency chromosomes and partially rescued by addition of a single copy of a Pak+ transgene. Compound heterozygotes with two truncation alleles (Pak11/Pak16) have full viability.

Examining each of the four alleles over the same deficiency chromosome (Df(3R)BSC744) reveals differences in the severity of the phenotypes. The most severely affected genotypes are alleles that produce no truncated protein (Pak14). Flies with alleles encoding truncated protein have relatively lower lethality (Pak16) and alleles producing truncated and some full-length protein have further reduced lethality (Pak11). The severity of the hemizygotes is higher than the compound heterozygotes, suggesting that these alleles retain partial function and are not acting as nulls. Combinations of alleles created in the same mutagenesis (Pak11/Pak6 = 76%, Pak14/Pak16 = 90%) enhance the penetrance of the lethality.

Hemizygotes with a larger deletion chromosome (Df(3R)Win11, Ki) also have enhanced lethality compared to hemizygotes with smaller deletions (Df(3R)BSC744).

We examined survival post-eclosion and found that Pak-mutant adults have a markedly reduced life span

(Fig. 3B). Pak11/+ heterozygous controls survive for longer than three weeks while both Pak-mutant genotypes have greatly reduced survival. Interpolating from the curve, 50% of Pak6/Pak14 animals died by

~ 2 days and 50% of Pak11/Pak16 animals were dead by ~4 days. The mean life span of Pak6/Pak14 was 3.2 days and 7.4 days for Pak11/Pak16. In other words, compound heterozygous adults with alleles encoding truncated Pak protein, (Pak11/Pak16) survived significantly longer than those without (Pak6/Pak14) (p < 10-

16 determined by Mann-Whitney rank sum test), suggesting that truncated Pak protein retains partial function for adult longevity.

Femur-selective leg morphogenesis defects of Pak mutants

In Pak-mutant pharate adults, the folded legs viewed through the pupal case were short and often deviated from the normal straight trajectory, splaying to one side of the animal (Fig. 4B, pg. 58). The specific defects contributing to the leg malformations included short segments (Fig. 4E, G, I), constrictions (Fig.

4E, F), bowing (Fig. 4F, I), distensions (Fig. 4H), and bends (Fig. 4J). Very rarely, we observed missing segments (Fig. 4G). Comparing the three sets of legs, we found a steep anterior-posterior gradient, with prothoracic (T1) legs being more commonly affected than meso- (T2) or metathoracic (T3) legs (Fig. 4K).

Comparing the leg segments, we also found a marked predominance of morphological defects in the femur compared with the more distal tibia and tarsi (Fig. 4L). Imaginal disc cells from the future femur region (Fig. 4L inset) undergo the greatest cell-shape changes and rearrangements to drive leg extension during leg morphogenesis (von Kalm, 1995).

For a number of genetic mapping experiments, we used two leg-phenotype measures (Fig. 3A), calculating penetrance, the percentage of animals of a given genotype with any leg defects, and expressivity, the percentage of legs with defects. The baseline frequency of any leg defects was 3% in each of three heterozygotes. The leg-morphogenesis phenotypes are highly but incompletely penetrant

(50-85% affected) in Pak compound heterozygotes with allele pairs from different mutageneses (Fig. 3A).

Compound heterozygotes producing no truncated proteins (Pak6/Pak14) had the highest penetrance, with

85% of flies affected, and the greatest expressivity (25% of legs). Two deletion chromosomes that remove the Pak gene, Df(3R)BSC744 and Df(3R)Win11 Ki pp, uncover the leg phenotypes. With the smaller deficiency, Df(3R)BSC744, penetrance and expressivity were lower in hemizygotes able to produce truncated Pak (compare Pak11/DfBSC744 and Pak6/DfBSC744). For both kinds of alleles, penetrance and/or expressivity are higher with the very large Win11 deletion, which uncovers >100 genes and carries

38 the semi-dominant marker Kinked (Ki). The penetrance of the leg defects is enhanced by 3rd chromosomes from the same mutagenesis (Pak11/Pak6 =95%, expressivity = 77%). When the Win11 deletion chromosome is used to provide a single copy of a wild-type Pak transgene, Pak11 is fully rescued to baseline heterozygote levels and Pak6 shows strong partial rescue, with the affected legs dropping from

60% to 14%. Thus, the leg morphogenesis defects are recessive phenotypes that map to Pak.

Pak regulates distal wing morphogenesis and maturation

In the wings of adult Pak mutants, we found evidence of folding within the dorsal and ventral layers and misalignment between the layers that was fully penetrant in all genotypes examined. The overall effect is that the wings appear small, dark, and “crumpled” like used tissue paper. The proximal portion of the wing is normal with ridges appearing in the distal portion of the wing blade, and a sharp demarcation between them (Fig. 5B-C, pg. 60). This often created a net dorsal or ventral curvature of the blade which, in turn caused a fold when the wing was mounted (Fig. 5C, arrows). Some wings had blisters with discrete margins (Fig. 5D) and some wings were entirely blistered with no points of adhesion between dorsal and ventral layers, giving the wing a ballooned appearance (Fig. 5E). The blisters phenotype was significantly more pronounced in compound heterozygotes with alleles that do not produce truncated protein (Pak6/Pak14) when compared to the allele pair producing truncated protein (Pak11/Pak16; Fig. 5L, p

= 0.004, based on Chi-squared analysis with 2 degrees of freedom). Additionally, the anterior and posterior cross veins had defects in length and width. Some cross veins are abnormally wide near their intersection with the longitudinal veins and relatively normal in the center, creating an hour-glass shape

(Fig. 5G, I). Cross veins could be completely missing (Fig. 5G, H, J, dots) or only partially formed (Fig.

5H, K, arrowheads). We quantified the cross-vein length defects and found that they were present in all

Pak-mutant genotypes examined (Fig. 5L). In agreement with the distal-proximal gradient of folding, the more distal posterior cross vein was more severely affected than the anterior cross vein (p = 0.02, based on Chi-square analysis with 2 degrees of freedom).

During wing maturation in the hours shortly after eclosion, the wing blade epithelial cells delaminate from the cuticle (Kiger et al., 2007), undergo programmed cell death, and are washed out of the wing and into the thorax with the aid of hemolymph pulses (Link et al., 2007). In Pak6/+ control wings, the only cells visible after wing maturation are those of the mechanosensory bristles along the anterior wing margin (Fig. 6A, B, pg. 62). Epithelial cells comprising the veins are also present (Blair, 2007), although we did not detect their nuclei in the veins of control wings with DAPI. However, Pak-mutant wings have cells remaining in the wing blade after maturation organized in rows (Fig. 6D), small clusters (Fig. 6E), large patches (Fig. 6G) and in patterns suggesting longitudinal and cross veins (Fig. 6J, K). Even in regions that appear relatively devoid of cells, individual nuclei were apparent at high magnification (Fig.

6H).

Mushroom body (MB) lobes have defects in axon branches and direction with negative effect of truncation alleles

We examined the α and β lobes of the MBs, mirror-image, symmetrical neuropil structures composed of axon-like α/β neuron fibers that bifurcates and projects medially and dorsally into both the α and β lobes.

These α/β axon fibers extend during metamorphosis and remain tightly bundled. The α and β lobes are readily visualized by FasII immunostaining, with weaker anti-FasII labeling present on the unbranched, medially projecting γ neurons (Crittenden et al., 1998), antennal lobes (AL), and ellipsoid body (eb). Pak- mutant brains have α- and β-lobe morphogenesis defects including short, thin (Fig. 7B, pg. 64) and the more severe defect, missing, lobes (Fig. 7C). Misdirected lobes were also present in Pak-mutant brains

(Fig. 7C). Midline fusion of the β-lobes is present in some of the mutant brains (Fig. 7D). Each α and β lobe, with two α and two β lobes per brain, was scored based on the following categories: normal, short, thin, missing, or misdirected. β-lobe fusion was based on the thickness of the bundle crossing the midline, which cannot be scored unless both β lobes extend fully.

The missing and misdirected α- and β- lobe phenotype is moderately penetrant (98% of compound heterozygous brains, 35-60% of lobes, Fig. 7E). Within a given genotype, there was no sexual dimorphism in penetrance or types of defects (data not shown). Simple heterozygotes of Pak-mutant alleles or deficiencies have rare, low-severity α- or β-lobe defects (Fig. 7F). This phenotype maps to Pak using multiple compound heterozygous allelic combinations, two different deficiency chromosomes, and rescue by the addition of a single copy of a Pak+ transgene on one of the deficiency chromosomes (Fig.

7G). Combinations of alleles encoding truncated Pak protein (Pak11/Pak16) cause defects with a higher penetrance with more lobes “missing”, the most severe category, compared to other compound heterozygotes (0.0002 < p < 0.02, based on Chi-squared analysis with three degrees of freedom).

Hemizygotes with alleles encoding truncated Pak (Pak11/DfBSC744) cause significantly more severe defects compared to hemizygotes producing no truncated protein (Pak6/DfBSC744; p < 0.0004).

Mushroom body defects worsen through development

To determine if Pak MB lobe phenotypes are due to a defects of outgrowth or maintenance, we examined mutant mushroom bodies through development. We used Pak11/Pak16 compound heterozygotes, the most severely affected genotype from the pharate adult MB analysis, as late-stage, wandering 3rd instar larva and at tightly-staged time points after head eversion. The 201Y-GAL4 driver was used in combination with the UAS-lacZ reporter (Brand and Perrimon, 1993) to label γ mushroom body neurons with β- galactosidase (Kraft et al., 1998). In both control and mutant brains, the γ neurons bifurcate to form the dorsal (DL) and medial lobes (ML) of the larval mushroom bodies (Fig. 8A, B, pg. 66), which degenerate and regrow into a single γ lobe by HE+24 (Lee et al., 1999). Pak-mutant larval MBs (Fig. 8B) and pupal

γ-lobes (not shown) appeared normal, suggesting normal γ neuron degeneration and regrowth. The α/β neurons are born and differentiate after HE, and by HE+24, α and β lobes are visible with anti-FasII staining (Fig. 8C). Over the next two days, these lobes thicken as newly born α/β neurons extend their processes within a tightly packed bundle until they reach their adult morphology by HE+72 (Fig. 8E, G).

The α and β lobe defects seen in pharate adult MBs were also present in pupae, including thin compared

41 to stage-matched controls and other lobes in the same brain (arrows, Fig. 8D), short (arrowhead, Fig. 8H), missing entirely (dotted outline, Fig. 8F), and misdirected lobes (not shown). Each α and β lobe was scored using the same categories as for the pharate-adult MB analysis (Fig. 8I).

Relatively early in pupal development, there were already some defects of the Pak11/Pak16 α and β lobes, and between HE+24 and HE+48, there was no difference in the severity of the defects. However, there is a significant increase in penetrance of α- and β-lobe defects between 48 and 72 hours post-HE (p < 0.04).

Between HE+72 and shortly before eclosion both penetrance and severity increase, with the majority of affected lobes being missing (p < 0.004, based on Chi-squared analysis). In addition, fewer misdirected lobes were seen at later stages of pupal development than at the earlier stages. There were no obvious defects in Pak11/+ α and β lobes during development (quantitative data not shown).

Discussion

Truncated and full-length protein from Pak nonsense mutations

We characterized the protein products of four Pak alleles and confirmed Pak11 and Pak16 produce truncated protein and that Pak6 and Pak14 do not. The truncated protein appears to retain partial function for leg development and adult survival. This is similar to salivary gland development where Pak6 homozygous embryos have more severe defects than Pak16, although this could also be due to recessive enhancers on the Pak6 chromosome (Pirraglia et al., 2010). In contrast, we also found that truncated Pak protein worsens MB defects. This truncated protein includes the autoinhibitory domain (AID) which is capable of inducing Pak phenotypes of dorsal closure defects and optic lobe axon guidance defects in an otherwise normal background (Conder et al., 2004). The AID can dimerize to inhibit Pak kinase activity and could be interacting with DmPakins (Combeau et al., 2012), the Group-2 Pak (mbt, mushroom bodies tiny), or full-length Pak generated by stop codon read through.

In all four of the alleles evaluated, some of the mutant mRNA evades nonsense-mediated decay and full- length protein is generated by read through. Read-through translation has been demonstrated in

Drosophila using ribosome-protected mRNA sequencing (Dunn et al., 2013). Furthermore, stop codons has been found within evolutionarily conserved open reading frames and constructs fused with GFP have demonstrated read through translation in vivo (Jungreis et al., 2011). We found that Pak6 and Pak11 alleles produce the most full-length protein in larval tissues. The amino acid substitutions that take place during read-through translation may alter the function of the protein and deserve further investigation. Based on the accumulation of protein products and evidence of partial function and negative functions from phenotypic analyses, we conclude that these alleles are not necessarily simple loss-of-function alleles.

Rather, they have with unique molecular consequences and cannot be treated as equivalent.

Mutations in Pak cause reduced survival and adult longevity

We found Pak mutations cause lethality before eclosion, but at lower rates than other reports (Hing et al.,

1999; Ihry and Bashirullah, 2014). This is probably because other studies used alleles from the same mutagenesis which enhances lethality, and could be due to other deleterious genes from the parental chromosome. This suggests that the Pak lethality phenotype is highly sensitive to second site enhancers and chromosomal deletions. Additionally, we found that Pak is required for adult longevity, with truncated protein retaining some function.

Pak is required for morphogenesis of selective regions of thoracic appendages

Pak regulates not only the length of leg extension, but also segment direction and width. A number of genes required for imaginal disc extension during remodeling cause short legs with segment defects, mostly affecting the femur and tibia, such as abnormally bent segments. These genes include transcription factors which mediate ecdysone response (i.e. broad, Kiss et al., 1988), small GTPase signaling (RhoA and RhoAGEF2, Halsell et al., 2000), regulation of myosin contraction (Drok, Drak, zipper, Bayer et al.,

2003; spaghetti squash, Edwards and Kiehart, 1996). Interestingly, the proximal>distal gradient of segments defects caused by mutations in Pak is reminiscent of the leg malformations caused by mutations in these genes. This suggests that Pak could play a role in regulating myosin contractility in addition to

43 organization of the actin cytoskeleton. These leg defects could potentially contributed to the uncoordinated behavior we observe, and was also briefly reported by Hing and colleagues (1999).

We found that Pak is required for wing development with mutations causing small wings, numerous microfolds in the distal portion, missing and partial cross veins, and blisters. Cross-vein formation requires extracellular transport of Dpp, a BMP ligand, with transport defects resulting in an accumulation of Dpp in the base of the future cross vein. This creates a characteristic hourglass shape (Matsuda et al.,

2013), similar to what we observed in some Pak-mutant cross veins. Dpp activates the Rho GAP

Crossveinless-c to suppress Rho signaling, which can be antagonized by Pak (Vlachos and Harden, 2011).

As described by Matsuda et al. (2013), defects in either the extracellular transport or signal transduction required for vein differentiation could explain the partial and missing cross veins in Pak mutants.

The process by which epithelial cells leave the wing during maturation requires Pak. Pak is a mediator of programmed cell death in the larval salivary glands during metamorphosis (Ihry and Bashirullah, 2014), with constitutively active Pak causing premature apoptosis of salivary gland cells in embryos (Pirraglia et al., 2010). Pak has been shown to transduce apoptosis signaling by activating MAPK p38 in response to stress forces sensed through integrin-ECM binding (Shifrin et al., 2012). Alternatively, Pak could be important for regulating integrin-based adhesion to the ECM of the wing cuticle since constitutively active integrin causes the persistence of epithelial cells and blister formation (Kiger et al., 2007).

Epithelial cells remaining in the wing blade due either to defects in programed cell death or adhesion could explain the defects of dorsal and ventral layer alignment, adhesion, and folding in Pak-mutant wings.

Expression of the gene four-jointed is localized in the developing wing in a distal>proximal gradient with mutations causing decreased wing size specific to the region distal to the anterior cross vein (Zeidler et al., 2000). It is also localized to the future femur region of leg discs with mutations causing dramatic shortening of leg length in addition to the fusing of the second and third tarsal segments (Villano and

Katz, 1995). The selectivity of wing and leg region regulated by Pak is unusual, but bears a resemblance to the localized expression of the four-jointed , suggesting that Pak activity is spatially restricted in the developing wing and leg.

Pak regulates MB development by stabilizing α and β lobes

Abnormal mushroom body development has been reported for mutants of Drosophila ID gene orthologs, such as highly penetrant β-lobe fusion in fragile X mental retardation 1 mutants (Michel et al., 2004) and in mutants of the CBP ortholog, nejire (R. Kraft, unpublished observations). We found that Pak-mutant α and β MB lobes were often short, thin, missing, and misdirected with less common β-lobe fusion. Some observers report a subtle decrease in lobe thickness of Pak11/+ heterozygotes compared to wild-type MBs.

However, small differences in lobe volumes are difficult to quantify with this immunostaining approach and may warrant further investigation to determine if there is a dominant-negative effect of the truncated protein created by Pak11 and Pak16 on brain development. If this is the case, this would suggest that the

AID domain is blocking other Pak proteins which may have partially overlapping roles with Pak.

Throughout metamorphosis, the MB defect phenotype progressed from relatively milder defects of short and thin lobes to the more severe defect of missing lobes. Therefore, we propose that missing lobes are due to branch retraction, with short and thin lobes potentially mid-retraction. Mutations in p190 RhoGAP, a negative regulator of Rho, ROCK, and myosin II-mediated contraction, also have progressive phenotypes, consistent with retraction of α/β axons during pupal development (Billuart et al., 2001).

Therefore, one possible mechanism for Pak-mediated lobe stability is inhibition of non-muscle myosin.

Since this study is focused on whole-lobe morphology, studies of individual neurons may reveal more widespread branching defects. Conversely, the β-lobe fusion seen in Pak mutants is probably due to

45 neurite overextension. Rac1 (Ng et al., 2002) and twinstar (Ng and Luo, 2004) mutants have short MB lobes due to reduced axonal growth; this phenotype can be suppressed by inhibiting Pak. Misdirected lobes indicate a pathfinding defect, which is similar to what has been seen in the photoreceptor axons projecting into the optic lobes (Hing et al., 1999) and olfactory neurons projecting into antennal lobe glomeruli (Ang et al., 2003). This suggests that missing/partial lobes, β-lobe fusion, and misdirected lobes may be due to distinct mechanisms.

We observed reduced staining of anti-FasII in Pak-mutant mushroom body lobes. Mutations in dpix, a

GEF activator of Rac/Cdc42 signaling that binds to Pak, cause an estimated 20% reduction in anti-FasII signal in the neuromuscular junction (Parnas et al., 2001) and Pak is important for the localization of

FasII at the NMJ (Albin and Davis, 2004). However, it is unlikely that the lobe defects are artifacts of reduced anti-FasII labeling because the short and thin lobes are suggestive of intermediate phenotypes and samples have normal background anti-FasII labeling in other regions of the brain, including at the base of the peduncle, antennal lobe, and γ lobe, in addition to other α and β lobes.

Concluding remarks

We have provided evidence of partial loss-of-function and in some cases, negative functions of the four nonsense mutations that have been most extensively used to study Pak to date. We have further characterized the role of Pak in morphogenesis of adult appendages of legs and wings. Additionally, we clarified previously published results on the requirement of Pak for survival. Finally, we demonstrated a novel role for Pak in brain-structure morphology, possibly through stabilization of neurite branches. The loss of neurites that make up brain structures could contribute to reduced brain size in other animal models of PAK3 and the human phenotype of microcephaly. For instance, mice with a double knockout of

PAK1 and PAK3 are born with normal brain size and normal neuron number, but have reduced brain size in the postnatal period with reduced dendrite arbors and synapse density (Huang et al., 2011). Since Pak is part of a genetic network with at least eighteen other genes with mutations causing ID (Ba et al., 2013;

Tuzovic et al., 2013; Rivière et al., 2012; Aoki 2008; Roelfsema et al., 2005; Ramakers, 2002; Pastural et al., 1999; Nagase et al., 1995), studies to understand how Pak regulates neuron-branch stability are needed and may provide targets for therapies.

