.3 .a
u
1

1
1;
.é

‘.

‘Rk

 

mum

. .-
.b‘c

 

 

LIBRARY
Michigan State
Univc . sity

 

 

 

This is to certify that the
dissertation entitled

STUDY OF TYROSINE KINASE LIKE PROTEIN KINASES: I.
A NOVEL ROLE OF MlXED-LINEAGE KINASE 3 IN
MITOCHONDRIA THROUGH ITS INTERACTION PROTEIN,
ADENINE NUCLEOTIDE TRANSLOCASE 2. II.
CHARACTERIZATION OF ROC DOMAIN OF PARKINSON'S
DISEASE-ASSOCIATED KINASE, LEUCINE RICH REPEAT
KINASE 2.

presented by

Geou-Yarh Liou

has been accepted towards fulfillment
of the requirements for the

Doctoral degree in Department of Biochemistry
and Molecular Biology

 

%2W4~ /°é%/A’ /

 

Major Professor’s Signature
1 2/ 1 0/2007

Date

MSU is an afiinnative-action, equal-opportunity employer

u-n---.-.-o-.-.-u-u---p---u-u-.-o--.-._.-,--.-.-.-..

 

PLACE IN RETURN BOX to remove this checkout from your record.
TO AVOID FINES return on or before date due.
MAY BE RECALLED with earlier due date if requested.

 

DATEDUE

DAIEDUE

DAIEDUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

6/07 p:/ClRC/DateDue.indd-p.1

 

 

STUDY OF TYROSINE KINASE LIKE PROTEIN KINASES:
I. A NOVEL ROLE FOR MIXED-LINEAGE KINASE 3 IN MITOCHONDRIA
THROUGH ITS INTERACTION PROTEIN, ADENINE NUCLEOTIDE
TRANSLOCASE 2.
II. CHARACTERIZATION OF ROC DOMAIN OF PARKINSON’S DISEASE-
ASSOCIATED KINASE, LEUCINE RICH REPEAT KINASE 2.

By

Geou-Yarh Liou

A DISSERTATION
Submitted to
Michigan State University

In partial fulfillment of the requirements
For the degree of

DOCTOR OF PHILOSOPHY
Department of Biochemistry and Molecular Biology

2007

ABSTRACT

STUDY OF TYROSINE KINASE LIKE PROTEIN KINASES:
I. A NOVEL ROLE FOR MIXED-LINEAGE KINASE 3 IN MITOCHONDRIA
THROUGH ITS INTERACTION PROTEIN, ADENINE NUCLEOTIDE
TRANSLOCASE 2.
II. CHARACTERIZATION OF ROC DOMAIN OF PARKINSON’S DISEASE-
' ASSOCIATED KINASE, LEUCINE RICH REPEAT KINASE 2.

By
Geou-Yarh Liou

Mixed Lineage Kinase 3 (MLK3) is a widely expressed mammalian mitogen
activated protein kinase kinase kinase (MAPKKK) that activates multiple mitogen-
activated protein kinase (MAPK) pathways. Since MLK3 is expressed at high levels in
human breast cancer MCF-7 cells, the role of MLK3 in breast cancer was investigated by
identifying its interacting proteins. Using immunoaffmity purification coupled with mass
spectrometry, adenine nucleotide translocase 2 (ANT2) was identified as one of the
components in the MLK3 protein complex.

ANT 2 is a mitochondrial ATP/ADP carrier protein and plays an important role in
regulating the cellular energetic balance. Co-immunoprecipitation experiments further
confirm the presence of expressed ANT 2 in MLK3 complexes. The kinase domain of
MLK3 was sufficient for MLK3-ANT2 interaction. In addition, ANT2 preferentially
interacted with the MLK3 mutants that have enhanced catalytic activity. Results from the
biochemical fractionation and confocal microspcopy have demonstrated MLK3 partially

localized to the mitochondria. Coexpression of MLK3 and ANT 2 increased total cellular

ATP level, whereas transiently expressing the kinase inactive MLK3 with AN T2
diminished it. Stable knockdown of MLK3 decreased total cellular ATP content. In
contrast, inducible expression of MLK3 enhanced total ATP production. Taken together,
these data open up the possibility that in addition to its role in MAPK pathway activation,
MLK3 regulation of cellular ATP levels may influence cell proliferation.

Emerging genetic studies have demonstrated that mutations in leucine-rich repeat
kinase (LRRK) 2 cause familial Parkinson’s disease. LRRK2 contains multiple predicted
catalytic and protein-interaction domains. Roc, a putative GTPase, always exists in
tandem with COR domain of unknown function. Since of solved GTPase structures, Roc
is most similar to the Rab7 GTPase, a molecular homology model of R00 was created
using the murine GTP-bound Rab7 crystal structure as a template to predict the
ftmctional consequence of PD mutation in the Roc. Roe and its predicted constitutively
active mutants, Roc A1396T as well as R1398L, specifically bound GTP, whereas Roc
T1348N, the predicted inactive form of Roc GTPase, failed to do so. Consistent with the
prediction from the Roc homology model, no apparent deficiency of GTP binding was
observed in the PD-associated variant, Roc R1441C. Using an enzymatic coupling assay,
weak GTPase activity of a recombinant Roc fusion protein was detected.

A Roc-COR fragment has been demonstrated to complex with the carboxyl
terminal region containing the Roc, COR, kinase domains and mm repeats of LRRK2.
The effect of the PD-associated mutations within the Roe-COR domains on this LRRK2
domain interaction was also studied. Data shown herein suggest that the Roc domain of
LRRK2 is a bonafide GTPase, and provide evidence that LRRK2 domains interact with

one other in ways that are altered by the introduction of PD-associated mutations.

ACKNOWLEDGEMENTS

First and most importantly, I would like to thank my advisor, Dr. Kathleen A.
Gallo for her guidance, support and encouragement during my doctoral years at MSU. I
do learn lots of lessons from her which will help me both in my science career and my
future life.

I would also like to thank the members of my committee, Dr. Jennifer Ekstrom,
Dr. Pamela Fraker, Dr. Katheryn Meek and Dr. Shelagh Ferguson-Miller for their
precious inputs, comments and suggestions for my search work. Especially, I want to
thank Dr. Jennifer Ekstrom for teaching me a lot for recombinant protein expression and
purification. In addition, I thank my former two committee members, Dr. Lee McIntosh

and Dr. Sara Courtneidge for their good opinions on my projects.

I especially thank the Gallo lab members, including Dr. Hua Zhang, Dr. Karen
Schachter, Dr. Yan Du, Dr. Julie Taylor, Eva Miller, J ian Chen and Meghann Devlin-
Durante for their friendships, collaborations, and invaluable discussionsas well as
providing a comfortable working environment. Especial thanks to Michelle Gilmer and

Steffany Kekstra for helping not only the lab work but also my communication skills.

I also thank people who work on the 3rd and 4th floor at Biomedical and Physical
Science Building and in the Biochemistry and Molecular Biology department for being a

continuous source of the reagents, equipments and advice.

Finally, I absolutely thank my family. My father and mother always encourage

me and support my decisions. My sister always cheers me up with different kinds of

iv

jokes. Big thanks to my brother and brother in law for their professional help on teaching

me the computer programs and dealing with any weird laptop troubles.

TABLE OF CONTENTS

Page
List of Tables ....................................................................................... ix
List of Figures ....................................................................................... x
Key to Abbrevivations ................................................................................ xiii
I. Literature Review .............................................................................. 1
1. Protein kinase classification, structure and function .................................. 1
2. GTPases ..................................................................................... 3
3. Mixed lineage kinases ..................................................................... 6
3.1 The MLK family members ........................................................... 6
3.1.1 MLK .............................................................................. 6
3.1.2 DLK .............................................................................. 8
3.1.3 ZAK .............................................................................. 10
3.1.4 MLK orthologs ................................................................. 10
3.2 MLK signaling activities ............................................................ 11
3.2.1 Activation of JNK pathway ................................................... 11
3.2.2 Activation of p38 pathway ................................................... 11
3.2.3 Activation of ERK pathway .................................................. 12
3.2.4 Activation of NF-KB ........................................................... 12
3.2.5 Other MLK substrates ......................................................... 13
3.3 Regulation of MLK ................................................................... 15
3.3.1 Regulation of MLKS by extracellular signals .............................. 15
3.3.2 Regulation of MLKS by autoinhibition ...................................... 17
3.3.3 Regulation of MLKS by dimerization ....................................... 18
3.3.4 Regulation of MLKS by small GTPases .................................... 20
3.3.5 Regulation of MLKS by scaffold proteins .................................. 21
3.3.6 Regulation of MLKS by chaperone proteins ................................ 24
3.3.7 Regulation of MLKS by phosphorylation .................................... 24
3.3.8 Regulation of MLKS by subcellular localization ........................... 26
3.4 The physiological roles of the MLKS ............................................. 27
3.4.1 The Drosophila MLK, slpr and Xenopus MLK ........................... 27
3.4.2 The mammalian MLKS ........................................................ 28
4. Leucine rich repeats kinases ............................................................... 30
4.1 LRRK family .......................................................................... 30
4.2 Physiological roles of LRRKS ...................................................... 31
4.3 LRRK signaling activity .............................................................. 33
4.3.1 LRRK in vitro kinase activity ................................................ 33
4.3.2 MAP kinases and NF-KB ...................................................... 34
4.4 Regulation of LRRK .................................................................. 35

vi

4.4.1 Regulation of LRRK by extracellular signals .............................. 35

4.4.2 Regulation of LRRK by GTPase ............................................. 37
4.4.3 Regulation of LRRK by chaperones ......................................... 38
4.4.4 Regulation of LRRK by subcellular localization ........................... 39
5. Mitochondria ................................................................................. 40
5.1 Mitochondria and human diseases .................................................. 40
5.2 Mitochondrial signaling .............................................................. 41
6. Objective of the thesis ...................................................................... 44
7. References ................................................................................... 46
II A Novel Role for MLK3 in Mitochondria through Its Interaction with ANT2. . . . . ....59
1. Abstract ...................................................................................... 59
2. Introduction ................................................................................. 61
3. Materials and Methods ...................................................................... 69
3.1 Reagents and antibodies .............................................................. 69
3.2 Plasmid constructs ..................................................................... 69
3.3 Cell culture and transfection ......................................................... 70
3.4 Construction of MLK3 stable knockdown cell line .............................. 70
3.5 Cell lysis and immunoprecipitation ................................................ 71
3.6 Gel electrophoresis and Western blot analysis .................................... 72
3.7 Identification of ANT peptides in the MLK3 immune complex by Mass
Spectrometry ........................................................................... 72
3.8 Immunofluorescence and confocal microscopy ................................... 72
3.9 Subcellular fractionation ............................................................. 73
3.10 Determination ofcellular ATP level74
4. Results ....................................................................................... 75
4.1 ANT was identified in affinity-associated MLK3 complexes .................. 75
4.2 Ectopically expressed ANT 1 and ANT 2 are targeted to the
Mitochondria .......................................................................... 75
4.3 MLK3 interacts with ANT2 through its kinase domain ........................ 79
4.4 The activation status of MLK3 regulates its association
with ANT 2 ............................................................................ 84
4.5 A portion of MLK3 is present in mitochondria .................................. 88
4.6 Mitochondrial MLK3 interacts with ANT 2 ........................................ 93
4.7 Effect of the MLK3-ANT2 interaction on total cellular
ATP level .............................................................................. 98
5. Discussion ................................................................................. 102
6. References ................................................................................. 1 11
III A Novel GTPase Domain of Parkinson’s Disease-Associated Kinase
LRRK2 .......................................................................................... 1 15
1. Abstract .................................................................................... 115
2. Introduction ................................................................................ 1 l7
3. Materials and Methods ................................................................... 124
3.1 Reagents and antibodies ............................................................ 124
3.2 Construction of expression vectors and mutagenesis .......................... 124

vii

3.3 Bacterial expression and purification of GST fusion proteins ................ 126

3.4 Cell culture, transfection and cell lysis .......................................... 127
3.5 Immunoprecipitation, gel electrophoresis and western
blot analysis ......................................................................... 128
3.6 GTP-binding assay .................................................................. 128
3.7 GTPase activity assay .............................................................. 129
4. Results ...................................................................................... 131
4.1 Recombinant GST-Roe variants, but not GST-Roc-COR are expressed as
soluble proteins under condition of auto-induction .............................. 131
4.2 Roc domain of LRRK2 binds GTP ................................................ 137
4.3 COR region affects the GTP binding of Roc .................................... 143
4.4 The GTP-binding specificity of Roc .............................................. 145
4.5 Hydrolysis of GTP by the Roc domain of LRRK2 .............................. 145
4.6 Interaction of ROC-COR region with the carboxyl terminal region of
LRRK2 ................................................................................. 145
4.7 Effect of PD mutations of Roe-COR on the protein interaction of
Roc-COR with RCKWE of LRRK2 .............................................. 151
5. Discussion ................................................................................. 156
6. References ................................................................................. 161
IV Concluding remarks ............................................................................ 165

viii

II.

III.

LIST of TABLES

Page
Literature Review
Table 1.1. Summary of the in vitro kinase activity of LRRK2 pathological
Mutations ........................................................................ 36

A Novel Role for MLK3 in Mitochondria through Its Interaction with AN T2

Table 2.1. Sequences of the ANT peptides identified by mass spectrometry and
their correspondence with ANT isoforrns .................................. 78

Table 2.2. Peptide sequences of the oxidative phosphorylation proteins identified by
mass Spectrometry ......................................................... 109

A Novel GTPase Domain of Parkinson’s Disease-Associated Kinase LRRK2

Table 3.1. Summary of the protein solubility of bacterial GST-Roe and GST-Roc-
COR variant proteins under various induced conditions ............... 139

ix

LIST OF FIGURES

Page
I. Literature Review

Figure 1.1. Diagram of the GTPases cycle ................................................. 4
Figure 1.2. Block digram structure of human mixed lineage kinases .................. 7
Figure 1.3. Schematic of MLK3 ............................................................. 9
Figure 1.4. Diagram of the activation of MAPK pathways by MLK3. . . . . . . . . ...14
Figure 1.5. Activation model of MLK3 .................................................... 19

II. A Novel Role for MLK3 in Mitochondria through Its Interaction with ANT2
Figure 2.1. Schematic of mitochondrial ATP synthesis ................................. 62
Figure 2.2. The architecture of the permeability transition pore complex. . . . . . ..64
Figure 2.3. ANT was identified in affinity-isolated MLK3 complexes ............... 77

Figure 2.4. Ectopically expressed ANTs are mainly present in mitochondria-
enriched fi'action ................................................................. 80
Figure 2.5. Ectopically expressed ANT 2 localized to the mitochondria .............. 81
Figure 2.6. ANT2, not ANTI, specifically interacts with MLK3 ..................... 83

Figure 2.7. The kinase domain of MLK3 is sufficient for the ANT 2 interaction. . .86

Figure 2.8. The regulation of MLK3 model and the diagram of the kinase activity of

different MLK3 mutations .................................................... 90
Figure 2.9. ANT2 preferentially associates with activated forms of MLK3 . . . . . .....92
Figure 2.10. A portion of expressed MLK3 is present in mitochondria .............. 94

Figure 2.11. A portion of endogenous MLK3 is present in a mitochondria-enriched
fraction ........................................................................ 96

Figure 2.12. Expressed MLK3 present in the mitochondria-enriched fraction
interacts with ANT 2 .......................................................... 97

Figure 2.13. Effect of MLK3 and ANT 2 on total cellular ATP level ................ 101

Figure 2.14. Alignment of the sequences containing the phosphorylation sites of
MLK3 in human MKK4 and MKK7 with human ANT isoforms. . ..107

Figure 2.15. N-Flag tagged MLK3 partially fractionates with mitochondria ....... 108

Figure 2.16. The proposed model of how the interaction of MLK3 with AN T2 in
elevation of the total cellular ATP level .................................. 110

III. A Novel GTPase Domain of Parkinson’s Disease Kinase-Associated LRRK2

Figure 3.1. Diagram of LRRK2 domains with mutations found in Parkinson’s
disease patients ................................................................ 118

Figure 3.2. Molecular model of GTPase domain of LRRK2 .......................... 121

Figure 3.3. Expression and solubility of GST-Roe variants and GST-Roc-COR
upon IPTG induction ........................................................... 134

Figure 3.4. Expressed GST-Roc variant proteins but not GST-Roc-COR are soluble
under the auto-induction condition ......................................... 136

Figure 3.5. Expression and solubility of GST-Roc-COR by IPTG induction with
coexpressed chaperones ..................................................... 138

Figure 3.6. Purification of recombinant GST-Roe variant proteins induced by auto-

induction condition ........................................................... 140
Figure 3.7. The GTP-binding ability of Roc variant proteins ........................ 142
Figure 3.8. The GTP-binding ability of Roe-COR variant proteins .................. 144
Figure 3.9. GTP-binding specificity of Roc variant proteins ......................... 146
Figure 3.10. The GTP hydrolysis ability of Roc ........................................ 147
Figure 3.11. Interaction of Roc-COR with different regions of LRRK2. . .. . . . . 1 50

Figure 3.12. Association of Roc with Roc-COR-kinase-WD 40 through the end of
LRRK2 ....................................................................... 152

xi

Figure 3.13. Effect of PD mutations found in the Roc and COR domains on their
interaction with the region of Roc-COR-kinase-WD 40 through the
end of LRRK2 .............................................................. 154

xii

AGC

AIF

Apaf- 1
Bsk
CAMK
CDK
CMBG
CNK1/2
COR
CREB
CRIB
CKl
COXIV
DLK
DMEM
EEAl
EGF
EGFR
ER

ERK

KEY TO ABBREVIATIONS

Protein kinase A, protein kinase G, protein kinase C families
Apoptosis inducing factor

A kinase anchor protein

Adenine nucleotide translocase
Apoptotic protease activating factor-1
Basket

Calcium/calmodulin protein kinase
Cyclin dependent potein kinase
CDK, MAPK, GSK3, CLK families
Connector enhancer of KSR1/2
C-terminal of Roc

c-AMP response element binding protein
Cdc42/Rac interactive binding
Casein kinase 1

Cytochrome c oxidase IV

Dual leucine zipper protein kinase
Dulbecco’s modified eagle’s medium
Early endosome antigen 1

Epidermal growth factor

Epidermal growth factor receptor
Endoplasmic reticulum

Extracellular signal regulated kinase

xiii

GAP
GEF
GluR6
GSK
GST
HA
HEK
Hep
HPKl
Hsp
IKB
IKK

J IP
JNK
KIF
KSR
LC-MS/MS
LDH
LRR
LRRK
LZK
MALDI-MS

MAPK

GTPase activating protein

Guanine exchange factor
Glutamate receptor 6
Glycogen synthase kinase
Glutathion S transferase
Hemagglutinin
Human embryonic kidney
Hemipterous
Hematopoietic progenitor kinase 1
Heat shock protein
Inhibitor of NF-KB
IKB kinase
JNK interacting protein
c-Jun N-terminal kinase
Kinesin superfamily motor protein
Kinase suppressor of Ras
Liquid chromatography-mass spectrometry/mass spectrometry
Lactate dehydrogenase
Leucine rich repeat
Leucine rich repeat kinase
Leucine zipper bearing kinase
Matrix-assisted laser desorption/ionization-mass spectrometry

Mitogen activated protein kinase

xiv

MAPKKK

MEF

MEK

MESG

MLK

MPTP

Msn

NGF

OXPHOS

PAS

PAGE

PD

PDGF

PI3K

PKA

PKC

PINK- 1

PMA

PNP

POSH

PSD-95

MAPK kinase kinase

Mouse embryonic fibroblast
Mitogen-activated protein/Erk kinase
2-amino-6-mercapto-7-methyl purine ribose
MAPK kinase

Mixed lineage kinase

Mitochondrial membrane perrneabilization
1 -methyl-4-phenyl-1 ,2,3 ,6-tetrahydropyridine
Misshapen

Nerve growth factor

Oxidative phosphorylation

Phospho-Akt substrate

Polyacrylarnide gel electrophoresis
Parkinson’s disease

Platelet derived growth factor
Phosphatidylinositol 3 kinase

protein kinase A

protein kinase C

PTEN induced kinase-1

Phorbol myristoyl acetate

Purine nucleotide phosphorylase

Plenty of SH3

Post-synaptic density protein 95

XV

PTPC

PTPDI

PTPMT— 1

RCKWE

RGC

RIPK

RNAi

Roc

ROS

SAM

SCG

SH3

ShRNA

SiRNA

Smac

STE

TAP

TCR

TGF

TK

TKL

TOR

Permeability transition pore complex

Protein tyrosine phosphatase 1

Phosphatase localized to the mitochondria
Roc-COR-kinase—WD40 through the end of LRRK2
Receptor guanylate cyclase

Receptor interacting protein kinase

RNA interference

Ras of complex protein

Reactive oxygen species

Sterile alpha motif

superior cervical ganglion

Src-homology 3

Short hairpin RNA

Small interference RNA

Second mitochondria-derived activator of apoptosis
homology of yeast sterile 7, 11, 20 kinases
Tandem affinity purification

T cell receptor
Transforming growth factor
Tyrosine kinase

Tyrosine kinase like

Tumor necrosis factor

Toll like receptor

xvi

TPA

TUNEL

VDAC

WE

ZAK

Phorbol ester 12-O—tetradecanoylphorbol-13-acetate
Terminal deoxynucleotidyl transferase mediated biotin-dUTP nick end
Labeling
Voltage dependent anion channel

WD40 through the end of LRRK2

Leucine zipper and sterile alpha motif kinase

xvii

I. Literature Review

1. Protein kinase classification, structure and function

Among many post-translational modifications, protein phosphorylation is best
known because of its broad involvement in cellular processes such as signal transduction,
transcription, metabolism, motility, cell cycle progression, proliferation, differentiation
and apoptosis. It is also pivotal in regulation of intercellular communication during

development and under various physiological conditions.

Protein phosphorylation is achieved by protein kinases which function to catalyze
the transfer of the yphosphate from ATP to protein substrates. Based on information
established in public and propriety genomic databases, 518 protein kinases have been
identified in the human genome, accounting for about 1.7% of all human genes [1]. The
human kinome is categorized into nine groups based on sequence similarity within the
catalytic domain. AGC (PKA, PKG and PKC families), CAMK (QlciunJ/leodulin
dependent protein kinases), CMGC (CDK, MAPK, _GSKB and _CLK families), TK
(tyrosine bnases), STE (homologs of yeast _Sfirile 7, 11, and 20 kinases), CKl (gasein
kinase _1), TKL (tyrosine kinase-like), RGC (receptor guanylate _c_yc1ase), and atypical
kinases [1]. Most mammalian protein kinases phosphorylate hydroxyl amino acids of the
substrate proteins. Therefore, they are also classified as Ser/Thr kinases, Tyr kinases, and
dual-specificity kinases which phosphorylate both Ser/Thr and Tyr residues [2, 3].
Mutations and dysregulation of protein kinases are tightly associated with various human
diseases including cancers and neurodegenerative diseases. Selective small molecule

inhibitors of protein kinases have proven to be valuable therapeutics. For instance,

Gleevec, which inhibits the oncogenic cytosolic tyrosine kinase Bcr-Abl is used to treat
chronic myelogenous leukemia; and receptor tyrosine kinase inhibitors such as the EGFR

family inhibitor, lapatinib, is approved for treatment of metastatic breast cancer.

Based on solved crystal structures of the catalytic domains, all protein kinases
are comprised of a small NH; terminal lobe and a large carboxyl terminal lobe [4]. An
active-site cleft where magnesium ion (Mgl’), ATP and protein substrate bind lies
between two lobes. The glycine loop or P loop with the Rossmann motif, GXGXXG, is
located at the top of the cleft and modulates the ATP binding through the interaction of
phosphates of ATP with backbone [5]. In the absence of ATP, the P-loop is unstructured.
A conserved Lys near the P-loop which plays a key role in kinase catalysis by properly
orientating the phosphates of ATP is stabilized by an invariant Glu residue through the
ionic interactions [6]. This conserved Lys is commonly mutated to create a catalytically
defective protein kinase that has been extensively applied in numerous research studies.
At the bottom of the cleft are the catalytic loop as well as the activation segment which is
a 20-35 residue sequence found between conserved motifs DF G and APE [7, 8]. Upon
phosphorylation of residues within the activation segment, the activation segment is
stabilized in an open and extended conformation, allowing substrate access to the

catalytic site [7, 9].

The substrate specificity of protein kinases is partially determined by the presence
of “consensus phosphorylation motif” in substrate proteins [5, 10] This has been widely
applied to identify new substrates of various protein kinases based on the results from in
vitro phosphorylation experiments with the synthetic peptides containing the recognition

motif of a given protein kinase. These consensus phosphorylation motifs are considered

as canonical phosphorylation sites. Docking sites and scaffolds are also important for

selection of substrates [11].

Besides the catalytic domain, most protein kinases also contain other domains to
regulate their functions in several manners, including controlling the subcellular
localization, binding activators or releasing inhibitors, associating with other proteins and
impacting the protein stability. These regulatory mechanisms are required for protein

kinase specificity in response to various extracellular stimuli.

2. GTPases

Small G proteins, also called GTPases are enzymes which hydrolyze GTP. GTP
hydrolysis allows the GTPase to function as a binary molecular switch to play crucial
roles in signal transduction and various cellular processes. In physiological settings,
GTPases are cycled between a GTP-bound (conformationally active) state and a GDP-
bound (conformationally inactive) state (Figure 1.1) with a conformational change.
Enzymatically active GTPases hydrolyze GTP with a weak intrinsic catalytic activity.
Only the GTP-bound conformation of GTPases which is conformationally active, bind
effector proteins, resulting in signal transduction processes. The GTPases are regulated
by guanine exchange factors (GEFS) and GTPase activating proteins (GAPS). GEFS
catalyze the release of GDP, allowing GTP to bind since the GTP concentration in cells is

higher than GDP. GAPS potentiate GTP hydrolysis to GDP, rendering GTPase inactive.

The Ras family of small GTPases is categorized into 5 subgroups, including Ras,

Rho, Ran, Rab and Arf subfamilies. Each subfamily is involved in different cellular

GAPs ‘ GEFs

GTP

Figure 1.1. Diagram of the GTPase cycle. GTPases cycling between a GDP-bound
state and a GTP-bound state is accomplished by a conformational change. Cycling of
GTPase is regulated by guanine exchange factors (GEFs) and GTPase activating proteins
(GAPS). GTP-bound (conformationally active) GTPases are competent to bind their

effector proteins.

processes. The Ras subfamily primarily participates in cell proliferation through
signaling to the Raf-MEK-ERK pathway. Ras is an oncogene and point mutations that
render Ras conformationally active including G12V, A59T and Q61L have been found in
the different types of human tumors [12]. It is estimated that 1/3 of solid human tumors
contain oncogenic Ras. Ran modulates the nucleo-cytoplasmic transport, spindle
assembly and post-mitotic nuclear envelope assembly [13]. The Rab and Arf subfamily
members are primarily involved in vesicular trafficking and cellular transports. Finally,
the Rho subfamily containing Cdc42, Rac and Rho mainly affects protein kinase

signaling and cytoskeletal remodeling [14].

A large body of work aimed at understanding the biological fimctions of GTPase
in different cellular processes as well as identifying their GEFS and GAPS is based largely
on overexpression of conformationally active and inactive mutants of GTPases. In the
paradigm GTPase, Ras, a substitution of Ser17 to Asn [15, 16] at the catalytic site renders
Ras GDP-bound and conformationally inactive. However, Gly12 to Val oncogenic
variant is conformationally active due to GTP binding but a failure of GTP hydrolysis
because of an inappropriately positioned magnesium ion [17, 18]. The conformationally
active Ras variants, G1n61 to Leu and A1359 to Thr, function in a mechanistically
analogous manner [17, 19]. Most small GTPases undergo post-translational prenylation,
with a farnesyl or geranylgeranyl group [20, 21]. These modifications render the
conformationally active GTPases hydrophobic and allow them to be targeted to cellular

membranes [22].