Materials and Methods

Drosophila culture and genetics

Fly stocks were maintained at room temperature (22-23°C) and experimental cultures were reared at

25ᵒC, 60-80% relative humidity, as previously described (Kraft et al., 2006), on corn flour- nutritional yeast-agar medium (Elgin and Miller, 1978). OreR-C (a wild-type laboratory strain), w(z) (a “Cantonized” strain of w1118), a second-chromosome UAS-lacZ reporter line, and the 201Y-GAL4 driver for γ mushroom body neurons have been in use in our laboratory for many years (Kraft et al., 1998, 2006). All Pak-related stocks were obtained from the Bloomington Drosophila Stock Center at Indiana University (funded by

NIH P40 OD018537), including mutant alleles (Pak6, Pak11, P{neoFRT, ry+}82B Pak14, P{neoFRT, ry+}82B Pak16), the wild-type Pak transgene, and chromosomal deletions, in descending size order:

Df(3R)Win11 Ki pp (Pattatucci et al., 1991), Df(3R)BSC744, and Df(3R)BSC745, P[w+] (Cook et al.,

2012). Wild-type Pak function was provided by a P-element construct with a Pak+ cDNA, inserted on a

Pak-deletion chromosome: T2.2 Df(3R)Win11 Ki pp (Hing et al., 1999), abbreviated P[Pak+]

Df(3R)Win11. All Pak stocks were maintained over a balancer, either TM6B, Tb e or TM3, Ser P[Act-

GFP]. Crosses, e.g., to create specific compound heterozygotes or hemizygotes, were performed by methods that optimize larval culture density to minimize competition and maximize mutant development and survival.

For developmental studies, bubble prepupae (stage P3-P4) were selected and incubated at 25°C in a humid Petri dish on wetted Whatman No. 42 ashless filter paper (VWR, Radnor, PA). They were monitored for head eversion and dissected at 24, 48, or 72 hours thereafter. Late-stage, slowly moving, wandering 3rd instar larvae were selected and dissected as a pre-metamorphosis time point. P14-15 pharate adults were used for the final pre-eclosion time point.

Adult survival was estimated by dividing the number of eclosed flies of the desired genotype by the expected number, based on balancer siblings. Lethality was calculated as (1 – survival) X 100%.

Longevity of adult flies was assayed by selecting animals as pharate adults and sequestering single individuals in small tubes with a plug of food in humidity-controlled incubators. The animals were observed daily for 21 days following eclosion (day 0) with flies transferred to fresh food every seven days. The number of days from eclosion to the day prior to death is the number of days survived.

Significant differences in number of days survived between genotypes calculated using Mann-Whitney rank sum test.

Antibodies

Table 2: Antibodies used for immunolabeling

Antigen/ Host and Type Antibody Immunogen (conjugate) Source Identifier Use Primary Antibodies N'-Pak peptide Rabbit polyclonal N. Harden, Univ. of anti-DPAK WB (first 15 AAs) British Columbia Beta-tubulin Mouse monoclonal Developmental Studies E7-s WB Hybridoma Bank Fasciclin II (FasII) Mouse monoclonal Developmental Studies 1D4 anti-FasII-s T from Drosophila Hybridoma Bank Beta-galactosidase Rabbit polyclonal Cappel 55976 T Secondary Antibodies Rabbit IgG Donkey polyclonal LI-COR 926-32213 WB (IRDye® 800CW) Mouse IgG Donkey polyclonal- LI-COR 926-32222 WB (IRDye® 680) Mouse IgG Donkey polyclonal Jackson 715-165-150 T (Cy3) ImmunoResearch Rabbit IgG Goat polyclonal Molecular Probes A11034 T (Alexa Fluor® 488) AAs: amino acids; T: tissue (whole-mount); WB: Western blot

Protein analysis by immunoblotting

Wandering 3rd instar larvae were selected, cleaned in ethanol and water, and used either whole or microdissected to explant the entire CNS. Tissue was homogenized on ice in lysis buffer with a broad- spectrum cocktail of protease inhibitors: either cOmplete® tabs (Roche, Basel, Switzerland) reconstituted in PBS, pH 7.4 or a lab-made solution of 62.5 mM Tris-HCl, pH 6.8, with 2% SDS w/v, 0.1 μg/μl

49 pepstatin A, 0.5 μg/μL leupeptin, and 10mM PMSF. The homogenates were centrifuged at 16,000 g, 4ᵒC, for 10 minutes, and the supernatants diluted 1:1 with 2X Laemmli sample buffer (Bio-Rad, Hercules,

CA), for a final concentration of lysis buffer with 2.1% w/v SDS, 26.3% v/v glycerol, 0.01% w/v bromophenol blue, plus 5% v/v β-mercaptoethanol (Thermo Fisher Scientific, Waltham, Massachusetts).

Protein samples were stored briefly at 4ᵒC and heated for 5 minutes at 95ᵒC immediately before loading.

Molecular weight standards (EZ-Run, Thermo Fisher Scientific), spanning 170 to 10 kDa were included on each gel. SDS-PAGE electrophoresis was performed using a 4% acrylamide stacking gel and either

7.5% or 12% acrylamide separating gel with running buffer (25 mM Tris, 192 mM glycine, 0.1% SDS, pH 8.3) at a constant current of 0.2 mA. Proteins from each gel were transferred onto 0.2-μm nitrocellulose membrane (GE Healthcare, Little Chalfont, UK) using TransBlot® Turbo (Bio-Rad) at a constant voltage of 25V for 30 minutes in Towbin buffer (25 mM Tris, 192 mM glycine, 0.1% v/v SDS,

20% v/v methanol, pH 8.3).

The membrane was first blocked with 5% dry milk in TBS for 60 minutes at RT, incubated to detect Pak proteins, rinsed, and imaged. The membrane was then briefly rinsed in blocking solution and stained for beta-tubulin (without stripping the membrane first), rinsed, and imaged. Primary antibodies (anti-Pak, diluted 1:2500-1:5000 and anti-β-tubulin diluted 1:90) were diluted in 5% dry milk in TBS and incubated overnight at 4ᵒC. Secondary antibodies (Donkey anti-Rabbit-800 and Donkey anti-Mouse-680, 1:20000) were diluted in 5% dry milk in TBS with 0.01% SDS and 0.1% Tween-20® and incubated for 1 hour at room temperature. All rinses were performed 3 times for 5 minutes each with TBST (TBS with 0.01%

Tween-20®) at room temperature with a final wash in TBS before visualization.

Western blots were visualized using Odyssey scanner (LI-COR, Lincoln, Nebraska) at wavelengths based on the secondary antibody. Exposures were optimized to give the highest sensitivity without saturation

(intensity setting between 4.5-8.0) at 169-μm resolution and images were exported as 600-dpi tiff files.

Quantification of fluorescent signal intensity was performed using Image Studio Lite (LI-COR) to subtract in-lane background and normalize to in-lane total protein loaded using the beta-tubulin signal.

The raw intensities correlating to the maximum and minimum display intensity were set using the LiCor curves function to optimize contrast for viewing; this does not alter raw pixel intensity values used for quantification.

Molecular weights of detected protein bands were estimated using the plot of log10 of the protein-standard molecular weights vs the Rf (distance migrated/distance of dye front). The means of the molecular weights were calculated from five western blot experiments with two replicates of whole-larval tissue and three replicates of CNS-only tissues. Predicted protein sizes were calculated based on amino acid sequence using a software tool developed by the Lab of Biological Mass Spectrometry at University of

Washington Department of Genome Sciences (http://proteome.gs.washington.edu/cgi-bin/aa_calc.pl).

Morphological analysis of appendages

For leg assessment, animals were selected as late pharate adults (stage P14-15 based on Bainbridge and

Bownes, 1981) and dissected from the puparium and pupal cuticle. Individual segments of all six legs were scored by two observers by stereomicroscopy for the presence of defects of the femur, tibia, and tarsi; the coxa and trochanter were not systematically scored. For photomicroscopy, legs were removed using iridectomy scissors and mounted in 100% glycerol. Images were acquired at an Olympus BX51TF compound microscope (Tokyo, Japan) with 4X objective (0.13 N.A.) and Olympus DP70 digital camera.

For wing assessment, adults were collected 1-4 days post eclosion and fixed in 70% ethanol for at least 24 hours. Both wings were removed using iridectomy scissors and scored individually by stereomicroscopy for the presence of blisters. Ballooning indicates complete failure of adhesion between the dorsal and ventral layers of the wing blade; the margin of the blister was in the wing hinge. Alternatively, smaller discreet blisters had margins contained within the wing blade; occasionally more than one blister was present on a wing. For cross-vein analysis, wings were dissected as above and mounted in 100% glycerol for examination of the veins by compound microscopy (Olympus BX51TF) with 10X and 20X objectives

(0.3 and 0.5 N.A., respectively). Lengths of the anterior and posterior cross veins were scored as complete, partial, missing, or not-scorable if wing folding obscured the vein.

For imaging of nuclei in adult wings, one-day-old adults were collected and fixed in 70% ethanol for 24 hours. Wings were removed with iridectomy scissors, mounted in VectaShield® DAPI mounting media

(Vector Laboratories, Burlingame, CA), coverslipped, and incubated at 4ᵒC for 24 hours to allow DAPI to penetrate. Wings were washed in PBS with gentle rocking for 2 hours at room temperature and remounted in 12% w/v PVA (polyvinyl alcohol; Sigma, St. Louis, MO) with 1.5% w/v DABCO (1,4-diazabicyclo

[2.2.2] octane; Sigma) to retard fading. Samples were viewed on an Olympus BX51TF microscope through 10X and 100X objectives (0.3 and 1.40 N.A., respectively) with epifluorescence (filter cube U with excitation filter BP330-385 nm, barrier filter 420 nm).

Fluorescence labeling of mushroom bodies and microscopy of whole-mount brain preparations

CNS tissue was fixed in EM-grade 4% formaldehyde (Ted Pella, Redding, CA) and immunostained as described in Michel et al., 2004. All samples were incubated in anti-FasII primary antibody diluted 1:25 in PTN (0.1 M sodium phosphate buffer, pH 7.4, 0.1% v/v Triton-X, 0.1% w/v sodium azide) and some were double-labeled with preabsorbed anti-β-galactosidase diluted 1:2500 in PTN overnight at 4°C and in secondary antibodies (Donkey anti-Mouse diluted 1:500 in PTN and Donkey anti-Rabbit diluted 1:500 in

PTN) for 3 hours at RT. When double-labeling for β-galactosidase and FasII, pairs of primary antibodies or secondary antibodies were added together. Samples were mounted in PVA with DABCO to minimize fading of fluorescent signals.

The mushroom bodies were visualized by laser-scanning confocal microscopy using a Zeiss 510 microscope with a 40X (1.3 N.A.) oil-immersion objective and LSM 4.2 SP1 image-acquisition software

(Zeiss, Jena, Germany). Samples were scanned with a helium-neon laser line with an excitation maximum of 543 nm and a long-pass filter at 565 nm for Cy3, and with an argon laser with an excitation max of 488 nm and a band-pass filter from 505-550 nm for Alexa 488. Double-labeled samples were imaged using

52 the multi-track line such that each channel’s signal is collected separately. Image projections were created in LSM Image Browser using a subset of sequential optical sections selected for optimal visualization of the MB lobe regions.

Scoring and classification of mushroom body phenotypes

MB morphology was assessed by examining individual sequential optic sections (z = 1.0 μm) and their projections of anti-FasII-labeled brains with special attention to the MBs and ellipsoid body regions. Each

α- and β-lobe was scored separately (four α/β lobes per brain). The α- and β-lobe defects were classified as “normal”, “thin/short” or “missing” based on the thickness and length of the lobes, if present, compared to stage-matched controls. Thin lobes were often variable in diameter along their length. Short lobes do not project the normal distance and appear truncated. Because a lobe could be both short and thin, and both of these defects seemed to represent an intermediate phenotype, the categories were combined. A missing lobe was not present, but often had anti-FasII staining at the base of the peduncle, indicating that the defect was not due to staining or loss of neurons. A separate category, “misdirected”, was used if the lobe projected in the wrong direction, usually one lobe following alongside the ipsilateral lobe containing α/β fibers.

The β-lobe midline-crossing phenotype was classified as “mild,” “moderate,” “severe,” or “normal,” as previously described (Michel et al., 2004). The γ mushroom bodies were assessed, but not scored, in CNS of 3rd instar wandering larvae and a subset of pupal samples. Statistical significance of differences in score distributions between genotypes was determined by Chi-squared testing using Sigma-Stat software

(San Jose, California).

Fig. 2: Pak proteins: evolutionary relationships, domain organization, mutations, and accumulation in Drosophila tissues

Fig. 2. Pak proteins: evolutionary relationships, domain organization, mutations, and accumulation in

Drosophila tissues. A: Phylogenetic tree of Drosophila melanogaster (Dm) Pak and human (Hs) PAK proteins based on kinase-domain amino-acid sequences (modified, with permission, from Mentzel and

Raabe, 2005). B: Diagrams of HsPAK3 and DmPak, showing mutations causing intellectual disability and alleles used in this study, respectively. PxxP, Dock/Nck binding site; CRIB, Cdc42/Rac interactive binding domain; PIX, PIX binding site; AID, autoinhibitory domain. C-G: Pak-protein accumulation revealed by immunoblotting. C-D: CNS proteins from wandering 3rd instar larvae (w3L) (7.5% acrylamide), analyzed by probing in series with anti-Pak 1:5000 (C), anti--tubulin, and anti-Pak 1:2500

(D), without stripping in between. Df, Df(3R)BSC745. Full-length Pak proteins, migrating at ~91- and

~113-kDa, correspond to the wild-type small and large isoforms. Both are detectable in Pak11/Df and

Pak6/Df samples. A smaller isoform of ~77-kDa is seen only in the presence of the Df chromosome.

Asterisk marks a Pak-related protein that may be insect-specific DmPak. Truncated Pak bands (boxes) are detected in Pak11/Df and Pak16/Df, migrating at ~30- and ~29-kDa, respectively, but not in Pak6/Df or

Pak14/Df. E: Quantification of full-length Pak-protein in larval CNS. Signal from D was normalized to - tubulin and compared with the 113- and 91-kDa bands in +/Df. F: CNS proteins from w3L (12% acrylamide). No truncated protein is detected in Pak6/Df; a very faint band migrating at ~17 kDa is barely detectable in the Pak14/Df lane. G: Whole-larval proteins, showing truncated Pak produced by Pak11 and

Pak16. Df, Df(3R)BSC744.

Fig. 3: Effects of Pak mutations on survival to adulthood, longevity, and leg morphology

Fig. 3: Effects of Pak mutations on survival to adulthood, longevity, and leg morphology. A: Quantitative summary of leg phenotype and lethality. Genotypes grouped by classes. Hemizygotes arranged with alleles producing truncated protein on top (Pak11 and Pak16) and alleles that do not produce truncated protein on the bottom (Pak6 and Pak14). Compound heterozygous combinations are arranged with the alleles encoding truncated protein on the top (Pak11/Pak16) and the non-truncation allele combination on bottom (Pak6/Pak14) and with combinations of allele types in the middle (Pak11/Pak14, Pak6/Pak16). For penetrance of leg defects, numbers in parenthesis indicate number of pharate adults scored. For percent legs affected, numbers in parenthesis indicate total legs scored. For lethality, numbers in parenthesis indicate and total progeny counted. n.d. = not determined. B: Adult longevity (post-eclosion survival) for heterozygous controls (Pak6/+ and Pak16/+) and compound heterozygous mutants (Pak6/Pak14 and

Pak11/Pak16). Numbers of animals scored daily are shown in parenthesis.

Fig. 4: Pak regulates leg morphogenesis

Fig. 4: Pak regulates leg morphogenesis. A-B: Ventral view of pharate adults. Scale bar = 100 μm. A:

Pak6/+. The legs extend relatively straight posteriorly, with the metathoracic leg terminating near the abdominal tip (arrowhead). B: Pak6/Pak14 legs appear short, curved, and splayed across the midline. C-J:

Legs of adult flies, all at same scale. Scale bar = 500 μm. C-D: Pak6/+. Wild-type morphology of prothoracic (T1) and metathoracic (T3) legs. Leg segments indicated in C. E-G: Pak-mutant T1 legs, with defects as labeled (arrowheads). E: Pak6/Pak14. F: Pak6/Df(3R)BSC744. G: Pak6/Pak14. H-J: Pak-mutant

T3 legs. H: Pak6/Df(3R)BSC744. I: Pak6/Pak14. J: Pak6/Df(3R)BSC744. K-L: Locations of Pak-mutant leg defects. Genotypes indicated at bottom of L; n = 20 animals of each. Df, Df(3R)BSC744. K:

Penetrance along the anterior-posterior axis. The T1 legs were most often affected. L: Penetrance along the proximal-distal axis. The femur was most frequently affected. Inset: Diagram of adult leg and the imaginal disc regions from which the segments derive; colors as indicated in the graph legend (adapted from Bryant, 1993).

Fig. 5: The crumpled-wing phenotype of Pak mutants

Fig. 5: The crumpled-wing phenotype of Pak mutants. A: Pak11/+ adult wing. L2-5, longitudinal veins; asterisk and PCV, anterior and posterior cross veins, respectively. Scale bar = 200 μm. B-D: Pak-mutant adult wings at same scale as A, showing reduced overall size, distortion of the distal blade margin and surface due to folds and blisters, with preservation of the proximal-most region. B: Pak11/Pak14. Abrupt transition between normal proximal and malformed distal regions (arrowheads). Compare with dotted line in A. C: Pak6/Pak14. Gross curvature of the blade caused the distal tip to fold (arrows). D: Pak11/Pak14.

White arrows indicate the margin of a large blister that collapsed. E: Pak6/Pak14. Ballooned wing, collapsed. The dorsal layer extends past the wing margin (arrows) and L3 is visible in both layers

(asterisks). F: High-magnification view of the central portion of A. Scale bar = 200 μm. G-H: Pak- mutant wings, shown at same scale as F, have cross-vein-morphogenesis defects. G: Pak6/Pak14.

Hourglass-shaped ACV, instead of narrow uniform width; missing PCV (dots). H: Pak11/Pak16. Partial

PCV (arrowheads), extending from L5 but not reaching L4; missing ACV (dots). I-K: Higher- magnification views of boxed regions of G (I,J) and H (K), Scale bar = 50 μm. L: Quantification of wing phenotypes, as indicated. Number of wings in parentheses. All Pak-mutant data combined for ACV and

PCV. ** p<0.004; * p<0.02.

Fig. 6: Persistent cells in the wing blades of Pak-mutant adults

Fig. 6: Persistent cells in the wing blades of Pak-mutant adults. DAPI staining of wings from flies 1-4 days after eclosion. A: Pak6/+. Cell bodies of mechanosensillae are present along the anterior wing margin and in the proximal-most region of longitudinal veins (arrow). There is no staining visible in the wing blade. h, nuclei of wing hinge. B: High-magnification view of boxed area in A, showing clustered nuclei of three mechanosensory bristles (arrows). C-K: Pak6/Pak14. C: Mutant wing with distal fold

(arrowheads) and scattered clusters of labeled nuclei. D-E: High-magnification view of boxed areas in C.