3. Mixed Lineage Kinases

Mixed lineage kinases (MLKS) belong to the serine/threonine kinase family and
are categorized in the tyrosine kinase like (TKL) branch of the human kinome. The
kinase domain of TKL family members share sequence similarity of the kinase domain
with both tyrosine kinases and serine/threonine kinases. Mixed lineage kinases are
known to function as MAPKKKS to activate MAPK signaling pathways. A large body of
work has demonstrated that MLKS induce apoptosis through the persistent activation of
JNK pathway in the neuronal system [23, 24]. Administration of CEP-1347, an MLK
inhibitor [25], in cultured cells and animal models rescued this apoptotic effect. Indeed,
CEP-l347 was tested in clinical trial for neuroprotection in Parkinson’s disease but

ultimately failed. Other biological functions of MLKS still need to be intensively studied.

3.1 The MLK family members

The mixed lineage kinases can be divided into three subgroups, including the
MLKS, the dual leucine zipper bearing kinases (DLKS) and the _z_ipper sterile-g motif
_lginases (ZAK) based on the sequence homology of the kinase domain and in the domain

arrangements (Figure 1.2)[26].

3.1.1 MLK

The MLK subfamily consists of MLK1 [27], MLK2 [28, 29], MLK3 [30-32] and

MLK4. MLK4 has two splicing forms, MLK4oz and MLK4/3. MLK1-4 have a NH;

SH3 kinase LZ CRIB

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

MLKI [ ET, , g 1 J fl I
1066
MLK2 I I f 1 I I I I
CW 954
MLK MLK3 [:1 L “I I I I I
mum ll an; J LI J 8"
mm l I r .r I ll 57"
1036
DLK [ L_ . 17 I? I
DLK [ 359
LZK I E '37 I:I I
SAM 9“
ZAK [ZAKa I l' '] l I I
ZAKI‘ [12221: 800

Figure 1.2. Block diagram structure of human mixed lineage kinases. The
relative position of the SH3 domain, leucine zipper (LZ), CRIB motif and sterile a motif
(SAM) is indicated (adapted from Gallo & Johnson, 2002)). Abbreviation, SH3: Src
homology 3, CRIB: Cdc42/Rac interactive binding.

terminal Src homology 3 (SH3) domain followed by a catalytic domain, a centrally
located leucine zipper region, a Cdc42/l_{acl interactive binding (CRIB) motif and a
proline rich carboxyl terminal region (Figure 1.3). MLK1-4 share a high sequence
identity, over 75% identity within the kinase domain and 60% identity from the SH3

domain through the CRIB motif.

The expression patterns of each MLK vary among MLK family members.
Messenger RNA of MLK1 is expressed in neuronal cells as well as epithelial cancer cell
lines of breast, colonic and esophageal origin [27]. Additionally, MLK] protein is
detected in an immature [3 cell line, RIN-5AH cells, but not in PIN-A12 cells, a more
mature 6 cell line, suggesting that MLK] play a role in regulation of B cell development
[27]. MLK2 mRNA is primarily expressed in brain, testis, skeletal muscle and at lower
levels in pancreas [28, 29]. MLK3 mRNA is ubiquitously expressed in human tissues,
but with a lower level in brain and heart [30-32]. To date, no expression pattern of

MLK4 has been reported.

3.1.2 DLK

The DLK subfamily contains DLK [33] and leucine-zipper _lginase (LZK) [34].
As compared to MLK subfamily, DLKs contain only a kinase domain, two small leucine
zipper motifs interrupted by 31 amino acids and a proline-rich carboxyl terminal region
with the undefined function. The kinase domain of DLK and LZK is 87% identical, but
shares only 35% sequence identity with MLK subfamily. DLK mRNA is predominantly

expressed in the adult mouse brain especially the hippocampus, cerebral cortex and the

Has Hagan 3325 3:9on

”EMU Emu—ow m 329:0: Em ”mzm .8983: 28 05:8 Bump—a2 EEmflv 05 E .2353: 2:. .mv—AE he uflafionom .m.— ensur—

Lanai

 
 

03 van mop ”v _.

Purkinje cells of the cerebellum [33]. Messenger RNA of LZK is widely expressed, with

the highest level in the pancreas [34].

3.1.3 ZAK

The ZAK subfamily only comprises ZAK (also known as MLTK) which has two
splicing isoforrns as ZAKa and ZAK/3 [35]. In addition to the kinase domain and a
leucine zipper region, a sterile-a motif (SAM) is also present in ZAKoz isoform. In
contrast, ZAKB does not possess the SAM structure due to the splicing process. It has
been demonstrated that SAM of the MAPKKK Stell mediates homo-oligomerization in
yeast [36]. ZAK mRNA is ubiquitously expressed in human tissues, with 3 highest level

in heart and skeletal muscle [35, 37].

3.1.4 MLK orthologs

MLKS are absent in yeast, but MLK orthologs are present in C. elegans,
Drosophila and Xenopus. The single ortholog of MLK subfamily in Drosophila is called
Slipper (Slpr). Disruption of Slpr which blocks cell movement during the process of
dorsal closure in the fly embryo is lethal [3 8]. The Xenopus MLK ortholog, which is
most similar to MLK2 with a 62% sequence identity, is mainly expressed in cement
gland, brain and pronephros [39]. Based on studies of antisense inactivation and
introduction of dominant negative constructs, the requirement of Xenopus MLK in
normal cement gland development and pronephros tubule formation has been
demonstrated[39]. Analysis of the amino acid sequences within kinase domain and

zipper regions of DLK suggests there is a DLK ortholog in Drosophila and C. elegans.

10

In addition, C. elegans also contains a ZAK ortholog and a distantly related SH3

containing MLK without other recognized structural domains.

3.2 MLK signaling activities

3.2.1 Activation of JNK pathway

All MLK family members, except MLK] and MLK4, have been demonstrated to
activate JNK pathway upon their overexpression in mammalian cells [35 , 40-42]. Both
MLK2 and MLK3 phosphorylate MKK4 and MKK7 in the in vitro kinase assays [43],
whereas DLK only phosphorylates MKK7 under the same conditions [44], suggesting
that DLK may activate JNK specifically through activation of MKK7 in vivo.
Ectopically expressed ZAK promotes the catalytic activity of MKK7 as well as MKK4 in
vitro [35]. However, only dominant negative MKK7 but not MKK4 is able to suppress
ZAK mediated JNK activation [45], suggesting that MKK7 might be the physiological

substrate of ZAK.

3.2.2 Activation of p38 pathway

Activation of the p38 pathway through phosphorylation of MKK3 and MKK6 is
also observed upon ectopic mammalian expression of MLK2 [40, 46], MLK3 [47], DLK
[41] and ZAKa [35, 37]. Upon TGF-B treatment, p38 is activated by endogenous MLK3
in hepatoma cells probably through an indirect mechanism since this activation occurs in
a very late stage after stimulation [48]. Knockdown MLK3 by RNAi attenuates TNFa-
induced p38 activation in human CCD-18Co colorectal fibroblasts and J urkat T

lymphocytes [49]. In addition, treatment of MLK3 depleted mouse embryonic fibroblasts

11

with TNF-a partially reduces p38 activation [50], suggesting a role for MLK3 in p38

activation. .

3.2.3 Activation of ERK pathway

Whether ERK can be activated by all MLK family members is still an ongoing
debate. It has been reported that ERK is activated upon overexpression of MLK2 [46, 51]
and ZAK [35]. Overexpression of MLK3 results in ERK activation in MCF-7 breast
cancer cells as well as HEK 293T cells (Gallo KA, unpublished data and [48]). On the
contrary, it has been demonstrated that ectopically expressed MLK3 impedes Phorbol
Myristoyl Acetate (PMA) and Platelet-Derived Growth Factor (PDGF)-induced ERK
activation in mesangial cells [52, 53]. Despite of the ability of MLK3 to phosphorylate
MEK] in vitro and in viva under certain circumstances, it has been proposed that
overexpressed MLK3 suppresses ERK activation because of the persistent activation of
JNK induced by MLK3 [54], suggesting negative cross-talk between the ERK and JNK

pathways when MLK3 is overexpressed.

A novel role for MLK3 in activation of ERK pathway has been proposed based on
the results of knockdown MLK3 by small interference RNA (SiRNA) in human and
murine cells (Figure 1.4). It has been demonstrated that knockdown of MLK3 impairs
ERK activation resulting in loss of cell proliferation in response to serum stimulation [49].
Furthermore, MLK3 was demonstrated to complex with B-Raf and Raf-1 and
introduction of the kinase defective MLK3 rescued the cell proliferation defect induced
by MLK3 SiRNA, suggesting that MLK3 functions as a scaffold protein that is required

for efficient signaling to ERK.

12

3.2.4 Activation of NF-KB

NF-KB is a transcription factor that participates in immune, inflammatory and
anti—apoptotic responses. Under normal condition, NF-KB is sequestered in the cytosol
through its interaction with the scaffold protein, IKB. Phosphorylation of IKB by
serine/threonine IKB kinases (IKKs) results in polyubiquitylation of IKB, targeting it to
the 26S proteasome for degradation thereby releasing NF-KB [55]. Free NF-KB
translocates to the nucleus where it activates transcription. In J urkat lymphocytes, T cell
receptor signaling can be induced by receptor cross linking with CD3/CD28 antibodies.
Expressed MLK3 induces phosphorylation and activation of IKKa and IKKB in Jurkat T
cells [56]. Expressed catalytically deficient MLK3 suppresses NF-KB-dependent
transcription upon CD3/CD28 costimulation. Similar to MLK3, LZK also activates and
phosphorylates IKKB through the interaction of its kinase domain with IKKB [57]. Both
LZK and ZAK have been demonstrated to activate NF-KB as judged by luciferase

reporter gene assays [42, 57].

3.2.5 Other MLK substrates

Most MLK substrates are protein kinases whose activation loops are
phosphorylated by MLKs, although MLK3 has been shown to phosphorylate golgin—16O
in vitro [58]. Golgin-16O is a peripheral membrane protein at the cytoplasmic face of the
Golgi complex [59]. In response to apoptotic stresses, cleavage of golgin-l60 by
caspases results in the disassembly of the Golgi complex during apoptosis [60].
Coexpression of MLK3 with golgin-160 reduces the gel mobility of golgin-160,

supporting the idea that MLK3 enhances golgin-16O phosphorylation in vivo [5 8].

l3

Stress, cytokines

. FFA I
Stimulus Growth factors / Cdc42,\R:IC
MLK3
MAPKKK Raf-1 MLKS MLK:
l l B ' Raf 1 l
MAPKK MKKI/Z MKK3/6 MKK4/7
MAPK Erkl/Z p38 JNK
Transcription Elk]. ATP - 2 C - Jun
factor
. . Growth AWPIOSiS
B'°'°9'c°' Development Inflammation
"3m" Differentiation Growth. differentiation

Figure 1.4. Diagram of the activation of MAPK pathways by MLK3. The best
characterized mammalian MAPK pathways are the ERK, JN K and p38 pathways.
Activation of MAPK is tightly regulated through a cascade which a MAPKKK
phosphorylates and activates a MAPKK which in turns phosphorylates and activates a
MAPK. Activation of different MAPK pathways in response to diverse stimuli leads to
different biological response. The diagram shows the activation of MAPK pathways by
MLK3 in response to different extracellular stimuli and the known biological effects of
MLK3-mediated MAPK activation. MLK3 is proposed to act catalytically as a
MAPKKK to activate the IN K and p38 pathways, and as a scaffold in B-Raf-mediated
ERK activation. Abbreviation, F FA: free fatty acid.

14

NeuroD is a transcription factor that belongs to the basic helix-loop-helix farnily
and plays a crucial role in neuronal development and survival in vertebrates [61].
NeuroD was identified as an MLK2-associating protein in a yeast two hybrid screen.
NeuroD and MLK2 can be coimmunoprecipitated from transfected neuroblastoma N2A
cells [62]. Purified recombinant MLK2 phosphorylates cellular NeuroD in vitro.
However, results fi'om in vitro kinase assay indicate that NeuroD is a poor substrate for
MLK2. It has been suggested that other cofactors or modifications of NeuroD in cells are
essential for its phosphorylation by MLK2. Coinj ection of NeuroD and MLK2 into

Xenopus embryos potentiates expressed NeuroD-mediated neurogenesis [62].

3.3 Regulation of MLK

3.3.1 Regulation of MLKS by extracellular signals

Although ectopically expressed MLKS have relatively high catalytic activities in
vitro, the extracellular signals that activate endogenous MLKS largely remain unknown.
The majority of MLK work has focused on MLK3, probably due to its broad tissue

expression as well as the availability of antibodies.

Early studies of the impact of extracellular signals on MLK3 rely on the alteration
of electrophoretic mobility in SDS-PAGE gels as a readout for MLK3 phosphorylation.
In Jurkat T lymphocytes, PMA treatment and co-stimulation with CD3/CD28 antibodies
induced an electrophoretic mobility shift of MLK3 [56], implying that PMA and
CD3/CD28 activate MLK3 in J urkat T lymphocytes. Treatment of HeLa cells with
sorbitol induces hyperosmotic stress and significantly alters the mobility of MLK3 in

SDS-PAGE (Karen Schachter, unpublished data). A phospho-specific antibody directed

15

against the posphopeptide containing phospho-Thr at 277 and phospho-Ser at 281 within
the activation loop of MLK3 has recently become commercially available. Using this
antibody, the activation loop phosphorylation of endogenous MLK3 has been detected in
A431 human epiderrnoid carcinoma cells treated with sorbitol [63]. In addition,
camptothecin, a topoisomerase inhibitor, induced activation loop phosphorylation of
endogenous MLK3 in HeLa cells as well as in PC12 neuronal cells [63]. These data

indicate that MLK3 plays a role in cellular responses in response to stress stimuli.

TNF-or stimulation has been shown to increase endogenous MLK3 catalytic
activity as judged by in vitro kinase assay in J urkat T lymphocytes [64] as well as in
MCF-7 breast cancer cells [65]. However, there is no significant change in the
electrophoretic mobility of MLK3 upon TNF-a treatment, suggesting that TNF-a induced
MLK3 phosphorylation does not result in alteration of the electrophoretic mobility of
MLK3. Lipid C6-ceramide treatment is reported to augment the endogenous MLK3 in
vitro kinase activity in J urkat T lymphocytes [64]. Ceramide is generated in response to
TNF-a because activation of sphingomyelinases converts sphingomyelin to ceramide and
phosphocholine. Ceramide potentiates apoptosis resulting from the permeabilization of
mitochondrial membranes [66]. The MLK inhibitor, CEP-11004, decreases TNF-or-
induced and ceramide-induced JNK activation [64]. Taken together, these data suggest in

response to TNF-a, MLK3 induces apoptosis.

Endogenous MLK3 is activated by TGF-B in rat hepatoma FatO cells and human
hepatoma Hep3B cells as judged by the reactivity with the phospho-MLK3 antibody [48].
However, MLK3 is activated by TGF-B after long time incubation, suggesting that

MLK3 is not directly activated through TGF-B receptor signaling. Expressed kinase

16

inactive MLKl, MLK2 or MLK3 abolishes TGF—6 induced apoptosis in these cells [48],

suggesting that the activation of MLKS is required for TGF-B mediated cell death.

Recent work has investigated the effect of grth factors on the MLK3 activation.
EGF increases the MLK3 catalytic activity in vitro as judged by the phosphorylation of
recombinant MEK] as the substrate in an in vitro kinase assay in immortal human colon
CCD-18Co cells [49]. Knockdown of MLK3 by small interference RNA eliminates
EGF-induced activation of ERK, JNK and P38 pathways, suggesting that MLK3 is

required for EGF induced signaling transduction.

Very recent work has linked MLKS to metabolic stress signaling pathways in
response to saturated fatty acids. Pahnitic acid (16:0) but not oleic acid (18:1) results in
activation of endogenous MLK3 and JNK as judged by the reactivity of phospho-MLK3
and phosphor-JNK antibodies in mouse embryonic fibroblasts and this activation is
abrogated in mlk3 konckout MEFs or upon treatment with PKC inhibitor[67]. These data

suggest that MLK3 participates in saturated fatty acid induced signaling pathways.

The mechanisms by which extracellular signals regulate the activation of different

MLKS are still not completely understood.

3.3.2 Regulation of MLKS by autoinhibition

SH3 domains are independently folding domains of about 60 amino acids. The
structural domain of SH3 binds to the proline rich sequence, PXXP, flanked by basic
amino acids [68]. The MLK subfamily members possess an SH3 domain at their amino
terminus. Hematopoietic progenitor kinase 1 (HPK1)phosphorylates kinase defective

MLK3 in vitro through the interaction of its proline-rich region with the SH3 domain of

17

MLK3 [69]. However, no alteration in MLK3 activity was observed upon coexpression

of HPKl and MLK3 (Gallo KA, unpublished data).

The most famous mechanism for the SH3 domain to control the activity of protein
kinases is the autoinhibitory regulation of Src tyrosine kinase [70]. In addition, the SH3
domain of the Src tyrosine kinase can also be modulated by inter-protein interactions.
Work fi'om our lab suggests that MLK3 is kept in an inactive state by an interaction of its
SH3 domain with a single proline-containing sequence located between the leucine
zipper region and the CRIB motif (Figure 1.5)[71]. Disruption of this interaction by
mutation of either a conserved tyrosine within the SH3 or of the single proline results in
an increase of MLK3 catalytic activity both in vitro and in vivo. The single proline
containing sequence of MLK3 is well conserved in all MLK subfamily and MLK
orthologs in fruit flies and Afiican clawed frogs, suggesting that this SH3 domain

mediated autoinhibiton applies to all MLK subfamily proteins.

3.3.3 Regulation of MLKS by dimerization

All MLKS contain a leucine zipper region which is a structural motif that creates
attractive forces in parallel or helices. The hydrophobic coiled coil is primarily stabilized
by placing leucine residues that repeat every 7 amino acids. Leucine zippers commonly
mediate protein dimerization or oligomerization [72, 73]. It has been demonstrated that
the leucine zipper region of MLK3 mediates oligomerization [74]. Even though MLK3

can coimmunoprecipitate with DLK from the total cellular lysates , the isolated DLK

18

GTP

 

C/Kinase- 7 H”. l t' .
1 C - {_SglflEiNHz ow ac IVIty
COOH
zipper Pro CRIB
GTP
HzN ‘ COOH
Pr° CRIB
zipper
. Pro high activity
H2“ z'pper CRIB
COOH

Figure 1.5. Activation model of MLK3. The proposed regulation model of MLK3.
Top, MLK3 is autoinhibited through a protein interaction mediated by the SH3 domain
and a proline located between the zipper region and Cdc42/Rac interactive binding
(CRIB) motif. Bottom, GTP-bound Cdc42/Rac binds to the CRIB motif of MLK3 and
disrupts the MLK3 autoinhibition, allowing the dimerization of MLK3 though the zipper
region. Dimerization of MLK3 brings two kinase domains of MLK3 close to each other,
resulting in the transphosphorylation of the activation loops of MLK3 and the fully active
form of MLK3.

leucine zipper region does not dimerize with other MLKS efficiently [75], suggesting that

heterodirnerization of MLKS is unlikely mediated by this region.

The leucine zipper region of MLK3 is necessary for MLK3 oligomerization [76].
Elimination of the entire leucine zipper region in MLK3 renders MLK3 catalytically
inactive and fails to activate JNK pathway [74]. Destabilization of the leucine zipper
region of MLK3 by introducing the mutation, L410P, impedes MLK3 dimerization and
its downstream signaling to JNK due to its inability to dually phosphorylate MKK4 [7 7].
Similar to MLK3, the leucine zipper mediation of DLK is necessary for DLK activation
and subsequent JNK activation. Induced dimerization of DLK by an artificial FK-
binding protein (FKBP) system potentiates DLK autophosphorylation and JNK activation
[78]. Ectopically expressed leucine zipper region of DLK, which competes for the
binding of the leucine zipper region between the full length DLKs, results in suppression
of DLK activation [78]. LZK has also been demonstrated to form oligomers. LZK with

a deleted leucine zipper region fails to activate JNK pathway [79].

In conclusion, leucine zipper region mediated dimerization is commonly used in
MLKS to regulate their activity and downstream pathways possibly by controlling their

autophosphorylation.

3.3.4 Regulation of MLKS by small GTPases

Small GTPases, also called G proteins, function as molecular switches in the
regulation of various cellular processes such as cell proliferation, cell movement,
cytoskeletal rearrangement and vesicular trafficking [80]. Under physiological

conditions, their cycling between GTP-bound and GDP-bound states is controlled by

20

guanine exchange factors (GEFs) and GTPase activating proteins (GAPS). Only GTP-
bound GTPases which are conformationally active bind to their effectors, resulting in
signal transduction. It has been demonstrated that conformationally active Rae and

Cdc42 activate the JNK and p38 pathways [81, 82].

Among MLKs, only MLK1-4 contain a CRIB GTPase binding motif. The CRIB
motif is composed of 14-16 amino acids containing eight conserved residues within the
motif which are critical for binding with activated Cdc42 and Rac [83]. It has been
shown that MLK3 binds to activated Rae and Cdc42 [76, 83, 84]. Knockdown of MLK3
by small interference RNA blocks ectopically expressed Cdc42 mediated JNK activation
in HEK 293T cells[85], further strengthening the notion that Cdc42 is an upstream
activator of MLK3. Coexpression of MLK3 with activated Cdc42 enhances MLK3
oligomerization, increases its catalytic activity, potentiates MLK3 -mediated JNK
activation and increases the in vivo phosphorylation of MLK3, as judged by the retarded
motility of MLK3 and in viva labeling [74, 76, 78, 86]. Additionally, activated
prenylated Cdc42 translocates MLK3 to cellular membrane compartments resulting in

full activation of MLK3 [85].

Coexpression of MLK3 with a constitutively active form of Racl increases
MLK3 catalytic activity in vitro [87]. Overexpression of dominant negative MLK3 or
DLK diminishes activated Racl mediated neuronal apoptosis[24], suggesting that Racl

induces neuronal apoptosis through activation of MLK family protein kinases.

21

3.3.5 Regulation of MLKS by scaffold proteins

JNK interacting proteins (J IPs) 1-4 are a group of scaffold proteins encoded from
four genes. J IPs are classified in 2 subgroups based on their sequence similarity. All
JIPs have common biochemical properties including a cargo function for kinesin motor
proteins and a scaffold firnction for JNK and P38 [88]. Other important signaling
molecules including the guanine nucleotide exchange factor (GEF), Tiarn, protein kinases,
Akt and Src, as well as MAPK phosphatase 7 (MKP7) have also been revealed to

associate with JIPs [88-90].

J 1P1 and JIP2 are structurally similar. J [P1 mainly participates in JNK signaling
and complexes with JNKs, MKK7 and several MLKS including MLK2, MLK3, DLK and
LZK. Coexpression of J IPl with MLK3 promotes JNK activation, suggesting that
scaffolds facilitate signaling by assembling the cascade module and ensuring the
signaling specificity. However, a different model of JIPl-mediated DLK and JNK
activation has also been suggested. DLK is sequestered in the monomeric, inactive state
by binding to J IPl [78, 91]. Upon stimulation with extracellular signals, JNK is recruited
to J [P] and phosphorylates J [P1 , allowing for DLK dissociation from J 1P1. Released
DLK dimerizes and subsequently becomes catalytically active resulting in
phosphorylation and activation of MKK7 and JNK in the JIPl complex [78, 91]. The
binding specificity of HP] and J 1P2 with other proteins is primarily controlled by the
binding affinity. J IPZ is mainly involved in p38 signaling. J [P2 associates with MLK3,
MKK3 and p38 isoforms [92, 93]. J LPZ also complexes with Tiaml and Ras-GRF, which
are the GEFs of Rac GTPase. Expression of either of these GEFs increases the

association of JIP2 with MLK3, MKK3 and p38 [92].

22

The other subgroup of J IPs is composed of JIP3 and JIP4. J IP3 interacts with
JNK, MKK7 and MLK family members. J 1P3 also complexes with MEKKI, MKK4 and
ASKl [94, 95]. JIP4 associates with MEKK3 and MKK4 [96]. Since all JIPs associate
with kinesin motor complexes, it is conceivable that JIP proteins may be involved in
kinesin-mediated axonal transport and regulate the subcellular localization of the JNK

signalsome.

In addition to J [P5, plenty gf SH3 (POSH) has been demonstrated to function as a
scaffold protein in JNK-induced neuronal apoptosis [97]. POSH interacts directly with
activated Racl, MLKS including MLK1-3 and DLK, and indirectly with MKK4/7 and
JNKs [23]. The neuronal apoptosis induced by overexpressed POSH can be rescued by
either dominant negative protein kinases in the POSH complex or MLK inhibitor, CEP-
1347 [23]. Binding of Akt2 to POSH disrupts the POSH-MLK-MKK-JNK complex and
impairs activation of JNK pathway [98]. Very recent data has demonstrated that POSH

interacts directly with J IP proteins, forming a signalsome [99].

Post-synaptic density protein 9_5 (PSD-95) is a neuronal scaffold protein
comprised of three PDZ domains, a SH3 domain and a catalytically inactive guanylate
kinase domain [100]. PSD-95 forms a complex with glgtamate receptor 6 (GluR6) and
MLK2 and/or MLK3 upon the interaction of its PDZ domain with GluR6 and of the SH3
domain with MLK2 and/or MLK3. Expressed kinase dead MLK2 or MLK3 abolishes
GluR6 mediated JNK activation and subsequent neuronal toxicity. Taken together, PSD-
95 functions as a scaffold protein to mediate signaling from ion channels to JNK pathway

by anchoring MLK family members to the GluR6 receptor complex [101].

23

Connector eghancer of KSR (CNK) was originally identified in a genetic screen
for genes which enhance the phenotype of flies with the overexpressed kinase domain of
kinase suppressor of _Ras (KSR) [102]. There are three isoforms of CNK: CNKl, CNK2
and CNK3. CNKl interacts with Raf-1 and c-Src, resulting in potentiation of Raf-1
activation [103]. Upon activation of the GTPase Rho, CNKl interacts with members of
the JNK pathway, including MLK2, MLK3 and MKK7, leading to c-Jun phosphorylation
[104]. Taken together, these data have suggested that CNKl fimctions as a scaffold

protein and may mediate cross talk among signaling pathways.

3.3.6 Regulation of MLKS by chaperone proteins

The function of chaperones is to maintain proteins in an appropriate folded
conformation, preventing aggregation. Hsp90 is one of the most abundant and highly
conserved proteins in eukaryotes. The co-chaperone Cdc37 is essential for the interaction
of Hsp90 with its client protein kinases, which occurs through their kinase domains. The
client proteins of Hsp90 include steroid hormone receptors and various protein kinases
including Src, Raf and CDK4 [105]. MLK3 associates with Hsp90/Cdc37 through its
kinase domain but this interaction is independent to the activation status of MLK3 [65].
Ablation of the ATPase activity of Hsp90 which is required for its chaperone function
reduces endogenous MLK3 protein levels [65], suggesting that MLK3 is stabilized with a

proper folded conformation which is maintained by Hsp90/Cdc37.

24

3.3.7 Regulation of MLKS by phosphorylation

Phosphorylation is the most common posttranslational modification that regulates
protein kinase activities. Phosphorylation of the activation loop which is flanked by the
conserved trinucleotide motifs, DFG and APE, is prerequisite for activation of most
protein kinases. Based on site—directed mutagenesis studies, Thr 277 and Ser 281 are the
phosphorylation sites within the MLK3 activation loop [106]. The commercially
available phospho-MLK3 antibody has been raised to recognize the phospho-Thr 277 and
phospho-Ser 281 of MLK3 and widely used to study MLK3 activity under various

physiological conditions.

Insulin stimulation activates phosphatidylinositol 3-kinase (PI3K), which in turns
activates protein kinase B, known as Akt, resulting in cell proliferation and survival. It
has been demonstrated that Akt is activated upon treatment of insulin in HepG2 liver
tumor cells. However, the catalytic activities of MLK3, MKK7 and JNK are reduced. It
has been suggested that Akt phosphorylates Ser 674 of MLK3, negatively regulating

MLK3 activity and JNK activation to block MLK3-mediated apoptotic effect [107].