D: Rows of cells suggestive of longitudinal-vein pattern. E: A dense patch of nuclei; solid band is out-of- focus signal due to a ridge in the uneven contour of the blade. F: Extensive labeling throughout wing blade; t, thoracic tissue. G-H: High-magnification view of boxed areas in F. G: Large, dense patch of nuclei, some out of focus. H: Individual nuclei (arrows) are present even in regions of the wing blade with low overall signal. I: Ballooned wing mounted with anterior wing margin running along center

(between arrowheads) and the ventral and dorsal layers seen above and below, respectively. J-K: High- magnification view of boxed areas in I showing nuclei arranged in the pattern of longitudinal (L) and cross veins (CV).

Fig. 7: Mushroom body α- and β-lobe defects in Pak-mutant brains

Fig. 7: Mushroom body α- and β-lobe defects in Pak-mutant brains. A-D Projections of sequential optical sections through the mushroom body (MB) lobe regions of pharate-adult whole-mount brain preparations, labeled by anti-FasII immunostaining, at same magnification. Frontal views; dorsal is at the top, midline indicated by arrowhead. Scale bar = 25 μm. AL, antennal lobe. eb, ellipsoid body. A: Pak11/+. The alpha

(), beta (), and gamma () lobes have normal shape, size, and trajectory. B-C: Pak6/Df(3R)BSC744. B:

The right MB has a short α lobe (triangle) and thin β lobe (arrows); the left β lobe is somewhat thin. C:

On the right, the α lobe is missing (dotted outline indicates normal lobe boundary) and the  lobe is very thick, suggesting that some -lobe fibers were misdirected. On the left, minimal β-lobe fibers are in the normal location (arrows), and many are misdirected (asterisk), projecting parallel to the ipsilateral α lobe.

D: Pak6/Pak16. Severe β-lobe fusion, with a large bundle of fibers crossing the midline (arrow); star indicates median-bundle fibers. E-G: Penetrance and expressivity of the MB phenotypes in pharate-adult brains. E: Compound heterozygotes. F: Simple heterozygotes. G: Mapping with deficiency chromosomes and wild-type Pak transgene. Deletions that remove Pak uncover α- and β-lobe defects, whereas the transgene, P[Pak+], rescues them. Numbers in parenthesis indicate number of lobes (left) or brains (right) scored.

Fig. 8: Progression of Pak-mutant α- and β-lobe defects during metamorphosis

Fig. 8. Progression of Pak-mutant α- and β-lobe defects during metamorphosis. A-H: Projections of serial optical sections through MB-lobe regions of whole-mount brain preparations. Stages as indicated at far left: w3L, wandering 3rd larval instar; HE+#, head eversion+# hours. Pairs of images compare Pak11/+

(left) with Pak11/Pak16 mutants (right). Scale bar = 25 μm. A-B: 201Y-GAL4-driven lacZ to label γ MB neurons visualized with anti-β-galactosidase (green). In both control and mutant larval MBs, the γ neurons bifurcate to form the dorsal (DL) and medial lobes (ML). C-H: Immunostaining with anti-FasII

(magenta). eb, ellipsoid body. D: Bilateral β lobes (arrows) and the left  lobe are thin compared to stage- matched heterozygote. F: The left  lobe is missing entirely (outline indicates its normal location). H:

The right  lobe is short (arrowhead). I: Penetrance and expressivity of α- and β-lobe defects of

Pak11/Pak16 mutants at different stages of pupal development. The pharate-adult data are those from Fig.

7. Number of lobes in parentheses. Statistical analysis performed using Chi-squared test with 3 degrees of freedom; * p < 0.04, ** p < 0.004, **** p < 0.0005.

Abstract

PAK3 mutations cause intellectual disability (ID) with microcephaly and craniofacial defects, but little is known about the cellular neuropathology. PAK3 mediates Cdc42 and Rac GTPase signaling by regulating activity of downstream proteins, such as non-muscle myosin and cofilin. The Drosophila ortholog of

PAK3, Pak, regulates the morphogenesis of the insect learning-and-memory brain structure, the mushroom bodies (MB, Chapter 2). When cultured in vitro for three days, Pak-mutant larval CNS neurons had small neurite arbors with reduced length and branching. These neurons had reductions in F- actin in regions important for adhesion and outgrowth. Initial neurite outgrowth and branch formation appeared normal in vitro, but were followed by reductions in branch number, suggesting loss of higher order branches. Pak-mutant growth cones seem to be reduced in number with a shift in the distribution of in F-actin and MT classes. We used genetic interaction studies to infer mechanisms by which Pak regulates neurite-arbor and MB-lobe morphology. Simultaneous genetic reductions in function of Pak and flw, which encodes myosin regulatory light chain (MRLC) phosphatase, synergistically reduced neurite- arbor size in vitro. Phosphomimic MRLC (sqhE21) synergistically enhances the Pak-mutant neurite-arbor and MB phenotypes. Conversely, non-phosphorylatable MRLC (sqhA20, A21) suppressed the Pak-mutant neurite-arbor and MB phenotypes. Together, our data support a working model whereby Pak stabilizes neuronal and MB morphogenesis by inhibition of non-muscle myosin.

Introduction

Neuronal morphogenesis is the process of establishing and maintaining appropriate neurite arbor shape, size, and position necessary for function in one or more circuits. Intrinsic genetic programs that determine polarity, branching, elongation, and stability in conjunction with external cues and activity (Parrish et al.,

2007; Puram and Bonni, 2013), determine neuron morphology. Correct morphogenesis is likely crucial for contacting appropriate synaptic partners as well as contributing the electrical and computational properties of dendrites (Häusser et al., 2000). In some well-studied disorders of the nervous system such as fragile X syndrome, Down syndrome, and Rett syndrome, alterations in dendritic spine density, shape, or dendrite arbor complexity are present (Kaufmann and Moser, 2000; Irwin et al., 2001; Dierssen and

Ramakers, 2006). Defective neuron morphology may be a feature of other developmental brain disorders as well.

Selective elimination of axons and dendrites via retraction and/or degeneration are mechanisms of both normal mammalian development and insect metamorphosis to generate adult connectivity and plasticity.

Similar cell-intrinsic mechanisms can also be triggered in neurodegeneration as either response to injury or in pathological states (Luo and O’Leary, 2005). These mechanisms involve cytoskeletal proteins, motor proteins, and the GTPase Rho (Billuart et al., 2001; Ng et al., 2004). Mutations in genes that encode cytoskeletal proteins can cause developmental brain disorders (Inlow and Restifo, 2004; Moon and Wynshaw-Boris, 2013; Verpelli et al., 2013; Restifo et al., manuscript in preparation). Among these genes are regulators of the Rho-family GTPases RhoA, Rac, and Cdc42 (Ba et al., 2013; Nagase et al.,

1995; Ramakers, 2002). RhoA is a negative regulator of dendrite arbor formation while Rac and Cdc42 are positive regulators of dendrite arbor growth and complexity (Newey et al., 2005). Together, these findings highlight a role for GTPase regulation of the cytoskeleton for neurite stabilization during brain development.

One mediator of Rho-family GTPase activity in neurons is the group 1 p21-activated serine/threonine kinase PAK3. PAKs are downstream effectors of Rac1 and Cdc42 signaling (Edwards et al., 1999; Van

Aelst and Symons, 2002). The carboxy-terminal kinase domain has multiple phosphorylation substrates, including other kinases with important signaling roles and regulators of the cytoskeleton (Rane and

Minden, 2015). PAK3 is essential for brain development and function; mutations in several different families cause X-linked intellectual disability with microcephaly and distinct craniofacial features (Peippo et al., 2007; Rejeb et al., 2008; Magini et al., 2014).

PAK3 and its orthologs may play a role in in morphogenesis of diverse tissues including the nervous system. In rodent slice cultures, PAK3-siRNA knockdown showed decreased number of dendritic spines with mature morphology and increased the number of spines with filopodia-like, i.e., immature morphology (Boda et al., 2004). Double-mutant mice with loss of both PAK1 and PAK3 are born with normal brain size, but have impaired postnatal growth. Further histological investigation of these brains revealed that the cell number appeared normal across the cortex, but that the neuronal processes were greatly reduced (Huang et al., 2011). In the Drosophila nervous system, the PAK3 ortholog, Pak, is essential for sensory axon elaboration and targeting into the primary visual (Hing et al., 1999) and olfactory (Ang et al., 2003) areas of the CNS. At the neuromuscular junction, Pak regulates the clustering of postsynaptic glutamate receptors and formation of the subsynaptic reticulum (Albin and Davis, 2004).

Mutations in Pak cause defects in the insect learning-and-memory brain structures, the mushroom bodies

(MB). These defects in length, width, and direction affect the bundled axonal projections that make up the

α and β lobes and became progressively worse during development, suggesting a loss of axonal branches and length (Chapter 2). Additionally, Pak (Mentzel and Raabe, 2005), is required for morphogenesis during oogenesis (Conder et al., 2007), dorsal closure (Conder et al., 2004) and salivary gland lumen formation (Pirraglia et al., 2010), as well as for leg and wing development (Chapter 2).

Candidates for molecular mechanisms of Pak-mediated neuron morphogenesis (Fig. 1, pg. 29) include myosin light chain kinase (MLCK; Sanders et al., 1999; Goeckeler et al., 2000; Wirth et al., 2003) and

LIM kinase (LIMK; Edwards et al., 1999). Phosphorylation of these Pak substrates results in inhibition of actomyosin contraction and inhibition of cofilin-based cleavage of filamentous actin (F-actin), respectively. Neuronal myosin can stabilize the growth cone for actin polymerization-mediated neurite extension towards attractive cues (Suter and Forscher, 2000; Vallee et al., 2009), mediate F-actin retrograde flow (Diefenbach et al., 2002) or retract neurites in response to repulsive cues (Wylie and

Chantler, 2003). The repulsive cue Semaphorin-3A activates Rho and ROCK and increases myosin II activity and axon retraction (Gallo, 2006). In Drosophila mushroom bodies, mutations in p190 Rho

GTPase activating protein (GAP), a negative regulator of Rho and myosin contraction, cause α and β lobe length and width phenotypes that become progressively worse during development. These defects, which resemble those seen in Pak-mutants, are consistent with excessive retraction during metamorphosis

(Billuart et al., 2001). Expression of constitutively active MLCK caused abnormal midline crossing within the commissures by intersegmental interneurons (pCC; posterior corner cells) axon fibers of the

Drosophila embryonic ventral nerve cord (Kim et al., 2002), suggesting regulation of MRLC phosphorylation is important for response to repulsive cues.

We use a combination of Drosophila genetics, primary neuron culture, and semi-automated software to investigate the cellular and molecular mechanisms of Pak regulation of neuron morphology. We used manipulations in gene dosage and the expression of transgenic constructs to demonstrate non-additive interactions between Pak and myosin regulatory light chain (MRLC) phosphorylation. We further showed that Pak regulation of mushroom body morphogenesis also requires MRLC. Together, our findings support a model whereby Pak regulates neuron-branch stability in vitro and in vivo through inhibition of myosin activity. This provides evidence of mechanisms disrupted in a pathway with multiple genetic causes of ID.

Results

Pak regulates neurite arbor size and complexity in vitro

We utilized the primary cell culture system to identify cellular and molecular mechanisms of neuronal defects that could underlie Pak-mutant mushroom body α/β lobe defects (Chapter 2Wild-type cultured neurons from late stage, wandering 3rd instar larva have neurons with elaborate arbors after three days in vitro (div) with smooth, tapering secondary neurites (Fig. 9A, pg. 89). Antibodies against HRP recognize an antigen in Drosophila neuron membrane that creates a strong, high-contrast signal that allows for automated quantification of arbor parameters with NeuronMetricsTM (Narro et al., 2007). Genetic mutations can cause obvious in vitro phenotypes. For example, fascin-mutant neurons in cell culture have dramatic increase in neurite curvature and irregular caliber (Kraft et al., 2006).

Neurons from Pak-mutant larval CNS cultured for 3 div had striking in vitro phenotypes including small, naked arbors with few higher-order branches. Additionally, neurite-caliber irregularities were present, such as ‘hooks’ appearing in regions of tight curvature (Fig. 9B,C, D). Quantitatively, compared to wild- type neurons and heterozygous controls, Pak-mutant neurons have small neurite arbor size with reduced branch number, total neurite length and neuron territory area (Fig. 9F). These results were verified in multiple independent replicates (data not shown). There were no changes in branch density

(branches/length) or length/area ratios (not shown). This phenotype is uncovered by a deficiency chromosome that deletes the Pak locus. Branch number, higher order branch number, and length were rescued by the addition of a single transgenic copy of the wild-type Pak gene o the same deficiency chromosome (Fig. 9E, G). The parameter of neuron territory area was not rescued by the addition of the

Pak+ transgene. The whole-CNS cultures are heterogeneous, which suggests that the Pak phenotype of reduced neurite arbor size isn’t limited to a specific cell type.

Truncated Pak protein retains partial function for regulating neurite length, but not branch number

The domains of Pak responsible for its activation and kinase activity are shown in Fig. 10A (pg. 91). The amino-terminal PxxP domain interacts with Dock/Nck for recruitment to the membrane. The PIX domain can recruit guanine exchange factor (GEF) proteins which can activate Rac and Cdc42 (Manser et al.,

1998). Binding of GTP-bound Rac/Cdc42 to the CRIB domain of Pak disrupts the blockage of the active site by the autoinhibitory domain (AID). The carboxy-terminal kinase domain phosphorylates enzymatic substrates, including MLCK and LIMK (Pirruccello et al., 2006).

We used four nonsense alleles to determine if kinase activity is required for Pak regulation of neuron arbor size (Fig. 10A) The alleles Pak6 and Pak14 do not create truncated protein product, but Pak11 and

Pak16 do. The truncated proteins, which would include the PxxP, PIX, CRIB, and AID domains, retain partial function for survival during development, but had negative effects on MB-lobe stability (Chapter

2). All compound heterozygous combinations cause a significant decrease in higher-order branch number

(Fig. 10B). Neurons from compound heterozygotes with no truncated protein also significant reductions in total neurite length and territory area. However, neurons from compound heterozygotes producing truncated protein were no different from controls for total neurite length or territory area (Fig. 10C).

These results suggest that the Pak kinase domain is required for higher-order branches, but not absolutely necessary for neurite length and territory area.

Filopodial, and focal adhesion F-actin distribution is downregulated in Pak-mutant neurons

Because Pak is predicted to stabilize the actin cytoskeleton, we examined the distribution of filamentous actin (F-actin) in Pak-mutant neurons. To visualize F-actin, we double-labeled cultured neurons with fluorescently-labeled phalloidin and anti-HRP. Phalloidin binds selectively to the plus-end of F-actin in a

1:1 stoichiometric ratio which provides a direct correlation between signal intensity and number of F- actin filaments in an area (De La Cruz and Pollard, 1994). In Pak6/+ control neurons at three div, phalloidin shows localized signal, indicated localization of F-actin, in small, distal processes that likely

73 represent growth cones with filopodia (Fig. 11B, D, pg. 93). F-actin also localizes to puncta in regions such as the soma (Fig. 11E), which probably represent focal adhesions with the substrate (Jester et al.,

1994). In Pak-mutant neurons, there was overall decreased intensity of phalloidin signal (Fig. 11G), indicating changes in distribution or levels of F-actin. Furthermore, there was an apparent decrease in filopodia, and decreases in concentrated F-actin at the distal tips of neurites (Fig. 11I). In addition, there are few puncta present; those present have low signal (Fig. 11J). This demonstrates disruptions in the overall organization of the F-actin cytoskeleton as well as locally in filopodia and focal adhesions, structures that are important for outgrowth and adhesion.

Branch stability and growth cone cytoskeleton organization defects in Pak-mutant neurons

In order to characterize the developmental trajectory of Pak-mutant neurons, we conducted a time-series experiment, examining cultured neurons from the same CNS at 6, 12, and 24 hours post-plating (Fig.

12A, pg. 95). Pak-mutant neurons had normal initial outgrowth and continued to arborize as much as heterozygous control neurons at 6 and 12 hours. However, between 12 and 24 hours, the Pak-mutant arbor size declined substantially, suggesting a loss of higher-order branches.

Because we predict that Pak regulates neuron morphology through the cytoskeleton, we examined the growth cones from Pak11/+ controls and Pak11/Pak6 compound heterozygous neurons at six hours post- plating. We double-labeled with fluorescent phalloidin to visualize F-actin and anti-alpha-tubulin immunostaining to visualize microtubules (MTs) (Fig. 12B-D). We used a classification scheme developed by Sánchez-Soriano and colleagues (2010) for types of cytoskeletal structures in cultured neurons. MT classes include bundled (a solid, unbroken line of tubulin; Fig. 12B), spread (MT bundles radiating into the growth cone; Fig. 12C), or looped (MT bundles form a circle in the distal-most process;

Fig. 12D). F-actin classes include lamellar (a spread-out, dense meshwork of F-actin in the end of the process; Fig. 12C), pointed (filopodial-like protrusions of F-actin from the MT core without lamellipodia;

Fig. 12B), or no extension (no F-actin protrusions past the MT core; Fig. 12D).

There appears to be a ~25% reduction in the number of growth cones per neuron in Pak-mutant neurons

(1.75 growth cones/neuron for Pak11/+ and 1.32 growth cones/neuron for Pak11/Pak6). There was a significant difference in the distribution in F-actin (p < 10-4) and MT (p = 0.01) classes between genotypes (Fig. 12E). In Pak-mutant growth cones, there was a decrease in the number of lamellar structures with more growth cones having no F-actin protrusions compared to controls. Additionally, there was a decrease in neurites with spread microtubules, a pattern thought to be important for axon elongation (Tanaka et al., 1995) and invading and stabilizing new branches (Dent et al., 1999). The changes in F-actin localization in Pak-mutant growth cones and cultured neurite arbors are consistent with a role of Pak in regulating the actin cytoskeleton.

The Pak-mutant neurite-arbor phenotype interacts with genes regulating myosin regulatory light chain phosphorylation

In order to elucidate targets of Pak signaling, we measured the effect of genetic interactions on neurite arbor size. We used reductions of gene dosage of Pak in combination with either reductions in gene dosage or expression of mutant transgenes of putative downstream targets to determine if they are in the same pathway. First, we established that the Pak phenotype of reduced neurite arbor size is recessive with no difference between neurons from heterozygotes and neurons from crosses of laboratory control strains

(results not shown).

Group 1 PAKs can phosphorylate MLCK and thereby inhibit its activity in cellular and cell-free protein assays (Goeckeler et al., 2000; Sanders et al., 1999). Although this function has been demonstrated in

HEK and HeLa cells, it has not previously been reported for neurons. Myosin ATPase and motor activities are greatly increased by phosphorylation of MRLC on residues 20 and 21 (Jordan and Karess,

1997) by MLCK (reviewed in Tan et al., 1992) or Rho kinase (Kosako et al., 2000; Ueda et al., 2002).

Phosphorylation is reversed by protein phosphatase 1 (PP1) (Vereshchagina et al., 2004). We hypothesized that Pak, the Rac/Cdc42 mediator, counteracts myosin activity (Fig. 13A, pg. 97), which is

75 responsible for neurite-retraction events such as axon retraction in response to repulsive cues such as semaphorin (Gallo, 2006).

The protein phosphatase responsible for removing phosphorylation from MRLC is produced by the flapwing gene in Drosophila. Genetic reductions in flw should increase the amount of MRLC phosphorylation. Neurons heterozygous for a mutation in flapwing (flwG0172/+) have normal neurite-arbor size, as do Pak6/+ heterozygotes (Fig. 13B). However, cultured neurons that are doubly heterozygous for

Pak and flapwing loss-of-function mutations have reduced branch number and total neurite length.

Pak6/Pak14 compound heterozygous neurons consistently have reduced branch number and total neurite length compared to Pak6/+ heterozygous controls. This synergistic interaction between Pak and flapwing in regulating neurite arbor size indicates that Pak and flapwing could function in the same pathway to control neurite arborization.

The phosphorylation target of Pak and flw could be MRLC (Vereshchagina et al., 2004). To test if MRLC regulates neurite arbor size, we examined constructs with mutations of the phosphorylation sites of the

Drosophila ortholog of MRLC, spaghetti squash (sqh) (Karess et al., 1991). Phosphomimic MRLC

(sqhE21) has the critical phosphorylation site mutated from serine to glutamate (E21) (Jordan and Karess,

1997). Non-phosphorylatable MRLC (sqhA20, A21) has both phosphorylation sites (serine-21 and threonine-

20) mutated to alanine (A20, A21). Both of these transgenes use the endogenous sqh promoter to produce mutant MRLC proteins at levels comparable to a single copy of wild type (Karess et al., 1991).

Since our model predicts that loss of Pak increases MRLC phosphorylation (Fig. 13A), we examined genetic reductions in Pak in combination with an increased ratio of phosphorylation MRLC for enhancement of the Pak-mutant phenotype of reduced neurite arbor size. We expressed one transgenic copy of P[sqhE21] in a background heterozygous for a null mutation of MRLC (sqhAX3/+; P[sqhE21]/+).

This substitution of ~50% of the MRLC for a phosphomimic form, and thereby, constitutively active form of myosin, does not decrease neurite arbor size (Fig. 13C). However, substituting ~50% of the total

MRLC with the phosphomimic form in Pak6/+ heterozygotes (sqhAX3/+; P[sqhE21]/+) reduces the parameters of both branch number and total neurite length. Since Pak heterozygotes and MRLC phosphomimic substitution alone does not reduce neurite-arbor size, this reveals a synergistic interaction.

These results were verified in two independent replicates (data from 2nd experiment not shown).This non- additive effect of combining MRLC phosphorylation and genetic reductions in Pak support the model where Pak inhibits MRLC phosphorylation.

Non-phosphorylatable MRLC (A20, A21) will compete with endogenous MRLC and have a dominant- negative effect on myosin-based contraction activity. Therefore, our model (Fig. 13A) predicts that the addition of non-phosphorylatable MRLC will suppress the reductions in neurite arbor size of Pak compound heterozygous neurons (Pak6/Pak14). Expression of non-phosphorylatable MRLC P[sqhA20,A21] in a sqh+ background somewhat reduced the parameters of branch number and total neurite length (sqh+;

P[sqhA20,A21]/+; Pak6/+) compared to Pak6/+ heterozygotes (Fig. 13D). However, expressing non- phosphorylatable MRLC in Pak-mutant neurons improved the parameters of both branch number and total neurite length compared to Pak6/Pak14 compound heterozygous neurons. These results were verified in two independent replicates (data from 2nd experiment not shown). Together, these results are consistent with the hypothesis that Pak regulates branch number and total neurite length via inhibition of MRLC phosphorylation.

Pak and MRLC phosphorylation interact to regulate MB morphology in vivo

Pak-mutant brains have α- and β-lobe defects including thin (Fig. 14B, pg. 99), short, missing (Fig. 14D), and misdirected lobes (not shown) (Chapter 2). Midline fusion of the β-lobes is present in some of the mutant brains (Fig. 14C). These defects are progressive, with increased penetrance and occurrence of

“missing”, the most severe category, increasing through metamorphosis (Chapter 2). We examined the α and β lobes of the adult MBs, paired neuropil structures composed of α/β neurons which extend a primary neurite that bifurcates and projects medially and ventrally into the two lobes. These axon fibers are tightly

77 bundled and visualized by anti-FasII immunostaining in both the lobes and proximal-most portion of the peduncle. Weaker anti-FasII labeling is present on the unbranched, medially projecting γ neurons

(Crittenden et al., 1998), antennal lobe (AL), and ellipsoid body (eb). Simple heterozygotes of Pak- mutant alleles do not have α- or β-lobe extension defects (Fig. 14E; Chapter 2). The α- and β-lobe phenotypes are highly penetrant in Pak6/Pak14 compound heterozygous brains (90% of brains, 42% of lobes, Fig. 14F), with missing and thin/short lobes as the most common defects.

We hypothesize that myosin overactivity might also be responsible for the in vivo Pak-mutant phenotype of MB lobe defects. To test this hypothesis, we used genetic manipulations to reduce Pak with expression of constructs with mutations at the MRLC phosphorylation sites to mimic phosphorylation (E21) or prevent phosphorylation (A20, A21). Substituting ~50% of endogenous MRLC for phosphomimic MRLC

(sqhAX3/+; sqhE21/+) caused α and β MB lobe defects with low penetrance (25% of brains, 9% of lobes).

As predicted by our hypothetical model (Fig. 13A, pg. 97), combining phosphomimic MRLC and reductions in Pak dosage (sqhAX3/+; P[sqhE21]/+; Pak6/+) increases the penetrance of α and β MB lobe defects (70% brains, 25% lobes; Fig. 14E; p = 0.04).

Since our model predicts that decreasing Pak subsequently increases MRLC phosphorylation, we also tested if adding non-phosphorylatable MRLC to Pak-mutant brains could improve α- and β-lobe defects

(Fig. 14F). The addition of non-phosphorylatable MRLC to Pak6/+heterozygotes (sqh+; P[sqhA20,A21]/+;

Pak6/+) causes MB lobe defects with low penetrance (25% brains, 6% lobes). However, adding non- phosphorylatable MRLC to Pak-mutant compound heterozygotes (sqh+; P[sqhA20,A21]/+; Pak6/Pak14) decreases the penetrance (61% of brains, 26% of lobes) and reduces the number of missing lobes compared to Pak6/Pak14 compound heterozygotes. Therefore, competition for assembly into myosin hexamers with non-phosphorylatable MRLC significantly suppresses the phenotype of MB α and β lobe defects (p = 0.02). There does not appear to be a change to the β-lobe fusion phenotype. These results show that Pak and MRLC phosphorylation interact to regulate morphogenesis of α and β lobe length and width, which supports our hypothetical molecular model.

Discussion

In this study, we show that Pak-mutant neurite arbors have reduced higher-order branches and total neurite length. Therefore, Pak regulates neurite arbor size and complexity. This finding parallels findings where boys with mutations in PAK3 have microcephaly (Peippo et al., 2007), PAK1/PAK3 double-mutant mice which have reduced brain volume (Huang et al., 2011), and Drosophila Pak mutants which have reduced terminal arborization of photoreceptors in the optic lobe (Hing et al., 1999). These defects appear after normal initial outgrowth in vitro, suggesting a loss of higher-order branches.

We demonstrated that mutations in Pak cause defects in the cytoskeleton, with reduced localization of F- actin to filopodia and focal-adhesion-like puncta in neurons cultured for 3 div. Pak mutations also cause reductions in the number of growth cones, with the fewer of the growth cones present exhibiting F-actin arranged in a lamellar-like organization. These results suggest that Pak contributes to the cytoskeletal organization of neurons, but need to be replicated in other Pak genotypes to confirm findings about growth cone number and actin organization. More detailed counts of filopodia, with information on rates of extension vs. retraction are also needed.

In contrast to the Pak-mutant phenotype of decreased phalloidin signal, neurons with mutations in singed, which encodes the actin-bundling protein fascin, have increased amounts of localized F-actin (Kraft et al.,

2006). Both Pak and singed mutations cause caliber defects with increased neurite curvature, reminiscent to the fascin ‘filagree’ phenotype. In both cases, F-actin could be failing to organize into the bundles important for normal neurite trajectory and caliber; however, in the case of Pak mutations, the F-actin is likely getting cleaved and in singed mutations, actin filaments are intact and accumulating elsewhere.

Overall, the organization of the cytoskeleton is important for neurite stabilization and maturation (Prokop et al., 2013; Mitchison and Kirschner, 1988), and problems in F-actin organization could be contributing to Pak-mutant morphogenesis defects.

We used manipulations to increase MRLC phosphorylation via genetic reductions in the phosphatase, flapwing, or substitution with a phosphomimic form of MRLC (E21). When in combination with genetic reductions in Pak, we saw non-additive decreases in neurite arbor size. We were able to suppress the severity of the Pak-neurite arbor size phenotype by addition of non-phosphorylatable MRLC (A20, A21), again demonstrating interactions between Pak and genes which regulate MRLC phosphorylation. These data support the hypothetical model where Pak inhibits MRLC phosphorylation to regulate neurite arbor size in vitro. In the pharate adult MBs, we saw non-additive increases in the penetrance of MB α and β lobe defects with the combination of phosphomimic MRLC substitution with genetic reductions in Pak.

The high penetrance of missing and short/thin lobe phenotypes of Pak-mutant compound heterozygotes is suppressed by addition of non-phosphorylatable MRLC. Therefore, Pak interacts with genes regulating

MRLC phosphorylation to regulate MB lobe length and width. These thin/short and missing lobes suggest defects in α/β fiber length and branching, similar to what is seen in cultured neurons. Therefore, these studies provide evidence that the molecular mechanisms of Pak regulation of the neurite arbor are recapitulated in an in vivo model of brain development.

We focused on candidate downstream proteins regulated by Pak phosphorylation activity, but the partial function of some Pak alleles in these preliminary studies suggests additional, non-kinase dependent roles of Pak. Reducing levels of the zebrafish ortholog of PAK3 cause jaw malformations and loss of axons connecting the brain hemispheres, (Magini et al., 2014). The expression of kinase-dead PAK3 protein in zebrafish enhanced these defects, and this enhancement could be blocked by an MEK1 inhibitor. This suggests that kinase-dead PAK3 protein has additional negative effects by increased activity of the

MAPK pathway. We found that truncated Pak11 and Pak16 proteins had additional negative effects that increase the penetrance and severity of MB lobe defects (Chapter 2), further demonstrating that Pak likely has kinase-independent functions with possible negative effects. This could be mediated by the AID inhibiting other Pak proteins. We also demonstrated that truncated Pak can retain partial function for phenotypes including survival and leg development (Chapter 2), as well as total neurite length and

80 territory area. This function could possibly be due to the ability to bind to both Rac/Cdc42 and a GEF activator of these GTPases, PIX. However, non-kinase dependent functions of Pak have not been reported and deserve further study.

The progressive loss of higher-order branches in vitro, which appears to be mediated at least in part by acto-myosin contractility, is consistent with excessive retraction. The length and width defects of the MBs include short/thin lobes, which appear to be similar, but less severe, compared to missing lobes. The penetrance and severity of defects increases during development (Chapter 2), and Pak interacts with genes that phosphorylation of MRLC, which is crucial for acto-myosin contraction. Additionally, the ends of some lobes have a ‘flamed’ appearance (Fig. 14D, pg. 99), suggesting the individual axons of the lobe are of different lengths. This was not apparent in control MB during development, suggesting this is consistent with variations between α/β fibers. These data support the model where Pak-mutant MB phenotypes are due to excessive retraction. Based on this, we hypothesize that Pak stabilizes the neurite arbor during morphogenesis.

The mechanisms responsible for overgrowth/fusion seen in β-lobe fusion and misdirected lobes may be separate from the mechanisms underlying the short/thin/missing MB lobe defects. The phenotype of β- lobe fusion was not suppressed by addition of non-phosphorylatable MRLC. A candidate mechanism is

Pak inhibition of cofilin (Fig. 1, pg. 29). MARCM clones of MB neurons with loss-of-function mutations in cofilin (twinstar) have short axon projections that don’t reach the end of the lobes, suggesting reduced neurite outgrowth, although the timing was not described in sufficient detail to eliminate the possibility of retraction (Ng and Luo, 2004). Overexpression of cofilin increases neurite extension of cells in culture

(Kuhn et al., 2000; Endo et al., 2003). Therefore, increasing cofilin activity may cause neurite overgrowth. The phenotype of misdirected lobes may be worsened by expressing non-phosphorylatable

MRLC. This suggests that interactions with MRLC phosphorylation may not be responsible for Pak- mutant pathfinding defects, such as misdirected MB lobes (Chapter 2), ectopic projections of olfactory receptor neurons (Ang et al., 2003), and disorganization of photoreceptors projecting into the optic lobe,

81 with regions of increased or decreased innervation (Hing et al., 1999). The downstream mechanism of

Pak for regulation of axon pathfinding deserves further study.

Although the small neurite arbor phenotype has been mapped to Pak using multiple alleles and two different deficiency chromosomes (data not shown), the parameter of neuron territory area was not rescued by the addition of the Pak+ transgene. This suggests that something else on the Pak+,

Df(3R)Win11, Ki chromosome could be contributing to the reduced territory area. One possible candidate is Kinked, a semi-dominant bristle marker since some bristle markers, such as singed and forked, encode proteins which bundle actin fibers. As singed mutations cause decreases in territory area of cultured neurons due to increased curvature of the neurites, it is possible that the Kinked mutation on that chromosome is contributing to reduced territory area in these neurons.

These findings may provide a cellular and molecular explanation for the features of reduced brain size growth in the post-natal period in a mouse model of this gene (Huang et al., 2011) and the clinical feature of microcephaly in affected boys (Peippo et al., 2007). This paves the way for future studies to determine the cellular signaling that regulates Pak-mediated neurite stability. Reductions in neurite arbor size resemble reduced neurite branching seen in post-mortem tissue from other genetic ID disorders, such as

Down syndrome, Fragile X syndrome, and Rett syndrome (Kulkarni and Firestein, 2012). The obvious

Pak-mutant primary neuron culture phenotype allows for screening for compounds that rescue the neurite arbor size; these compounds could be of therapeutic benefit to individuals with mutations in PAK3 or other ID genes of the same pathway (Fig. 1, pg. 29).

Materials and Methods

Drosophila culture and genetics

Fly stocks were maintained at room temperature (22-23°C) and experimental cultures were reared at

25°C, 60-80% relative humidity, as previously described (Kraft et al., 2006), both on corn flour- nutritional yeast-agar medium (Elgin and Miller, 1978). OreR-C (a wild-type laboratory strain) and w(z)

(a “Cantonized” strain of w1118) were used for wild-type chromosomes. Crosses, e.g., to create specific compound heterozygotes or hemizygotes, were performed using methods that optimize larval culture density to minimize competition and optimize mutant development and survival.

All Pak-related stocks were obtained from the Bloomington Drosophila Stock Center at Indiana

University (funded by NIH P40 OD018537), including mutant alleles (Pak6, Pak11, P{neoFRT, ry+}82B

Pak14, P{neoFRT, ry+}82B Pak16), the wild-type Pak transgene, and chromosomal deletion Df(3R)Win11

Ki pp (Pattatucci et al., 1991). Wild-type Pak function was provided by a P-element construct

(“T2.2”)with a Pak+ cDNA, inserted on a Pak-deletion chromosome: T2.2 Df(3R)Win11 Ki pp (Hing et al., 1999), abbreviated P[Pak+] Df(3R)Win11. All Pak stocks were maintained over a balancer, either

TM6B, Tb e or TM3, Ser P[Act-GFP]. The sqh and flw stocks were also obtained from Bloomington including mutant alleles: sqhAX3, P[sqhA20,A21], P[sqhE21] (Jordan and Karess, 1997), and flwG0172 (Peter et al., 2002). sqhAX3 and flw chromosomes maintained over FM7i-GFP; P[sqhA20,A21] and P[sqhE21] maintained as homozygotes.

Preparation of primary neuronal cell cultures

Dissociated neuronal cultures were prepared from the CNS of 3rd instar larvae as previously described

(Kraft et al., 1998, 2006, 2013). Animals were selected as late-stage wandering 3rd instar larvae from optimum density cultures and matched for sex; females were used for every neuronal culture except growth cone and time-series experiments. Explanted CNS were treated with Liberase Blendzyme 1

(purified collagenase I, collagenase II, and neutral protease dispase, Roche, Indianapolis, IN) at a concentration of 40 μg/mL (0.21 Wünsch units) in Rinaldini’s saline (140 mM NaCl, 2.7 mM KCl, 0.35 mM NaH2PO4, 12 mM NaHCO3, 55 µM D-glucose) before issue dissociation by mechanical trituration by pipetting in culture media. Culture media composed of Schneider’s Insect medium (Invitrogen, San

Diego, CA) with 10% fetal bovine serum of United States origin (Hyclone Lot APC20860 or

AZA180873, Logan, UT or Optima lot #D0118, Atlanta Biologicals, Flowery Branch, GA) and 50 μg/mL bovine insulin (Sigma, St. Louis, MO). Cells plated at the density of 1/5-1/6 of a CNS on photoetched

(“gridded”) glass cover slip (Bellco, Vineland, NJ) with an 8-mm-diameter well coated with

Concanavalin A (Sigma) and laminin (BD Biosciences, Franklin Lakes, NJ). Cells were then incubated at

25°C, usually for three days in vitro (div) (~74 hours post-plating) before fixation, except for the time series experiment, when dishes were fixed at six, twelve, or 24 hours post-plating.

Immunostaining and microscopy of primary neuronal cultures

Fixation and immunohistochemical methods performed as described previously (Kraft et al., 1998, 2006).

Cells were washed twice with ice-cold, calcium-free Ikeda Ringer’s saline (130 mM NaCl, 4.7 mM KCl,

1.8 mM MgCl2, 0.74 mM KH2PO4, 0.35 mM NaHPO4) and fixed using 4% EM-grade formaldehyde (Ted

Pella, Redding, CA) in PBS for 20-30 minutes at 4˚C. Cells were washed three times at 4˚C in PTN (0.1

M sodium phosphate buffer, pH 7.4, 0.1% Triton (v/v), 0.1% sodium azide (w/v)). For staining cultured neurons, neuronal membranes were labeled using a polyclonal goat anti-horseradish peroxidase (HRP) antiserum (Cappel, 55973) which recognizes the Drosophila nervana 2 protein (Sun and Salvaterra,

1995). Goat anti-HRP was diluted to 1:500 in PTN and incubated between 6-16 hours at 4°C with shaking. Cells were washed with PTN six times for 15 minutes each at room temperature. Neuronal membranes were visualized with Donkey anti-Goat Alexa 488 IgG (Molecular Probes, Life Technologies,

A-11055) or Donkey anti-Goat IgG Alexa 568 (Molecular Probes, A-11057) diluted 1:500 in PTN and incubated at room temperature for three hours before performing six washes with PTN for 15 minutes each at room temperature. Cells received a final wash in 0.1 M Tris-HCl buffer, pH 6.8 and were then

84 coverslipped with 12% polyvinyl alcohol (PVA; Sigma, St. Louis, MO) with 1.5% DABCO (1,4- diazabicyclo [2.2.2] octane; Sigma). For colabeling of neurons cultured 3 div with anti-HRP and phalloidin, Acti-stainTM 488 phalloidin (Cytoskeleton Inc., Denver, CO), diluted to 100 nM in PBS pH

7.4, was incubated for 30 minutes at room temperature after the final wash. Cells were then washed briefly with PBS before mounting with <1 week-old PVA/DABCO.

For the growth-cone analysis, cells at six hours post-plating were fixed in 0.5% glutaraldehyde (Ted

Pella) in PIPES, pH 6.8 (100 mM PIPES, 5 mM EGTA, 2 mM MgCl, 0.02% sodium azide (w/v)) for 30 minutes at room temperature before washing in PTN three times for 15 minutes. Cells were incubated with anti-alpha tubulin (12G10, Developmental Studies Hybridoma Bank, deposited by Joseph Frankel and E.M. Nelson), diluted 1:500 in PTN, for four hours at room temperature. After washing five times for

15 minutes each in PTN at room temperature, neurons were incubated in Oregon Green ® 488 phalloidin

(100 nM, 0.5 units/dish, Molecular Probes) and Donkey anti-Mouse Cy3 (1:500, Jackson

ImmunoResearch) in PTN for 14 hours at 4˚C without rocking. Neurons were washed four times for five minutes with PBS and coverslipped as above. High-contrast overlays of the phalloidin and anti-tubulin images were created using Photoshop merge channels and curves functions. These images were processed using NeuronMetricsTM to quantify the neurite arbor parameters for the six-hour time point instead of anti-HRP staining. Categories for scoring growth cone cytoskeletal features adapted from Sanchez-

Soriano et al., 2010. Only microtubule-containing processes were scored to ensure we were examining growth cones with neurite potential. F-actin classes: lamellar, (spread out sections of F-actin at the end of the neurite/microtubule core), pointed (linear extensions of F-actin present at end of neurite), or no extensions (no protrusions of F-actin were present at the end of the neurite). Microtubule (MT) organization classes: bundled (microtubules form a solid line with no gaps in between them, blunt or pointed MTs at the neurite tip), spread (growth cones contain dispersed arrangements of MTs, often curved), looped (bundled MT at the axon tip display a curved or looped ‘constellation’). Differences between genotypes determined using Chi-squared analysis using 2 degrees of freedom.