Eleven serines located in the proline-rich carboxyl terminal region of MLK3 have
been identified as Cdc42 activated in viva phosphorylation sites of MLK3 by mass
spectrometry coupled with phosphopeptide mapping analysis [86]. Seven of them are
proline directed kinase sites, which appear to be phosphorylated by JNK through a
positive feedback loop mechanism. Ablation of MLK3 phosphorylation mediated by
JNK reduces MLK3 activity, as judged by activation loop phosphorylation, and
redistributes MLK3 to a Triton-insoluble fraction. These data have suggested that

phosphorylation of the proline-rich carboxyl terminus of MLK3 by JNK maintains

25

MLK3 in the Triton soluble compartments where MLK3 signaling pathway is

constitutively activated through positive feedback regulation [108].

3.3.8 Regulation of MLKS by subcellular localization

Signaling specificity can also be achieved by regulation of the subcellular
localization, allowing the modulation of the substrate accessibility. Studies of
immunohistochemical staining indicate that endogenous DLK localizes to the Golgi
apparatus. Upon biochemical fractionation, DLK is present in a crude membrane-
enriched fi'action and can only be solubilized with Triton-X114, indicating that DLK is
associated with the cytosolic face of the Golgi membranes [109]. Similar to DLK, MLK3
has been shown to be located at the perinuclear region [85] and colocalized with the
Golgi marker [58]. Introduction of catalytic inactive MLK3 does not change its
localization upon the biochemical fractionation [85], suggesting that the presence of
MLK3 in the perinuclear region/Golgi is not regulated by MLK3 activity. Expressed
MLK3 is targeted to cellular membranes by prenylated, activated Cdc42, resulting in full
activation of MLK3 [85]. MLK3 has been reported to locate at the centrosome with the
retarded gel motility during G2/M phase [110], suggesting that MLK3 is phosphorylated
at that stage. As mentioned previously, inhibition of JNK-mediated MLK3
phosphorylation within its proline-rich carboxyl terminus redistributes MLK3 to a Triton

insoluble fraction and attenuates MLK3 activity [108].

Expressed MLK2 localizes along microtubular structures with a punctate staining
pattern [51]. Furthermore, kinesin superfamily proteins (KIFs) have been identified as

MLK2-interacting proteins by yeast two hybrid assays [51]. KIFs are motor proteins

26

mainly involving in vesicular trafficking. MLK2 and MLK3 interact with KIF3 and the
putative cargo recognition protein, KAP3 [51]. Taken together, these data have

suggested that MLK2 and MLK3 might participate in regulation of vesicular trafficking.

3.4 The physiological roles of the MLKS

The function of MLKS is tightly associated with their upstream activators and
downstream effector pathways under different physiological conditions. A large body of
research work has demonstrated MLKS in development, cancer cell proliferation and

neuronal apoptosis

3.4.1 The Drosophila MLK, slpr and Xenopus MLK

The Drosophila MLK ortholog, slpr, is required for JNK mediated dorsal closure
during embryogenesis. During dorsal closure, the dorsal ectoderm moves from a lateral
position to the dorsal midline to cover the embryo in a continuous protective epidermis
[111]. GTPase dRacl, a MAP4K named Misshapen (Msn),Slpr, the MKK7 ortholog
Hemipterous (Hep), the JNK ortholog Basket (Bsk) and the AP-l transcription factors
containing dJun and dFos comprise the JNK pathway in Drosophila in guidance of dorsal
closure process [112]. Slpr mutant embryos have a dorsal open cuticle phenotype and

fail to maintain the stretched morphology [112].

The Xenopus MLK ortholog, most similar to MLK2 with a 62% sequence identity,
is expressed in the cement gland which is a mucus-secreting epithelium that forms from
the extreme anterior ectoderm during the early development [39]. The expression of

xMLK2 is associated with cell elongation and the apoptotic phase. at later stage,

27

expressed xMLK2 in the pronephros correlates with differentiation and opening of the
nephric tubules [39]. These data suggest that MLKS function to regulate the

differentiation of the epithelial layers.

3.4.2 The mammalian MLKS

Using RNAi-mediated MLK3 silencing, MLK3 has been connected to cell
proliferation in tumor cells [49]. Recent reports suggest that MLK3 complexes with B-
Raf and Raf-1 and is required for B-Raf- resulting in activation of ERK and promotion of
cell proliferation. Furthermore, the catalytic activity of MLK3 is not required for EGF-
induced ERK activation [113]. These data, suggest a scaffolding role, rather than
catalytic role for MLK3 in growth factor mediated cell proliferation. The mlk3 knockout
mice are viable without any obvious defects [50]. TNF-a induced JNK activation was
reduced two fold in mlk3 knockout mouse embryonic fibroblasts (MEFs) [50], suggesting
that MLK3 is essential for TNF-a mediated JNK activation and other MLKS may
particularly compensate for the lack of MLK3. The mlk3" MEFS retain intact MAPK
activation pathways in response to various growth factor stimulations [50], whereas
knockdown MLK3 by small interference RNA abolishes mitogen-induced and cytokines-
induced MAPK activations [49]. It is possible that the deficiencies in MAPK activation
are compensated by other MLKS during development. With transient knockdown by
RNAi, the cellular phenotype observed will reflect effects of acute MLK3 depletion.
Recently, the mutations of MLK4 have been found in clinical samples of gastric cancers,
hepatocellular carcinomas, and colorectal cancers in Asian population [114-116]. The

MLK4 gene mutation is present in 6.8% colorectal cancers and 50% mutations occur

28

within the kinase domain [114]. It has been proposed that MLK4 is a tumor suppressor

gene.

Recent work has placed MLKS in metabolic stress signaling pathway which is
important in insulin resistance and the development of type 2 diabetes. The saturated
fatty acid, pahnitic acid has been demonstrated to activate endogenous MLK3 in MEFs
which in turns activates MKK7, MKK4 and JNK, resulting in phosphorylation of insulin
receptor adaptor protein, IRS 1 , which causes the insulin resistance [67]. These
demonstrated effects are abolished in mlk3"', mkk4'l' and mkk7"' MEFs. Additionally,
treatment of MEFs with TPA, an inhibitor of all PKC isoforms, blocks MLK3 activation
induced by palrnitate, suggesting that PKC is the upstream activator of MLK3 [67].
Taken together, these data suggest that MLK3 activation in response to saturated fatty

acids could contribute to insulin resistance and diabetes.

Some studies have implicated the MLKs in JNK-mediated neuronal apoptosis
relying primarily on ectopic expression. It has been demonstrated that gerve growth
factor (NGF) deprivation induces apoptosis through the substantial activation of JNK
[117]. Expressed MLK2, MLK3 or DLK induces neuronal apoptosis in neuronal-like
PC12 cells [24]. Overexpression of MLK3 in superior cervical ganglion (SCG)
sympathetic neurons activates JNK and induces apoptosis, whereas expressed
catalytically inactive MLK3 abrogates apoptosis in response to NGF deprivation [118].
Treatment with the MLK inhibitor CEP-1347 protects the cultured rat motoneurons and
mice from neuronal apoptosis in response to various neurotoxins and neuronal stresses
[25, 119]. Taken together, these data suggest that MLKs are activated by various

neuronal stresses, resulting in potentiation of neuronal apoptosis through the persistent

29

activation of JNK. However, CEP-l347 failed to Show efficacy in treating PD in a phase

II/III clinical trial [120].

4. Leucine rich repeat kinases

Leucine rich repeat kinases (LRRKS) are members of the tyrosine kinase like
(TKL) subfamily. These sequences of the kinase domain of TKL family members are
similar to both tyrosine kinases and serine/threonine kinases. LRRKS are most similar to
MLKs and receptor interacting protein kinases (RIPKs) which are crucial sensors of
cellular stress connecting the extracellular stimuli to intracellular signal pathways. To
date, LRRK2 mutations have been tightly associated with familial Parkinson’s disease.

No biological functions of wild type LRRKS have been reported.

4.1 LRRK family

The human genome encodes two LRRKS, LRRKl and LRRK2. LRRKS are
extremely large proteins. LRRK] contains 2014 amino acids, giving a calculated
molecular weight around 225 kDa. LRRK2 has 2527 amino acids, resulting in a
predicted protein size of nearly 286 kDa. Based on amino acid sequence prediction,
LRRKl and LRRK2 share the same domain arrangements which are an amino terminal
ankyrin (Ank) domain, a leucine rich repeat (LRR) domain, a putative Roe GTPase
domain followed by a C—terminal gf Roe (COR) region, a protein kinase domain and a
carboxyl terminal WD-40 repeat domain [121]. The sequence identity within each

structural domain between LRRK] and LRRK2 is: 36% in the ankyrin repeat domain,

30

62% in the LRR domain, 47% in the Roc domain, 46% in the COR domain, 50% in the

kinase domain and 59% in the WD40 repeat domain [121].

In human tissues, LRRKl mRNA is primarily expressed in testis, with lower
levels found in brain, prostate, lymph node, placenta and pancreas [122]. LRRK2 mRNA
is highly expressed in lung, heart and putamen, at a lower level in substantial nigra [123,
124]. In mouse, both LRRKs are widely expressed in all brain regions. Murine LRRKl
mRNA is predominantly expressed in olfactory bulb and murine LRRK2 mRNA is
mainly expressed in striatum, cortex, cerebellum and brainstem, at a lower level in
hippocampus, midbrain and olfactory bulb [121]. In addition, mRNA of murine LRRK2
is also highly expressed in lung, spleen, and kidney tissues [125, 126]. However, murine
LRRK2 protein is only detected in kidney and brain perhaps due to the low sensitivity of

the used antibody [125].

4.2 Physiological roles of LRRKS

Mutations of LRRK2 have been segregated with familial Parkinson’s disease
(PD). PD is the most common neurodegenerative motor disorder in the western countries
affecting 1-2% of the population aged 65 and older. The clinical symptoms of PD
include resting tremor, bradykinesia (slowness of movement), muscle rigidity and
impaired balance. The loss of the dopaminergic neurons in the substantia nigra and the
presence of protein aggregates named Lewy bodies in the surviving neurons of the
brainstem are the pathological hallmarks of PD [127]. To date, all available treatments
are to ameliorate the clinical symptoms but do not cure the disease. Many environmental

or genetic factors have been demonstrated to increase the risk of PD. To date, the

31

etiology of PD is ambiguous. It has been generally thought to correlate with the
mitochondrial dysfunction, abnormal protein aggregation/degradation and defects in

vesicular trafficking [128].

LRRK2 mutations cause autosomal dominant late-onset Parkinsonism, and are
estimated to be responsible for about in 7% of familial PD and 04-16% in idiopathic PD
cases [123, 124]. At present, more than 20 LRRK2 mutations have been reported from
clinical genetic studies and these mutations are found in multiple LRRK2 structural
domains (review in [129]), suggesting that each domain of LRRK2 which has a different
predicted ftmction which is somehow involved in the etiology of PD. Pathological
examination of postmortem PD human brains that harbor LRRK2 mutations has revealed
the loss of dopaminergic neurons in the substantia nigra, as expected, but other
uncommon heterogeneous pathological features are sometimes present. While some
cases retain oz-synuclein positive Lewy body intracytoplasmic aggregates, other cases
either do not have Lewy body aggregates, instead displaying widespread Lewy body

pathology in the cerebral cortex or the presence of tau-positive axonal inclusions.

Interestingly, LRRK2 mutations within certain structural segments are prevalent
in populations with different ethnic backgrounds. G2019, the most prevalent LRRK2
mutation within the kinase domain, is predominantly responsible for familial and
sporadic PD in Arabs from North Africa; R1441C/G/H within the Roe domain accounts
for LRRK2-associated familial PD in Spanish population and G2385R of the WD40

repeats is the major familial PD mutation found in Asian populations.

Although LRRK2 is widely expressed in various tissues, very little is known

about LRRK2 function. Because LRRK2 mutations increase the risk of getting PD,

32

recent studies have intensively focused on the impact of pathological mutations of
LRRK2 on neuronal toxicity, as judged by the aberrant neuronal morphology, viability,
germinal deoxynucleotidyl transferase mediated biotin-dUTP pick grid labeling (TUNEL)
positive staining and caspase3 activation. Overexpression of kinase-defective LRRK2 in
rat primary cortical neurons was demonstrated to prolong the neurite length and
branching [130], suggesting a role of normal LRRK2 in controlling of neurite process.
All expressed LRRK2 pathological mutants diminish the neurite length and complexity,
with a most severe impairment in G2019S and IZOZOT-transfected neurons and a minor
defect in R1441G and Y1699C expressed neurons [130]. Knockdown of LRRK2 by
short hairpin RNA rescues this axonal deterioration mediated by LRRK2 mutations.
Furthermore, expressed LRRK2 pathological mutants including G2019S, R1441C,
Y1699C and G2358R enhanced caspase 3 activation, nuclear DNA fragmentation and
TUNEL positive stained cells [130-132], resulting in neuronal apoptosis. Overexpression
of the kinase-defective LRRK2 or mutants that disrupt GTP binding site of Roc
significantly increases neuronal viability [130-132], suggesting that activation of LRRK2

or Roc results in neuronal apoptosis.

4.3 LRRK signaling activity

4.3.1 LRRK in vitro kinase activity

Recent emerging biochemical studies have demonstrated that both LRRKl and
LRRK2 possess in vitro kinase activity as judged by autophosphorylation [122, 126, 131-
137]. In addition, LRRK2 has been shown to phosphorylate myelin basic protein (MBP),

myosin light chain, moesin, creatine kinase and pollapsing response rn_ediator protein

33

(CRMP) 2 in vitro [126, 130, 135-137]. Interesting, overexpressed LRRKl failed to
phosphorylate histone H1, casein and MBP which are commonly used as the exogenous
kinase substrates in vitro, suggesting that the ovexpressed LRRKl is not very active
[122]. Recently, LRRK2 purified from transgenic mouse brain has a higher in vitro

catalytic activity than that fi'om lung and transfected cultured cells [126].

Various amino acid substitutions in LRRK2 segregate with familial Parkinson’s
disease. These pathological mutations fall through all LRRK2 structural domains.
G2019S, the most prevalent mutation which is located in the kinase domain, slightly
increases LRRK2 catalytic activity in vitro [130-135, 137]. The impact of other LRRK2

pathological mutations on the in vitro kinase activity is still controversial (Table 1.1).

The conformationally active (GTP-bound) GTPases including Ras, Cdc42 and
Rac ([76, 83, 84]and data not shown) have been demonstrated to bind their effector
protein kinases, resulting in phosphorylation and activation of the protein kinases .
Similarly, incubation of non-hydrolysable GTP, GTP'yS, enhances LRRKl and LRRK2
in vitro autophsophorylation levels [122, 131, 132], suggesting that conformationally

active GTPases increase LRRKl and LRRK2 in vitro kinase activities.

4.3.2 MAP kinases and NF-xB

Mitogen activated protein (MAP) kinases basically include ERK, JNK and p38
protein kinases that participate in the regulation of cell growth, differentiation, immune
responses and apoptosis. The kinase sequence of LRRK2 is most closely relatedto that of
MLKs and of RIP kinases in the TKL subfamily of protein kinases. MLKs function as

MAPKKKS to activate MAP kinases and overexpressed MLKs also activate NF—KB in

34

response to immune stimulation. RIP kinases mediate extracellular immune stresses
through rumor pecrosis factor or (TNFa) receptor, rpll-like receptor (TOR) and I gell
receptor (TCR) to cellular signaling pathways, MAP kinase pathways and NF-KB (review
in [138]. In spite of the high sequence similarity of the LRRK2 kinase domain with
MLKs and RIP kinases, none of MAP kinase pathways including ERK, JNK, p38 and
ERK5 is directly activated by overexpressed LRRK2 in human embryonic kidney 293T
and human neuroblastoma SH-SY5Y cells (data not shown and [132]). Furthermore,
upon overexpression of LRRK2, no NF-KB activation was detected as judged by the

luciferase reporter gene assay (data not shown).

4.4 Regulation of LRRK

Several recent biochemical studies have mainly focused on the impacts of LRRK2
pathological mutations for regulation of the LRRK2 kinase activity and neuronal

apoptosis.

4.4.1 Regulation of LRRK by extracellular signals

Reactive oxygen species (ROS) are largely produced from the mitochondrial
respiratory chain in cells. In animal experiments, administration of 1-methyl-4-phenyl-
l,2,3,6-retrahydrop_yridine (MPTP), a neurotoxin that inhibits the electron transport chain
complex I of mitochondria which increases ROS level, induces symptoms similar to these
in human PD. Moreover, 25-30% inhibition of complex I is observed in the postmortem

samples of substantial nigra from PD patients. These data have revealed a key role of

35

Table 1.1. Summary of the in vitro kinase activity of LRRK2 pathological

 

 

 

 

 

 

 

 

 

 

 

 

mutations.
amino acid location kinase substrate8| kinase references
substitution (domain) activityb

11122V LRR LRRK2 +/- [132]
11371V Roc LRRK2 +/- [132]

R1441C Roc LRRK2, MBP + [132, 137]
LRRK2 - [136]
LRRK2, MBP +/- [136]
LRRK2* + [135]

R1441G Roe LRRK2 + [132, 137]
myosirli/Iigglht chain + “301
LRRK2* +/- [135]
R1514Q COR LRRK2 + [132]
Y1699C COR LRRK2 ++ [132]
LRRK2 - [136]
LRRKZ“ + [135]
12012T kinase LRRK2 - [132]
LRRK2* -- [135]

G2019S kinase LRRK2 [131, 113352], 134,

LRRK2“ [135]
myosirri/Illiaglht chain + [130]
MBP + [137]

12020T kinase LRRK2 -I—+ [132, 133]
LRRK2* +/- [135]
T2356I kinase LRRK2* +/- [135]
G2358R WD-4O LRRK2 +/- [132]
LRRK2* - [135]

 

 

 

 

 

 

a, the exogenous substrate used in the in vitro kinase assay.

b, in vitro kinase activity of LRRK2.

LRRKZ”: truncated LRRK2 with AA1326-2527.
Abbreviations: LRR, leucine rich repeat; COR, C-terminal of ROC; MBP, myelin basic protein.
+: increased in vitro kinase activity, -: decreased in vitro kinase activity, +/-: no difference in in
vitro lcinase activity as compared to wild type LRRK2

36

 

ROS in PD. Treatment of the LRRK2-overexpressed mouse primary cortical neurons
with hydrogen peroxide significantly induced neuronal death as judged by propidium-
iodide positive nuclei and formation of nuclear fragmentation [132]. Whereas kinase
defective LRRK2, D1994A, abolished hydrogen peroxide induced neuronal death,
implicating that activation of LRRK2 induced by hydrogen peroxide results in neuronal
death [132]. However, whether hydrogen peroxide modulates LRRK2 catalytic activity

has not been tested.

4.4.2 Regulation of LRRK by GTPase

GTPases are molecular switches that regulate numerous cellular processes
including activation of protein kinases in signal transduction. Roc, a putative GTPase
evolutionally distinct from Ras GPase superfamily, is always found in tandem with COR
of unknown function in the ROCO proteins. ROCO proteins widely exist in prokaryotes,

Dictyastelium, plants and metazoa but little is known about their functions.

LRRKl and LRRK2 are ROCO proteins. Both have been shown to be competent
to bind GTP [122, 126, 131, 132, 136, 139]. Disruption of the nucleotide binding pocket
of Roc domain in full length LRRKS abolished the GTP binding ability in vitro,
suggesting that GTP binds to the Roc domain of LRRKS [122, 126, 131, 132, 136, 139].
Purified murine LRRK2 from transgenic brain tissues has a 2-fold higher GTPase activity
than that from lung and transfected HEK 293T cells [126]. It is possible that the
interaction of LRRK2 with specific proteins within brain cells results in this higher
GTPase activity of LRRK2. Moreover, expression of Roc domain alone is capable to

bind and hydrolyze GTP, suggesting that Roc domain functions as a GTPase [126].

37

Addition of GTPyS, the non-hydrolysable GTP, in the LRRK kinase assays potentiates
the in vitro catalytic activity of LRRK about 2.5 fold [122, 131, 132]. Theses data
suggest that Roc domain of LRRKS functions as a GTPase to regulate LRRK kinase

activity in vitro.

Recent biochemical studies have mainly concentrated on the relationship of the
pathological LRRK2 mutations and its in vitro kinase activity mediated by Roc domain.
R1441C/G, the PD mutation within Roc domain, resides on the Roc surface which is far
way from the nucleotide catalytic pocket based on the Roc structural model and is
predicted to influence its protein interaction rather than the catalytic activity [129].
Recent results from other groups have demonstrated that this mutation decreases the GTP
hydrolysis efficiency of both Roc and LRRK2, but does not affect their GTP-binding
ability (unpublished data, [126, 136]). However, another contradictory report has
demonstrated the bacterially expressed Roc domain containing R1441C has a similar
GTPase activity as wild type [140]. Notably, there is a discrepancy from the published
studies of R1441C/G and R1699C, which are the PD-associated mutations within the Roc

and COR segments, in activation of LRRK2 in vitro (Table 1.1).

4.4.3 Regulation of LRRK by chaperones

Recently, the tandem affinity purification (TAP) procedure coupled with MALDI
mass spectrometry analysis was used to identify proteins interacting with the LRRK2
kinase domain and full length LRRK2 in transient transfected HEK 293T cells. Hsp90
and its co-chaperone, Cdc37, were found to complex with LRRK2 kinase domain and full

length LRRK2 under this condition [133]. Hsp90, one of the most abundant proteins in

38

eukaryotes, functions to assist protein folding and maintain the conformation of the client
proteins. Many protein kinases including Raf-1 [141], MLK3 [65]and Src [141] have
been demonstrated to interact with Hsp90 and Cdc37 through the catalytic domains of
protein kinases. In these cases, the catalytic activity of protein kinases is not required for
their interactions, suggesting that Hsp90 and Cdc37 are not substrates of protein kinases
but serve as molecular chaperones. It is conceivable that Hsp90 and Cd037 are required

as chaperones function for LRRK2 because of its tremendously large size.

4.4.4 Regulation of LRRK by subcellular localization

In general, LRRK2 seems to be present in the cytosol and also associated with the
membrane surface of various organelles. Upon biochemical fractionations, expressed
LRRK2 in HEK 293T cells appears in an organell-enriched fraction which includes
mitochondria, peroxisomes and endoplasmic reticulum (ER) and the microsomal
membrane-enriched fraction [133]. Upon treatment of the microsomal membrane
fraction with basic carbonate buffer which releases most peripheral proteins fi'om
membranes, expressed LRRK2 is recovered in the supernatant portion [133], suggesting
that LRRK2 associates with membranes rather than is integrated into membranes.
Meanwhile, in addition to the membranous organelles, mitochondria, Golgi and ER,
endogenous LRRK2 is also detected in the synaptic vesicle-enriched and synpatosomal
cytosolic fractions from rodent brain tissues [125]. Submitochondrial fractionation
experiments using purified mitochondria isolated from rodent brain have demonstrated
that endogenous LRRK2 primarily locates on the outer membrane and matrix of

mitochondria at a similar level [125]. It has been proposed that LRRK2 possibly binds

39

phospholipids through the LR and WD40 domains due to their positive charge surfaces

[129].

From analysis of irnmunofluorescent staining, expressed LRRK2 displays a
diffuse cytoplasmic distribution and partially colocalizes with mitochondria, ER, Golgi
and B-tubulin in HEK 293T cells [133]. In rat primary cortical neurons, endogenous
LRRK2 colocalizes with mitochondria, lysosomes, at a less level with microtubular B-
tubulin and with a negligible amount associated with synaptic vesicles[125].
Furthermore, in human and rat brain tissues, endogenous LRRK2 is ubiquitously
distributed to cerebral cortex, substantial nigra, caudate putamen, olfactory bulb and
brainstem. It also localizes at dopaminergic neurons with a punctate staining pattern
around the cell body, at a lower content in dendrites and axons [125]. Same distribution
of LRRK2 in the central nerve system has also been reported in LRRK2 transgenic
mouse brain specimen [126]. However, whether the catalytic activity of LRRK2 is
involved in regulation of its subcellular localizations and whether pathological mutations

of LRRK2 are associated with the alteration of subcellular localization are still unknown.

5. Mitochondria
5.1 Mitochondria and human diseases

Mitochondria are the major sites to produce ATP through oxidative
phosphorylation, and also serve as pivotal centers in determining programmed cell death.

Aberrant mitochondrial function is tightly associated with various human diseases

including cancer, cardiac disorders, diabetes and neurodegenerative disease [142-144].

40

Several lines of evidence have tightly linked the dysfunctional mitochondria to
neurodegenerative diseases such as Parkinson’s disease, Alzheimer disease, and
Huntington disease. Biochemical studies demonstrated that defects in enzymes of the
electron transport chain in postmortem human brain tissues of patients. The ganglia and
substantial nigra are vulnerable to the accumulation of age dependent mitochondrial
DNA deletions. Administration of MPTP, a neurotoxin which inhibits mitochondrial
complex I of the electron transport chain, in animals, the subsequent symptoms resemble
those observed in PD patients. In addition, the electron transport chain complex 11
inhibitor, 3-nitropropionic acid reconstitutes the clinical hallmarks of Huntington disease

in animal models.

Under aerobic condition, the majority of cellular ATP is from the oxidative
phosphorylation (OXPHOS) of mitochondria, with a slight level fi'om glycolysis. In
cancers, especially the solid tumor tissues which are under hypoxia environment, tumor
cells primarily rely on glycolysis for ATP production rather than through OXPHOS, as
known as the “Warburg effect” [145]. In this regard, suppression of glycolysis pathway
should arrest tumor progression. Consequently, drugs that inhibit glycolysis have been
applied in clinical cancer therapy to abate the growth of several carcinomas. For instance,
2-deoxy-D-glucose is in phase I clinical trial for lung, breast, pancreatic, gastric and head

and neck cancers (Clinical study, 2007).

5.2 Mitochondrial signaling

Recently, mitochondrial signaling has been attracted intensive attentions. Upon

development of the phospho-specific and phosphor-motif antibodies, many mitochondrial

41

proteins have been demonstrated as the phosphor-proteins. In addition to the metabolic
labeling with radioisopes, an invention of the pro-Q diamond dye for detecting the
steady-state phosporylated proteins in SDS-PAGE allows more than 60 mitochondrial
proteins from all mitochondrial compartments were identified (review in [146]).
Enormous emerging evidence has been reported that more than 40 well known kinases
and phosphatases which are involved in various signaling pathways and their downstream

transcription factors were found to impact the mitochondrial functions.

Many well known protein kinases that play crucial roles in cell survival and
tumorigenesis are also found in mitochondria, manipulating cellular energetics to
potentiate cell proliferation and protecting cells against stress-induced cell death.
Ectopically expressed, catalytic active Raf-l is located at the outer membrane of
mitochondria through its direct interaction with the BH4 domain of Bcl-2 which is a
mitochondrial protein [147]. Mitochondrial Raf—1 phosphorylates Bad and cooperates
with Bcl-2 to prevent staurosporine-induced apoptosis. Src tyrosine kinase is targeted to
the mitochondria upon forming a protein complex with c-AMP dependent protein kinase
(PKA), A-lrinase gnchor proteins (AKAP) 121 and protein tyrosine phosphatase 1
(PTPDl) [148]. AKAP121 increases Src-dependent phosphorylation of mitochondrial
substrates, activity of oxidative respiratory chain and ATP production, resulting in an
elevation of mitochondrial potential and the protective effect from apoptosis [148]. In
response to gpidermal grth factor (EGF) stimulation, tyrosine 845 of the expressed
EGFR is phosphorylated by overexpressed Src kinase, allowing to be translocated to
mitochondria where its associates with cytochrome c gridase subunit fl (COXH) [149].

Substitution of tyrosine 845 to phenylalanine sensitizes casapse-3 activation induced by

42

adriamycin, an anticancer drug commonly used in chemotherapy, in highly invasive
human MDA-MB-231 breast cancer cells under EGF stimulation. These data suggest that
translocation of EGFR to mitochondria interacting with COXII may positively control

cell survival pathways and enhance oncogenesis.

A large body of work has also demonstrated that numerous protein kinases
regulate mitochondrial proteins including Bcl-2 family proteins and components in
electron transport chain reaction as well as metabolism, resulting in promotion of cellular
apoptosis. JNK retranslocates to the mitochondria and phosphorylates Bel-XL at Thr 47
and 115 upon ionizing radiation [150]. Mutating these two phosphorylation sites in Bcl-
XL reduces ionizing radiation-induced apoptosis. These data suggest that the translocation
of JNK to mitochondria regulates Bel-XL, the antiapoptotic protein, through its kinase
activity, resulting in genotoxic induced apoptosis. Similarly, with phorbol ester 12-O-
tetradecanoylphorbol-l3-acetate (TPA) treatment, protein kinase C (PKC) 6 is
translocated to the mitochondria and induces apoptosis by releasing cytochrome c [151].
Expressed kinase inactive PKC6 fails to target to the mitochondria, suggesting that the
kniase activity of PKC6 is required for TPA-mediated PKC mitochondrial redistribution

to the mitochondria.

43

6. Objective of the thesis

Protein kinases are pivotal regulators in numerous cellular processes and the
aberrant regulation of protein kinases results in hmnan diseases. My thesis work has
primarily concentrated on understanding the role of MLK3 protein kinase in human
MCF-7 breast cancer cell line through its interacting protein, ANT 2. The other part of
my thesis is to characterize the Roc domain of LRRK2 whose mutations are segregated

with familial Parkinson’s disease.

As mentioned previously, MLK3 is widely expressed in most human tissues.
Survey of various human epithelial cell lines has revealed the highest expression of
MLK3 in breast cancer cell lines. To decipher the role of MLK3 in human breast cancers,
an inducible F lag-MLK3 human breast cancer MCF-7 cell line and the immunoaffinity
purification coupled with mass spectrometry were used to identify MLK3 interacting
proteins. Data in Chapter II demonstrates the role of MLK3 interaction with the
mitochondrial adenine nucleotide translocase 2 (ANT2) in regulation of cellular
energetics in a breast cancer cell line. This study also unveils the mechanism of MLK3

associating with ANT 2.

Various LRRK2 mutations cause the autosomal dominant late-onset familial
Parkinson’s disease with diverse pathological phenotypes. The pathological mechanisms
resulted from these mutations are obscure. In addition to the kinase domain, LRRK2 also
contains a putative GTPase domain, Roc, a COR region which is always in tandem with
Roe and several protein-interaction domains including ankyrin, LR and WD40. LRR2
mutations are found in all structural segments, suggesting that each domain is crucial in

PD etiology. Data in Chapter III reveals the Roc domain of LRRK2 is a banafide

44

GTPase. Introduction of R1441C/G, the PD-associated mutation within the Roc domain,
has no effect on the GTP binding of the Roc domain. We have previously hypothesized
that the PD-associated mutation, R1441C/G disrupts protein interaction rather impacts
the GTP binding and hydrolysis of the Roc domain. Another part of Chapter III has also
described studies of the Roe-COR domain-mediated protein interactions of LRRK2. In
addition, the effect of the PD-associated mutations within the Roe-COR region on the
RoccCOR mediated LRRK2 protein interaction has been assessed. In Chapter IV, the
results presented in this work are summarized and final remarks as well as future

directions are also included.

45

7. References

10.

ll.

12.

13.

Manning, G., et al., The protein kinase complement of the human genome. Science,
2002. 298(5600): p. 1912-34.

Dhanasekaran, N. and E. Premkumar Reddy, Signaling by dual specificity kinases.
Oncogene, 1998. 17(11 Reviews): p. 1447-55.

Lindberg, R.A., A.M. Quinn, and T. Hunter, Dual-specificity protein kinases: will
any hydroxyl do? Trends Biochem Sci, 1992. 17(3): p. 114-9.

Knighton, D.R., et al., Crystal structure of the catalytic subunit of cyclic adenosine
monaphasphate-dependent protein kinase. Science, 1991. 253(5018): p. 407-14.

Kemp, BE. and RB. Pearson, Protein kinase recognition sequence motifs. Trends
Biochem Sci, 1990. 15(9): p. 342-6.

Hanks, S.K., A.M. Quinn, and T. Hunter, The protein kinase family: conserved

features and deduced phylogeny of the catalytic domains. Science, 1988. 241(4861):
p. 42-52.

Huse, M. and J. Kuriyan, The conformational plasticity of protein kinases. Cell,
2002. 109(3): p. 275-82.

Morgan, DO. and H.L. De Bondt, Protein kinase regulation: insights from crystal
structure analysis. Curr Opin Cell Biol, 1994. 6(2): p. 239-46.

Johnson, L.N., M.E. Noble, and DJ. Owen, Active and inactive protein kinases:
structural basis for regulation. Cell, 1996. 85(2): p. 149-58.

Pearson, RB. and BE. Kemp, Protein kinase phosphorylation site sequences and
consensus specificity motifs: tabulations. Methods Enzymol, 1991. 200: p. 62-81.

Biondi, RM. and AR. Nebreda, Signalling specificity of Ser/Thr protein kinases
through docking-site-mediated interactions. Biochem J, 2003. 372(Pt l): p. 1-13.

Bos, J .L., ras oncogenes in human cancer: a review. Cancer Res, 1989. 49(17): p.
4682-9.

Joseph, J ., Ran at a glance. J Cell Sci, 2006. 119(Pt 17): p. 3481-4.

46

I4.

15.

l6.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

Rafiopoulou, M. and A. Hall, Cell migration: Rho GT Poses lead the way. Dev Biol,
2004. 265(1): p. 23-32.

Famsworth, CL. and LA. Feig, Dominant inhibitory mutations in the Mg(2+)-
binding site of RasH prevent its activation by GT P. Mol Cell Biol, 1991. 11(10): p.
4822-9.

F eig, LA. and GM. Cooper, Inhibition of NIH 3T3 cell proliferation by a mutant
ras protein with preferential affinity for GDP. Mol Cell Biol, 1988. 8(8): p. 3235-43.

Gibbs, J .B., et al., Intrinsic GT Pase activity distinguishes normal and oncogenic ras
p21 molecules. Proc Natl Acad Sci U S A, 1984. 81(18): p. 5704-8.

Manne, V., E. Bekesi, and HF. Kung, Ha-ras proteins exhibit GT Pase activity:
point mutations that activate Ha-ras gene products result in decreased GT Pase
activity. Proc Natl Acad Sci U S A, 1985. 82(2): p. 376-80.

Der, C.J., T. Finkel, and GM. Cooper, Biological and biochemical properties of
human rasH genes mutated at codon 61. Cell, 1986. 44(1): p. 167-76.

Seabra, M.C., Membrane association and targeting of prenylated Ros—like GT Pases.
Cell Signal, 1998. 10(3): p. 167-72.

Silvius, J .R., Mechanisms of Ras protein targeting in mammalian cells. J Membr
Biol, 2002. 190(2): p. 83-92.

Leung, K.F., R. Baron, and MC. Seabra, Thematic review series: lipid

posttranslational modifications. geranylgeranylatian of Rab GT Pases. J Lipid Res,
2006. 47(3): p. 467-75.

Xu, Z., N.V. Kukekov, and LA. Greene, POSH acts as a scaflaldfor a multiprotein
complex that mediates JNK activation in apoptosis. Embo J, 2003. 22(2): p. 252-61.

Xu, Z., et al., The MLK family mediates c-Jun N-terminal kinase activation in
neuronal apoptosis. Mol Cell Biol, 2001. 21(14): p. 4713-24.

Maroney, A.C., et al., Motoneuron apoptosis is blocked by CEP-134 7 (KT 7515), a
novel inhibitor of the JNK signaling pathway. J Neurosci, 1998. 18(1): p. 104-11.

Gallo, K.A. and G.L. Johnson, Mixed-lineage kinase control of JNK and p38 MAPK
pathways. Nat Rev Mol Cell Biol, 2002. 3(9): p. 663-72.

47

27.

28.

29.

30.

31.

32.

33.

34.

35.

36.

37.

Dorow, D.S., et al., Identification of a new family of human epithelial protein

kinases containing two leucine/isoleucine-zipper domains. Eur J Biochem, 1993.
213(2): p. 701-10.

Dorow, D.S., et al., Complete nucleotide sequence, expression, and chromosomal
localisation of human mixed-lineage kinase 2. Eur J Biochem, 1995. 234(2): p. 492-
500.

Katoh, M., et al., Cloning and characterization of MST, a novel (putative)
serine/threonine kinase with SH3 domain. Oncogene, 1995. 10(7): p. 1447-51.

Ezoe, K., et al., PT K I , a novel protein kinase required for proliferation of human
melanocytes. Oncogene, 1994. 9(3): p. 935-8.

Gallo, K.A., et al., Identification and characterization of SPRK, a novel src-

homology 3 domain-containing proline-rich kinase with serine/threanine kinase
activity. J Biol Chem, 1994. 269(21): p. 15092-100.

Ing, Y.L., et al., MLK-3: identification of a widely-expressed protein kinase bearing

an SH3 domain and a leucine zipper-basic region domain. Oncogene, 1994. 9(6): p.
1745-50.

Holzman, L.B., S.E. Merritt, and G. Fan, Identification, molecular cloning, and
characterization of dual leucine zipper bearing kinase. A novel serine/threanine

protein kinase that defines a second subfamily of mixed lineage kinases. J Biol
Chem, 1994. 269(49): p. 30808-17.

Sakuma, H., et al., Molecular cloning and functional expression of a cDNA

encoding a new member of mixed lineage protein kinasefi'om human brain. J Biol
Chem, 1997. 272(45): p. 28622-9.

Gotoh, I., M. Adachi, and E. Nishida, Identification and characterization of a novel
MAP kinase kinase kinase, MLT K. J Biol Chem, 2001. 276(6): p. 4276-86.

Bhattacharjya, S., et al., Polymerization of the SAM domain of MAPKKK Ste] 1
from the budding yeast: implications for eflicient signaling through the MAPK
cascades. Protein Sci, 2005. 14(3): p. 828-35.

Gross, EA, et al., MRK, a mixed lineage kinase-related molecule that plays a role
in gamma-radiation-induced cell cycle arrest. J Biol Chem, 2002. 277(16): p.
13873-82.

48

38.

39.

40.

41.

42.

43.

44.

45.

46.

47.

48.

Stronach, B. and N. Perrimon, Activation of the JNK pathway during dorsal closure
in Drosophila requires the mixed lineage kinase, slipper. Genes Dev, 2002. 16(3): p.
377-87.

Poitras, L., et al., A tissue restricted role for the Xenopus Jun N-terminal kinase

kinase kinase MLK2 in cement gland and pronephric tubule difi'erentiatian. Dev
Biol, 2003. 254(2): p. 200-14.

Cuenda, A. and D.S. Dorow, Diflerential activation of stress-activated protein
kinase kinases SKK4/MKK 7 and SKKI/MKK4 by the mixed-lineage kinase-2 and
mitogen-activated protein kinase kinase (MKK) kinase-1. Biochem J, 1998. 333 (Pt
1): p. 11-5.

Fan, G., et al., Dual leucine zipper-bearing kinase (DLK) activates p4 6SAPK and
p38mapk but not ERKZ. J Biol Chem, 1996. 271(40): p. 24788-93.

Liu, T.C., et al., Cloning and expression of ZAK, a mixed lineage kinase-like
protein containing a leucine-zipper and a sterile-alpha motif. Biochem Biophys Res
Commun, 2000. 274(3): p. 811-6.

Rana, A., et al., The mixed lineage kinase SPRKphospharylates and activates the
stress-activated protein kinase activator, SEK-I. J Biol Chem, 1996. 271(32): p.
19025-8.

Merritt, S.E., et al., The mixed lineage kinase DLK utilizes MK 7 and not MKK4 as
substrate. J Biol Chem, 1999. 274(15): p. 10195-202.

Yang, J .J ., Mixed lineage kinase ZAK utilizing MK 7 and not MKK4 to activate the
c-Jun N-terminal kinase and playing a role in the cell arrest. Biochem Biophys Res
Commun, 2002. 297(1): p. 105-10.

Hirai, S., et al., MST /MLK2, a member of the mixed lineage kinase family, directly
phosphorylates and activates SEKI, an activator of c-Jun N-terminal kinase/stress-
activated protein kinase. J Biol Chem, 1997. 272(24): p. 15167-73.

Tibbles, L.A., et al., MLK-3 activates the SAPK/flVK and p38/RX pathways via
SEKI and MKK3/6. Embo J, 1996. 15(24): p. 7026-35.

Kim, K.Y., et al., Mixed lineage kinase 3 (MLK3)-activatedp38 MAP kinase
mediates transforming growth factor-beta-induced apoptosis in hepatoma cells. J
Biol Chem, 2004. 279(28): p. 29478-84.

49

49.

50.

51.

52.

53.

54.

55.

56.

57.

58.

59.

60.

61.

Chadee, D.N. and J .M. Kyriakis, MLK3 is required for mitogen activation of B-Raf
ERK and cell proliferation. Nat Cell Biol, 2004. 6(8): p. 770-6.

Brancho, D., et al., Role of MLK3 in the regulation of mitogen-activated protein
kinase signaling cascades. Mol Cell Biol, 2005. 25(9): p. 3670-81.

Nagata, K., et al., Activation of G1 progression, JNK mitogen-activated protein
kinase, and actinfilament assembly by the exchange factor F GD] . J Biol Chem,
1998. 273(25): p. 15453-7.

Parameswaran, N., et al., Mixed lineage kinase 3 inhibits phorbol myristoyl acetate-

induced DNA synthesis but not osteopontin expression in rat mesangial cells. Mol
Cell Biochem, 2002. 241(1-2): p. 37-43.

Parameswaran, N., et al., Mixed lineage kinase 3 inhibits platelet-derived growth

factor-stimulated DNA synthesis and matrix mRNA expression in mesangial cells.
Cell Physiol Biochem, 2002. 12(5-6): p. 325-34.

Shen, Y.H., et al., Cross—talk between JNK/SAPK and ERK/MAPK pathways:
sustained activation of JNK blocks ERK activation by mitogenic factors. J Biol
Chem, 2003. 278(29): p. 26715-21.

Israel, A., The IKK complex: an integrator of all signals that activate NF -kappaB?
Trends Cell Biol, 2000. 10(4): p. 129-33.

Hehner, S.P., et al., Mixed-lineage kinase 3 delivers CD3/CD28-derived signals
into the IkappaB kinase complex. Mol Cell Biol, 2000. 20(7): p. 2556-68.

Masaki, M., et al., Mixed lineage kinase LZK and antioxidant protein-1 activate
NF-kappaB synergistically. Eur J Biochem, 2003. 270(1): p. 76-83.

Cha, H., et al., Phosphorylation of galgin-I 60 by mixed lineage kinase 3. J Cell Sci,
2004. 117(Pt 5): p. 751-60.

Hicks, SW. and CE. Machamer, Isoform-specific interaction of golgin-I 60 with
the Golgi-associated protein PIST. J Biol Chem, 2005. 280(32): p. 28944-51.

Mancini, M., et al., Caspase-Z is localized at the Golgi complex and cleaves golgin-
160 during apoptosis. J Cell Biol, 2000. 149(3): p. 603-12.

Lee, J .E., et al., Conversion of Xenopus ectoderm into neurons by NeuroD, a basic
helix-loop-helix protein. Science, 1995. 268(5212): p. 836-44.

50

62.

63.

64.

65.

66.

67.

68.

69.

70.

71.

72.

73.

74.

Marcora, E., K. Gowan, and J .E. Lee, Stimulation of NeuroD activity by huntingtin
and huntingtin-associated proteins HAP] and MLK2. Proc Natl Acad Sci U S A,
2003. 100(16): p. 9578-83.

Xu, Z., N.V. Kukekov, and LA. Greene, Regulation of apoptotic c-Jun N-terminal
kinase signaling by a stabilization-basedfeed-forward loop. Mol Cell Biol, 2005.
25(22): p. 9949-59.

Sathyanarayana, P., et al., Activation of the Drosophila MLK by ceramide reveals
TNF-alpha and ceramide as agonists of mammalian MLK3. Mol Cell, 2002. 10(6):
p. 1527-33.

Zhang, H., et al., Hsp90/p50cdc3 7 is required for mixed-lineage kinase (MLK) 3
signaling. J Biol Chem, 2004. 279(19): p. 19457-63.

Ding, W.X. and X.M. Yin, Dissection of the multiple mechanisms of TNF-alpha-
induced apoptosis in liver injury. J Cell Mol Med, 2004. 8(4): p. 445-54.

J aeschke, A. and R.J. Davis, Metabolic stress signaling mediated by mixed-lineage
kinases. Mol Cell, 2007. 27(3): p. 498-508.

Yu, H., et al., Structural basis for the binding of proline-rich peptides to SH3
domains. Cell, 1994. 76(5): p. 933-45.

Kiefer, F., et al., HPKI, a hematopoietic protein kinase activating the SAPK/JNK
pathway. Embo J, 1996. 15(24): p. 7013-25.

Hubbard, S.R., M. Mohammadi, and J. Schlessinger, Autoregulatory mechanisms in
protein-tyrosine kinases. J Biol Chem, 1998. 273(20): p. 11987-90.

Zhang, H. and K.A. Gallo, Autoinhibitian of mixed lineage kinase 3 through its Src
homologr 3 domain. J Biol Chem, 2001. 276(49): p. 45598-603.

Hu, J .C., et al., Sequence requirements for coiled-coils: analysis with lambda
repressor-GCN4 leucine zipper fusions. Science, 1990. 250(4986): p. 1400-3.

O'Shea, E.K., et al., X-ray structure of the GCN4 leucine zipper, a two-stranded,
parallel coiled coil. Science, 1991. 254(5031): p. 539-44.

Leung, I.W. and N. Lassarn, Dimerization via tandem leucine zippers is essential
for the activation of the mitogen-activated protein kinase kinase kinase, MLK-3. J
Biol Chem, 1998. 273(49): p. 32408-15.

51

75.

76.

77.

78.

79.

80.

81.

82.

83.

84.

85.

Nihalani, D., S. Merritt, and LB. Holzman, Identification of structural and
fimctional domains in mixed lineage kinase dual leucine zipper-bearing kinase

required for camplexformatian and stress-activated protein kinase activation. J
Biol Chem, 2000. 275(10): p. 7273-9.

Bock, B.C., et al., Cdc42-induced activation of the mixed-lineage kinase SPRK in
viva. Requirement of the Cdc42/Rae interactive binding motif and changes in
phosphorylation. J Biol Chem, 2000. 275(19): p. 14231-41.

Vacratsis, PD. and K.A. Gallo, Zipper-mediated oligomerization of the mixed
lineage kinase SPRK/MLK-3 is not required for its activation by the GT Pase cdc 42
but Is necessary for its activation of the JNK pathway. Monomeric SPRK L410P
does not catalyze the activating phosphorylation of Thr258 of murine MIT OGEN-
ACTIVATED protein kinase kinase 4. J Biol Chem, 2000. 275(36): p. 27893-900.

Nihalani, D., et al., Mixed lineage kinase-dependent JNK activation is governed by
interactions of scaffold protein JIP with MAPK module components. Embo J, 2001.
20(13): p. 3447-58.

Ikeda, A., et al., Identification and characterization of functional domains in a
mixed lineage kinase LZK. FEBS Lett, 2001. 488(3): p. 190-5.

Bar-Sagi, D. and A. Hall, Ras and Rho GTPases: a family reunion. Cell, 2000.
103(2): p. 227-38.

Coso, O.A., et al., The small GT P—binding proteins Rae] and Cdc42 regulate the
activity of the JNK/SAPK signaling pathway. Cell, 1995. 8 1 (7): p. 1137-46.

Minden, A., et al., Selective activation of the JNK signaling cascade and c-Jun
transcriptional activity by the small GT Pases Rae and Cdc42Hs. Cell, 1995. 81(7):
p. 1147-57.

Burbelo, P.D., D. Drechsel, and A. Hall, A conserved binding motif defines
numerous candidate target proteins for both Cdc42 and Rac GT Pases. J Biol Chem,
1995. 270(49): p. 29071-4.

Teramoto, H., et al., Signaling from the small GT P-binding proteins Rae] and
Cdc42 to the c-Jun N—terminal kinase/stress—activated protein kinase pathway. A
role for mixed lineage kinase 3/protein-tyrosine kinase 1, a novel member of the
mixed lineage kinase family. J Biol Chem, 1996. 271(44): p. 27225-8.

Du, Y., et al., Cdc42 induces activation loop phosphorylation and membrane
targeting of mixed lineage kinase 3. J Biol Chem, 2005. 280(52): p. 42984-93.

52

86.

87.

88.

89.

90.

91.

92.

93.

94.

95.

96.

97.

Vacratsis, P.O., et al., Identification of in viva phosphorylation sites of MLK3 by
mass spectrometry and phosphopeptide mapping. Biochemistry, 2002. 41(17): p.
5613-24.

Lambert, J.M., et al., Role of MLK3-mediated activation of p 70 S6 kinase in RacI
transformation. J Biol Chem, 2002. 277(7): p. 4770-7.

Morrison, D.K. and R.J. Davis, Regulation of MAP kinase signaling modules by
scaflold proteins in mammals. Annu Rev Cell Dev Biol, 2003. 19: p. 91-118.

Whitrnarsh, A.J., et al., A mammalian scaflold complex that selectively mediates
MAP kinase activation. Science, 1998. 281(5383): p. 1671-4.

Yasuda, J., et al., The JIP group of mitogen-activated protein kinase scaffold
proteins. Mol Cell Biol, 1999. 19(10): p. 7245-54.

Nihalani, D., H.N. Wong, and LB. Holzman, Recruitment of JNK to JIPI and JNK-
dependent JIPI phosphorylation regulates JNK module dynamics and activation. J
Biol Chem, 2003. 278(31): p. 28694-702.

Buchsbatun, R.J., B.A. Connolly, and LA. Feig, Interaction of Rac exchange
factors T iamI and Ras-GRFI with a scafloldfor the p38 mitogen-activated protein
kinase cascade. Mol Cell Biol, 2002. 22(12): p. 4073-85.

Schoorlemmer, J. and M. Goldfarb, Fibroblast growth factor homologous factors
are intracellular signaling proteins. Curr Biol, 2001. 11(10): p. 793-7.

Kelkar, N., et al., Interaction of a mitogen-activated protein kinase signaling
module with the neuronal protein JIP3. Mol Cell Biol, 2000. 20(3): p. 1030-43.

Matsuura, H., et al., Phosphorylation-dependent scaflolding role of JSAPI/JIP3 in
the ASKI-JNK signaling pathway. A new mode of regulation of the MAP kinase
cascade. J Biol Chem, 2002. 277(43): p. 40703-9.

Lee, C.M., et al., JLP: A scaflolding protein that tethers JNK/p38MAPK signaling
modules and transcription factors. Proc Natl Acad Sci U S A, 2002. 99(22): p.
141 89-94.

Tapon, N., et al., A new rac target POSH is an SH3-containing scaffold protein
involved in the JNK and NF -kappaB signalling pathways. Embo J, 1998. 17(5): p.
1395-404.

53

98.

99.

100.

101.

102.

103.

104.

105.

106.

107.

108.

109.

Figueroa, C., et al., Akt2 negatively regulates assembly of the POSH-MLK-JNK
signaling complex. J Biol Chem, 2003. 278(48): p. 47922-7.

Kukekov, N.V., Z. Xu, and LA. Greene, Direct interaction of the molecular
scaffolds POSH and JIP is required for apoptotic activation of JNKs. J Biol Chem,
2006. 281(22): p. 15517-24.

Garcia, E.P., et al., SAP90 binds and clusters kainate receptors causing incomplete
desensitization. Neuron, 1998. 21(4): p. 727-39.

Savinainen, A., et al., Kainate receptor activation induces mixed lineage kinase-
mediated cellular signaling cascades via post-synaptic density protein 95. J Biol
Chem, 2001. 276(14): p. 11382-6.

Kolch, W., Coordinating ERK/MAPK signalling through scaflalds and inhibitors.
Nat Rev Mol Cell Biol, 2005. 6(11): p. 827-37.

Ziogas, A., K. Moelling, and G. Radziwill, CNKI is a scaffold protein that
regulates Src-mediated Raf-1 activation. J Biol Chem, 2005. 280(25): p. 24205-11.

Jaffe, AB, A. Hall, and A. Schmidt, Association of CNKI with Rho guanine
nucleotide exchange factors controls signaling specificity downstream of Rho. Curr
Biol, 2005. 15(5): p. 405-12.

Pearl, L.H., Hsp90 and Cdc3 7 -- a chaperone cancer conspiracy. Curr Opin Genet
Dev, 2005. 15(1): p. 55-61.

Leung, I.W. and N. Lassarn, The kinase activation loop is the key to mixed lineage

kinase-3 activation via both autophosphorylation and hematopoietic progenitor
kinase 1 phosphorylation. J Biol Chem, 2001. 276(3): p. 1961-7.

Barthwal, M.K., et al., Negative regulation of mixed lineage kinase 3 by protein
kinase B/AKT leads to cell survival. J Biol Chem, 2003. 278(6): p. 3897-902.

Schachter, K.A., et al., Dynamic positive feedback phosphorylation of mixed
lineage kinase 3 by JNK reversibly regulates its distribution to T riton-soluble
domains. J Biol Chem, 2006. 281(28): p. 19134-44.

Douziech, M., et al., Localization of the mixed-lineage kinase DLK/MUK/ZPK to
the Golgi apparatus in NIH 3 T 3 cells. J Histochem Cytochem, 1999. 47(10): p.
1287-96.

54

110.

111.

112.

113.

114.

115.

116.

117.

118.

119.

120.

121.

Swenson, K.I., K.E. Winkler, and AR. Means, A new identity for MLK3 as an
NIMA-related, cell cycle-regulated kinase that is localized near centrasomes and
influences microtubule organization. Mol Biol Cell, 2003. 14(1): p. 156-72.

Noselli, S., flVK signaling and morphogenesis in Drosophila. Trends Genet, 1998.
14(1): p. 33-8.

Stronach, B., Dissecting JNK signaling, one KKKinase at a time. Dev Dyn, 2005.
232(3): p. 575-84.

Chadee, D.N., et al., Mixed-lineage kinase 3 regulates B-Raf through maintenance
of the B-Raf/Raf-I complex and inhibition by the NF 2 tumor suppressor protein.
Proc Natl Acad Sci U S A, 2006. 103(12): p. 4463-8.

Bardelli, A., et al., Mutational analysis of the tyrosine kinome in colorectal cancers.
Science, 2003. 300(5621): p. 949.

Lin, W., et al., Tyrosine kinases and gastric cancer. Oncogene, 2000. 19(49): p.
5680-9.

Shao, R.X., et al., Absence of tyrosine kinase mutations in Japanese colorectal
cancer patients. Oncogene, 2007. 26(14): p. 2133-5.

Mielke, K. and T. Herdegen, JNK and p38 stresskinases--degenerative efl'ectors of

signal-transduction-cascades in the nervous system. Prog Neurobiol, 2000. 61(1): p.
45-60.

Mota, M., et al., Evidence for a role of mixed lineage kinases in neuronal apoptosis.
J Neurosci, 2001. 21(14): p. 4949-57.

Saporito, M.S., et al., CEP—I347/KT- 7515, an inhibitor of c-jun N-terminal kinase
activation, attenuates the 1-methyl-4-phenyl tetrahydropyridine-mediated loss of

nigrostriatal dopaminergic neurons In vivo. J Pharmacol Exp Ther, 1999. 288(2): p.
421-7.

Waldmeier, P., et al., Recent clinical failures in Parkinson 's disease with apoptosis
inhibitors underline the need for a paradigm shift in drug discovery for
neurodegenerative diseases. Biochem Pharmacol, 2006. 72(10): p. 1197-206.

Taylor, J .P., et al., Leucine-rich repeat kinase 1 : a paralog of LRRK2 and a
candidate gene for Parkinson 's disease. Neurogenetics, 2007. 8(2): p. 95-102.

55

122.

123.

124.

125.

126.