Neurons were examined on a Diaphot 300 inverted microscope (Nikon, Tokyo, Japan) with a 60X oil- immersion objective (numerical aperture, 1.4); neurons at six hours post-plating were imaged using a

100X oil-immersion objective (N.A., 1.25). Epifluorescence illumination achieved with filter cube

Chroma #41001 (exciter, 460-500 nm; dichroic, 505 nm; bandpass emitter, 510-560 nm) for Alexa Fluor

488 and with a Nikon G-2A filter cube (exciter, 510-560 nm; dichroic, 580 nm; long-pass emitter, 590 nm) for Alexa 568. Images were collected with Hamamatsu (Hamamatsu Photonic Systems, Bridgewater,

NJ) ORCA285 camera and image-acquisition software. Images were converted from16-bit tiff to 8-bit images for processing and were stitched if needed using PanaVue software (Québec, Canada). Images of phalloidin-labeled neurons cultured for 3 div were captured with all imaging parameters held constant including exposure time, sequence of image capture to minimize fading, and gain.

In most experiments, dishes were coded by a lab member not involved with the experiment to allow for

“blind” imaging and analysis. For each experiment, a minimum of two independent replicates were used to verify results. For most experiments, 100+ neurons per dish were selected using a staircase search pattern on the gridded dish. For deficiency-mapping experiment, ~100 neurons from each of two independent replicates were sampled using a horizontal-switchback pattern on non-gridded dishes, with care taken to avoid repeat sampling of the same region. For the transgenic rescue experiment, ~70 neurons were imaged from 3 dishes from the same sample for a combined total of ~210 neurons.

Quantitative morphometric analysis using NeuronMetricsTM and curvature analysis

See Narro et al., 2007 for detailed explanations of the development of and algorithms used by

NeuronMetricsTM. NeuronMetricsTM converts each neurite arbor into a one-pixel-wide skeleton to compute parameters of primary neurites (neurites coming from the cell body), estimate branch number, total neurite length, and neuron territory area; these are used to calculate higher-order branches (branches

− primary neurites) and the branch density (branches/1000 μm2 territory area). Sigma Stat (Systat

Software, San Jose, CA) or R (R Foundation, Vienna, Austria) used for statistical tests, including Mann-

Whitney non-parametric rank sum test. Graphs created using R with modifications to the boxplot defaults to place whiskers at the 90th and 10th percentiles.

Immunostaining and confocal imaging of mushroom body structure

MB labeling of female, late-stage pharate adults P14-15 was performed as described in Chapter 2 and

Michel et al., 2004. CNS tissue was fixed in EM-grade 4% formaldehyde (Ted Pella, Redding, CA), washed three times in PTN (0.1 M sodium phosphate buffer, pH 7.4, 0.1% v/v Triton-X, 0.1% w/v sodium azide) for 15 minutes each. The samples were incubated in primary antibody (anti-FasII, ID4,

Developmental Studies Hybridoma bank, deposited by Corey Goodman), diluted 1:25 in PTN, overnight at 4°C, then washed six times for 20 minutes each in PTN at room temperature. Brains were incubated in secondary antibody (Donkey anti-Mouse conjugated with Cy3, Jackson ImmunoResearch, 715-165-150), diluted 1:500 in PTN, for three hours at room temperature and washed six times for 20 minutes each at room temperature, followed by a final wash in 0.1 M Tris-HCl, pH 8.0. Samples were mounted in PVA with DABCO to minimize fading of fluorescent signals. The mushroom bodies were visualized by laser- scanning confocal microscopy using a Zeiss 510 microscope with a 40X (1.3 N.A.) oil-immersion objective and LSM 4.2 SP1 image-acquisition software (Zeiss, Jena, Germany). Samples were scanned with a helium-neon laser line with an excitation maximum of 543 nm and a long-pass filter at 565 nm for

Cy3 (z = 1.0 μm). Image projections were created in LSM Image Browser using a selected subset of sequential optical sections for optimal visualization of the α and β lobes.

Scoring and classification of adult mushroom body phenotypes

Brain morphology was assessed by examining stacks and individual optic sections of anti-FasII-labeled brains with special attention to the MBs and ellipsoid body regions. Each α- and β-lobe was scored separately (four lobes per brain). The α- and β-lobe defects were classified as “normal”, “thin/short” or

“missing” based on the thickness and length of the lobes, if present. A thin lobe had an unusually thin diameter which was often variable along the length of the lobes. Short lobes do not project the normal

87 distance and appear truncated. Since a lobe could be both short and thin and both of these defects seemed to resemble an intermediate phenotype, the categories were combined. A missing lobe was not present, but often had anti-FasII staining at the base of the peduncle, indicating that the defect was not due to staining or loss of neurons. A separate category of “misdirected” was used if the lobe projected in the wrong direction, usually one lobe following alongside its complementary branch.

The β-lobe midline-crossing phenotype was classified as “mild,” “moderate,” “severe,” or “normal,”

(Michel et al., 2004). Statistical testing for differences in distribution of α- and β-lobe defects

(misdirected, missing, thin/short, or normal) between genotypes determined by Chi-squared analysis with

3 degrees of freedom using Sigma-Stat software (San Jose, California).

Fig. 9: Pak regulates neurite arbor size and complexity

Figure 9: Pak regulates neurite arbor size and complexity. A: A typical Pak11/+ control neuron after three div with many higher-order branches and tapering neurites. B-D: Pak-mutant neurons have decreased neurite-arbor size with irregular caliber. B: Pak11/Pak6, Compound heterozygous Pak-mutant neurite arbor has fewer higher-order branches with overall reduction in size. Arrow indicates a ‘hook’, a characteristic feature of Pak neurons caused by irregular caliber and regions of high curvature. C-D:

Pak11/Df(3R)Win11 Ki. A deficiency chromosome uncovers the in vitro phenotypes. E: Pak11/Pak+,

Df(3R)Win11 Ki. Neurons with a transgenic copy of wild-type Pak have restored neurite arbor size and complexity. Scale bar A-E, = 20 μm. F: The phenotypes of decreased branch number, total neurite length, and neuron territory area seen in compound heterozygous neurons are also uncovered by a deficiency chromosome that deletes the Pak locus. Pak11/+ (light grey); Pak11/Pak6 (blue); Pak11/Df(3R) Win11 Ki

(dark blue). Box and whisker plots display the 90th and 10th percentiles as the ends of the whiskers and

75th and 25th percentiles define the top and bottom of the box. The middle line denotes the 50th percentile

(median). Lowercase letters indicate different statistical groups determined using Mann-Whitney rank sum test. a-b P < 10-16. Number of neurons in parentheses. G, Pak11/+, (light grey); Pak11/Df(3R) Win11

Ki, (dark blue); Pak11/Pak+, Df(3R) Win11 Ki, (light blue). Addition of a wild-type Pak transgene rescues branch number and total neurite length, but not territory area. a-b, p < 10-5, a-b p < 0.02, a-c, n.s.

Fig. 10: Truncated Pak protein retains partial function for neurite length and territory area, but not higher- order branch number

Figure 10: Truncated Pak protein retains partial function for neurite length and territory area, but not higher-order branch number. A: Drosophila Pak protein domains and alleles used in these studies.

Pak6 and Pak14 are nonsense mutations near the region encoding the N’-terminus. Pak11 and Pak16 are nonsense mutations following the PIX-coding domain which produce truncated protein which accumulates in larval CNS (Chapter 2). Pak11 and Pak6 alleles also produce a small amount of full-length protein. PxxP = Dock/Nck binding site; CRIB = Cdc42/Rac interactive binding domain; PIX = alpha PIX

GEF (guanine exchange factor) binding site; AID = autoinhibitory domain. Scale bar = 50 amino acids.

B-D: Neurite-arbor size parameters of Pak11/+ (light grey), Pak11/Pak16 (light purple), Pak11/Pak14

(medium purple) and Pak6/Pak14 (dark purple) neurons cultured 3 div. B: All Pak-mutant genotypes showed a significant decrease in higher-order branch number compared to heterozygous controls.

Lowercase letters indicate different statistical groups as determined by Mann-Whitney rank sum test. a-b, p < 0.01. Number of neurons in parentheses. C: Pak6/Pak14 neurons also have significantly reduced neurite length while Pak11/Pak16 neurons do not have reduced length. Pak11/Pak14 neurons have an intermediate phenotype. a-b, p < 0.009; a-b′, p < 10-16. D: Pak6/Pak14 neurons also have significantly reduced territory area while Pak11/Pak16 neurons do not have reduced territory area. Pak11/Pak14 neurons have an intermediate phenotype. a-b, p < 0.01; a-c, p< 10-16; b-c, p < 0.03, determined by Mann-Whitney rank sum test.

Fig. 11: Abnormalities of the actin cytoskeleton in Pak-mutant cultured neurons

Figure 11: Abnormalities of the actin cytoskeleton in Pak-mutant cultured neurons. Neurons from wandering 3rd instar larva, cultured for three div and double labeled with anti-HRP and fluorescently- conjugated phalloidin. A-E: Pak6/+. F-J: Pak6/Pak14. A,F: neuronal membrane visualized with anti-HRP.

B,G: Filamentous actin (F-actin) visualized with fluorescently-conjugated phalloidin; images taken with the same exposure time and other imaging parameters. Pak-mutant neurons have a decreased phalloidin signal. C,H: Overlay of anti-HRP (magenta) and phalloidin (green) staining showing overlap (white). D,

E: High magnification views of phalloidin staining from boxed regions in B. There is strong signal in small, distal processes (D) and punctate staining in regions such as the soma (E) which probably represent focal adhesions with the substrate. I,J: High magnification views of boxed regions in G shows no increased in phalloidin signal in filopodia-like structures or in regions likely to contain focal adhesions.

Scale bar A-C, F-H, = 20 μm; D-E, I-J, = 5 μm.

Fig. 12: Defects in neurite growth over time and changes in F-actin and microtubule classes of Pak- mutant growth cones

Figure 12: Defects in neurite growth over time and changes in F-actin and microtubule classes of

Pak-mutant growth cones. A: Higher-order branch number for Pak11/+ (light grey) and Pak11/Pak6 neurons (blue) from dishes fixed at 6, 12, or 24 hours post-plating. Number of neurons in parentheses.

Lowercase letters indicate different statistical groups; all differences p < 10-16 determined by Mann-

Whitney Rank sum test. B-D: Examples of Pak11/+ growth cones at 6 hours in vitro stained with fluorescent phalloidin (magenta) to visualize F-actin (F-act) and anti-alpha-tubulin (green) to visualize microtubules (MT). White shows overlap between phalloidin and anti-alpha tubulin. B: pointed F-actin and bundled microtubules. C: lamellar F-actin and spread microtubules. D: no F-actin extensions and looped microtubules. Scoring scheme adapted from Sanchez-Soriano and colleagues (2010). Scale bar = 5

μm. E: Quantification of cytoskeletal classes for Pak11/+ and Pak11/Pak6 growth cones. Numbers in parenthesis refer to number of neurites scored; number of neurons Pak11/+ = 135, Pak11/Pak6 = 132. * p =

0.01, *** p < 10-4. Significance based on Chi-squared analysis.

Fig. 13: Pak and genes regulating MRLC phosphorylation interact to regulate neurite arbor size in vitro

Figure 13: Pak and genes regulating MRLC phosphorylation interact to regulate neurite arbor size in vitro. A: Proposed mechanism of Pak regulation of neurite arbor size through myosin-based branch retraction. MRLC = myosin regulatory light chain (sqh). MLCK = myosin light chain kinase. PP1/flw = protein phosphatase 1, B-D: Higher-order branch number and total neurite length from neurons after three div. B: Reductions in Pak and flw. Higher-order branch number, top; total neurite length, bottom.

Numbers in parenthesis refer to number of neurons. Lowercase letters indicate different significant groups. a′-b, b-c, c-d P < 10-16; a-b P < 0.02 determined by Mann-Whitney rank sum test. C: Substitution of phosphomimic P[sqhE21] with and without reductions in Pak. a-b, b-c, c-d, b-d, p < 10-16; a-c, n.s. D:

Addition of non-phosphorylatable MRLC P[sqhA20,A21]and reductions in Pak. a-d, p < 10-16; a-b, b-c, c-d, a-c, b-d, n.s; e-f, p< 10-16; f-g P < 0.006.

Fig. 14: Pak and genes regulating MRLC phosphorylation interact to regulate MB morphology in vivo

Figure 14: Pak and genes regulating MRLC phosphorylation interact to regulate MB morphology in vivo. A-D: Projections of confocal optical sections showing frontal views of α and β MB lobes from pharate adults, visualized by anti-FasII immunostaining. Dorsal is at the top. A: Pak6/+ heterozygote shows normal morphology. γ lobe and antennal lobe (AL) are also weakly stained. Scale bar A-D, = 25

μm. B: sqhAX3/+; P[sqhE21]/+; Pak6/+. α and β lobes on the right side of the brain are somewhat thin

(arrows). C: sqh +; P[sqhA20, A21]/+; Pak6/Pak14. Moderate β-lobe fusion. Arrow shows where fibers cross midline; star shows median bundle. D: Pak6/Pak14. One α and one β lobe are missing. The contralateral α lobe is short with a “flamed” end (triangle). E-F: Graphs of α/β lobe defects (left) and β-lobe fusion

(right). Numbers in parenthesis are number of lobes (α and β defects) and number of brains (β-lobe fusion). E: Substitution of ~50% of endogenous MRLC with a phosphomimic form (P[sqhE21]) with and without reductions in Pak. * p = 0.04 as determined by Chi-square analysis. F: Addition of non- phosphorylatable MRLC (P[sqhA20,A21]) with reductions in Pak. * p = 0.02.

100

Chapter IV: Summary and Future directions

Summary of data

In this body of work, I explored how Pak, the Drosophila ortholog of intellectual disability gene PAK3 regulates tissue and cell morphogenesis, neurite-arbor shape and size in vitro, and the organization of the cytoskeleton in neurons. I have shown that Pak is important for morphogenesis of the adult thoracic appendages, the leg and wing. In the wing, the epithelial cells persist after wing maturation, indicating a defect in either programmed cell death or delamination from the cuticle. I found that Pak causes lethality at a much lower penetrance than reported in prior studies. Pak-mutant MB α and β lobes have defects in thickness, length, and trajectory. These defects become more severe during development. I demonstrated that of four supposedly null alleles, three have some combination of read-through translation and accumulation of truncated protein. Truncated Pak protein retains partial function in leg morphogenesis and survival. Conversely, truncated protein increases the severity of MB lobe defects. Additionally, full- length protein may also provide some function (Chapter 2).

Pak regulates the neurite-arbor size and complexity of primary cultured neurons. Initial outgrowth is no different from controls, but is followed by a reduction in branch number, suggesting a loss of branches.

There are changes in the organization of F-actin in neurons cultured for 3 div, with reduced filopodia and localization to puncta likely to be focal adhesion compared to control neurons. In the growth cones, there is reduction in F-actin protrusions and spread microtubules, the organizations associated with branch initiation, invasion, and stabilization. Alleles producing truncated protein had normal neurite length and territory area, while all Pak-mutant genotypes caused significant reductions in higher-order branch number. Therefore, truncated protein retains partial function in regulating neurite length, but not branch number, suggesting non-kinase dependent functions for Pak. I used genetic interaction studies to determine that increased MRLC phosphorylation and reductions in Pak cause decreased in vitro neurite arbor size and defects in MB lobe morphology. Together, this body of evidence suggests that Pak-mutant

101 neuron and brain structure phenotypes are due to excessive acto-myosin retraction of neuronal processes

(Chapter 3).

Significance of research

Although defects in legs and wings of Pak-mutant Drosophila have been briefly noted previously (Hing et al., 1999; Ihry and Bashirullah, 2014), this is the first time these phenotypes have reported in any detail.

The missing and partial cross veins suggest defects in either extracellular transport of Dpp or cell differentiation in response to a morphogen (Blair, 2007). Additionally, persistent wing epithelial cells are likely due to defects in either programmed cell death or adhesion. Pak has important roles in regulating adhesion and is important for compartmental localization of E-cadherin (Pirraglia et al., 2010) and formation of septate junctions during dorsal closure (Bahri et al., 2010). Additionally, localization of F- actin in Pak-mutant neurons suggests possible defects in focal adhesions (Chapter 3). The defects of the leg segments are reminiscent of those caused by genes that regulate the coordinated cell-shape changes important for morphogenesis. Interestingly, in one family with mutations in PAK3, affected boys also have ichthyosis, a skin disorder (Magini et al., 2014). Correlates between Drosophila Pak regulation of epithelial structures, the legs and wings, and ichthyosis suggest that PAKs regulate epithelial morphogenesis, and studies in the fly may help elucidate the underlying cause of ichthyosis.

These studies build on prior studies demonstrating a role of Pak in axon pathfinding and terminal arborization in the optic lobe (Hing et al., 1999) and antennal lobe (Ang et al., 2003). Here, I have demonstrated that Pak regulates neurite arbor morphology at the level of individual cells as well. These findings suggest that the small brain size of individuals with PAK3 mutations could be due to reduced neurite arbor size. I also showed that Pak is crucial for MB structure in the developing Drosophila nervous system. The length, thickness, and size defects of the mushroom body lobes, or bundled axon tracts, is reminiscent of the missing corpus callosum characterized recently in one family (Magini et al.,

102

2014). This data supports a prediction that individuals with PAK3 mutations causing ID may have defects in axonal tracts, such as the corpus callosum.

Defects in the actin cytoskeleton suggest that Pak stabilizes filamentous actin; this stabilization could subsequently regulate neuron morphology. This is reminiscent of oogenesis, when Pak is required to organize F-actin, with mutations causing decreased and disorganized actin in the basal part of follicle cells which interferes with elongation of the chamber (Conder et al., 2007). Pak also stabilizes neurite branches, possibly by inhibiting excessive retraction. In mice with double mutations in PAK1 and PAK3, animals do not develop defects in brain size until postnatal development (Huang et al., 2011). This time- course corroborates with this finding that Pak may play a role in neurite arbor stability. These findings provide clues to the cellular pathology leading to reduced neurite arbor size in individuals with PAK3 mutations. Although cell signaling mechanisms that inhibit neurite retraction have been predicted to be important for both brain development and neurodegeneration (Luo and O’Leary, 2005), the mechanisms have not been well-described. These studies provide evidence that Pak is part of a mechanism capable of inhibiting neurite retraction since loss-of-function phenotypes are consistent with excessive retraction.

Furthermore, this identifies Pak as a candidate gene for future studies to better understand this process.