127.

128

129.

130.

131.

132.

133.

134.

Korr, D., et al., LRRK] protein kinase activity is stimulated upon binding of GT P to
its Roc domain. Cell Signal, 2006. 18(6): p. 910-20.

Paisan-Ruiz, C., et al., Cloning of the gene containing mutations that cause PARK8-
linked Parkinson ’s disease. Neuron, 2004. 44(4): p. 595-600.

Zimprich, A., et al., Mutations in LRRK2 cause autosomal-dominant parkinsanism
with pleomarphic pathology. Neuron, 2004. 44(4): p. 601-7.

Biskup, S., et al., Localization of LRRK2 to membranous and vesicular structures in
mammalian brain. Ann Neurol, 2006. 60(5): p. 557-69.

Li, X., et al., Leucine-rich repeat kinase 2 (LRRK2)/PARK8 possesses GT Pase
activity that is altered in familial Parkinson's disease R1441 C/G mutants. J
Neurochem, 2007. 103(1): p. 238-47.

Fahn, 8., Description of Parkinson ’s disease as a clinical syndrome. Ann N Y Acad
Sci, 2003. 991: p. 1-14.

. Cookson, M.R., The biochemistry of Parkinson ’s disease. Annu Rev Biochem, 2005.

74: p. 29-52.

Mata, I.F., et al., LRRK2 in Parkinson ’s disease: protein domains and fimctional
insights. Trends Neurosci, 2006. 29(5): p. 286-93.

MacLeod, D., et al., The familial Parkinsonism gene LRRK2 regulates neurite
process morphology. Neuron, 2006. 52(4): p. 587-93.

Smith, W.W., et al., Kinase activity of mutant LRRK2 mediates neuronal toxicity.
Nat Neurosci, 2006. 9(10): p. 1231-3.

West, A.B., et al., Parkinson ’s disease-associated mutations in LRRK2 link

enhanced GT P-binding and kinase activities to neuronal toxicity. Hum Mol Genet,
2007. 16(2): p. 223-32.

Gloeckner, C.J., et al., The Parkinson disease causing LRRK2 mutation I2020T is
associated with increased kinase activity. Hum Mol Genet, 2006. 15(2): p. 223-32.

Greggio, E., et al., Kinase activity is required for the toxic effects of mutant
LRRK2/dardarin. Neurobiol Dis, 2006. 23(2): p. 329-41.

56

135.

136.

137.

138.

139.

140.

141.

142.

143.

144.

145.

146.

147.

J aleel, M., et al., LRRK2 phosphorylates moesin at threonine-558: characterization
of how Parkinson's disease mutants aflect kinase activity. Biochem J, 2007. 405(2):
p. 307-17.

Lewis, P.A., et al., The R1441 C mutation of LRRK2 disrupts GT P hydrolysis.
Biochem Biophys Res Commun, 2007. 357(3): p. 668-71.

West, A.B., et al., Parkinson 's disease—associated mutations in leucine-rich repeat
kinase 2 augment kinase activity. Proc Natl Acad Sci U S A, 2005. 102(46): p.
16842-7.

Meylan, E. and J. Tschopp, The RIP kinases: crucial integrators of cellular stress.
Trends Biochem Sci, 2005. 30(3): p. 151-9.

Ito, G., et al., GT P binding is essential to the protein kinase activity of LRRK2, a

causative gene product for familial Parkinson 's disease. Biochemistry, 2007. 46(5):
p. 1380-8.

Guo, L., et al., The Parkinson's disease-associated protein, leucine-rich repeat
kinase 2 (LRRK2), is an authentic GT Pase that stimulates kinase activity. Exp Cell
Res, 2007. 313(16): p. 3658-70.

Helrnbrecht, K., E. Zeise, and L. Reusing, Chaperones in cell cycle regulation and
mitogenic signal transduction: a review. Cell Prolif, 2000. 33(6): p. 341-65.

Carew, J .S. and P. Huang, Mitochondrial defects in cancer. Mol Cancer, 2002. 1: p.
9.

DiMauro, S. and EA. Schon, Mitochondrial respiratory-chain diseases. N Engl J
Med, 2003. 348(26): p. 2656-68.

MacKenzie, J .A. and RM. Payne, Mitochondrial protein import and human health
and disease. Biochim Biophys Acta, 2007. 1772(5): p. 509-23.

Kim, J .W. and CV. Dang, Cancer’s molecular sweet tooth and the Warburg ejfect.
Cancer Res, 2006. 66(18): p. 8927-30.

Pagliarini, DJ. and IE Dixon, Mitochondrial modulation: reversible
phosphorylation takes center stage? Trends Biochem Sci, 2006. 31(1): p. 26-34.

Wang, H.G., U.R. Rapp, and J .C. Reed, Bel-2 targets the protein kinase Raf-1 to
mitochondria. Cell, 1996. 87(4): p. 629-38.

57

148.

149.

150.

151.

Livigni, A., et al., Mitochondrial AKAP121 links CAMP and src signaling to
oxidative metabolism. Mol Biol Cell, 2006. 17(1): p. 263-71.

Boerner, J .L., et al., Phosphorylation of Y845 on the epidermal growth factor
receptor mediates binding to the mitochondrial protein cytochrome c oxidase
subunit 11. Mol Cell Biol, 2004. 24(16): p. 7059-71.

Kharbanda, S., et al., T ranslocation of SAPK/JNK to mitochondria and interaction
with Bcl-x(L) in response to DNA damage. J Biol Chem, 2000. 275(1): p. 322-7.

Majumder, P.K., et al., Mitochondrial translocation of protein kinase C delta in

phorbol ester-induced cytochrome c release and apoptosis. J Biol Chem, 2000.
275(29): p. 21793-6.

58

II. A N ovel Role for MLK3 in Mitochondria through Its Interaction
with AN T2

1. Abstract

Mixed Lineage Kinase 3 (MLK3) is a widely expressed mammalian mitogen
activated protein kinase kinase kinase (MAPKKK) that activates multiple mitogen-
activated protein kinase (MAPK) pathways. In addition to its catalytic domain, MLK3
also contains multiple protein binding domains that participate in the regulation of MLK3
activity and signaling. Immunoblotting of cellular lysates from a variety of human
epithelial cell lines revealed that MLK3 is expressed at high levels in breast cancer cell
lines. To identify MLK3-interacting proteins in breast cancer cells, human breast cancer
MCF-7 cells were engineered to inducibly express Flag-tagged MLK3. Using
immunoaffinity purification coupled with mass spectrometry, adenine nucleotide
translocase 2 (ANT 2) was identified as one of the components of the Flag-MLK3

complex.

ANT 2 functions as an ATP/ADP translocase in the mitochondria and plays an
important role in regulating the energetic balance of cells. Co-immunoprecipitation
experiments further confirmed the association of ectopically expressed AN T2 and MLK3.
The kinase domain of MLK3 was sufficient for its interaction with ANT 2. In addition,
ANT 2 preferentially bound to mutant forms of MLK3 that have enhanced catalytic
activity. A portion of MLK3 biochemically fractionated with mitochondria. Confocal
microscopy results demonstrated that MLK3 partially localized to the mitochondria.
Ectopic coexpression of MLK3 and ANT 2 resulted in elevation of the total cellular ATP

level. In contrast, cellular ATP content was diminished upon transient expression of

59

catalytically inactive MLK3 with ANT2. Stable knockdown of MLK3 by short hairpin
RNA decreased the total cellular ATP content whereas stable expression of MLK3
enhanced total ATP production. Taken together, these data open up the possibility that in
addition to its role in MAPK pathway activation, MLK3 regulation of cellular ATP levels

may influence cell proliferation.

6O

2. Introduction

Mitochondria are cellular organelles with double membrane structures, giving rise
to the different compartments including the outer membrane, intermembrane space, inner
membrane and matrix (Figure 2.1). Mitochondria, a major site of oxidative metabolism
in the cell, contain pyruvate dehydrogenase, the Krebs cycle enzymes, lipid-oxidizing
enzymes as well as redox proteins involved in electron transport and oxidative
phosphorylation (Figure 2.1). Aberrant mitochondrial function is associated with
numerous human diseases including cancer, diabetes, cardiac disorders and
neurodegenerative diseases [1-3]. Under aerobic conditions, the majority of cellular ATP
is produced through oxidative phosphorylation mediated by electron chain transport

complexes in the inner membrane of the mitochondria.

Mitochondria are also the decisive centers in apoptosis. Apoptosis, the name for
energy-dependent programmed cell death, is critical in development. However,
dysregulation of apoptosis is involved in pathologies, including cancer and
neurodegenerative diseases. Apoptosis is accomplished through release of
proapoptogenic factors, including cytochrome c, apoptosis inducing factor (AIF),_second
mitochondria-derived aptivator of apoptosis (Smac) and endo G nuclease, from the
mitochondria. Released cytochrome c interacts with the apoptotic protease pctivating
factor f (Apaf-l) and dATP to form the apoptosome, leading to the activation of caspase
cascade. Caspases are proteases responsible for cleavage and inactivation of various
proteins involved in DNA repair, cytoskeleton and chromosome integrity during
apoptosis. In addition, released AIF [4, 5] and endo G nuclease [6] cause DNA

fragmentation.

61

Outer membrane

Inner

Matrix

metabolites

Krebs
cycle

oxidation

+

 

Figure 2.1. Schematic of mitochondrial ATP synthesis. Substrates, from
pyruvates, electron-rich metabolites and the degradation of lipids through the [3 oxidation,
are oxidized to produce NADH used by complex I, II, III and IV of the respiratory chain,
generating a proton gradient across the inner mitochondrial membrane. ATP synthase
uses the proton gradient to produce ATP from phosphate and ADP. The mitochondrial
ATP is exported by ANT. ANT: adenine nucleotide translocase, VDAC: voltage
dependent anion channel, BA: bongkrekic acid, CATR: carboxyatractyloside. Figure

used is from Dahout-Gonzalez C. et a1. Physiology, 2006, 21 :242-249 with permission.

62

The release of proapoptogenic factors is accomplished by mitochondrial
membrane perrneabilization (MMP). MMP is regulated by Bcl-2 family proteins. The
Be] family includes the pro-survival Bel-2 proteins such as Bel-2 and Bel-XL as well as
the pro-apoptotic BH-3-only proteins such as Bax, Bad and Bak. Pro-survival Bel-2
proteins complex with the pro-apoptotic BH-3-only proteins to maintain the
mitochondrial membrane integrity [7]. In response to apoptotic insults, pro-survival Bcl-
2 proteins are sequestered fiom the heterocomplex, allowing the BH-3-only proteins to
oligomerize and form nonspecific pores in the mitochondrial outer membrane. Formation
of the nonspecific pores by the oligomeric BH3 -only proteins results in an increase of
MMP [8, 9], considered to be the irreversible step of apoptosis. Elevation of MMP is
believed to partially result from the opening of different channels including the
permeability transition pore complex (PTPC) (Figure 2.2.). The PTPC is an assembled
protein complex built up at the mitochondrial outer and inner membranes and consists of
three central components: the outer membrane yoltage gependent pnion phannel (VDAC),
the inner membrane gdenine pucleotide franslocase (ANT) and the matrix cyc10phi1in D

[10].

In addition to its role as a major component of the PTPC, ANT also functions to
exchange ADP and ATP across the mitochondrial membrane. Two conformational states
of ANT have been described which exist in equilibrium in the mitochondrial membrane.
In the c conformation, ANT faces the intermembrane space and in the m conformation,
ANT faces the matrix. Based on studies fi'om structure, functional characterization, and

native gel electrophoresis, the current transport model of ANT has been proposed in

63

Bax BcLZ

   

Figure 2.2. The architecture of the permeability transition pore complex. The
precise components of the permeability transition pore complex (PTPC) are not
completely known yet. The PTPC is primarily composed of the VDAC in the outer
mitochondrial membrane and the ANT in the inner mitochondrial membrane. In addition,
other proteins, including HK (interacting with VDAC from the cytosol), PBR (interacting
within PTPC from the outer membrane), CK (interacting with PTPC from the
intermembrane space) and Cpr (interacting with ANT from the mitochondrial matrix),
interact to bridge the outer and inner mitochondrial membranes of the PTPC. Both Bel-2
and Bax proteins can directly interact and regulate the PTPC. Abbreviations, HK:
hexokinase, VDAC: voltage dependent anion channel, PBR: peripheral-type
benzodiazepine receptor, CK: creatine kinase, Cpr: cyclophilin D, OM: outer
membrane, IM: inner membrane, IMS: intermembrane space. Figure used is from

Kroemer G et al. Physiological review, 2007, 87:99-163 with permission.

which two nucleotides bind the ANT monomer at the same time, one binding a
nucleotide coming fi'om the intermembrane space and the other one from the matrix.

Cyclophilin D cooperates with the two monomers of ANT for synchronous exchange of

ADP and ATP [11].

ANT isoforrns are found in all organisms containing mitochondria but the number
of ANT isoforrns varies among different organisms. Yeast has three ANT isofonns from
different genes but only ANT 3 is required for cell growth under anaerobic condition [12].
Mouse and rat contain three ANT isoforms. In human, four isoforrns of ANT are
encoded by four different genes. ANT 1, ANT 2 and ANT 3 share 88% sequence identity
whereas ANT 4 is less related, with about 70% sequence identity. The expression
patterns of ANT isoforrns seem to be determined by tissue type. The highest levels ANTI
mRNA are found in terminally differentiated tissues that do not undergo mitotic division
such as heart, skeletal muscles and brain [13, 14]. ANT 2 mRNA is mainly expressed in
tissues capable of proliferation and regeneration including liver, kidney and spleen.
Furthermore, it is also highly expressed in transformed and tumor cell lines and is down-
regulated in quiescent cells [15]. ANT 3 mRNA is ubiquitously expressed in various
tissues [13, 14]. The transcripts of ANT 4 are predominantly present in brain, liver and
testis [14]. The expression of different ANT isoforrns is also associated with
developmental stage. Studies of muscle cell differentiation have demonstrated an
increase in ANT 1 mRNA and a decrease in ANT 2 mRNA through the process of muscle
cell differentiation [16]. The level of ANT3 mRNA is relatively low and stable during

differentiation.

65

Alterations of levels of ANT isoforms lead to different biological outcomes in
some experimental settings. For instance, transient overexpression of ANT 1 or ANT 3 in
HEK 293T and HeLa cells induces a rapid apoptotic response through activation of
caspase 9 and 3 [17], whereas overexpression of ANT 2 has no effect on apoptosis [17].
A cumulative body of work has demonstrated that purified ANT (probably ANT 1 since
ANT is isolated fi'om rat heart or brain tissues and ANT 1 is the predominant isofOrrn of
ANT in these tissues) in the reconstituted liposomes forms nonspecific pores in response
to diverse proapoptotic stimuli including pro-oxidants, calcium, nitric oxide and the
ceramide derivative ganglioside GD3 [18]. Knockdown of ANT 2 by small interference
RNA (SiRNA) in HeLa cells resulted in an elevation of mitochondrial membrane
potential, increased ROS content and the decrease of total cellular ATP level, but had no
effect on glycolysis and cell cycle [19]. Moreover, ANT 2 knockdown sensitized HeLa
cancer cells to Lonidarnine, a clinical antitumor compound targeting mitochondrial ANT,
leading to the induction of MMP [19]. Since ANT is involved in cellular energetics
through ATP exchange as well as MMP regulation by forming the PTPC, ANT 'l' mice
have been created to further assess these functions. ANTI knockout mice are viable,
fertile and have normal growth characteristics [20]. However, ANT 1‘/' mice exhibit
exercise intolerance and collapse during the incremental exercise stress test. Notably, the
numbers of mitochondria present in the skeletal muscle of ANT 1 knockout mice are
increased and the respiration rate of complex III is also decreased. To date, ANT 2 has
been conditionally depleted in the murine liver tissue but no ANT 24 ' mice have been

reported yet.

66

Until recently most research in signal transduction has focused on cytosolic and
nuclear processes. However, recent results from several labs are giving rise to the idea
that the rnitochondrion and its constituent proteins are regulated by phosphorylation and
other signaling events. Several labs have used phospho-specific and phospho-motif [21]
antibodies such as the phospho-Akt substrate (PAS) antibody to identify the
mitochondrial phosphoproteins under physiological conditions. In addition, pro-Q
diamond staining[22] of phosphoproteins has been used in conjunction with SDS-PAGE
to demonstrate steady state phosphorylation of mitochondrial proteins. Based on
published reports, along metabolic labeling experiments using inorganic phosphate, more
than 60 mitochondrial proteins fiom different mitochondrial compartments were

identified as phosphoproteins [23] so far.

Since it is now known that mitochondrial phosphoproteins exist, it is reasonable
that protein kinases may exist in association with or within the mitochondrion. Indeed,
several signaling proteins have been discovered that are constitutively found in the
mitochondria, including the protein kinase PINK-1 [24] and the protein phosphotase
PTPMTI [25]. One challenge has been to determine how these signaling proteins can be
translocated to the mitochondria. Most mitochondrial proteins are synthesized in the
nucleus and transported to the mitochondria through the canonical N-terminal
mitochondrial targeting sequence. However, some resident mitochondrial proteins,
including ANT and VDAC, lack canonical mitochondrial targeting sequences. The
precise mechanism of translocation for these proteins precisely to the mitochondria is still

a mystery.

67

Several signaling proteins lacking the canonical mitochondrial targeting sequence
have been shown to be translocated to mitochondria through interaction with the
mitochondrial (adaptor) proteins. For instance, in response to ionizing irradiation, JNK is
translocated to mitochondria through its interaction with Bcl-xr [26]. Expressed Src
tyrosine kinase is targeted to mitochondria through its protein interaction with A-kinase
_anchor proteins (AKAP) 121, which contains the canonical mitochondria targeting
sequence [27]. However, the mechanisms of how these signaling molecules regulate

resident mitochondrial proteins and their specific mitochondrial substrates remain unclear.

In the present study, using immunoaffinity purification coupled with mass
spectrometry to identify MLK3 interacting proteins in breast cancer cells, ANT 2 was
discovered as one of the components in the MLK3 complexes. The data presented in this
chapter Show that the interaction of MLK3 with ANT 2 is ANT isoform-specific and
regulated by MLK3 activation status. Results from biochemical fractionation and
confocal microscopy demonstrate that expressed MLK3 partially colocalizes with
mitochondrial cytochrome c oxidase IV (COXIV). Coexpression of MLK3 and ANT 2 in
HEK 293T cells resulted in an elevation of the total cellular ATP level. Depletion of
MLK3 by short hairpin RNA in MCF-7 breast cancer cells decreased the total cellular
ATP content whereas induced expression of active MLK3 in MCF-7 cells enhanced it.
Taken together, these data presented here suggest an upregulation of cellular energetics
by MLK3 in a human breast cancer cell line and sheds light on the development of

possible therapeutic targets for breast cancer.

68

3. Materials and Methods

3.1 Reagents and antibodies

TRIzol and SuperScript one-step RT-PCR for long templates were obtained from
Invitrogen. The antibodies against phospho-JNK (G9), phospho-MLK3 and cleaved
PARP were fiom Cell Signaling Biotechnology Inc. The sodium/potassium ATPase orl
(N15) antibody was purchased fi'om Santa Cruz Biotechnology Inc. The antibodies
against cytochrome c oxidase subunit IV (COXIV), Flag-FITC and Alexa Fluor 488- or
Alexa Fluor 546-conjugated secondary antibody were from Molecular Probes. The
platelet derived growth factor receptor (PDGFR) B antiserum was obtained from BD
PharMingen Biotechnology Inc. The Flag M2 and actin monoclonal antibodies were from
Sigrna-Aldrich. Other antibodies used were the MLK3 antibody, hemagglutinin (HA)
antibody (BAbCO), and horseradish peroxidase-conjugated secondary antibodies (Bio-

Rad).

3.2 Plasmid constructs

Mammalian expression constructs of wild type MLK3 and its mutations as well as
truncation variants have been described previously [28-30]. The cDNAs encoding ANTI
or ANT 2 were obtained by RT-PCR using total RNA isolated fiom HeLa cells or human
MCF-7 breast cancer cells and the following oligonucleotides:

ANT 1: 5'-CTTCTAGAATGGGTGATCACGCTTGGAGC-3',
5'-TAACTAGTGACATATTTTTTGATCTCATC-3';
ANT2: 5'-TGACCTTCTAGAATGACAGATGCCGCTGTGTCC-3',

5'-TAACTAGTTGTGTACTTCTTGATTTCATC-3'.

69

The amplified ANT 1 and ANT 2 cDNAs were individually subcloned into the pCF 1
mammalian expression vector (a gift from Dr. Joan Brugge, Harvard Medical School)
using Xbal and SpeI restriction sites with the HA tag at the C-terrninus. All constructs

were subsequently verified by DNA sequencing.

3.3 Cell culture and transfection

Human embryonic kidney (HEK) 293T cells were cultured in Ham’s F 12/ low
glucose DMEM media (1:1) supplemented with 8% fetal bovine serum, 2 mM glutamine
and 100 U/ml penicillin/streptomycin and transfected using either Lipofectamine 2000
reagent (Invitrogen) following the manufacturer’s instructions, or using the calcium
phosphate method. Human breast cancer MCF-7 cells, human neuroblastoma SH-SY5Y
cells and HeLa cells were cultured in high glucose Dulbecco's modified Eagle's medium
supplemented with 10% fetal bovine serum, 2 mM glutarnine, and 100 U/ml
penicillin/streptomycin. All cell lines were maintained at 37 °C in a humidified incubator
containing 5% C02 in air. MCF-7/iFLag-MLK3 cells inducibly expressing Flag-MLK3

have been described elsewhere [31].

3.4 Construction of MLK3 stable knockdown cell line

The oligonucleotides for short hairpin RNAs predicted to target human MLK3,
5’-GATCCCCGCAGTGACGTCTCCAGTI‘TTTCAAGAGAAAACTCCAGACGT
CACTGCTTTTTA-B’ and 5’-AGCTTAAAAAGCAGTGACGTCTGGAGTTTTCT
CTTGAAAAACTCCAGACGTCACTGCGGG-3’ were designed using OligoEngine 2.0

program, synthesized, annealed and then ligated into the pSuper—retro-puro vector

70

(OligoEngine) with HindIII and Bgl 11 sites. 293-GPG packaging cells were transfected
with pSuper-shMLK3 plasmid using Lipofectamine 2000. Viral supernatant was
collected on days 4, 5, 6 and 7 post transfection, pooled virus was filtered and added to
MCF-7 cells for 24 h. After removal of the retrovirus-containing media, infected MCF-7
cells were selected with 2 [.1ng puromycin for 1 week. The MCF-7 cell line stably
expressing the short hairpin RNA for MLK3 was maintained in DMEM growth media
containing 1 rig/ml puromycin for further experiments. This cell line was established by

Eva Miller in the Gallo lab.

3.5 Cell lysis and immunoprecipitation

After transfection, cells were rinsed with ice-cold PBS and lysed in lysis buffer
(50 mM HEPES (pH 7.5), 150 mM NaCl, 1.5 mM MgC12, 2 mM EGTA, 1% Triton X-
100, 10% glycerol, 1 mM Na4PPi, 100 M fl-glycerophosphate, 1 mM Na3V04 and 1X

protease inhibitor cocktail (Sigma Aldrich)) on ice.

Antibodies against the proteins of interest were prebound to protein A agarose
beads for 30 min at room temperature: MLK3 antiserum (0.25 ug/ul slurry), M2
monoclonal Flag antibody (0.45 rig/u] slurry), HA antibody (0.45 rig/u] slurry). Clarified
cellular lysates were incubated with 20 ul of antibody-bound protein A agarose for 90
min at 4°C. Irnmunoprecipitates were washed twice with HNTG buffer (20 mM HEPES

(pH 7.5), 150 mM NaCl, 0.1% Triton X-100, and 10% glycerol).

71

3.6 Gel electrophoresis and Western blot analysis

Clarified cellular lysates and immunoprecipitates were resolved by SDS-PAGE.
Proteins were transferred to nitrocellulose membranes and immunoblotted using

appropriate antibodies. Irnmunoblots were developed by the chemiluminescence method.

3.7 Identification of ANT peptides in the MLK3 immune complex by Mass

Spectrometry

F lag-MLK3 expression was induced by the addition of 50 nM AP21967 for 20 h
to MCF-7/iFlag-MLK3 cells. F lag-MLK3 protein complexes were isolated as described
previously [31] and further subjected to an in-gel digestion with trypsin.
Nanoelectrospray liquid chromatography/tandem mass spectrometry (LC/MS/MS) was
performed to identify the proteins co-purified with MLK3 [31]. The identification of
MLK3 interacting proteins by mass spectrometry was performed by Hua Zhang and Yan

Du in the Gallo lab in conjunction with the MSU Proteomics facility.

3.8 Immunofluorescence and confocal microscopy

Transfected HeLa cells grown on cover slips were incubated with 0.1 uM
MitoTracker Red for 30 min, rinsed twice with PBS and fixed in 2% formaldehyde
(Polyscience, Inc.) for 10 min at room temperature. Cells were permeabilized in PBS
containing 0.15% Triton X-100 for 10 min. For COXIV staining, transfected HeLa cells
were directly fixed and permeabilized with cold' methanol for 15 min at -20°C. Samples

were blocked with 2% bovine serum albumin in PBS for 1 h at room temperature. After

72

incubation with primary antibody (Flag M2 monoclonal antibody, F ITC-F lag antibody
and COXIV mouse monoclonal antibody) in PBS containing 2% bovine serum albumin
for 1 h at room temperature, cells were washed three times with PBS and incubated with
the appropriate Alexa Fluor 48 8- or/and Alexa Fluor 546-conjugated secondary

antibodies (Molecular Probes) for 1 h at room temperature. After extensive washing with
PBS, cover slips were mounted onto glass slides. The fluorescently labeled cells were
examined with aZeiss LSM Pascal confocal laser scanning microscope. Three separate
tracks were used for capturing fluorescence and phase contrast images; excitation was
done using 488 nm- and 543 nm-lasers with BP 500-530 nm, BP 560-615 nm and phase 3,

respectively, as emission filters.

3.9 Subcellular fractionation

Cells were washed with phosphate-buffered saline (PBS) after trypsinization and
resuspended in ice-cold Buffer A (10 mM Tris-HCl (pH7.6), 0.15 mM MgC12, 10 mM
KCl) supplemented with protease inhibitor cocktail (Sigma), Na4PPi and Na3VO4. Cells
were incubated on ice for 1 h and homogenized with a Dounce homogenizer. Nuclei and
unbroken cells were eliminated after centrifugation at 900 x g for 3 min at 4°C.
Subsequent centrifugation of the supernatant at 6800 x g for 10 min at 4°C generates a
soluble fraction (86.8) and a mitochondrial-enriched pellet fraction (P68). The pellet
fi'action was resuspended with Buffer A. Equal cellular equivalents from the 86.8 and

P68 fractions were resolved by SDS-PAGE and analyzed by immunoblotting.

73

3.10 Determination of cellular ATP level

Twenty four hours post-transfection, HEK 293T cells transiently expressing the
proteins of interest were lysed in lysis buffer containing 1% trichloroacetic acid. MCF-7
cells either stably expressing short hairpin RNA of MLK3 or inducibly expressing MLK3
upon addition of AP21967 for 24 h were also lysed in the same lysis buffer as described
previously. The total cellular ATP level from clarified cellular lysates was determined
using the ENLITEN bioluminescence ATP detection kit (Promega, Inc.), according to the
manufacturer’s instructions. Beta-galactosidase activity, measured using a luminescent
beta-galactosidase detection kit (BD Biosciences), was used as a transfection control.
Protein concentration was determined using the Bradford method and equal amounts of

proteins were used for the assays involving the stable cell lines.

74

4 Results

4.1 ANT was identified in MLK3 complexes

The MCF-7/iFlag-MLK3 cell line, engineered to inducibly express F lag-tagged
MLK3, was previously established in our laboratory to study the role of MLK3 in breast
cancer through the identification of its interacting proteins [31]. Inducibly expressed
Flag-MLK3 and its associated proteins were isolated by affinity purification with the Flag
antibody. MLK3 -associated proteins were visualized by Coomassie Blue staining. In
addition to Flag-MLK3, a Coomassie blue-stained 32 kDa protein band was consistently
present only in the induced sample in several independent experiments, one of which is
shown in Figure 2.3. Tryptic digestion followed by mass spectrometry analysis identified
this 32 kDa band as adenine nucleotide translocase (ANT). All of the identified peptides
from the ANT protein sequence from different experiments are listed in Table 2.1. In
human cells, there are four isoforms of ANT encoded by four different genes. All of the
ANT peptides identified by LC/MS/MS can be found in the protein sequence of ANT2,
although a subset of them is present in the protein sequences of ANT 1 , ANT 3 or ANT4.
Therefore, we can reasonably conclude that ANT2 is present in the MLK3 complex.
However, we cannot exclude the possibility that ANT 1 , ANT 3 and/or ANT 4 may interact

with MLK3 as well.

4.2 Ectopically expressed ANTI and ANT2 are targeted to the mitochondria

ANT is a six trans-membrane protein located in the inner membrane of the
mitochondria. The firnction of ANT is to export ATP from the mitochondria to the

cytoplasm, in exchange for ADP. The sequences of ANT 1, ANT 2 and ANT3 share 90%

75

identity, ANT 4 is ~70% identical to ANTI. As a result, no antibody commercially
available is known to discriminate between the different ANT isoforms. Furthermore, no
commercially available antibodies reliably recognize ANT2. Indeed, it has been proven
difficult to generate antibodies that recognize AN T2 (Dr. Catherine Brenner at Universite
de Versailles/St. Quentin, France, personal communication). ANT 1 is the best
characterized of the ANT isoforms and the most abundant in many tissue types. To study
the isoform selectivity of the MLK3-ANT interaction, ANT 1 and ANT2 were cloned
with an expressed HA epitope tag. ANT 1 cDNA was amplified from total RNA of HeLa
cells and ANT 2 cDNA was amplified from total RNA of MCF-7 cells. The ANT cDNAs
were cloned into expression vectors with the coding sequence for the HA epitope-tag at
their carboxyl termini. The localization of HA-ANTI and HA-ANT 2 was assessed by
biochemical fractionation and confocal microscopy. To generate a mitochondria-
enriched fiaction, post-nuclear lysates were centrifuged at 6,800g to generate a soluble
fiaction ($6.8) and a pellet fraction (P6.8). As shown in Figure 2.4, the 86.8 fraction
contains cytosolic proteins like Cdc37 as well as light membrane proteins such as early
endosome antigen 1 (EEAI). The P6.8 fi'action contains the mitochondrial protein, COX
IV, but is devoid of cytosolic and other light membrane proteins. Ectopically expressed
HA-ANT 1 and HA-ANT 2 were found in the mitochondria-enriched fraction (Figure 2.4).
Moreover, as shown in Figure 2.5, ANT 2 transiently expressed in HeLa cells colocalized
with MitoTracker Red, a mitochondrial dye that stains intact mitochondria. Taken
together, these data indicate that the ectopically expressed ANT proteins are correctly

targeted to the mitochondria, suggesting that they are properly folded and fiinctional.

76

AP21967

  

FLAG-MLK3

Hsp90

Cdc37

Figure 2.3. ANT was identified in affinity-isolated MLK3 complexes. MCF-
7/iFlag-MLK3 cells were incubated in the presence or absence of AP21967 for 20 h.
Flag-MLK3 and its associated protein complexes were immunopurified as described
previously [31], resolved by SDS-PAGE on a 4—12% acrylamide gradient gel and stained
with Coomassie Blue. Molecular weight markers are indicated on the left in kDa.
Selected proteins identified by LC/MS/MS are indicated (experiment performed by Hua

Zhang).

77

Table 2.1. Sequences of the ANT peptides identified by mass spectrometry
and their correspondence with ANT isoforms. The total numbers of ANT peptides

obtained in 5 independent experiments are shown.

 

 

Peptide sequence # of times ANT
observed lsoforms
DFLAGGVAAAISK 4 2
DFLAGGVAAAISKTAVAPIER 1 2
TAVAPIER 1 1,2,3,4
LLLQVQHASK 1 1,2,3
EQGVLSFWR 1 2,3
YFPTQALNFAF K 5 1.2.3.4
QIFLGGVDKR 2 2
AAYFGIYDTAK 4 2
NTHIVISWMIAQTVTAVAG LTSYPFDTVR 1 2
MMMQSGR 1 1,2,3

78

4.3 MLK3 interacts with ANT2 through its kinase domain

The mass spectrometry data indicate the presence of ANT2 in the MLK3 complex.
It is possible that MLK3 may complex specifically with the ANT2 isoform. Alternatively,
ANT2 may be present in the MLK3 complex because it is the predominant isoform
expressed in the MCF-7 breast cancer cell line. To confirm the MLK3-ANT interaction
and to determine whether MLK3 exhibits selectivity for ANT isoforms, co-
immunoprecipitation experiments were performed. HEK 293T cells were co-transfected
with expression vectors for MLK3 and HA-ANT l or HA-ANT 2. MLK3 was
immunopurified from total cellular lysates and the presence of AN T1/2 was detected by
irnmunoblotting. As shown in Figure 2.6, HA-ANT 2, but not HA-ANTI, can interact
with MLK3. These data suggest that MLK3 is competent to complex with ANT 2, but not
ANTI. In agreement with published reports, overexpression of HA-ANT 1, but not HA-
ANT 2, induced apoptosis as judged by the cleavage of poly ADP ribose polymerase
(PARP), a substrate for activated caspase 3 and 7. These data suggest that the ANT 1 and
ANT 2 constructs used here are firnctionally competent. Notably, coexpression of MLK3
with ANT 2 in HEK 293T cells did not result in PARP cleavage, suggesting that neither

MLK3 nor ANT 2 is proapoptotic under these conditions.

To determine which region of MLK3 is required for its interaction with ANT 2,
the NHz-terminal or COOH-terminal halves of MLK3 were used in co-
immunoprecipitation experiments. MLK3 (1-399) encompasses the NHz-terminal
glycine-rich region, the SH3 domain and the kinase domain; MLK3 (400-847) includes

the leucine zipper region, the CRIB motif and the carboxyl-terminal Pro/Ser/Thr-rich

79

HA: ANT1 ANT2

cyto mito cyto mito
($6.8) (P63) ($6.8) (P6.8)

 

.1. ----o HA-ANT WB
-- - Cdc37 WB

.' - COXIV WB
..- Oll- EEA‘I WB

Figure 2.4. Ectopically expressed ANTs are mainly present in a
mitochondria-enriched fraction. HEK 293T cells transiently expressing either HA-
ANTl or HA-ANT2 were biochemically fractionated as described in Materials and
Methods. Proteins of interest from each fraction were analyzed by Western blotting with
appropriate antibodies. Equal cellular equivalent proteins from each fraction were loaded.
The efficiency of the fractionation was assessed by Western blotting with antibodies
directed against the cytosolic Cdc37, the mitochondrial cytochrome c oxidase (COX) IV
and early endosomal antigen 1 (EEAI). Data shown is representative of 3 independent

experiments.

80

Flag-AN T2 MitoTracker Red

 

phase contrast

      

Figure 2.5. Ectopically expressed ANT2 localizes to the mitochondria. HeLa
cells transiently expressing Flag-ANT2 were incubated with the mitochondrial probe,
MitoTracker Red for 30 min, followed by formaldehyde fixation. The images of Flag-
ANT2 (green) and mitochondria (red) were visualized with a Zeiss LSM Pascal confocal
laser scarming microscope. A phase contrast image was taken as a control.
Representative images shown here are fi‘om 3 independent experiments. Scale bar

indicated is 10 um.

81

region. MLK3 (1-399), but not MLK3 (400-847), associated with ANT 2 (Figure 2.7A).
Various Flag-tagged MLK3 truncations were used to confirm this result. MLK3 (1-598)
lacks the Pro/Ser/Thr-rich region; MLK3 (1-485) terminates just after the zipper domain;
and MLK3 (1-3 86) contains from the NH; terminus through the kinase domain. All of
these MLK3 truncations associated with HA-ANT 2 (Figure 2.7B), indicating that amino

acids 1-3 86 of MLK3 are sufficient for its interaction with ANT 2.

To further dissect which domain within MLK3 (1-3 86) is responsible for the
interaction with ANT 2, the binding ability of the individual regions or domains within
this portion of MLK3 was evaluated by coirnmunoprecipitation experiments. The
isolated kinase domain, MLK3 (115-399) is able to associate with ANT 2 whereas the
region including the glycine-rich region and SH3 domain, MLK3 (1-114), fails to do so
(Figure 2.7C). However, the expression level of MLK3 (1-114) is much lower than that
of the kinase domain in transfected cells, which prevents a definitive conclusion. To rule
out the possibility that MLK3 may also interact with ANT 2 through its SH3 domain, a
recombinant GST-SH3 fusion protein was used in a pull-down assay. ANT2 is
incompetent to bind the SH3 domain of MLK3. MLK3 Y52A which previously shown
to interact with the SH3 domain of MLK3 through the proline at 469, was used as a
positive control and MLK3 P469A was used as a negative control (Figure 2.7D). Taken
together, these data indicate that the kinase domain of MLK3 is necessary and sufficient

for its interaction with ANT 2.

82

MLK3

K (I- K ‘1’
15“ :5“ v9 .v‘“
45" ~25“ - e“ 4"

.- HA WB

in i.

MLK3 IP[ - .
' "‘3 W" i " MLK3 we

..ICI .

-- -- HA WB

cellular lysates m MLK3 WB

*- W———-
PARP WB

Jar-L“? 0— ~ *_ ~

 

"" o-I-

~J~~~fl ACtiflWB

Figure 2.6. ANT2, not ANT1 specifically interacts with MLK3. HA-ANTI,
HA-ANT2 with or without MLK3 were transiently expressed in HEK293T cells for 20-
24 h. Tap panel, MLK3 was immunoprecipitated using a MLK3 antibody from the
clarified cellular lysates. The presence of ANT in the MLK3 immunoprecipitates was
assessed by western blot with the HA antibody. Bottom panel, the protein level of MLK3,
ANT in the total cellular lysates was assessed by western blot with appropriate antibodies.
Actin was used as a loading control. Representative data shown is from 5 independent

experiments.

83

4.4 The activation status of MLK3 regulates its association with ANT2

Since the kinase domain of MLK3 mediates its interaction with ANT 2, it is
conceivable that the activation state of MLK3 may influence this association. To
investigate this possibility, a series of MLK3 mutants with various levels of catalytic
activity were tested for their ability to associate with ANT 2. A MLK3 leucine zipper
mutant form, MLK3 L410P, has very low MLK3 kinase activity and fails to activate the
JNK pathway ([29], Figure 2.8B). As shown in Figure 2.9A, MLK3 L410P fails to co-
immunoprecipitate with ANT 2. The association of a kinase-defective form of MLK3,
MLK3 K144A (Figure 2.8B), with ANT 2 was also dramatically decreased (Figure 2.9B).
Conversely, MLK3 variants that have higher basal catalytic activity through disruption of
MLK3 autoinhibition, MLK3 Y52A and MLK3 P469A ([30], Figure 2.8A and 2.8C) ,
were more efficiently complexed with ANT 2 than wild type MLK3. It has been
postulated that relieving MLK3 autoinhibition results in a conformational change that
opens and extends MLK3, exposing the kinase domain for catalysis. If the interaction of
MLK3 with AN T2 depends on this conformational alternation, a MLK3 variant
conformationally open but catalytically inactive would be expected to associate with
ANT 2. To test this hypothesis, MLK3 P469A and MLK3 K144A/P469A, a kinase
dead/open conformer of MLK3, were tested for their catalytic activity first and then
assessed for their ability to interact with ANT 2 in the co-immunoprecipitation
experiment. Since MLK3 activates JNK through activation and phosphorylation of
MKK4/7, the JNK activity was used as a readout for MLK3 activity of MLK3 P469A and
MLK3 K144R/P469A. MLK3 P469A strongly activates JNK whereas MLK3

K144A/P469A did not (Figure 2.9C), suggesting that, as expected, MLK3 P469A but not

84

Figure 2.7. The kinase domain of MLK3 is sufficient for the ANT 2
interaction. Different MLK3 truncation proteins were transiently coexpressed with or
without ANT 2 in HEK293T cells for 20-24 h. The interaction of ANT 2 with different
MLK3 truncation variants including (A) the N- and C-terrninal halves, (B) the C—terminal
truncations, (C) the glycine-rich sequence and SH3 domain (1-114) as well as the kinase
domain (115-399) of MLK3 was assessed by co-immunoprecipitation assay. (D) T op
panel, cellular lysates containing Flag-ANT 2, Flag-MLK3 Y52A or Flag-MLK3 P469A
from HEK 293T cells were used in the GST pull-down experiments as described in
Materials and Methods. The presence of Flag-ANT2, F lag-MLK3 Y52A or Flag-MLK3
P469A in the pulldown was assessed by western blotting. Equal amounts of GST and
GST-SH3 in the assays was confirmed by Coomassie staining. Bottom panel, levels of
expressed proteins and of endogenous actin as a loading control, were detected by
western blotting with the indicated antibodies. The numbers in the figure represent
amino acid numbers in MLK3. IP, immunoprecipitation. Data shown is representative of

3 independent experiments.

85

(A)

HA-MLK3 1-399 400-847
1"“ :—

Flag-ANT2r + - + - +

HA-MLK3 WB

Flag lP f

 

\ 9‘ "’ "" Flag-ANT2 W3

.3. i...
Cellular . . .. HA—MLK3 WB
lysates .5...-

"" "’ -- Flag-ANT2 W3

(3)
1 ‘6 s e
MLK3 (at .1529 (if: (:55

HHr—rr—n

HA'ANT2+-+-+-+_4,

 

I .. .. .- HA-ANT2 WB
H e
Flag IP
H H
Flag-MLK3 WB
[ can 4
i i
cellular '- 1“ Flag-MLK3 W8
lysates .-
.0

- 4 HA-ANTZ WB

86

(C)
HA-MLK3 1.143115.” 15.39
Flag-ANT2 +_: + - +5 - +

." HA-MLK3 we
Flag IPI a...gl m H +196

 

A“ - u .- Flag-ANT2 we

a i 2. ~ -
cellular - ~ 9‘

lysates HA-M LK3 WB
(D) MLK3 MLK3

Flag-tagged: ANT2 Y52A P469A
r—-\ rfi (fl

GST+s+ +

GST-SH3 - + .. +

+
-

Flag WB

*6qy-5Qfi9v:
«1 +2 (-
Flag-tagged: P" «(\- «y

Flag WB

—- - -- Actin WB

87

MLK3 K144A/P469A is catalytically active. Moreover, MLK3 K144R/P469A
associated very poorly with ANT 2 as compared to MLK3 P469A, suggesting that the
MLK3 catalytic activity rather than its open conformation regulates MLK3 interaction
with ANT 2. Taken together, these data suggest that MLK3 activity is crucial for

formation of the MLK3-ANT 2 complex.

4.5 A portion of MLK3 is present in mitochondria

Several groups have analyzed the subcellular localization of both endogenous and
overexpressed MLK3. Expressed MLK3 is primarily found in the perinuclear region and
partial colocalization has been demonstrated with Golgi marker, GM130 [32, 33] and in
the centrosomal regions during G2fM cell cycle stage [34]. In addition, upon
coexpression with activated Cdc42, MLK3 translocates to the plasma membrane [35]. To
test whether MLK3 can localize to mitochondria, confocal microscopy experiments were
performed using HeLa cells transiently expressing F lag-MLK3. Flag-MLK3 partially
colocalized with cytochrome oxidase IV (COXIV), a mitochondrial inner membrane
protein that participates in the oxidative phosphorylation process (Figure 2.10A),
suggesting that a portion of MLK3 localized to mitochondria. To corroborate these
results, crude biochemical fractionation experiments were performed. The presence of
different organellar protein markers including lactate dehydrogenase (LDH) and COXIV
in each fraction was assessed by immunoblotting of equal cellular equivalent of proteins
from each fraction. LDH and COXIV proteins were found almost exclusively in the $6.8

and P68 fractions, respectively (Figure 2.10B), indicating that mitochondria has been

88

Figure 2.8. The regulation of MLK3 model and the diagram of the kinase

activity of different MLK3 mutations. (A) The proposed regulation model of
MLK3. T op, MLK3 is autoinhibited through a protein interaction mediated by the SH3
domain and a proline between the zipper region and Cdc42/Rae interactive binding
(CRIB) motif. Bottom, GTP-bound Cdc42/Rac binds to the CRIB motif of MLK3 and
disrupts the MLK3 autoinhibition, allowing the dimerization of MLK3 though the zipper
region. Dimerization of MLK3 brings two kinase domains of MLK3 close to each other,
resulting in the transphosphorylation of the activation loop of MLK3 and fully activation
of MLK3. (B) The kinase defective mutant of MLK3 (K144A) and the leucine zipper
mutant of MLK3 (L410P) that fails to mediate MLK3 dimerization have lower MLK3
catalytic activity. (C) T op, MLK3 is auto-inhibited through its SH3 domain and a single
proline containing sequence. Bottom, amino acid substitution of Tyr52 in the SH3
domain with Ala or of Pro469 with Ala disrupts MLK3 autoinhibitory conformation

resulting in a higher MLK3 catalytic activity.

89

GTP

   

 

A .
I I p.
Kinase __ NH
.C“”’\CQS$$‘ 2
p
zipper ro CRIB
:1 er
zrpper CRIB '
(B)
K144A
Hthrf". Mel V
€11.51 Xase P'° CRIB
\1..-_ . zipper
H NF" L410P
2 L I “aim
SEEK Kinase X Pro CRIB
"____ ./- zipper
(C)
COOH
H N" 152A
21¢ ' .’
\- .. Kinase P'° CRIB
, zipper

P469A

Hszfl'r. ' .r"‘ '-
SSQinase PX CRIB
c. ’ zipper

90

low activity

COOH

COOH

high activity

COOH

COOH

COOH

Low activity

Low activity

COOH
High Activity

COOH

Figure 2.9. ANT2 preferentially associates with activated forms of MLK3.
Different MLK3 mutants, (A) zipper mutant (L410P), (B) kinase dead (K144A), mutants
that disrupt autoinhibition of MLK3 (Y 52A, P469A) and (C) kinase dead with disrupted
autoinhibition (Kl44R/P469A) were transiently coexpressed with or without HA-ANT 2
in HEK 293T cells. Co-immunoprecipitation assays were performed using total cellular
lysates with the MLK3 antibody. The immunoprecipitates were subjected to SDS-PAGE
electrophoresis followed by immunoblotting using the antibodies indicated. Levels of
expressed proteins and of proteins of interest were accessed by immunoblotting with the
indicated antibodies. Actin level was used as loading control. Data shown is

representative of independent experiments.

91

(A)

MLK3 WT L410P
HA-ANT2 + - + - +
.- HA-ANT2 WB
MLK3", Li a U u MLK3 we
«_- - «— HA-ANT2 WB
cellular
lysates g 3 gr, hi __, MLK3 WB
B
( ) MLK3 WT K144A Y52A P469A
F“ a—\ r—a .—
HA-ANT2 + - + - + - + - +

"" 'I- -' HA-ANT2 WB

.q.gfi~3~ MLK3 we

cellular - - — up ‘T HA-ANTZ WB
lysates

MLK3|P [
--~ on.“ ao-d MLK3WB

(091.63.21.69?
“we 9 933?
MLK3 . 959150915
HA-ANT2 - + - . + _+
:3 a ig, HA WB
‘31:“?! MLK3 we
.Hl‘efl MLK3 we

’ .- HAWB

MLK3 IP [

Phospho-JNK WB
-

_—---.". Actin WB

92

 

efficiently isolated. The biochemical fractionation further demonstrated that a portion of
expressed MLK3 is distributed to a mitochondria—enriched fraction (Figure 2.10B). A
portion of the MLK3 mutant forms including MLK3 L410P, MLK3 K144A and MLK3
P469A also partitioned to mitochondria-enriched fractions, suggesting that the

localization of MLK3 to mitochondria is independent of its activation status.

The subcellular localization of endogenous MLK3 in HEK 293T cells and MCF-7
human breast cancer cells was also evaluated by biochemical fractionation. A portion of
endogenous MLK3 was observed in the mitochondria-enriched fraction (Figure 2.11A) in
both HEK 293T and MCF-7 cells. The same result was obtained in HeLa human cervical
carcinoma cells and SH-SYSY human neuroblastoma cells (Figure 2.11B). Since the
previous fractionation generates the crude mitochondrial-enriched fiaction which also
contains other heavy membrane organelles including Golgi and endoplasmic reticulum
(ER) (data not shown), MCF-7 cells were biochemically fractionated using the sucrose
gradient to obtain a pure mitochondria-enriched fraction. As shown in Figure 2.11C, a
portion of endogenous MLK3 from MCF-7 cells was also present in the pure
mitochondria-enriched fraction. Taken together, the data suggest that a portion of MLK3

is associated with the mitochondria in multiple cell lines.

4.6 Mitochondrial MLK3 interacts with ANT2

Based on these results, one would predict that MLK3 interacts with ANT 2 in the
mitochondria. To determine if this is indeed the case, HEK 293T expressing MLK3

and/or ANT 2 were subjected to biochemical fractionation. As previously shown, nearly

93

 

   

COX 1V phase contrast
(A)
Flag-MLK3
(B) .
Cyto ($6.8) Mlto (P6.8) Cyto ($6.8) Mito(P6.8)
r—\ r—\ r—fi r—fl
MLK3 - WT K144A - WT K144A MLK3 - WT P469A - WT P469A
-_ — - MLK3 we “ Q "‘ Q MLK3 we
- - - LDH we 8 ‘-‘ = LDH we
- _ / cox“, WB - - - COXIVWB

Figure 2.10. A portion of expressed MLK3 is present in mitochondria. (A)
Flag-MLK3 was transiently expressed in HeLa cells. Cells were directly fixed and
permeabilized with cold methanol. The subcellular localization of MLK3 was detected
using a Flag-FITC-conjugated antibody (green). Mitochondria were detected using a
cytochrome c oxidase IV (COXIV) antibody and an Alexa Fluor 546-conjugated
secondary antibody (red). The images were obtained using a confocal laser scanning
microscope. A phase contrast image was taken as a control. Scale bar is 10 um. (B) Post-
nuclear HEK 293T cell lysates expressing different MLK3 mutants were fractionated by
centrifugation at 6800g. Proteins of interest from the soluble ($6.8) and pellet (P6.8)
fractions were analyzed by western blotting with appropriate antibodies. Equal cellular
equivalents from each fraction were loaded. The efficiency of the fractionation was
assessed by western blotting with antibodies directed against cytosolic lactate

dehydrogenase (LDH) and mitochondrial cytochrome c oxidase (COX) IV.

94

 

Figure 2.11. A portion of endogenous MLK3 is present in a mitochondria-
enriched fraction. Fractionation by differential centrifugation was performed on
various cell lines (A, HEK 293T and MCF-7; B, HeLa and SH-SYSY), as described
previously. Equal cellular equivalents from each fraction were analyzed by
immunoblotting using the indicated antibodies. Potential contamination with other
organelles in each fraction was assessed by western blotting using antibodies against
different organelle markers. LDH was used as cytosol protein marker. COXIV is a
mitochondrial protein marker. Both soditun potassium ATPase and platelet derived
growth factor receptor (PDGF R) were used as the markers of plasma membrane. The
presence of MLK3 in each fraction was assessed by western blotting using the MLK3
antibody. (C) Mitochondria from MCF-7 cells were fractionated through a sucrose
gradient as described in “Materials and Methods”, and fractions were analyzed as for A
and B. Cdc37 was the cytosolic protein marker. In addition to COXIV, AIF was also

used as another mitochondrial protein marker.

95

(A)

HEK293 MCF7

Cyto Mito Cyto Mito
($6.8) (P63) ($6.8) (P6.8)

_ - .- MLK3 we
.- - cox1vwe
- LDH we
.1: . -‘ “'1'; Nat/KtATPaseWB
h E n PDGFRB we
(3)
Cyto Mito

($6.8) (P6.8)

0 . e 4
Q9 9* ~89, e}?
....... .- MLK3 WB Ion ex osure
=3 " ‘ 9 P ’
H. _ '- MLK3 WB (short exposure)
"" "" LDH WB

— -— COXIVWB

96

(C)

Cyto Mito
($6.8) (P6.8)

- ----

MLK3 WB
Cdc37 WB

AIF WB
COX IV WB

 

 

(A) Cyto ($6.8) Mito (P6.8)
f N f N
MLK3 - - + + - - + 4»
HA-ANT2 - + - + - + - +
- .. _ .- HA we
- — -— ~ — LDH we
- - - «- cox1v we
(3) MLK3 - - + +
HA-ANT2 - + - +
M IP[ . . .. HA we
LK3 .
-...... ..A a :13 JIM-“3 “"3

* non-specific band

Figure 2.12. Expressed MLK3 present in the mitochondria-enriched fraction

interacts with ANT 2. HEK 293T cells expressing MLK3 with or without HA-ANT2

were subjected to a biochemical fractionation as previously described. (A) The presence

of MLK3 and HA-ANT2 proteins in each fraction was assessed by western blotting using

the MLK3 and HA antibodies. Equal cellular equivalent from each fraction were loaded.

The efficacy of the fractionation was assessed by western blotting with LDH and COXIV

antibodies. (B) MLK3 from the mitochondrial fraction was immunoprecipitated using the A

MLK3 antibody. The presence of ANT2 in the MLK3 immunoprecipitates was assessed

by western blotting with the HA antibody. Data shown is representative from 5

independent experiments.

97

all LDH was found in the $6.8 cytosolic fraction and mitochondrial COXIV was in the
P68 fiaction, suggesting that P6.8 mitochondria-enriched fraction is devoid of cytosol
(Figure 2.12A). The majority of ANT 2 and a portion of MLK3 were present in the
mitochondria-enriched fraction. MLK3 was then immunopurified from the
mitochondria-enriched fraction and the presence of associated ANT 2 was analyzed by
western blotting. As expected, the MLK3 present in the nritochondria-enriched fraction

was capable of complexing with ANT 2 (Figure 2.12B).

4.7 Effect of the MLK3-ANT2 interaction on total cellular ATP level

The mitochondrion is a major site of ATP production. Knockdown of AN T2 in
HeLa cells was recently shown to decrease the cellular ATP content [19]. To test
whether the interaction of MLK3 with ANT2 might influence cellular ATP, levels of
ATP were measured in cells expressing MLK3 and/or ANT 2. The expression vectors
encoding MLK3 and/or ANT2 were transfected into HEK 293T cells. A beta-
galactosidase construct was used as a transfection control. The cellular ATP level was
determined using a luciferin-based assay. The total cellular ATP level, normalized for
beta-galactosidase transfection efficiency, in cells coexpressing MLK3 and ANT 2 is 4-
fold higher than in cells transfected with empty vector (Figure 2.13A). The normalized
total cellular ATP level in ANT 2 expressing cells increased 1.5 fold and in MLK3
expressed cells increased 2.5 fold over control. Notably, the total cellular ATP level was
synergistically enhanced by coexpression of MLK3 and ANT 2, suggesting that the

interaction of MLK3 with ANT 2 may contribute to cellular ATP increasement.

98

To address whether catalytic activity is required for MLK3’s ability to enhance
total cellular ATP, the kinase inactive MLK3 K144A and hyperactive MLK3 P469A
variants were used. Cellular ATP content was diminished upon the expression of the
catalytically inactive MLK3 alone or with ANT 2 (Figure 2.13A), suggesting that this
effect was dependent on MLK3 kinase activity because the catalytically inactive MLK3
K144A decreased cellular ATP level. The cellular ATP level was synergistically
increased 5 fold in cells coexpressing MLK3 P469A and ANT 2 as compared to MLK3

P469A or ANT 2 alone.