Genetic interactions support a molecular model of Pak inhibiting phosphorylation of myosin regulatory light chain and subsequently, non-muscle myosin contractility. There was no change found in the ratio of phosphorylated MLCK in mouse PAK1, PAK3 or PAK1+PAK3 knockout models (Meng et al., 2005;

Asrar et al., 2009; Huang et al., 2011), which may be missing important information about regulated localization and timing of MRLC phosphorylation that is crucial for function (Vasquez et al., 2014). My findings demonstrate that Pak and MRLC phosphorylation interact in Drosophila neurons both in vivo and in vitro. Therefore, ways of decreasing MRLC phosphorylation, such as inhibiting Rho kinase or

MLCK, may be a potential therapeutic target for improving the cognitive defects caused by mutations in this pathway. Alternatively, the small neurite arbor phenotype in vitro could be used in a drug screen to find compounds which restore normal growth.

103

Future directions

With time, more families with mutations in PAK3, or in other genes of this pathway, will probably be identified. Future characterizations of the PAK3 clinical phenotype should include brain imaging studies in order to elucidate if brain structure defects are common. This would allow for comparisons between individuals with different types of mutations to ask if the disorder is worse with the accumulation of kinase-dead protein.

Correlations between human and fly phenotype suggest that this is an excellent model for studying this disorder. Pak might be a promising gene for studying potential developmental neurotoxins if a genetically sensitive population may be more sensitive, demonstrating gene-X-environment interactions.

Limitations of studies to-date

Although the body of data I've collected is consistent with retraction, I was not able to do serial imaging of the mushroom bodies to directly demonstrate excessive retraction. There are protocols for culturing explanted pupal brains that can be imaged live using an adapted chamber and inverted microscope

(Prithviraj et al., 2013; Medioni et al., 2015), and Gal4 drivers can be used to express fluorescent protein specifically in the α/β neurons (Aso et al., 2009). Live imaging of cells in culture was attempted, but failed due to massive cell death (S.A. Lewis and L. A. O’Neill, data not shown). Further investigation revealed that the laminin and Schneider’s media are both light-sensitive reagents and high-intensity light exposure can harm cells, for instance through generation of reactive oxygen species and free radicals

(Halliwell, 2014). In order to demonstrate that the cell-culture phenotype is due to excess retraction, images of the same neurons can be captured at successive times between 12-24 hours to see if neurites that were present at earlier time points are lost in later ones. Labs that have published data on serial imaging of Drosophila neurons may have protocols designed to limit damage due to light exposure, possibly by addition of a free radical scavenger to the media. Furthermore, future genetic interaction studies can identify the specific isoform of myosin that is responsible for the Pak-mutant phenotypes of

104 reduced neurite arbor size and MB lobe defects. Myosin II is a strong candidate, due to mutations causing leg defects similar to those caused by Pak (Bayer et al., 2003), the role in retracting neurites in cell culture in response to repulsive cues such as semaphorin (Wylie and Chantler, 2003), and defects in an upstream regulator of Rho kinase, which positively regulates Myosin II activity (Gallo, 2006) causing MB alpha and beta lobe defects that become progressively worse over time like Pak.

Preliminary studies examining whether decreasing cofilin could suppress the Pak phenotype indicated that cofilin and Pak interact to regulate branch number, but not length (Appendix A). This needs additional replication. Then, if this interaction does replicate, test in vivo to see if MB lobe phenotypes are also regulated via cofilin. In particular, it would be interesting to see if β-lobe fusion phenotype of Pak mutants could be suppressed by reducing tsr/cofilin.

105

Appendix A: Cofilin (twinstar) and Pak interact to regulate neurite branches, but not total neurite length in vitro

Cofilin, or actin depolymerizing factor/ADF, cleaves filamentous actin and is necessary for actin turnover and subsequent actin polymerization (Bamburg, 1999). Loss of function mutations in cofilin causes small, collapsed growth cones (Endo et al., 2003) with a reduced rate of neurite extension (Kuhn et al., 2000;

Endo et al., 2003) and disrupts cell rearrangements of terminal filaments in ovary development and border cell fail to migrate during oogenesis changes driving morphogenesis (Chen et al., 2001). Conversely, overexpression of cofilin increases neurite extension of cells in culture (Kuhn et al., 2000; Endo et al.,

2003) and increasing cofilin activity by blocking phosphorylation also blocks border cell migration

(Zhang et al., 2011). Therefore, cofilin is likely to be important for neuron outgrowth and cell-shape changes during morphogenesis.

One of the targets of Pak is LIMK (Fig. 1, pg. 29; Edwards et al., 1999). Phosphorylation increases LIMK activity which phosphorylates cofilin which decreases its activity. Cofilin has been demonstrated to function downstream of Pak in the nervous system by genetic experiments (Ang et al., 2006; Ng and Luo,

2004). Therefore, I hypothesized that mutations in Pak could cause cofilin overactivity, which may contribute to the reduced neurite arbor size.

I performed two replicates of a preliminary experiment to examine interactions between the Drosophila ortholog of cofilin, twinstar, (tsr), and Pak. If cofilin overactivity is contributing to the Pak phenotype of decreased neurite arbor size, then genetic reductions in cofilin in Pak-mutant neurons should increase their size. I used tsrN121 which is a null allele of the gene encoding cofilin; heterozygotes have a ~50% decrease in cofilin. Double heterozygotes of Pak6 and tsrN121 have decreased neurite arbor size. In these experiments, Pak- mutant neurons have reduced neurite arbor size compared to Pak6/+ heterozygotes, as expected (Chapter 3). Pak-mutant neurons that are also heterozygous for tsrN121, which decreases cofilin by ~50%, had significantly improved higher-order branch number, at levels comparable to Pak6/+ control

106 neurons. However, the Pak-mutant reduction in total neurite length was not suppressed by reducing cofilin in either replicate (Fig. A1, pg. 107).

Although this shows an interesting interaction between cofilin and Pak to regulate a specific aspect of the

Pak-mutant in vitro phenotype, more studies are needed. These results are difficult to interpret since one of the controls, the tsrN121/+; Pak6/+ double heterozygote, significantly reduces neurite arbor size compared to Pak6/+ control and tsrN121/+; Pak6/Pak14 neurons. An additional control of tsrN121/+ heterozygous neurons is needed to determine if the small size of the tsrN121/+; Pak6/+ double heterozygotes are due to interactions from reducing both genes, or if reducing tsr alone decreases neurite arbor size in a dominant fashion. I expect this to be a dominant phenotype since decreasing cofilin in cultured rat hippocampal neurons (Kuhn et al., 2000) and chick spinal neurons (Endo et al., 2003) does reduce neurite outgrowth. Future studies substituting ~50% constitutively active cofilin in combination with reductions in Pak could further demonstrate interactions between cofilin and Pak.

107

Fig. A1: Interaction between tsr and Pak

108

Fig. A1: Interaction between tsr and Pak. A: Hypothesized molecular pathway. B-C: Branch number and neurite length of neurons at three div. Genotypes: Pak6/+ (grey) tsrN121/+; Pak6/+ (yellow) tsrN121/+;

Pak6/Pak14 (brown) Pak6/Pak14 (purple). Numbers of neurons in parenthesis. B: Higher-order branch number. a-b, p < 0.02. C: Total neurite length. a-b, 5x10-5

109

Webpages referenced Brain Span Consortium of the Allen Human Brain Atlas (2015). http://brainspan.org. Accessed 7/2015. Human Brain Transcriptome. http://hbatlas.org. Accessed 11/2015. Amino Acid Calculator (2004). http://proteome.gs.washington.edu/cgi-bin/aa_calc.pl Accessed 12/14.

NCBI Protein (2015) http://www.ncbi.nlm.nih.gov/protein Accessed 12/14.

References

Abuelo D (2007) Microcephaly syndromes. Semin Pediatr Neurol 14:118-127.

Adelstein RS, Conti MA (1975) Phosphorylation of platelet myosin increases actin-activated myosin

ATPase activity. Nature 256:597-598.

Albin SD, Davis GW (2004) Coordinating structural and functional synapse development: postsynaptic

p21-activated kinase independently specifies glutamate receptor abundance and postsynaptic

morphology. J Neurosci 24:6871-6879.

Allen KM, Gleeson JG, Bagrodia S, Partington MW, MacMillan JC, Cerione RA, Mulley JC, Walsh CA

(1998) PAK3 mutation in nonsyndromic X-linked mental retardation. Nat Genet 20:25-30.

Alverez VA and Sabatini BL (2007) Anatomical and physiological plasticity of dendritic spines. Annu

Rev Neurosci 30:79-97.

American Pediatric Association (2013) Diagnostic and statistical manual of mental disorders (5th ed.).

Washington, DC.

Andersen R, Li Y, Resseguie M, Brenman JE (2005) Calcium/calmodulin-dependent protein kinase II

alters structural plasticity and cytoskeletal dynamics in Drosophila. J Neurosci 25:8878-8888.

Ang LH, Chen W, Yao Y, Ozawa R, Tao E, Yonekura J, Uemura T, Keshishian H, Hing H (2006) Lim

kinase regulates the development of olfactory and neuromuscular synapses. Dev Biol 293:178-

190.

Ang LH, Kim J, Stepensky V, Hing H (2003) Dock and Pak regulate olfactory axon pathfinding in

Drosophila. Development 130:1307-1316.

110

Aoki Y, Niihori T, Narumi Y, Kure S, Matsubara Y (2008) The RAS/MAPK syndromes: novel roles of

the RAS pathway in human genetic disorders. Hum Mutat 29:992-1006.

Arber S, Barbayannis FA, Hanser H, Schneider C, Stanyon CA, Bernard O, Caroni P (1998) Regulation

of actin dynamics through phosphorylation of cofilin by LIM-kinase. Nature 393:805-809.

Ashwal S, Michelson D, Plawner L, Dobyns WB (2009) Practice parameter: Evaluation of the child with

microcephaly (an evidence-based review): report of the Quality Standards Subcommittee of the

American Academy of Neurology and the Practice Committee of the Child Neurology Society.

Neurology 73:887-897.

Aso Y, Grübel K, Busch S, Friedrich AB, Siwanowicz I, Tanimoto H (2009) The mushroom body of

adult Drosophila characterized by GAL4 drivers. J Neurogenet 23:156-172.

Asrar S, Meng Y, Zhou Z, Todorovski Z, Huang WW, Jia Z (2009) Regulation of hippocampal long-term

potentiation by p21-activated protein kinase 1 (PAK1). Neuropharmacology 56:73-80.

Azzarelli R, Kerloch T, Pacary E (2015) Regulation of cerebral cortex development by Rho GTPases:

insights from in vivo studies. Front Cell Neurosci 8:445.

Ba W, van der Raadt J, Nadif Kasri N (2013) Rho GTPase signaling at the synapse: implications for

intellectual disability. Exp Cell Res 319:2368-2374.

Bahri S, Wang S, Conder R, Choy J, Vlachos S, Dong K, Merino C, Sigrist S, Molnar C, Yang X, Manser

E, Harden N (2010) The leading edge during dorsal closure as a model for epithelial plasticity:

Pak is required for recruitment of the Scribble complex and septate junction formation.

Development 137:2023-2032.

Bai Y, Xiang X, Liang C, Shi L (2015) Regulating Rac in the nervous system: molecular function and

disease implication of Rac GEFs and GAPs. Biomed Res Int 2015:632450.

Bainbridge SP, Bownes M (1981) Staging the metamorphosis of Drosophila melanogaster. J Embryol

exp Morph 66:57-80.

Bamburg JR (1999) Proteins of the ADF/cofilin family: essential regulators of actin dynamics. Annu Rev

Cell Dev Biol 15:185-230.

111

Bayer CA, Halsell SR, Fristrom JW, Kiehart DP, von Kalm L (2003) Genetic interactions between the

RhoA and Stubble-stubbloid loci suggest a role for a type II transmembrane serine protease in

intracellular signaling during Drosophila imaginal disc morphogenesis. Genetics 165:1417-1432.

Bernard O (2007) Lim kinases, regulators of actin dynamics. Int J Biochem Cell Biol 39:1071-1076.

Bienvenu T, des Portes V, McDonell N, Carrie A, Zemni R, Couvert P, Ropers HH, Moraine C, van

Bokhoven H, Fryns JP, Allen K, Walsh CA, Boue J, Kahn A, Chelly J, Beldjord C (2000)

Missense mutation in PAK3, R67C, causes X-linked nonspecific mental retardation. Am J Med

Genet 93:294-298.

Billuart P, Winter CG, Maresh A, Zhao X, Luo L (2001) Regulating axon branch stability: the role of

p190 RhoGAP in repressing a retraction signaling pathway. Cell 107:195-207.

Blair SS (2007) Wing vein patterning in Drosophila and the analysis of intercellular signaling. Annu Rev

Cell Dev Biol 23:293-319.

Boda B, Alberi S, Nikonenko I, Node-Langlois R, Jourdain P, Moosmayer M, Parisi-Jourdain L, Muller

D (2004) The mental retardation protein PAK3 contributes to synapse formation and plasticity in

hippocampus. J Neurosci 24:10816-10825.

Boguski MS, McCormick F (1993) Proteins regulating Ras and its relatives. Nature 366:643-654.

Bolduc FV, Bell K, Cox H, Broadie KS, Tully T (2008) Excess protein synthesis in Drosophila Fragile X

mutants impairs long-term memory. Nat Neurosci 11:1143-1145.

Bolognin S, Lorenzetto E, Diana G, Buffelli M (2014) The potential role of rho GTPases in Alzheimer's

disease pathogenesis. Mol Neurobiol 50:406-422.

Boquet I, Boujemaa R, Carlier MF, Préat T (2000) Ciboulot regulates actin assembly during Drosophila

brain metamorphosis. Cell 102:797-808.

Bosch M, Castro J, Saneyoshi T, Matsuno H, Sur M, Hayashi Y (2014) Structural and molecular

remodeling of dendritic spine substructures during long-term potentiation. Neuron 82:444-459.

Boyle CA, Sheree B, Schieve L, Cohen RA, Blumberg S, Yeargin-Allsopp M, Visser S, Kogan MD

(2011) Trends in the prevalence of developmental disabilities in US children, 1997-2008.

112

Pediatrics 127:1034-1042.

Braga VM, Machesky LM, Hall A, Hotchin NA (1997) The small GTPases Rho and Rac are required for

the establishment of cadherin-dependent cell-cell contacts. J Cell Biol 137:1421-1431.

Brand AH, Perrimon N (1993) Targeted gene expression as a means of altering cell fates and generating

dominant phenotypes. Development 118:401-415.

Bridgman PC, Dave S, Asnes CF, Tullio AN, Adelstein RS (2001) Myosin IIB is required for growth

cone motility. J Neurosci 21:6159-6169.

Bromberg-White JL, Andersen NJ, Duesbery NS (2012) MEK genomics in development and disease.

Brief Funct Genomics 11:300-310.

Brown MD, Cornejo BJ, Kuhn TB, Bamburg JR (2000) Cdc42 stimulates neurite outgrowth and

formation of growth cone filopodia and lamellipodia. J Neurobiol 43:352-364.

Bryant PJ (1975) Pattern formation in the imaginal wing disc of Drosophila melanogaster: fate map,

regeneration and duplication. J Exp Zool 193:49-77.

Chazeau A, Garcia M, Czöndör K, Perrais D, Tessier B, Giannone G, Thoumine O (2015) Mechanical

coupling between transsynaptic N-cadherin adhesions and actin flow stabilizes dendritic spines.

Mol Biol Cell 26:859-873.

Chen J, Godt D, Gunsalus K, Kiss I, Goldberg M, Laski FA (2001) Cofilin/ADF is required for cell

motility during Drosophila ovary development and oogenesis. Nat Cell Biol 3:204-209.

Chew TL, Masaracchia RA, Goeckeler ZM, Wysolmerski RB (1998) Phosphorylation of non-muscle

myosin II regulatory light chain by p21-activated kinase (gamma-PAK). J Muscle Res Cell Motil

19:839-854.

Chong C, Tan L, Lim L, Manser E (2001) The mechanism of PAK activation. Autophosphorylation

events in both regulatory and kinase domains control activity. J Biol Chem 276:17347-17353.

Clark EA, King WG, Brugge JS, Symons M, Hynes RO (1998) Integrin-mediated signals regulated by

members of the rho family of GTPases. J Cell Biol 142:573-586.

Combeau G, Kreis P, Domenichini F, Amar M, Fossier P, Rousseau V, Barnier J (2012) The p21-

113

activated kinase PAK3 forms heterodimers with PAK1 in brain implementing trans-regulation of

PAK3 activity. J Biol Chem 287:30084-30096.

Conder R, Yu H, Ricos M, Hing H, Chia W, Lim L, Harden N (2004) dPak is required for integrity of the

leading edge cytoskeleton during Drosophila dorsal closure but does not signal through the JNK

cascade. Dev Biol 276:378-390.

Conder R, Yu H, Zahedi B, Harden N (2007) The serine/threonine kinase dPak is required for polarized

assembly of F-actin bundles and apical-basal polarity in the Drosophila follicular epithelium. Dev

Biol 305:470-482.

Condic ML, Fristrom D, Fristrom JW (1991) Apical cell shape changes during Drosophila imaginal leg

disc elongation: a novel morphogenetic mechanism. Development 111:23-33.

Crittenden JR, Skoulakis EMC, Han K-A, Kalderon D, Davis RL (1998) Tripartite mushroom body

architecture revealed by antigenic markers. Learn & Mem 5:38-51.

Davies J (2013) Mechanisms of Morphogenesis. London, UK: Academic Press. de Belle JS, Heisenberg M (1994) Associative odor learning in Drosophila abolished by chemical

ablation of mushroom bodies. Science 263:692-695.

De La Cruz E, Pollard TD (1994) Transient kinetic analysis of rhodamine phalloidin binding to actin

filaments. Biochemistry 33:14387-14392.

Dent EW, Callaway JL, Szebenyi G, Baas PW, Kalil K (1999) Reorganization and movement of

microtubules in axonal growth cones and developing interstitial branches. J Neurosci 19:8894-

8908.

Diefenbach TJ, Latham VM, Yimlamai D, Liu CA, Herman IM, Jay DG (2002) Myosin 1c and myosin

IIB serve opposing roles in lamellipodial dynamics of the neuronal growth cone. J Cell Biol

158:1207-1217.

Dierssen M, Ramakers GJ (2006) Dendritic pathology in mental retardation: from molecular genetics to

neurobiology. Genes Brain Behav 5 Suppl 2:48-60.

Donnelly AJ, Partington MW, Ryan AK, Mulley JC (1996) Regional localisation of two non-specific X-

114

linked mental retardation genes (MRX30 and MRX31). Am J Med Genet 64:113-120.

Dunn JG, Foo CK, Belletier NG, Gavis ER, Weissman JS (2013) Ribosome profiling reveals pervasive

and regulated stop codon readthrough in Drosophila melanogaster. Elife 2:e01179.

Eaton S, Auvinen P, Luo L, Jan YN, Simons K (1995) Cdc42 and Rac1 control different actin-dependent

processes in the Drosophila wing disc epithelium. J Cell Biol 131:151-164.

Edwards AC, Zwarts L, Yamamoto A, Callaerts P, Mackay TF (2009) Mutations in many genes affect

aggressive behavior in Drosophila melanogaster. BMC Biol 7:29.

Edwards DC, Sanders LC, Bokoch GM, Gill GN (1999) Activation of LIM-kinase by Pak1 couples

Rac/Cdc42 GTPase signalling to actin cytoskeletal dynamics. Nat Cell Biol 1:253-259.

Edwards KA, Kiehart DP (1996) Drosophila nonmuscle myosin II has multiple essential roles in imaginal

disc and egg chamber morphogenesis. Development 122:1499-1511.

Elgin SR, Miller DW (1978) Mass rearing of flies and mass production and harvesting of eggs. In: The

Genetics and Biology of Drosophila (Ashburner M, Wright TRF, eds), pp 112-121. New York:

Academic Press.

Endo M, Ohashi K, Sasaki Y, Goshima Y, Niwa R, Uemura T, Mizuno K (2003) Control of growth cone

motility and morphology by LIM kinase and Slingshot via phosphorylation and

dephosphorylation of cofilin. J Neurosci 23:2527-2537.