Kinase inactive MLK3 can also localize to mitochondria (Figure 2.10B) but binds
ANT 2 only weakly (Figure 2.9B). Therefore it is possible that either MLK3 activity or
its interaction with ANT 2 is required for modulation of total cellular ATP cellular levels.
As an alternative approach, the impact of silencing of endogenous MLK3 on ATP levels
was determined. A stable population of MCF-7 cells in which MLK3 expression has
been largely reduced showed decreased total cellular ATP content compared with a

control vector population (Figure 2.13B).

We previously engineered MCF-7 cells to inducibly express F lag-MLK3 [31].
Addition of the small molecular inducer, AP21967, to the parental inducible cells
containing an empty expression vector had no effect on cellular ATP levels (Figure
2.13C). However, induction of Flag-MLK3 increased total cellular ATP level about 2
fold. Taken altogether, these data suggest that catalytically active MLK3 modulates

cellular ATP, possibly through its interaction with ANT 2.

99

Figure 2.13. Effect of MLK3 and ANT2 on total cellular ATP level. (A)
Expression vectors encoding MLK3 variants, ANT2 and beta-galactosidase were
transfected using Lipofectamine 2000 into HEK 293T cells for 20-24 h. Cells were lysed
in a lysis buffer containing 1% trichloroacetic acid. The total cellular ATP level from the
clarified cellular lysates was measured as described in Material and Methods. Beta-
galactosidase was used as an internal control; activity was measured using a luminescent
beta-galactosidase detection kit. The cellular ATP levels were normalized to beta-
galactosidase activity and plotted in arbitrary units with empty vector control (pRK) set
as one. Error bars represent standard deviation. (B) MCF-7 cells stably expressing a
MLK3 short hairpin RNA interference vector were lysed, and total cellular ATP levels
were measured as described in A. MLK3 levels after knockdown were assessed by
immunoblotting. The total cellular ATP content was determined from cellular lysates
containing equal amounts of proteins and plotted as arbitrary units with control cells
(pSuper) set as one. (C) MCF-7/iFlag—MLK3 cells were incubated with various
concentrations of AP21967 for 20 h to induce Flag-MLK3 expression. The total cellular
ATP content was determined from cellular lysates containing equal amounts of proteins
as described in A. Data from 3 independent experiments is plotted in arbitrary units with

control cells set as one. Error bars represent standard deviation.

100

(A)

 

 

 

 

 

 

 

(B)
7 1.4
E e E 1.2
g 5 g- 1
D 4 =
e 3 5 °'8
2 a. 0.6
g 1 g 0.4
< o g 0.2
o
pSuper pSuper-shMLK3
pSuper pSuper-ehMLKa
r———\
.._._.__ MLK3 WB
(C)
2.5
6‘ 2 r
E
g 1.5 »
a
E
< 0.5 -
0
AP21967(nM): - 5 5O - 5 50

 

Neo MLK3

101

5 Discussion

In this chapter, a novel interaction between the protein kinase MLK3 and the
mitochondrial translocase ANT 2 is characterized. Interestingly, despite the high amino
acid identity that is shared among ANT isoforms, the interaction with MLK3 appears to
be isoform specific. One possible explanation for this unexpected finding is that different

isoforms of ANT may have distinct submitochondrial localization patterns. The well

".A"
1
a
h-

characterized ANT 1 isoform is localized in the peripheral inner membrane, near the outer
membrane, where it can interact with VDAC and cyclophilin D [36, 37]. The localization I

of ANT 2 is less well defined. The highly folded regions of the inner membrane of the

 

mitochondria are known as the cristae. Based on isolation of cristae-enriched membranes
and a multiwell detergent screening assay, ANT 2 has been shown to physically interact
with the Pi carrier and FoFl-ATP synthase [3 8] forming the ATP synthasome that is
confined to the cristae [36, 37]. Thus, ANT isoform selectivity of MLK3 may be
explained by the distinct submitochondrial localization of ANT 2 and by inference of

MLK3.

Characterization of the MLK3-ANT 2 interaction revealed that the kinase domain
is sufficient for its association with ANT2 and that MLK3 mutants with high catalytic
activity preferentially bound ANT 2. Given the association of active MLK3 with ANT 2,
one might predict that MLK3 could phosphorylate ANT 2 or another protein involved in
the MLK3-ANT 2 interaction. No well-defined consensus phosphorylation sequence
within MLK3 substrates exists. MLK3 is known to phosphorylate MKK4 and MKK7 in
their activation loops (Figure 2.14). Intriguingly, examination of the sequences of the

ANT isoforms revealed that a sequence similar to the known MLK3 substrate

102

phosphorylation sites is present in AN T2. Preliminary metabolic labeling did not
conclusively show ANT 2 phosphorylation in HEK 293T cells coexpressing ANT 2 and
active MLK3. However, this intriguing but speculative hypothesis can be further
examined using potential phosphorylation site mutants of AN T2 in conjunction with in

vitro metabolic labeling and co-immunoprecipitation assays.

Consistent with previous reports, overexpressed ANT1, but not ANT2, induced
,.
apoptosis, as judged by PARP cleavage. Overexpression of ANT 2 with MLK3 neither
induced apoptosis nor impaired MLK3 signaling integrity, suggesting that this interaction
primarily affected ANT 2 function. It is tempting to speculate that interaction with MLK3
or phosphorylation by MLK3 might enhance ANT2 activity to accelerate ADP/ATP

exchange. To test this hypothesis, an ANT2 activity assay using isolated intact

mitochondria or ANT 2-reconstituted liposomes will be required.

Most mitochondrial proteins are translated in the cytoplasm and their
translocation to mitochondria is mediated through a mitochondrial targeting sequence at
their NH; termini. MLK3 contains no predicted mitochondrial targeting sequence based
upon analysis using the Predotar computer program (ExPASy proteomics server). N-
terminal epitope tags have been shown to mask mitochondrial targeting sequences,
abolishing protein translocation to mitochondria [39]. Furthermore, a portion of MLK3
was present in a mitochondria-enriched fraction even through it contains a Flag epitope at
its NH; terminus (Figure 2.15), consistent with the idea that MLK3 lacks a classical

mitochondrial targeting sequence.

Several studies have demonstrated that, in response to particular cellular stimuli,

certain protein kinases that are generally found in the cytosol and/or nucleus, such as the

103

tyrosine kinase Src [40, 41][27], JNK [26] and B-Raf [42] are targeted to mitochondria

through interaction with proteins which contain canonical mitochondrial targeting

sequences. For instance, cotransfection experiments have shown that the scaffold protein

AKAP121 promotes Src translocation to mitochondria where Src phosphorylates

unknown mitochondrial proteins and increases oxidative phosphorylation, causing an

elevation of mitochondrial membrane potential and increased ATP production [27].

Intriguingly, cytochrome c has been demonstrated to be tyrosine-phosphorylated in i ..
bovine heart [43]. Upon TNF-a stimulation, Src is targeted to mitochondria through
interaction with the mitochondrial adaptor protein, Dok4, leading to decreased levels of
complex I through an as yet undefined mechanism [41]. Expressed kinase inactive
MLK3 variants bind only poorly to ANT 2, yet they are present in the mitochondria-
enriched fraction. It is unclear how MLK3 mitochondrial localization is controlled. It is
possible that in addition to ANT 2, there are other mitochondrial (adaptor) proteins target
a portion of MLK3 to mitochondria and this interaction is kinase activity independent. It

is also unclear whether the interaction between MLK3 and ANT 2 is direct or mediated

through an intervening protein.

In most normal cells, the majority of cellular ATP is derived through oxidative
phosphorylation rather than through glycolysis. Under hypoxic conditions, tumor cells
rely primarily on glycolysis for energy production. This idea has re-emerged in cancer
research but whether this is a universal property of all cancer cells remains under debate.
Our data indicate that coexpression of MLK3 and ANT 2 dramatically increases the total
cellular ATP levels and that cellular ATP levels correlate with MLK3 activity and protein

levels. It is quite possible that MLK3 modulates mitochondrial proteins of oxidative

104

phosphorylation and Krebs cycle which is known as TCA cycle, resulting in higher level
of total ATP. Indeed, peptides corresponding to these mitochondrial proteins crucial for
ATP production, especially ATP synthase, were detected in the mass spectrometry
experiments (Table 2.2) Recent studies combining knockdown of individual ANT
isoforms and treatment with oligomycin, the ATP synthase inhibitor, in HeLa cancer cells
suggest that AN T2 cooperates with ATP synthase [19] to increase cellular ATP levels. It
will be interesting to determine the impact of MLK3 and ANT 2 on ATP synthase activity

in tumor cell lines grown under different conditions including norrnoxia and hypoxia.

Given that phosphorylation is a reversible modification, the effect of
mitochondrial phosphatases on cellular energy regulation is postulated being opposite to
that of mitochondrial kinases. Depletion of PTPMTI, a mitochondrial PTEN-like
phosphatase, by short hairpin RNA has been demonstrated to promote mitochondrial

ATP production [25].

Unpublished work from our lab demonstrated that transient knockdown of MLK3
by SiRNA in MCF-7 breast cancer cells decreased DNA synthesis in response to the
stimulation of grth factors and the steroid hormones, estrogen and progesterone. A
previous study showed that depletion of MLK3 by SiRNA in human immortalized colon
fibroblasts, colon adenocarcinoma cells with the oncogenic Res and neurofibromatosis 1
(NF -1) schwannoma cells reduced serum-induced cell proliferation as judged by counting
of cell numbers [44]. It has been suggested that MLK3 scaffolds the Ras-Raf-MEK-ERK
pathway upon growth factor stimulation [45] to contribute to cell proliferation. The
experiments presented in this Chapter show a correlation between reduced total cellular

ATP levels and decreased cell proliferation upon knockdown of MLK3 in breast cancer

105

 

cells. Notably, coexpression of MLK3 and ANT 2 synergistically enhances cellular ATP
levels, suggesting that the MLK3-ANT 2 interaction contributes the increased ATP levels.
It will be important to evaluate the biological function of the MLK3-ANT 2 interaction by
directly testing whether depletion of ANT 2 by SiRNA blocks the MLK3-induced increase
in cellular ATP levels and affecs cellular proliferation. Taken altogether, these data
suggest a novel function for MLK3 and open up the possibility that, in addition to its role
in the MAPK pathway, MLK3 may promote cancer cell proliferation and/or survival by
facilitating the energetic metabolism through regulation of a MLK3-ANT 2 containing

synthasome (Figure 2.16).

106

MKK4 257 SIAKT 261
MKK7 271 SKAKT 275

ANT1 41 SKQIS 45
ANT2 41 SKQIT 45
ANT3 41 SKQIA 45
ANT4 s4 SKQIS 58

Figure 2.14. Alignment of the sequences containing the phosphorylation sites
of MLK3 in human MKK4 and MKK7 with human ANT isoforms. The Ser
and Thr of human MKK4 and MKK7 phosphorylated by MLK3 are labeled with
phosphate symbols. The phospho-Ser and phospho-Thr within MKK4 and MKK7 are
separated by 3 amino acids, giving the phosphorylation consensus motif as SXAKT. The
phosphorylation consensus motif of MLK3 is not clearly defined. The potential
phosphorylation motif of MLK3 in human ANT isoforms is listed. The numbers shown

correspond to the amino acids in each protein. X: any amino acid residues.

107

Cyto ($6.8) Mito (P63)

’ 3 r———.a

x I

_I ..l

E E

n U n O

. i '5 , 5 a?

E 2 E 2
. - MLK3WB
_H: .— FlagWB

8 8 ..- LDH we
., - «- cox1v we

Figure 2.15. N-Flag tagged MLK3 partially fractionates with mitochondria.
MLK3 or N-Flag MLK3 expression vector was transfectd in HEK 293T cells. Twenty
four hours after transfection, cells were subjected to a biochemical fractionation as
described in Figure 2.2.10. Equal cellular equivalents of each fraction were loaded.
Ectopically expressed MLK3 in each fraction was detected by immunoblotting with the
MLK3 and Flag antibodies, respectively. The efficacy of the fractionation was assessed
by immunoblotting with LDH (cytosolic marker) and COXIV (mitochondrial marker)

antibodies. Data shown is representative of 3 independent experiments.

108

Table 2.2. Peptide sequences of the oxidative phosphorylation proteins and

the voltage dependent anion channel 3 (VDAC 3) identified by mass

 

 

 

 

 

 

 

 

 

 

 

spectrometry.
protein complex subunit 0f identified peptide sequence
oxidative phosphorylation

NADH dehydrogenase Complex Ia, subunit 4 FYSVNVDYSK

NADH dehydrogenase Complex Ia, subunit 9 SSVSGIVATVFGATGFLGR

NADH dehydrogenase Complex 101 subunit 13 IALLPLLQAETDR

NADH dehydrogenase Complex 13, subunit 4 TLPETLDPAEYNISPETR

Ub‘qmnd'cymhmme Complex 111, subunit 7 HVISYSLSPFEQR

c reductase

Ub‘qum01‘cy‘°°h’°me Complex 111, subunit 10 AFDQGADAIYDHINEGK

c reductase

Cytochrome c oxidase Complex IV, subunit 7a2 LFQEDDEIPLYLK

ATP synthase Complex V, F chain DF SPSGIFGAF QR
LTLDTIFVPNTGK

VDAC 3 LTLSALIDGK

 

 

 

109

 

 

   

Intermembrane space

W13

ATP
' \4 matrix
III

I : ATP synthase

Figure 2.16. The proposed model of how the interaction of MLK3 with ANT 2
in elevation of the total cellular ATP level. The model is proposed based on the
data shown in this Chapter, the findings from other labs and the preliminary data from
mass spectrometry. It has been demonstrated that ANT2 interacts and cooperates with
ATP synthase for ATP production. The identified peptides of VDAC3 and the electron
transport chain complexes from mass spectrometry (listed in Table 2.2) were present in
the MLK3 protein complex. MLK3 could interact with the VDAC3-ANT2-
ATPsynthase-the complexes of the electron transport chain from the mitochondrial
surface or/and the intermembrane space of mitochondria or/and the matrix through the
direct or indirect interaction. The interaction of MLK3 with this protein complex leads to
an elevation of the total cellular ATP level. Abbreviation, VDAC: voltage dependent

anion channel.

110

t—5

10.

11

12.

13.

References

. Carew, J.S. and P. Huang, Mitochondrial defects in cancer. Mol Cancer, 2002. 1: p. 9.

DiMauro, S. and EA. Schon, Mitochondrial respiratory-chain diseases. N Engl J
Med, 2003. 348(26): p. 2656-68.

MacKenzie, J .A. and RM. Payne, Mitochondrial protein import and human health
and disease. Biochim Biophys Acta, 2007. 1772(5): p. 509-23.

Daugas, E., et al., Apoptosis-inducing factor (AIF ).' a ubiquitous mitochondrial
oxidoreductase involved in apoptosis. FEBS Lett, 2000. 476(3): p. 118-23.

Susin, $.A., et al., Two distinct pathways leading to nuclear apoptosis. J Exp Med,
2000. 192(4): p. 571-80.

. Saelens, X., et al., Toxic proteins released fiom mitochondria in cell death. Oncogene,

2004. 23(16): p. 2861-74.

J acotot, B., et al., Control of mitochondrial membrane permeabilization by adenine
nucleotide translocator interacting with HIV-1 viral protein rR and Bcl-2. J Exp Med,
2001. 193(4): p. 509-19.

Vieira, H.L., et al., Cell permeable BH3-peptides overcome the cytoprotective efl'ect
of Bel-2 and Bcl-X(L). Oncogene, 2002. 21(13): p. 1963-77.

Zamzami, N., et al., Bid acts on the permeability transition pore complex to induce
apoptosis. Oncogene, 2000. 19(54): p. 6342-50.

Verrier, F., et al., Study of PT PC composition during apoptosisfor identification of
viral protein target. Ann N Y Acad Sci, 2003. 1010: p. 126-42.

. Dahout-Gonzalez, C., et al., Molecular, functional, and pathological aspects of the

mitochondrial ADP/A T P carrier. Physiology (Bethesda), 2006. 21: p. 242-9.

Sokolikova, B., et al., A carbon-source—responsive element is required for regulation
of the hypoxic ADP/A T P carrier (AA C3) isoform in Saccharomyces cerevisiae.
Biochem J, 2000. 352 Pt 3: p. 893-8.

Doemer, A., et al., T issue-specific transcription pattern of the adenine nucleotide
translocase isoforms in humans. FEBS Lett, 1997. 414(2): p. 258-62.

111

 

14. Dolce, V., et al., A fourth ADP/A T P carrier isoform in man: identification, bacterial

expression, fimctional characterization and tissue distribution. FEBS Lett, 2005.
579(3): p. 633-7.

15. Barath, P., et al., The growth-dependent expression of the adenine nucleotide
translocase-2 (AN T 2) gene is regulated at the level of transcription and is a marker
of cell proliferation. Exp Cell Res, 1999. 248(2): p. 583-8.

16. Stepien, G., et al., Diflerential expression of adenine nucleotide translocator isoforms
in mammalian tissues and during muscle cell differentiation. J Biol Chem, 1992.
267(21): p. 14592-7.

17. Zamora, M., et al., Adenine nucleotide translocase 3 (ANT 3) overexpression induces
apoptosis in cultured cells. FEBS Lett, 2004. 563(1-3): p. 155-60.
18. J acotot, E., et al., Mitochondrial membrane permeabilization during the apoptotic .,
process. Ann N Y Acad Sci, 1999. 887: p. 18-30. .!
Ii

 

19. Le Bras, M., et al., Chemosensitization by knockdown of adenine nucleotide
translocase-2. Cancer Res, 2006. 66(18): p. 9143-52.

20. Graham, B.H., et al., A mouse model for mitochondrial myopathy and
cardiomyapathy resulting fiom a deficiency in the heart/muscle isoform of the
adenine nucleotide translocator. Nat Genet, 1997. 16(3): p. 226-34.

21. Berwick, DC. and J .M. Tavare, Identijyingprotein kinase substrates: hunting for the
organ-grinder's monkeys. Trends Biochem Sci, 2004. 29(5): p. 227-32.

22. Schulenberg, B., et al., Analysis of steady-state protein phosphorylation in
mitochondria using a novel fluorescent phosphosensor dye. J Biol Chem, 2003.
278(29): p. 27251-5.

23. Pagliarini, DJ. and J .E. Dixon, Mitochondrial modulation: reversible
phosphorylation takes center stage? Trends Biochem Sci, 2006. 31(1): p. 26-34.

24. Valente, E.M., et al., Hereditary early-onset Parkinson '3 disease caused by mutations
in PINK]. Science, 2004. 304(5674): p. 1158-60.

25. Pagliarini, D.J., et al., Involvement of a mitochondrial phosphatase in the regulation
of A T P production and insulin secretion in pancreatic beta cells. Mol Cell, 2005.
19(2): p. 197-207.

112

26. Kharbanda, S., et al., T ranslocation of SAPK/JNK to mitochondria and interaction
with Bcl-x(L) in response to DNA damage. J Biol Chem, 2000. 275(1): p. 322-7.

27. Livigni, A., et al., Mitochondrial AKAP121 links CAMP and src signaling to oxidative
metabolism. Mol Biol Cell, 2006. 17(1): p. 263-71.

28. Book, B.C., et al., Cdc42-induced activation of the mixed-lineage kinase SPRK in
vivo. Requirement of the Cdc42/Rae interactive binding motif and changes in
phosphorylation. J Biol Chem, 2000. 275(19): p. 14231-41.

29. Vacratsis, PO. and K.A. Gallo, Zipper-mediated oligomerization of the mixed lineage
kinase SPRK/MLK-3 is not required for its activation by the GT Pase cdc 42 but Is ....
necessary for its activation of the MK pathway. Monomeric SPRK L410P does not

catalyze the activating phosphorylation of Thr258 of marine MI T OGEN-A C T I VA T ED
protein kinase kinase 4. J Biol Chem, 2000. 275(36): p. 27893-900.

30. Zhang, H. and K.A. Gallo, Autoinhibitian of mixed lineage kinase 3 through its Src
homology 3 domain. J Biol Chem, 2001. 276(49): p. 45598-603.

31. Zhang, H., et al., Hsp90/p50cdc37 is required for mixed-lineage kinase (MLK) 3
signaling. J Biol Chem, 2004. 279(19): p. 19457-63.

32. Cha, H., et al., Inhibition of mixed-lineage kinase (MLK) activity during G2-phase
disrupts microtubule formation and mitotic progression in HeLa cells. Cell Signal,
2006. 18(1): p. 93-104.

33- Cha, H., et al., Phosphorylation of golgin—I 60 by mixed lineage kinase 3. J Cell Sci,
2004. 117(Pt 5): p. 751-60.

34. Swenson, K.I., K.E. Winkler, and A.R. Means, A new identityfor MLK3 as an NIMA-
related, cell cycle-regulated kinase that is localized near centrosomes and influences
microtubule organization. Mol Biol Cell, 2003. 14(1): p. 156-72.

35- Du, Y., et al., Cdc42 induces activation loop phosphorylation and membrane
targeting of mixed lineage kinase 3. J Biol Chem, 2005. 280(52): p. 42984-93.

36- Faustin, B., et al., Mobilization of adenine nucleotide translocators as molecular

bases of the biochemical threshold eflect observed in mitochondrial diseases. J Biol
Chem, 2004. 279(19): p. 20411-21.

37' V3’ssokikh, M.Y., et al., Adenine nucleotide translocator isoforms 1 and 2 are
difi'erently distributed in the mitochondrial inner membrane and have distinct
affinities to cyclophilin D. Biochem J, 2001. 358(Pt 2): p. 349-58.

113

38. Ko, Y.H., et al., Mitochondrial ATP synthasome. Cristae—enriched membranes and a
multiwell detergent screening assay yield dispersed single complexes containing the
ATP synthase and carriers for Pi and ADP/A T P. J Biol Chem, 2003. 278(14): p.
12305-9.

39. Ventura, A., eta1., A cryptic targeting signal induces isoform-specific localization of
p46Shc to mitochondria. J Biol Chem, 2004. 279(3): p. 2299-306.

40. Cardone, L., et al., Mitochondrial AKAP121 binds and targets protein tyrosine
phosphatase D], a novel positive regulator of src signaling. Mol Cell Biol, 2004.
24( 11): p. 4613-26.

41. Itoh, S., et al., Mitochondrial Dok-4 recruits Src kinase and regulates NF -kappaB
activation in endothelial cells. J Biol Chem, 2005. 280(28): p. 26383-96.

42. Wang, H.G., U.R. Rapp, and J .C. Reed, Bel-2 targets the protein kinase Raf-1 to
mitochondria. Cell, 1996. 87(4): p. 629-38.

 

43. Lee, I., et al., New prospects for an old enzyme: mammalian cytochrome c is tyrosine-
phosphorylated in viva. Biochemistry, 2006. 45(30): p. 9121-8.

44. Chadee, D.N. and J .M. Kyriakis, MLK3 is required for mitogen activation of B-Raf
ERK and cell proliferation. Nat Cell Biol, 2004. 6(8): p. 770-6.

45. Chadee, D.N., et al., Mixed-lineage kinase 3 regulates B-Rafthraugh maintenance of
the B-Raf/Raf-I complex and inhibition by the NF2 tumor suppressor protein. Proc
Natl Acad Sci U S A, 2006. 103(12): p. 4463-8.

114

III. A Novel GTPase Domain of the Parkinson’s Disease-Associated
Kinase LRRK2

1. Abstract

Parkinson’s disease (PD) is the most common neurodegenerative movement
disorder in western countries. Emerging evidence from genetic studies has demonstrated
that mutations in PARK8, which encodes the leucine-rich repeat kinase (LRRK) 2, causes
late-onset familial PD. LRRK2 contains multiple predicted catalytic and protein-
interaction domains; and pathogenic amino acid substitutions are located throughout

LRRK2.

Roc, a distinct, putative GTPase of the Ras superfamily, always exists in tandem
with so-called COR domain of unknown function. No representative structures of either
Roc or COR domains have been solved to date. Of solved GTPase structures, Roc is
most similar to the Rab7 GTPase; therefore a molecular homology model of ROC was
created using the murine GTP-bound Rab7 crystal structure as a template. This model
was used in conjunction with the GTPase literature to predict the functional consequence

of PD mutation in the Roc.

To investigate the GTPase activity of the Roe and Roe-COR domains,
recombinant bacterial expression of Glutathione S-Transferase (GST) fusion proteins of
Roe and Roe-COR, as well as their mutants that are predicted to be conformationally
active and inactive, has been studied. Most GST-Roe variant proteins were optimally

soluble under autoinducible conditions and can be fiirther purified by glutathione affinity

115

agarose. However, GST-Roc-COR remained in inclusion bodies under all conditions,

including co-expression with chaperone proteins.

Roe and its predicted constitutively/conformationally active mutants, Roe
A1396T as well as R1398L, specifically bound GTP, whereas Roc T1348N, the predicted
inactive form of ROC GTPase, failed to do so. Consistent with the prediction from the
Roc homology model, no apparent deficiency of GTP binding was observed in the PD-
associated variant, Roc R1441C. Using an enzymatic coupling assay, a weak GTPase

activity of a recombinant Roc fusion protein was detected.

The idea that LRRK2 might adopt a conformation that involves interactions
among its domain was tested by coirnmunoprecipitation experiments of expressed
fragments of LRRK2. A Roc-COR fragment complexed with the carboxyl terminal
region consisting of Roc, COR, kinase domains and WD 40 repeats of LRRK2, but not
the N-terminal region containing the ankyrin domain. Introduction of the PD-associated
COR mutation, Y1699C, in the Roe-COR domain enhanced its association with carboxyl
terminal region of LRRK2. However, the R1441C/G mutation in the context of Roc or
Roe-COR resulted in reduction of protein complex formation with the carboxyl terminal
region of LRRK2. Taken altogether, the data presented in this chapter suggest that Roc
domain of LRRK2 is a bonafide GTPase, and provide evidence that LRRK2 domains
interact with one other in ways that are altered by the introduction of PD-associated

mutations.

116

 

2. Introduction
Parkinson’s disease (PD) is the most common neurodegenerative movement
disorder in western countries. The clinical symptoms of PD include resting tremor,
bradykinesia, muscle rigidity and impaired balance. The loss of the dopaminergic
neurons in the substantia nigra pars compacta and the presence of protein aggregations
named Lewy bodies in the surviving neurons of the brainstem are the pathological
hallmarks of PD [1]. To date, all available treatments aim to ameliorate the clinical :—

symptoms but do not cure the disease.

The etiology of PD remains ambiguous so far. At the cellular level, PD has been

generally thought to correlate with the mitochondrial dysfunction, abnormality of protein

 

aggregation/degradation and/or defects in vesicular trafficking [2]. Many environmental
or genetic factors have been demonstrated to increase the risk of PD such as rotenone
which is a commonly used pesticide and inhibits complex I of the oxidative

phosphorylation [3].

Over the past several years, mutations in several genes including alpha-synuclein,
parkin, UCH-LI, PINK 1 , and DJ-I have been tightly linked to Parkinsonism [4]. More
recently, evidence from genetic studies has indicated that mutations of leucine-rich repeat
kinase (LRRK) 2 gene cause autosomal dominant, late-onset familial Parkinsonism [5, 6].
LRRK2 is an extraordinarily large protein which contains 2,527 amino acids, giving an
estimated molecular weight of around 286 kDa. LRRK2 contains multiple predicted
catalytic and protein-interaction domains (Figure 3.1). The Roc domain of LRRK2 is a

putative GTPase and, like all other Roe containing proteins, is found in tandem with the

117

 

Se .8 85:55 “moo €82 nauseous ”my: .555 ”M72
6:625 NEE anorerénem Bang: 28 05:3 3 teammate”. gen. 05 323 23:82 605325 0.8 038% Page.“
55 Bumbag 8 33530806 :03 26: 35 803852 Sofia—cow 28 285 m: 89a 938555 05 no women 3388 032:8

"”585... 3324 .3553 83m... Fugue—an.— E 258 23:33: .53 2.22:2. and”: we Banana .3” «Earn

nova «EN—Warn ass. as: Fwnwcvnrnnmw_chN—. 3a 08 goes 8..