Evans LL, Hammer J, Bridgman PC (1997) Subcellular localization of myosin V in nerve growth cones

and outgrowth from dilute-lethal neurons. J Cell Sci 110 (Pt 4):439-449.

Fristrom D, Chihara C (1978) The mechanism of evagination of imaginal discs of Drosophila

melanogaster. V. Evagination of disc fragments. Dev Biol 66:564-570.

Fristrom D, Fristrom JW (1975) The mechanism of evagination of imaginal discs of Drosophila

melanogaster. I. General considerations. Developmental Biology 43.

Fristrom D, Fristrom JW (1993) The metamorphic development of the adult epidermis. In: The

Development of Drosophila melanogaster (Bate M, Martinez Arias A, eds), pp 843-897. Cold

Spring Harbor, N.Y.: Cold Spring Harbor Laboratory Press.

115

Fristrom JW, Logan WR, Murphy C (1973) The synthetic and minimal culture requirements for

evagination of imaginal discs of Drosophila melanogaster in vitro. Dev Biol 33:441-456.

Gallo G (2006) RhoA-kinase coordinates F-actin organization and myosin II activity during semaphorin-

3A-induced axon retraction. J Cell Sci 119:3413-3423.

Gallo G, Yee HF, Jr., Letourneau PC (2002) Actin turnover is required to prevent axon retraction driven

by endogenous actomyosin contractility. J Cell Biol 158:1219-1228.

Gedeon AK, Nelson J, Gécz J, Mulley JC (2003) X-linked mild non-syndromic mental retardation with

neuropsychiatric problems and the missense mutation A365E in PAK3. Am J Med Genet A

120A:509-517.

Gervasi N, Tchénio P, Preat T (2010) PKA dynamics in a Drosophila learning center: coincidence

detection by rutabaga adenylyl cyclase and spatial regulation by dunce phosphodiesterase.

Neuron 65:516-529.

Goeckeler ZM, Masaracchia RA, Zeng Q, Chew TL, Gallagher P, Wysolmerski RB (2000)

Phosphorylation of myosin light chain kinase by p21-activated kinase PAK2. J Biol Chem

275:18366-18374.

Halliwell B (2014) Cell culture, oxidative stress, and antioxidants: avoiding pitfalls. Biomed J 37:99-105.

Halsell SR, Chu BI, Kiehart DP (2000) Genetic analysis demonstrates a direct link between rho signaling

and nonmuscle myosin function during Drosophila morphogenesis. Genetics 155:1253-1265.

Hammer JA, 3rd, Wagner W (2013) Functions of class V myosins in neurons. J Biol Chem 288:28428-

28434.

Han PL, Levin LR, Reed RR, Davis RL (1992) Preferential expression of the Drosophila rutabaga gene

in mushroom bodies, neural centers for learning in insects. Neuron 9:619-627.

Harden N, Lee J, Loh HY, Ong YM, Tan I, Leung T, Manser E, Lim L (1996) A Drosophila homolog of

the Rac- and Cdc42-activated serine/threonine kinase PAK is a potential focal adhesion and focal

complex protein that colocalizes with dynamic actin structures. Mol Cell Biol 16:1896-1908.

Harden N, Ricos M, Ong YM, Chia W, Lim L (1999) Participation of small GTPases in dorsal closure of

116

the Drosophila embryo: distinct roles for Rho subfamily proteins in epithelial morphogenesis. J

Cell Sci 112 ( Pt 3):273-284.

Häusser M, Spruston N, Stuart GJ (2000) Diversity and dynamics of dendritic signaling. Science

290:739-744.

Hawrylycz MJ, Lein ES, Guillozet-Bongaarts AL, Shen EH, Ng L, Miller JA, van de Lagemaat LN,

Smith KA, Ebbert A, Riley ZL, Abajian C, Beckmann CF, Bernard A, Bertagnolli D, Boe AF,

Cartagena PM, Chakravarty MM, Chapin M, Chong J, Dalley RA, Daly BD, Dang C, Datta S,

Dee N, Dolbeare TA, Faber V, Feng D, Fowler DR, Goldy J, Gregor BW, Haradon Z, Haynor

DR, Hohmann JG, Horvath S, Howard RE, Jeromin A, Jochim JM, Kinnunen M, Lau C, Lazarz

ET, Lee C, Lemon TA, Li L, Li Y, Morris JA, Overly CC, Parker PD, Parry SE, Reding M,

Royall JJ, Schulkin J, Sequeira PA, Slaughterbeck CR, Smith SC, Sodt AJ, Sunkin SM, Swanson

BE, Vawter MP, Williams D, Wohnoutka P, Zielke HR, Geschwind DH, Hof PR, Smith SM,

Koch C, Grant SG, Jones AR (2012) An anatomically comprehensive atlas of the adult human

brain transcriptome. Nature 489:391-399.

Hayashi ML, Choi SY, Rao BS, Jung HY, Lee HK, Zhang D, Chattarji S, Kirkwood A, Tonegawa S

(2004) Altered cortical synaptic morphology and impaired memory consolidation in forebrain-

specific dominant-negative PAK transgenic mice. Neuron 42:773-787.

Heisenberg M (2003) Mushroom body memoir: from maps to models. Nat Rev Neurosci 4:266-275.

Heisenberg M, Borst A, Wagner S, Byers D (1985) Drosophila mushroom body mutants are deficient in

olfactory learning. J Neurogenet 2:1-30.

Hing H, Xiao J, Harden N, Lim L, Zipursky SL (1999) Pak functions downstream of Dock to regulate

photoreceptor axon guidance in Drosophila. Cell 97:853-863.

Huang W, Zhou Z, Asrar S, Henkelman M, Xie W, Jia Z (2011) p21-Activated kinases 1 and 3 control

brain size through coordinating neuronal complexity and synaptic properties. Mol Cell Biol

31:388-403.

Ihry RJ, Bashirullah A (2014) Genetic control of specificity to steroid-triggered responses in Drosophila.

117

Genetics 196:767-780.

Ikebe M, Inagaki M, Naka M, Hidaka H (1988) Correlation of conformation and phosphorylation and

dephosphorylation of smooth muscle myosin. J Biol Chem 263:10698-10704.

Impey S, Smith DM, Obrietan K, Donahue R, Wade C, Storm DR (1998) Stimulation of cAMP response

element (CRE)-mediated transcription during contextual learning. Nat Neurosci 1:595-601.

Inlow JK, Restifo LL (2004) Molecular and comparative genetics of mental retardation. Genetics

166:835-881.

Irwin SA, Patel B, Idupulapati M, Harris JB, Crisostomo RA, Larsen BP, Kooy F, Willems PJ, Cras P,

Kozlowski PB, Swain RA, Weiler IJ, Greenough WT (2001) Abnormal dendritic spine

characteristics in the temporal and visual cortices of patients with fragile-X syndrome: a

quantitative examination. Am J Med Genet 98:161-167.

Ito K, Hotta Y (1992) Proliferation pattern of postembryonic neuroblasts in the brain of Drosophila

melanogaster. Dev Biol 149:134-148.

Iyer SC, Wang D, Iyer EP, Trunnell SA, Meduri R, Shinwari R, Sulkowski MJ, Cox DN (2012) The

RhoGEF trio functions in sculpting class specific dendrite morphogenesis in Drosophila sensory

neurons. PLoS One 7:e33634.

Jaffer ZM, Chernoff J (2002) p21-activated kinases: three more join the Pak. Int J Biochem Cell Biol

34:713-717.

Jay DG (2000) The clutch hypothesis revisited: ascribing the roles of actin-associated proteins in

filopodial protrusion in the nerve growth cone. J Neurobiol 44:114-125.

Jester JV, Barry PA, Lind GJ, Petroll WM, Garana R, Cavanagh HD (1994) Corneal keratocytes: in situ

and in vitro organization of cytoskeletal contractile proteins. Invest Ophthalmol Vis Sci 35:730-

743.

Jordan P, Karess R (1997) Myosin light chain-activating phosphorylation sites are required for oogenesis

in Drosophila. J Cell Biol 139:1805-1819.

Jungreis I, Lin MF, Spokony R, Chan CS, Negre N, Victorsen A, White KP, Kellis M (2011) Evidence of

118

abundant stop codon readthrough in Drosophila and other metazoa. Genome Res 21:2096-2113.

Kamisoyama H, Araki Y, Ikebe M (1994) Mutagenesis of the phosphorylation site (serine 19) of smooth

muscle myosin regulatory light chain and its effects on the properties of myosin. Biochemistry

33:840-847.

Kang HJ, Kawasawa YI, Cheng F, Zhu Y, Xu X, Li M, Sousa AM, Pletikos M, Meyer KA, Sedmak G,

Guennel T, Shin Y, Johnson MB, Krsnik Z, Mayer S, Fertuzinhos S, Umlauf S, Lisgo SN,

Vortmeyer A, Weinberger DR, Mane S, Hyde TM, Huttner A, Reimers M, Kleinman, JE, Sestan

N (2011) Spatio-temporal transcriptome of the human brain. Nature 478:483-489.

Karess RE, Chang XJ, Edwards KA, Kulkarni S, Aguilera I, Kiehart DP (1991) The regulatory light chain

of nonmuscle myosin is encoded by spaghetti-squash, a gene required for cytokinesis in

Drosophila. Cell 65:1177-1189.

Kaufman L, Ayub M, Vincent JB (2010) The genetic basis of non-syndromic intellectual disability: a

review. J Neurodev Disord 2:182-209.

Kaufmann WE, Moser HW (2000) Dendritic anomalies in disorders associated with mental retardation.

Cereb Cortex 10:981-991.

Kiger JA, Jr., Natzle JE, Kimbrell DA, Paddy MR, Kleinhesselink K, Green MM (2007) Tissue

remodeling during maturation of the Drosophila wing. Dev Biol 301:178-191.

Kim CH, Lisman JE (1999) A role of actin filament in synaptic transmission and long-term potentiation. J

Neurosci 19:4314-4324.

Kim YS, Fritz JL, Seneviratne AK, VanBerkum MF (2002) Constitutively active myosin light chain

kinase alters axon guidance decisions in Drosophila embryos. Dev Biol 249:367-381.

Kisiel M, McKenzie K, Stewart B (2014) Localization and mobility of synaptic vesicles in Myosin VI

mutants of Drosophila. PLoS One 9:e102988.

Kiss I, Beaton AH, Tardiff J, Fristrom D, Fristrom JW (1988) Interactions and developmental effects of

mutations in the Broad-Complex of Drosophila melanogaster. Genetics. 118:247-259.

Kneussel M, Wagner W (2013) Myosin motors at neuronal synapses: drivers of membrane transport and

119

actin dynamics. Nat Rev Neurosci 14:233-247.

Kosako H, Yoshida T, Matsumura F, Ishizaki T, Narumiya S, Inagaki M (2000) Rho-kinase/ROCK is

involved in cytokinesis through the phosphorylation of myosin light chain and not

ezrin/radixin/moesin proteins at the cleavage furrow. Oncogene 19:6059-6064.

Kozma R, Sarner S, Ahmed S, Lim L (1997) Rho family GTPases and neuronal growth cone remodelling:

relationship between increased complexity induced by Cdc42Hs, Rac1, and acetylcholine and

collapse induced by RhoA and lysophosphatidic acid. Mol Cell Biol 17:1201-1211.

Kraft R, Escobar MM, Narro ML, Kurtis JL, Efrat A, Barnard K, Restifo LL (2006) Phenotypes of

Drosophila brain neurons in primary culture reveal a role for fascin in neurite shape and

trajectory. J Neurosci 26:8734-8747.

Kraft R, Kahn A, Medina-Franco JL, Orlowski ML, Baynes C, López-Vallejo F, Barnard K, Maggiora

GM, Restifo LL (2013) A cell-based fascin bioassay for multi-purpose compound screening. Dis

Model Mech 6:217-235.

Kraft R, Levine RB, Restifo LL (1998) The steroid hormone 20-hydroxyecdysone enhances neurite

growth of Drosophila mushroom body neurons isolated during metamorphosis. J Neurosci

18:8886-8899.

Kreis P, Thévenot E, Rousseau V, Boda B, Muller D, Barnier JV (2007) The p21-activated kinase 3

implicated in mental retardation regulates spine morphogenesis through a Cdc42-dependent

pathway. J Biol Chem 282:21497-21506.

Kuhn TB, Meberg PJ, Brown MD, Bernstein BW, Minamide LS, Jensen JR, Okada K, Soda EA,

Bamburg JR (2000) Regulating actin dynamics in neuronal growth cones by ADF/cofilin and rho

family GTPases. J Neurobiol 44:126-144.

Kulkarni VA, Firestein BL (2012) The dendritic tree and brain disorders. Mol Cell Neurosci 50:10-20.

Lecuit T, Lenne PF, Munro E (2011) Force generation, transmission, and integration during cell and

tissue morphogenesis. Annu Rev Cell Dev Biol 27:157-184.

Lee T, Lee A, Luo L (1999) Development of the Drosophila mushroom bodies: sequential generation of

120

three distinct types of neurons from a neuroblast. Development 126:4065-4076.

Lee T, Winter C, Marticke SS, Lee A, Luo L (2000) Essential roles of Drosophila RhoA in the regulation

of neuroblast proliferation and dendritic but not axonal morphogenesis. Neuron 25:307-316.

Lei M, Lu W, Meng W, Parrini MC, Eck MJ, Mayer BJ, Harrison SC (2000) Structure of PAK1 in an

autoinhibited conformation reveals a multistage activation switch. Cell 102:387-397.

Leverrier Y, Ridley AJ (2001) Requirement for Rho GTPases and PI 3-kinases during apoptotic cell

phagocytosis by macrophages. Curr Biol 11:195-199.

Lewis TL, Jr., Mao T, Arnold DB (2011) A role for myosin VI in the localization of axonal proteins.

PLoS Biol 9:e1001021.

Lewis TL, Jr., Mao T, Svoboda K, Arnold DB (2009) Myosin-dependent targeting of transmembrane

proteins to neuronal dendrites. Nat Neurosci 12:568-576.

Li BX, Satoh AK, Ready DF (2007) Myosin V, Rab11, and dRip11 direct apical secretion and cellular

morphogenesis in developing Drosophila photoreceptors. J Cell Biol 177:659-669.

Lindsay AJ, McCaffrey MW (2014) Myosin Va is required for the transport of fragile X mental

retardation protein (FMRP) granules. Biol Cell 106:57-71.

Link N, Chen P, Lu WJ, Pogue K, Chuong A, Mata M, Checketts J, Abrams JM (2007) A collective form

of cell death requires homeodomain interacting protein kinase. J Cell Biol 178:567-574.

Liu YZ, Thomas NS, Latchman DS (1999) CBP associates with the p42/p44 MAPK enzymes and is

phosphorylated following NGF treatment. Neuroreport 10:1239-1243.

Luo L, O'Leary DD (2005) Axon retraction and degeneration in development and disease. Annu Rev

Neurosci 28:127-156.

Lv K, Chen L, Li Y, Li Z, Zheng P, Liu Y, Chen J, Teng J (2015) Trip6 promotes dendritic

morphogenesis through dephosphorylated GRIP1-dependent myosin VI and F-actin organization.

J Neurosci 35:2559-2571.

Ma QL, Yang F, Calon F, Ubeda OJ, Hansen JE, Weisbart RH, Beech W, Frautschy SA, Cole GM (2008)

p21-activated kinase-aberrant activation and translocation in Alzheimer disease pathogenesis. J

121

Biol Chem 283:14132-14143.

Maekawa M, Ishizaki T, Boku S, Watanabe N, Fujita A, Iwamatsu A, Obinata T, Ohashi K, Mizuno K,

Narumiya S (1999) Signaling from Rho to the actin cytoskeleton through protein kinases ROCK

and LIM-kinase. Science 285:895-898.

Magini P, Pippucci T, Tsai IC, Coppola S, Stellacci E, Bartoletti-Stella A, Turchetti D, Graziano C,

Cenacchi G, Neri I, Cordelli DM, Marchiani V, Bergamaschi R, Gasparre G, Neri G, Mazzanti L,

Patrizi A, Franzoni E, Romeo G, Bordo D, Tartaglia M, Katsanis N, Seri M (2014) A mutation in

PAK3 with a dual molecular effect deregulates the RAS/MAPK pathway and drives an X-linked

syndromic phenotype. Hum Mol Genet 23:3607-3617.

Malenka RC, Nicoll RA (1999) Long-term potentiation--a decade of progress? Science 285:1870-1874.

Manser E, Leung T, Salihuddin H, Zhao ZS, Lim L (1994) A brain serine/threonine protein kinase

activated by Cdc42 and Rac1. Nature 367:40-46.

Manser E, Loo TH, Koh CG, Zhao ZS, Chen XQ, Tan L, Tan I, Leung T, Lim L (1998) PAK kinases are

directly coupled to the PIX family of nucleotide exchange factors. Mol Cell 1:183-192.

Mao Z, Roman G, Zong L, Davis RL (2004) Pharmacogenetic rescue in time and space of the rutabaga

memory impairment by using Gene-Switch. Proc Natl Acad Sci U S A 101:198-203.

Martin JR, Ernst R, Heisenberg M (1998) Mushroom bodies suppress locomotor activity in Drosophila

melanogaster. Learn Mem 5:179-191.

Matsuda S, Blanco J, Shimmi O (2013) A feed-forward loop coupling extracellular BMP transport and

morphogenesis in Drosophila wing. PLoS Genet 9:e1003403.

Maulik PK, Mascarenhas M, Mathers C, Dua T, Saxena S (2011) Prevalence of intellectual disability: a

meta-analysis of population-based studies. Res Dev Disabil 32:419-436.

McBride SM, Giulianai G, Choi C, Krause P, Correale D, Watson K, Baker G, Siwicki KK (1999)

Mushroom body ablation impairs short-term memory and long-term memory of courtship

conditioning in Drosophila melanogaster. Neuron 24:967-977.

McMichael G, Bainbridge MN, Haan E, Corbett M, Gardner A, Thompson S, van Bon BW, van Eyk CL,

122

Broadbent J, Reynolds C, O'Callaghan ME, Nguyen LS, Adelson DL, Russo R, Jhangiani S,

Doddapaneni H, Muzny DM, Gibbs RA, Gecz J, MacLennan AH (2015) Whole-exome

sequencing points to considerable genetic heterogeneity of cerebral palsy. Mol Psychiatry 20:176-

182.

Medioni C, Ephrussi A, Besse F (2015) Live imaging of axonal transport in Drosophila pupal brain

explants. Nat Protoc 10:574-584.

Melzig J, Rein K-H, Schäfer U, Pfister H, Jäckle H, Heisenberg M, Raabe T (1998) A protein related to

p21-activated kinase (PAK) that is involved in neurogenesis in the Drosophila adult central

nervous system. Curr Biol 8:1223-1226.

Meng J, Meng Y, Hanna A, Janus C, Jia Z (2005) Abnormal long-lasting synaptic plasticity and cognition

in mice lacking the mental retardation gene Pak3. J Neurosci 25:6641-6650.

Menges CW, Sementino E, Talarchek J, Xu J, Chernoff J, Peterson JR, Testa JR (2012) Group I p21-

activated kinases (PAKs) promote tumor cell proliferation and survival through the AKT1 and

Raf-MAPK pathways. Mol Cancer Res 10:1178-1188.

Mentzel B, Raabe T (2005) Phylogenetic and structural analysis of the Drosophila melanogaster p21-

activated kinase DmPAK3. Gene 349:25-33.

Menzel N, Schneeberger D, Raabe T (2007) The Drosophila p21 activated kinase Mbt regulates the actin

cytoskeleton and adherens junctions to control photoreceptor cell morphogenesis. Mech Dev

124:78-90.

Michel CI, Kraft R, Restifo LL (2004) Defective neuronal development in the mushroom bodies of

Drosophila fragile X mental retardation 1 mutants. J Neurosci 24:5798-5809.

Milner MJ (1977) The eversion and differentiation of Drosophila melanogaster leg and wing imaginal

discs cultured in vitro with an optimal concentration of beta-ecdysone. J Embryol Exp Morphol

37:105-117.

Milner MJ, Bleasby AJ, Kelly SL (1984) The role of the peripodial membrane of the leg and wing

imaginal discs of Drosophila melanogaster during evagination and differentiation in vitro. Roux's

123

Arch Dev Biol 193:180-186.

Mitchison T, Kirschner M (1988) Cytoskeletal dynamics and nerve growth. Neuron 1:761-772.

Moon HM, Wynshaw-Boris A (2013) Cytoskeleton in action: lissencephaly, a neuronal migration

disorder. Wiley Interdiscip Rev Dev Biol 2:229-245.

Moorman JP, Luu D, Wickham J, Bobak DA, Hahn CS (1999) A balance of signaling by Rho family

small GTPases RhoA, Rac1 and Cdc42 coordinates cytoskeletal morphology but not cell survival.

Oncogene 18:47-57.

Nagase T, Seki N, Tanaka A, Ishikawa K, Nomura N (1995) Prediction of the coding sequences of

unidentified human genes. IV. The coding sequences of 40 new genes (KIAA0121-KIAA0160)

deduced by analysis of cDNA clones from human cell line KG-1. DNA Res 2:167-174, 199-210.

Narro ML, Yang F, Kraft R, Wenk C, Efrat A, Restifo LL (2007) NeuronMetrics: software for semi-

automated processing of cultured neuron images. Brain Res 1138:57-75.

Neubueser D, Hipfner DR (2010) Overlapping roles of Drosophila Drak and Rok kinases in epithelial

tissue morphogenesis. Mol Biol Cell 21:2869-2879.

Newey SE, Velamoor V, Govek EE, Van Aelst L (2005) Rho GTPases, dendritic structure, and mental

retardation. J Neurobiol 64:58-74.

Newsome TP, Åsling B, Dickson BJ (2000) Analysis of Drosophila photoreceptor axon guidance in eye-

specific mosaics. Development 127:851-860.

Ng J, Luo L (2004) Rho GTPases regulate axon growth through convergent and divergent signaling

pathways. Neuron 44:779-793.

Nighorn A, Healy MJ, Davis RL (1991) The cyclic AMP phosphodiesterase encoded by the Drosophila

dunce gene is concentrated in the mushroom body neuropil. Neuron 6:455-467.

Nöthiger R, Schubiger G (1966) Developmental behaviour of fragments of symmetrical and asymmetrical

imaginal discs of Drosophila melanogaster (Diptera). J Embryol Exp Morphol 16:255-268.

Okamoto K, Nagai T, Miyawaki A, Hayashi Y (2004) Rapid and persistent modulation of actin dynamics

regulates postsynaptic reorganization underlying bidirectional plasticity. Nat Neurosci 7:1104-

124

1112.

Pan L, Zhang YQ, Woodruff E, Broadie K (2004) The Drosophila fragile X gene negatively regulates

neuronal elaboration and synaptic differentiation. Curr Biol 14:1863-1870.

Park ER, Eblen ST, Catling AD (2007) MEK1 activation by PAK: a novel mechanism. Cell Signal

19:1488-1496.

Parnas D, Haghighi AP, Fetter RD, Kim SW, Goodman CS (2001) Regulation of postsynaptic structure

and protein localization by the Rho-type guanine nucleotide exchange factor dPix. Neuron

32:415-424.

Parrini MC, Lei M, Harrison SC, Mayer BJ (2002) Pak1 kinase homodimers are autoinhibited in trans and

dissociated upon activation by Cdc42 and Rac1. Mol Cell 9:73-83.

Parrish JZ, Emoto K, Kim MD, Jan YN (2007) Mechanisms that regulate establishment, maintenance,

and remodeling of dendritic fields. Annu Rev Neurosci 30:399-423.

Pascual A, Préat T (2001) Localization of long-term memory within the Drosophila mushroom body.

Science 294:1115-1117.

Pastural E, Barrat FJ, Dufourcq-Lagelouse R, Certain S, Sanal O, Jabado N, Seger R, Griscelli C, Fischer

A, de Saint Basile G (1997) Griscelli disease maps to chromosome 15q21 and is associated with

mutations in the myosin-Va gene. Nat Genet 16:289-292.

Pattatucci AM, Otteson DC, Kaufman TC (1991) A functional and structural analysis of the Sex combs

reduced locus of Drosophila melanogaster. Genetics 129:423-441.

Pauls D, Selcho M, Gendre N, Stocker RF, Thum AS (2010) Drosophila larvae establish appetitive

olfactory memories via mushroom body neurons of embryonic origin. J Neurosci 30:10655-

10666.

Peabody NC, White BH (2013) Eclosion gates progression of the adult ecdysis sequence of Drosophila. J

Exp Biol 216:4395-4402.

Peippo M, Koivisto AM, Särkämö T, Sipponen M, von Koskull H, Ylisaukko-oja T, Rehnström K,

Froyen G, Ignatius J, Järvelä I (2007) PAK3 related mental disability: further characterization of

125

the phenotype. Am J Med Genet A 143A:2406-2416.

Penzes P, Rafalovich I (2012) Regulation of the actin cytoskeleton in dendritic spines. Adv Exp Med Biol

970:81-95.

Peter A, Schottler P, Werner M, Beinert N, Dowe G, Burkert P, Mourkioti F, Dentzer L, He Y, Deak P,

Benos PV, Gatt MK, Murphy L, Harris D, Barrell B, Ferraz C, Vidal S, Brun C, Demaille J,

Cadieu E, Dreano S, Gloux S, Lelaure V, Mottier S, Galibert F, Borkova D, Minana B, Kafatos

FC, Bolshakov S, Siden-Kiamos I, Papagiannakis G, Spanos L, Louis C, Madueno E, de Pablos

B, Modolell J, Bucheton A, Callister D, Campbell L, Henderson NS, McMillan PJ, Salles C, Tait

E, Valenti P, Saunders RD, Billaud A, Pachter L, Klapper R, Janning W, Glover DM, Ashburner

M, Bellen HJ, Jackle H, Schafer U (2002) Mapping and identification of essential gene functions

on the X chromosome of Drosophila. EMBO Rep 3:34-38.

Pirraglia C, Walters J, Myat MM (2010) Pak1 control of E-cadherin endocytosis regulates salivary gland

lumen size and shape. Development 137:4177-4189.

Pirruccello M, Sondermann H, Pelton JG, Pellicena P, Hoelz A, Chernoff J, Wemmer DE, Kuriyan J

(2006) A dimeric kinase assembly underlying autophosphorylation in the p21 activated kinases. J

Mol Biol 361:312-326.

Pitman JL, McGill JJ, Keegan KP, Allada R (2006) A dynamic role for the mushroom bodies in

promoting sleep in Drosophila. Nature 441:753-756.

Podkowa M, Christova T, Zhao X, Jian Y, Attisano L (2013) p21-Activated kinase (PAK) is required for

Bone Morphogenetic Protein (BMP)-induced dendritogenesis in cortical neurons. Mol Cell

Neurosci 57:83-92.

Poodry CA, Schneiderman HA (1970) The ultrastructure of the developing leg of Drosophila

melanogaster. W Roux Arch 166:1-44.

Prithviraj R, Trunova S, Giniger E (2012) Ex vivo culturing of whole, developing Drosophila brains. J

Vis Exp:e4270.

Prokop A, Beaven R, Qu Y, Sánchez-Soriano N (2013) Using fly genetics to dissect the cytoskeletal

126

machinery of neurons during axonal growth and maintenance. J Cell Sci 126:2331-2341.

Puram SV, Bonni A (2013) Cell-intrinsic drivers of dendrite morphogenesis. Development 140:4657-

4671.

Ramakers GJ (2002) Rho proteins, mental retardation and the cellular basis of cognition. Trends Neurosci

25:191-199.

Rane CK, Minden A (2014) P21 activated kinases: structure, regulation, and functions. Small GTPases 5.

Rejeb I, Saillour Y, Castelnau L, Julien C, Bienvenu T, Taga P, Chaabouni H, Chelly J, Ben Jemaa L,

Bahi-Buisson N (2008) A novel splice mutation in PAK3 gene underlying mental retardation with

neuropsychiatric features. Eur J Hum Genet 16:1358-1363.

Restifo LL (2005) Mental retardation genes in Drosophila: New approaches to understanding and treating

developmental brain disorders. Ment Retard Dev Disabil Res Rev 11:286-294.

Ridley AJ, Hall A (1992) The small GTP-binding protein rho regulates the assembly of focal adhesions

and actin stress fibers in response to growth factors. Cell 70:389-399.

Ridley AJ, Paterson HF, Johnston CL, Diekmann D, Hall A (1992) The small GTP-binding protein rac

regulates growth factor-induced membrane ruffling. Cell 70:401-410.

Rivière JB, van Bon BW, Hoischen A, Kholmanskikh SS, O'Roak BJ, Gilissen C, Gijsen S, Sullivan CT,

Christian SL, Abdul-Rahman OA, Atkin JF, Chassaing N, Drouin-Garraud V, Fry AE, Fryns JP,

Gripp KW, Kempers M, Kleefstra T, Mancini GM, Nowaczyk MJ, van Ravenswaaij-Arts CM,

Roscioli T, Marble M, Rosenfeld JA, Siu VM, de Vries BB, Shendure J, Verloes A, Veltman JA,

Brunner HG, Ross ME, Pilz DT, Dobyns WB (2012) De novo mutations in the actin genes ACTB

and ACTG1 cause Baraitser-Winter syndrome. Nat Genet 44:440-444, S441-442.

Roberts PJ, Der CJ (2007) Targeting the Raf-MEK-ERK mitogen-activated protein kinase cascade for the

treatment of cancer. Oncogene 26:3291-3310.

Roche AF, Mukherjee D, Guo SM, Moore WM (1987) Head circumference reference data: birth to 18

years. Pediatrics 79:706-712.

Rochlin MW, Itoh K, Adelstein RS, Bridgman PC (1995) Localization of myosin II A and B isoforms in

127

cultured neurons. J Cell Sci 108 (Pt 12):3661-3670.

Roelfsema JH, White SJ, Ariyürek Y, Bartholdi D, Niedrist D, Papadia F, Bacino CA, den Dunnen JT,

van Ommen GJ, Breuning MH, Hennekam RC, Peters DJ (2005) Genetic heterogeneity in

Rubinstein-Taybi syndrome: mutations in both the CBP and EP300 genes cause disease. Am J

Hum Genet 76:572-580.

Rossman KL, Der CJ, Sondek J (2005) GEF means go: turning on RHO GTPases with guanine

nucleotide-exchange factors. Nat Rev Mol Cell Biol 6:167-180.

Royal I, Lamarche-Vane N, Lamorte L, Kaibuchi K, Park M (2000) Activation of cdc42, rac, PAK, and

rho-kinase in response to hepatocyte growth factor differentially regulates epithelial cell colony

spreading and dissociation. Mol Biol Cell 11:1709-1725.

Ruppert C, Kroschewski R, Bähler M (1993) Identification, characterization, and cloning of myr 1, a

mammalian myosin-I. J Cell Biol 120:1393-1403.

Sakai T, Kitamoto T (2006) Differential roles of two major brain structures, mushroom bodies and central

complex, for Drosophila male courtship behavior. J Neurobiol 66:821-834.

Sánchez-Soriano N, Gonçalves-Pimentel C, Beaven R, Haessler U, Ofner-Ziegenfuss L, Ballestrem C,

Prokop A (2010) Drosophila growth cones: a genetically tractable platform for the analysis of

axonal growth dynamics. Dev Neurobiol 70:58-71.

Sanders LC, Matsumura F, Bokoch GM, de Lanerolle P (1999) Inhibition of myosin light chain kinase by

p21-activated kinase. Science 283:2083-2085.

Santiago-Medina M, Gregus KA, Gomez TM (2013) PAK-PIX interactions regulate adhesion dynamics

and membrane protrusion to control neurite outgrowth. J Cell Sci 126:1122-1133.

Schalock RL, Luckasson RA, Shogren KA, Borthwick-Duffy S, Bradley V, Buntinx WH, Coulter DL,

Craig EM, Gomez SC, Lachapelle Y, Reeve A, Snell ME, Spreat S, Tassé MJ, Thompson JR,

Verdugo MA, Wehmeyer ML, Yeager MH (2007) The renaming of mental retardation:

understanding the change to the term intellectual disability. Intellect Developmental Disabilities

45:116-124.

128

Shifrin Y, Pinto VI, Hassanali A, Arora PD, McCulloch CA (2012) Force-induced apoptosis mediated by

the Rac/Pak/p38 signalling pathway is regulated by filamin A. Biochem J 445:57-67.

Shivalkar M, Giniger E (2012) Control of dendritic morphogenesis by Trio in Drosophila melanogaster.

PLoS One 7:e33737.

Simões S, Denholm B, Azevedo D, Sotillos S, Martin P, Skaer H, Hombría JC, Jacinto A (2006)

Compartmentalisation of Rho regulators directs cell invagination during tissue morphogenesis.

Development 133:4257-4267.

Star EN, Kwiatkowski DJ, Murthy VN (2002) Rapid turnover of actin in dendritic spines and its

regulation by activity. Nat Neurosci 5:239-246.

Suter DM, Espindola FS, Lin CH, Forscher P, Mooseker MS (2000) Localization of unconventional

myosins V and VI in neuronal growth cones. J Neurobiol 42:370-382.

Suter DM, Forscher P (2000) Substrate-cytoskeletal coupling as a mechanism for the regulation of growth

cone motility and guidance. J Neurobiol 44:97-113.

Suzuki M, Morita H, Ueno N (2012) Molecular mechanisms of cell shape changes that contribute to

vertebrate neural tube closure. Dev Growth Differ 54:266-276.

Tamada A, Kawase S, Murakami F, Kamiguchi H (2010) Autonomous right-screw rotation of growth

cone filopodia drives neurite turning. J Cell Biol 188:429-441.

Tan JL, Ravid S, Spudich JA (1992) Control of nonmuscle myosins by phosphorylation. Annu Rev

Biochem 61:721-759.

Tanaka E, Ho T, Kirschner MW (1995) The role of microtubule dynamics in growth cone motility and

axonal growth. J Cell Biol 128:139-155.

Taylor J, Adler PN (2008) Cell rearrangement and cell division during the tissue level morphogenesis of

evaginating Drosophila imaginal discs. Dev Biol 313:739-751. ten Klooster JP, Jaffer ZM, Chernoff J, Hordijk PL (2006) Targeting and activation of Rac1 are mediated

by the exchange factor beta-Pix. J Cell Biol 172:759-769.

Tuzovic L, Yu L, Zeng W, Li X, Lu H, Lu HM, Gonzalez KD, Chung WK (2013) A human de novo

129

mutation in MYH10 phenocopies the loss of function mutation in mice. Rare Dis 1:e26144.

Ueda K, Murata-Hori M, Tatsuka M, Hosoya H (2002) Rho-kinase contributes to diphosphorylation of

myosin II regulatory light chain in nonmuscle cells. Oncogene 21:5852-5860.

Vallee RB, Seale GE, Tsai JW (2009) Emerging roles for myosin II and cytoplasmic dynein in migrating

neurons and growth cones. Trends Cell Biol 19:347-355.

Van Aelst L, Symons M (2002) Role of Rho family GTPases in epithelial morphogenesis. Genes Dev

16:1032-1054.

Vasquez CG, Tworoger M, Martin AC (2014) Dynamic myosin phosphorylation regulates contractile

pulses and tissue integrity during epithelial morphogenesis. J Cell Biol 206:435-450.

Vereshchagina N, Bennett D, Szöőr B, Kirchner J, Gross S, Vissi E, White-Coper H, Alphey L (2004)

The essential role of PP1β in Drosophila is to regulate nonmuscle myosin. Mol Biol Cell 15:

4395-4405.

Verpelli C, Montani C, Vicidomini C, Heise C, Sala C (2013) Mutations of the synapse genes and

intellectual disability syndromes. Eur J Pharmacol 719:112-116.

Villano JL, Katz FN (1995) four-jointed is required for intermediate growth in the proximal-distal axis in

Drosophila. Development 121:2767-2777.

Vlachos S, Harden N (2011) Genetic evidence for antagonism between Pak protein kinase and Rho1

small GTPase signaling in regulation of the actin cytoskeleton during Drosophila oogenesis.

Genetics 187:501-512. von Kalm L, Fristrom D, Fristrom J (1995) The making of a fly leg: a model for epithelial

morphogenesis. Bioessays 17:693-702.

Wang Z, Edwards JG, Riley N, Provance DW, Jr., Karcher R, Li XD, Davison IG, Ikebe M, Mercer JA,

Kauer JA, Ehlers MD (2008) Myosin Vb mobilizes recycling endosomes and AMPA receptors

for postsynaptic plasticity. Cell 135:535-548.

Weiss A, Leinwand LA (1996) The mammalian myosin heavy chain gene family. Annu Rev Cell Dev

Biol 12:417-439.

130

Wells CM, Jones GE (2010) The emerging importance of group II PAKs. Biochem J 425:465-473.

Wirth A, Schroeter M, Kock-Hauser C, Manser E, Chalovich JM, De Lanerolle P, Pfitzer G (2003)

Inhibition of contraction and myosin light chain phosphorylation in guinea-pig smooth muscle by

p21-activated kinase 1. J Physiol 549:489-500.

Wylie SR, Chantler PD (2003) Myosin IIA drives neurite retraction. Mol Biol Cell 14:4654-4666.

Yang N, Higuchi O, Ohashi K, Nagata K, Wada A, Kangawa K, Nishida E, Mizuno K (1998) Cofilin

phosphorylation by LIM-kinase 1 and its role in Rac-mediated actin reorganization. Nature

393:809-812.

Yin JC, Wallach JS, Del Vecchio M, Wilder EL, Zhou H, Quinn WG, Tully T (1994) Induction of a

dominant negative CREB transgene specifically blocks long-term memory in Drosophila. Cell

79:49-58.

Zars T, Fisher M, Schulz R, Heisenberg M (2000) Localization of a short-term memory in Drosophila.

Science 288:672-675.

Zeidler MP, Perrimon N, Strutt DI (2000) Multiple roles for four-jointed in planar polarity and limb

patterning. Dev Biol 228:181-196.

Zenke FT, King CC, Bohl BP, Bokoch GM (1999) Identification of a central phosphorylation site in p21-

activated kinase regulating autoinhibition and kinase activity. J Biol Chem 274:32565-32573.

Zhang L, Luo J, Wan P, Wu J, Laski F, Chen J (2011) Regulation of cofilin phosphorylation and

asymmetry in collective cell migration during morphogenesis. Development 138:455-464.

Zhao L, Ma QL, Calon F, Harris-White ME, Yang F, Lim GP, Morihara T, Ubeda OJ, Ambegaokar S,

Hansen JE, Weisbart RH, Teter B, Frautschy SA, Cole GM (2006) Role of p21-activated kinase

pathway defects in the cognitive deficits of Alzheimer disease. Nat Neurosci 9:234-242.

Zhao ML, Wu CF (1997) Alterations in frequency coding and activity dependence of excitability in

cultured neurons of Drosophila memory mutants. J Neurosci 17:2187-2199.

Zhao ZS, Manser E, Chen XQ, Chong C, Leung T, Lim L (1998) A conserved negative regulatory region

in alphaPAK: inhibition of PAK kinases reveals their morphological roles downstream of Cdc42

131 and Rac1. Mol Cell Biol 18:2153-2163.

132