Realm Se: 4; 8:5. moo 8m: 5... xz<i :5 T

 

 

 

 

 

 

 

 

 

 

 

 

: _ lI—I _ - _ _ _
ensue 33"» 53E 3%; 032: BN5 58.“.88 :82.
FIRE 38.: >5: >33.
MEG :53:
HEN. 93.5. 435..
082m

 

118

so called C-terminal of Roc (COR) region. The armadillo, ankyrin, leucine rich repeat
(LRR) found in the N-terminal half of LRRK2 and the C-terminal WD-40 domains are
likely to be involved in protein—protein interactions. PD associated mutations resulted in
amino acid substitutions scattered throughout the LRRK2 protein [7-9], suggesting

alterations in each structural domain contribute in some way to PD.

Several recent biochemical studies have demonstrated a weak in vitro protein
kinase activity of LRRK2 [10-18]. The most prevalent PD associated mutation has a Ser
substituted for Gly at 2019 near the predicited activation loop in the kinase domain.
LRRK2 G2019S exhibits slightly enhanced LRRK2 catalytic activity in vitro [10, ll, 13,
16-18]. Transiently expressed LRRK2 620198 increased neuronal apoptosis as judged
by capase 3 activation in primary rat nigral dopamine neurons [19]and positive TUNEL
staining in SH-SYSY human neuroblastoma cells and mouse primary cortical neurons
[16]. Modulation of LRRK2 catalytic activity in vitro by other PD mutations has not been

conclusive so far.

The Ras superfamily consists of the R33, Rho/Rae, Ran/Rab, and Arf subfamilies.
Bosgraaf et al. have constructed a dendrogram comparing the relatedness of Roc domains
to GTPases of the Ras superfamily. According to a dendrogram of GTPase domains, Roc
is distantly related to the other 4 subgroups [20]. LRRK2 belongs to the ROCO family of
proteins since it contains both Roc and COR domains. Notably, Roc-COR tandem
domains are found in a variety of organisms including prokaryotes, Dictyostelium, plants
and metazoa, but are absent in yeast and Plasmodium. Roc domains are always found in
tandem with COR regions, suggesting that their functions may be interdependent.

However, no function has been assigned to any COR domain in the ROCO proteins. Nor

119

has any representative structure of a COR domains been solved to date. Like COR, no
Roc structures have been solved but, based on its amino acid sequence similarity with
GTPases of determined structures, the Roc domain is most similar to Rab7, a GTPase
involved mainly in vesicular trafficking. Based on the known structure of murine Rab7,
the Roc domain of LRRK2 has been modeled by Dr. William Wedemeyer (BMB,

Michigan State University) (Figure 3.2).

GTPases play crucial roles in signal transduction through regulation of protein
kinases and numerous other effector proteins [21][22, 23]. Based on the sequence of the
kinase domain, LRRK2 and MLK3 are closely related and both reside in the tyrosine
kinase like (TKL) subfamily. Members of the TKL subfamily, though they share
sequence similarity with tyrosine kinases, appear to function as either serine/threonine
kinases or, in some cases, as catalytically inactive scaffold proteins [24]. Our lab has
suggested that LRRK2 may be regulated in manner [25] somewhat similar to that which
we have described for MLK3 [26, 27]. In our model for MLK3 activation, binding of
GTP-bound Cdc42 or Racl through the CRIB motif disrupts an autoinhibitory interaction,
resulting in elevation of MLK3 kinase activity and subsequent activation of its
downstream pathways [26-28]. LRRK2 is interesting in that it contains both a putative
GTPase and kinase in the same polypeptide chain. It is conceivable that Roc domain
might activate LRRK2 protein kinase activity as Cdc42/Rae activates MLK3 as we
suggested [25]. Therefore, determining whether the Roc domain of LRRK2 functions as
a bonafide GTPase is a key to understanding LRRK2 regulation and this knowledge may

contribute to the development of therapeutic drugs for treating Parkinson’s disease.

120

 

LRRK2 Roc 1336 KLMIVGNTGSGKTTL ------- VWDFAGREEFMOI

RAB? 10 KVIILGDSGVGKTSL ------- IWDTAGQERF 7o
CD042 5 KCVWGDGAVGKTCL ------- LFDTAGQEDY 64
MS 5 KLWVGAGGVGKSAL ------- ILDTAGQEEY 64

Figure 3.2. Molecular model of GTPase domain of LRRK2 A homology model
of Roc domain of LRRK2 was generated by Dr. William Wedemeyer based on the
murine Rab7 structure (1VG8). TOP, residues mutated to generate the predicted active
forms of LRRK2 GTPase and dominant negative form of LRRK2 GTPase are shown in
green stick side chains and red stick side chain, respectively. R1441 is shown as a blue
stick side chain. GTP analog and Mg2+ are labeled as a sky blue side chain and an orange
sphere, respectively. Bottom, alignment of human Roc domain of LRRK2 with other
small GTPases. Identical residues are highlighted in solid yellow boxes and similar
residues are highlighted in the sky blue boxes. Residues usually mutated to generate a

conformationally active or a dominant negative GTPase are labeled with green and red.

121

Three LRRK2 R1441C/G/H mutations, two of which (R1441C/G) are known to
segregate with disease, result in amino acid substitutions of Arg 1441 within the Roc
domain. Based on the model of Roc, Arg 1441 is found on the surface of Roc, distant
from the predicted site of catalysis (Figure 3.2). Thus we have hypothesized that PD
associated substitutions, R1441C/G/H, disrupt a protein-protein interaction leading to
pathology associated with disease [25]. Tremendous effort from many labs has been
focused on LRRK2 since its recent association with PD was discovered. Work from
some labs, as well as the data presented in this Chapter, indicates that LRRK2 can bind
GTP through its Roc domain [12, 14-16, 18, 29]. Two studies have demonstrated that
LRRK2 in vitro kinase activity was slightly increased in the presence of non-
hydrolysable GTP [14, 16, 18, 29]. Under different experimental conditions and
contexts, different labs have reported different effects upon introduction of R1441C or
R14410 in the Roc domain on GTP hydrolysis [14, 15, 29]. In addition, whether PD
mutations within the Roc-COR tandem domains affect LRRK2 in vitro catalytic activity

remains controversial.

In the first part of Chapter 3, I present optimization of conditions for production
of recombinant GST-Roc in E. Coli. In addition, the recombinant GST-Roc is shown to
function as a bonafide GTPase, albeit with the low intrinsic GTPase activity, consistent
with other recent reports [14-16, 29]. The GTP-binding ability of Roc variants, expressed
in both mammalian and bacterial systems, was examined. Notably, I found that the PD-
associated mutation, R1441C, had no effect on GTP binding, consistent with our
molecular model which shows this residue is distant from the GTP binding site. This

finding led me to test the idea that the Roe-COR tandem domains might be involved in

122

 

intra- or inter-molecular interactions within LRRK2. The Roc-COR region was found to
mediate the intra or inter molecular protein interactions of LRRK2 through interaction
with the carboxyl terminal region of LRRK2 containing Roc, COR, kinase and WD40
domains. Introduction of R1441C/G reduced the interaction of Roc-COR with the C-
terrninal domains of LRRK2, whereas Y1699C enhanced it. The results shown herein
support the idea that LRRK2 domains interact with one another and that PD associated

mutations impact these protein-protein interactions.

123

3. Materials and Methods

3.1 Reagents and antibodies

The actin monoclonal antibody, V5 monoclonal antibody, hemagglutinin (HA)
rabbit polyclonal antibody, GTP conjugated agarose resin, GTP, ATP,CTP and Protein A
resin were from Sigma-Aldrich. Protein G resin and isopropyl-fi-D-thiogalacto-
-pyranoside (IPTG) were from Roche. Glutathione sepharose 4B affinity agarose beads
were purchased from GE Healthcare. Other antibodies used were the MLK3 antibody,
HA mouse monoclonal antibody (BAbCO), glutathione S—transferase (GST) mouse
monoclonal antibody (Santa Cruz), c-myc monoclonal antibody (Santa Cruz) and
horseradish peroxidase-conjugated secondary antibodies (Bio-Rad). Autoinduction

media was from Novagen.

3.2 Construction of expression vectors and mutagenesis

For bacterial expression vectors, the Roc (amino acid 1331-1521 of LRRK2),
Roc-COR (amino acid 1331-1862 of LRRK2) regions were amplified by PCR with

following primers and the pCDNA3.1-LRRK2-V5/His as the template.

Roe: 5 '-GCTGGATCCTACCTTATAACCGAATGAAAC-3 ',

5 '-CTCCTCGAGTCACTGTCCAACAACAAGCTG-3 '.

Roe-COR: 5'-GCTGGATCCTACCTTATAACCGAATGAAAC-3 ',
5'-CCGCTCGAGTCAGTCAGCCAAAATCAAGTA-3 '. The amplified Roe and Roc-
COR cDNAs were each inserted into the pGEX-SX vector (GE healthcare, NJ) using the

BamHI and XhoI restriction sites.

124

 

 

Mammalian expression constructs, for Roc (amino acid 1327-1521 of LRRK2),
Roe-COR (amino acid 1327-1836 of LRRK2), N-terminal half of LRRK2 (amino acid 1-
984), C-terminal half of LRRK2 containing Roc-COR-kinase-WD 40 through the end
(amino acid 1335-2527), LRR-Roc-COR-kinase (amino acid 984-2160), and kinase
domain of LRRK2 (amino acid 1879-2160) were similarly generated with the following

primers.

Roc: 5'-CGCTCGAGAAAAAGGCTGTGCCTTATAACC-3 ', F”

5 '-ATAAGAAATGCGGCCGCTCACTGTCCAACAACAAGCTGATC-3 '.

Roe-COR: 5 '-CGCTCGAGAAAAAGGCTGTGCCTTATAACC-3 ',

 

5 '-ATAAGAAATGCGGCCGCTCATTCCTCTGCTTTCTTCATCAA-3 '.

if

N-terminal half of LRRK2: 5'-GGGGCTCGAGGTATGGCTAGTGGCAGCTGTC-3 ',

5 '-ATAAGAATGCGGCCGCCTCTTACTCAACAGATGTTCG-3 '.

Roc-COR-kinase-WD 40 through the end of LRRK2:
5 ’-CGCTCGAGAAAAAGGCTGTGCCTTATAACC—3 ',

5 '-ATAAGAATGCGGCCGCCTCTTACTCAACAGATGTTCG-3 '

LRR-Roc-COR-kinase: 5 '-GTTCTCGAGCAATTACATCACTAGACCTTTCAG—3 ',
5 '-ATAAGAAATGCGGCCGCTCAGTGATGTGTAGCAACCATGC-3 '.
Kinase: 5 '-GTTCTCGAGGTCAAGCTCCAGAGTTTCTCCTAG-3 ',

5 '-ATAAGAAATGCGGCCGCTCAGTGATGTGTAGCAACCATGC-3 '.

The amplified cDNAs were individually inserted into pCMV-HA vector and pCMV-myc

vector (Clontech Laboratories, CA) using XhoI and NotI restriction sites.

125

The COR (amino acid 1522-1857 of LRRK2), Roe-COR (amino acid 1331-1857
of LRRK2) and WD40-end (amino acid 2197-2527 of LRRK2) cDNAs were amplified
by PCR and individually inserted into the pCDNA3.1/V5—His TOPO expression vector (a

gift from Dr. Julie Taylor, Michigan State University).

Variants of Roc and Roc-COR constructs were generated by the QuickChange
site-directed mutagenesis (Stratagene) using the following oligonucleotides and their
reverse complements. For Roc A1396T: 5’-TGGGATTTTACCGGTCGTGAG-3 ’;
Roc R1398L: 5 '-TTTGCAGGTCTGGAGGAATTC-3 ';

Roc T1348N: 5 '-AGTGGTAAAAACACCTTATTG-3 ';

Roc R1441 G: 5 '-TAAAGGCTGGCGCTTCTTC-3 ';

 

Roc R1441C: 5 '-TAAAGGCTTGCGCTTCTTC-3 ';
Roe-COR Y1699C: 5’-AAATGCCTTGTTTTCCAATG-3’. All constructs were

subsequently verified by DNA sequencing (MSU-GTSF facility).

3.3 Bacterial expression and purification of GST fusion proteins

GST-fusion proteins were expressed in Rostta (DE3) pLysS bacteria using either
the auto-induction method or IPTG induction. For chaperone coexpression, GST—fusion
proteins were induced in BL21 (DE3) bacteria containing pGTfZ or pGJKE8 plasmids (a
gift from Dr. Jennifer Ekstrom, Michigan State University) by IPTG induction. For auto-
induction, Rosetta (DE3) pLysS derivatives of E. coli containing the various recombinant
plasmids were grown at 37 °C in LB medium supplemented with arnpicillin (100 rig/ml)
and chloramphenicol (37 ug/ml ). Proteins were expressed in overnight express instant

TB auto-induction media (Novagen, CA) at 30°C for 16-18 h according to the

126

manufacturer’s instructions. For IPTG induction, Rosetta (DE3) pLysS bacteria
containing the pGEX—SX-l-Roc variant plasmids and pGEX-SX-l -Roc-COR plasmid
were cultured at 37 °C in the LB media supplemented with ampicillin (100 ug/ml) and
chloramphenicol (37 ug/ml ) overnight. The BL21 (DE3) bacteria containing the pGEX-
SX-l-Roc-COR recombinant plasmid and pGTf-2 or pGJKE8 plasmids were cultured at
37°C overnight in LB media supplemented with ampicillin (100 ug/ml) and
chloramphenicol (37 ug/ml) and inducers including 10 ng/ml tetracycline and/or 2 mM C"
L-arabinose for chaperone expression. When OD600 reached 0.6, protein expression was
induced by addition of 1 mM IPTG for 2 h. Cells were harvested by centrifugation at

4000 x g for 20 min, resuspended in 30 mL cold Tris buffer (50 mM Tris, pH 8)

 

containing the protease inhibitor cocktail and lysed by three passes through a French ;
pressure cell. GST fusion proteins were purified by affinity chromatography using

Glutathione sepharose 4B affinity agarose beads and eluted with the elution buffer (50

mM Tris-HCl, pH 8, 250 mM NaCl, 0.5 mM DTT, 10% glycerol, 20 mM reduced

glutathione). The concentration of eluted protein was determined by the Bradford method

using Bio-Rad protein assay kit with BSA as standard according to manufacturer’s

protocol.

3.4 Cell culture, transfection and cell lysis

Human embryonic kidney (HEK) 293T cells were transfected using
Lipofectamine 2000 (Invitrogen) following the manufacturer’s instructions. Cells were
maintained at 37 °C in a humidified incubator containing 5% C02 in air. Cells were

harvested 20—24 h after transfection and lysed either in lysis buffer G (100 mM Tris-HCl

127

pH 7.5, 50 mM KCl, 0.1 mM EDTA, 0.1 mM DTT, 5 mM MgC12, 1% Triton X-100) for
GTP binding assay or in the lysis buffer (50 mM HEPES (pH 7.5), 150 mM NaCl, 1.5
mM MgC12, 2 mM EGTA, 1% Triton X-100, 10% glycerol, 1 mM Na4PPi, 100 uM B-
glycerophosphate, 1 mM Na3VO4 and 1X protease inhibitor cocktail (Sigma Aldrich))

and kept on ice for immunoprecipitation .

3.5 Immunoprecipitation, gel electrophoresis and western blot analysis I?“

Antibodies were prebound to protein A or G agarose beads for 30 min at room
temperature: HA rabbit polyclonal antibody (0.15 [Lg/pl slurry), c-myc monoclonal

antibody (0.04 ug/ul slurry). Clarified cellular lysates from an xmm plate were incubated

 

with 20 pt] antibody-bound protein A or G agarose resin for 90 min at 4°C.
Irnmunoprecipitates were washed twice with HNTG buffer (20 mM HEPES (pH 7 .5),
150 mM NaCl, 0.1% Triton X-100, and 10% glycerol). Cellular lysates and
immunoprecipitates were resolved by SDS-PAGE. Proteins were transferred to
nitrocellulose membranes and immunoblotted using appropriate antibodies. Irnmunoblots

were developed by the chemiluminescence method.

3.6 GTP-binding assay

Aliquots (5 ug) of affinity isolated GST-Roe variants, except GST-Roe T1348N,
were incubated with 60 ul GTP-conjugated agarose beads in the GTP-binding buffer (50
mM Tris-HCI pH 7.6, 50 mM KCl, 0.1 mM DTT, 5 mM MgC12, 1% Triton X-100) at 4°C
for 1 h. Since GST-Roe T1348N was likely contaminated with dnaK (see Results, Figure

3.7A)), 50 pg of total protein was used to ensure that each sample contained the same

128

amount of. The supernatant was removed and GTP-conjugated beads were washed 3
times with ice-cold HNT G buffer. Proteins were eluted from GTP-conjugated beads by
boiling in SDS loading buffer and resolved by SDS-PAGE followed by immunoblotting

using a GST antibody.

In mammalian cells, the GTP-binding experiment was performed according to
Korr et a1. [30]. Briefly, HEK 293T cells transiently expressing HA-Roc or HA-Roc-
COR variant proteins were lysed in lysis buffer G containing protease inhibitor cocktail P
and clarified. Clarified total cellular lysates were incubated with GTP-conjugated agarose
beads in the lysis buffer G and bound HA-Roc variants were detected by Western blot as

described above.

 

'F p

Afier binding of Roc variants to GTP-conjugated resins, nucleotide competition
experiments were performed by adding the indicated concentration of GTP, ATP or CTP
for additional 1 h incubation at 4°C. GTP beads were then washed 3 times with ice-cold

HNTG buffer followed by Western blot analysis as described above.

3.7 GTPase activity assay

The GTPase assays were performed using the EnzChek phosphate assay kit
(Molecular Probes) according to the manufacturer’s instructions. This assay measures
phosphate released fi'om GTP hydrolysis by using purine nucleoside phosphorylase (PNP)
as coupling enzyme and 2-amino-6-mercapto-7-methylpurine riboside (MESG) as a co-
substrate. Typically, purified GST or GST-Roc variant proteins were incubated in 50 mM

Tris, pH 7.5 buffer including 5 mM MgC12, 0.2 mM MESG substrate and 1 U/ml PNP at

129

room temperature for 10 min. GTP hydrolysis was initiated by addition of 0.5 mM GTP
at 30°C and the reaction was followed by monitoring the absorbance at 360 nm in every

10 seconds using a 96—well plate spectrophotometer (SpectraMax plus, Molecular

Devices).

130

 

4. Results

4.1 Recombinant GST-Roc variants, but not GST-Roc-COR are expressed as

soluble proteins under condition of auto-induction

Studies of Ras farnily GTPases have demonstrated that a single amino acid
substitution within the catalytic site renders the GTPase either GTP-bound
(conformationally active) or GDP-bound (conformationally inactive). These GTPase
variants are commonly used to study their physiological roles in various cellular
processes and to identify their regulators, guanine exchange factors (GEFs) and GTPase
activating proteins (GAPS) (see Chapter 1). A single amino acid substitution of Gly at

codon 12 with Val [31—34], Ala at codon 59 with Thr [3l]or Gln at condon 61 [35, 36]

 

with Leu in Ras reduces the rate of GTP hydrolysis by the activated Ras GTPase and
constitutively activates Raf-MEK-ERK pathway. In contrast, substituting Asn for Ser at
codon 17 in Ras displaces the Mg” that is required for GTP nucleotide binding, resulting

in a higher affinity of binding GDP [37, 38].

To evaluate whether Roc is a bona fide GTPase and to test the proposed
hypothesis that the R1441C PD-associated LRRK2 mutation impacts protein interactions
rather than GTPase activity, mutations predicted to generate conformationally active and
inactive forms of Roc were introduced to the Roc cDNA in bacterial expression
constructs. For creation of a conformationally active form of Roc (GTPase defective),
the sequence of the LRRK2 Roc domain was compared with representative members of
the R33 superfamily: Rab7, Cdc42 and Ras. Notably, substitution of Gly 12 with Val
renders both Ras and Cdc42 conformationally active and competent to bind effector

proteins including protein kinases. Unfortunately, this critical Gly residue is not

131

conserved in LRRK2 or in Rab7, a closer relative of the Roc domain of LRRK2. In Rab7,
Cdc42 and Ras, mutation of the conserved Gln residue to Leu, Q67L, Q60L and Q6OL,
respectively, renders each of these GTPases conformationally active. Instead of Gln at
this position, the Roc domain has an Arg residue, R1398. However, a less well-known
Ras variant, A59T, renders Ras conformationally active. This Ala residue, A1396, is

conserved within Roc as well as the other representative GTPases (Figure 3.2).

In order to create conformationally active variants of Roc, two different mutations, I
A1396T and R1398L, were introduced into the Roc domain. Conforrnationally inactive

forms of Ras, Cdc42 and Rab7 have been created by substituting the conserved Ser/Thr

 

residue, corresponding to Ser17 in Ras to Asn. Therefore the corresponding conserved
Thr at 1348 in LRRK2 was substituted with Asn in order to create a predicted

conformationally inactive variant.

GST-Roe variants, GST-Roe A1396T and R1398L, predicted to be
conformationally active and the predicted inactive variant, GST-Roe T1348N, were
overexpressed in E. Cali upon IPTG induction. All GST-Roe variants expressed well but
were found mainly in the insoluble fraction (Figure 3.3). Wild type GST-Roc-COR
exhibited similar insolubility. Massive expression of protein in a short time as takes
place during IPTG induction may result in protein aggregation. Gradual expression of
proteins often renders them more soluble, allowing time for proper folding, resulting in a
higher enzymatic activity. In the autoinduction system, the growth media contain a
mixture of carbon source including lactose. After glucose is consumed by E. Cali,
lactose will be used, thus automatically inducing protein expression but at a much slower

rate than occuring upon acute addition of IPTG.

132

The solubility of expressed GST-Roe variants and GST-Roc-COR was
determined under autoinduction conditions. As shown in Figure 3.4A and 3.4C, all GST-
Roc variants as well as GST-Roc-COR expressed well under autoinduction conditions.
Wild type and predicted constitutively active GST-Roe R1398L were highly soluble
(Figure 3.4B), whereas the predicted inactive GST-Roe T1348N had limited solubility
and GST-Roc-COR was still completely insoluble. The PD associated variant R1441G
was introduced into wild type GST-Roc and GST-Roe R1398L. Both variants expressed
well upon autoinduction as shown in Figure 3.4C (two different clones shown for each
construct). A portion of the expressed variants was insoluble, but over 50% remained in

the soluble portion.

In a final effort to facilitate the solubility of Roe-COR, the expression vectors
encoding bacterial chaperones, pGTt2 and pGJKE8, were coexpressed with GST-Roc-
COR. The pGTt2 expression vector encodes three chaperones including groES, groEL
and tig and these chaperones are expressed upon tetracyclin induction [39]. The pGJKE8
plasmid encodes five chaperones including dnaK, dnaJ, grpE, groES and groEL and the
expression of these five chaperones is induced by addition of L-Arabinose and
tetracycline [39]. In the presence of chaperones, the protein expression and solubility of
GST-Roc-COR were assessed by SDS-PAGE followed by Coomassie blue staining. As
shown in Figure 3.5, GST-Roc-COR was expressed in the presence of different
chaperones upon IPTG induction. However, almost all of GST-Roc-COR coexpressed
with several chaperones still remained in the insoluble fraction. A summary of the

protein solubility

133

 

060
GST-Roe WT RL TN «9* WT
f—fi r—n I—\ r—\ (‘99 r—\
WTATRLTN PSPSPS WT PS
171kb—
110kD— :-
79kb—
60kb-
‘L in»
47kD-“uuu a “*‘h'u
35kD— .
25kD— :

 

Figure 3.3. Expression and solubility of GST-Roe variants and GST-Roc-
COR upon IPT G induction. Left panel, GST-Roe variant proteins and right panel,
GST-Roc-COR were individually expressed in Rosetta (DE3) pLysS cells upon IPTG
induction for 2 h at 37°C. The expression level and solubility of GST-Roe and GST-
Roc-COR were evaluated as described previously by Coomassie staining. WT: wild type,

RL: R1398L, TN: T1348N, P: pellet fraction (insoluble), S: supernatant fraction (soluble).

Equal cellular equivalents from each fi'action were loaded.

134

Figure 3.4. Expressed GST-Roc variant proteins but not GST-Roc-COR are
soluble under the auto-induction condition. pGEX-SX-l-Roc and pGEX-SX-l-
Roc-COR variant plasmids were individually transformed into Rosetta (DE3) pLysS cells.
GST-Roe and GST-Roc-COR variant proteins were induced by auto-induction media at
30°C for 16-18 h. A, bacteria were spun down and lysed in SDS-sample loading buffer.
Levels of expressed GST-Roe variants and GST-Roc-COR were evaluated by Coomassie
staining. B, bacteria were collected followed by sonication on ice as described in
Materials and Methods. Equal cellular equivalents of protein in supernatant (S) and
pellet (P) fractions were subjected to SDS-PAGE followed by Coomassie staining for
solubility assessment. Arrows indicated the GST-Roe variants and GST-Roc-COR,
respectively. C, Level of expressed GST-Roe harboring PD mutation was assessed as A.
D, as described in B, equal cellular equivalents of protein in each fraction was loaded and
the presence of GST-Roe with PD mutation in each fraction was checked by SDS-PAGE

with a Coomassie staining. WT: wild type, RL: R1398L, TN: T1348N, RG: R1441G.

135

c’0‘?" GST-Roe GST-Roc-COR
(A) 0’

 

9‘0 (B) WT RL TN WT
GST-ROC 098 {_1 I_‘ {—1 {—1
P S P S P S p S
WT RL TN WT
171kD— 171kb—
110kD— 110kD—
79kD— 79kD— " ‘ s'
fl ‘- 9" 1 r
GORD- soro— - - <-
l ' ' « "‘
47kD- 47ko— « a - .-
a on m . + ’* P - P .- ., g.
35m- . Q ; 35kD- . M t . - ‘
. , ..
25kD— f ' 25kD— ~ ._ f "f
u—u'. .- a.
(C) GST-Roe (D) GST-Roe
R" R6 RL RG
(fl r—fi P S ‘ ‘ '—‘
Cl 1 2 P S P s P
one 1 2 Clone 1 2 1 2
180kD— 13mm-
13°kD— 130kb-
79kD— 79kg-
62kD- 62kD—
. - - -
47kD— . u - - 47kD— .0 ~ ~ .
35kD— 351(0—
- ..- ” 'I‘ u —
25kD— 25'“).

136

of bacterially expressed GST-Roc variants and GST-Roc-COR under various conditions

is presented in Table 3.1.

Soluble GST-Roc variant proteins were purified from cellular lysates by
glutathione affinity chromatography and checked for their purity by SDS-PAGE followed
by Coomassie blue staining. The control proteins, GST and GST-Cdc42 G12V, were
efficiently eluted from glutathione agarose beads. About half of the GST-Roe variants

were obtained from an eluted fi'action (Figure 3.6).

4.2 Roc domain of LRRK2 binds GTP

Purified GST-Roe variant proteins were tested for their ability to bind GTP-
conjugated resin. Wild type as well as predicted conformationally active variants,
A1396T and R1398L of, GST-Roe bound GTP-conjugated agarose, suggesting that
bacterially expressed Roc is a GTP-binding protein. As predicted, GST-Roc T1348N
failed to bind GTP (Figure 3.7A). However, it is unclear whether the bacterially
expressed GST-Roe T1348N was correctly folded since a massive 60 kD protein band,
likely the bacterial dnaK heat shock protein which often copurifies with unfolded GST
fusion proteins in bacteria (Figure 3.7A). To confirm these results, HA-tagged Roc
variants were also transiently expressed in mammalian HEK 293T cells and tested for
their ability to bind GTP conjugated agarose. In HEK 293T cells, the HA-Roc proteins
were all expressed at similar levels (Figure 3.7B). Consistent with the bacterially
expressed GST-Roe variants, both wild type and its predicted conformationally active

variant, R1398L bound GTP-conjugated beads. Intriguingly, one of the predicted

137

chaperone: pGTf2 pGJKE8
chaperone: pGTf2 pGJKE8 IPTG (h): 1
F——\ r———\ rfl

2 1 2
(_\ {—1 {fl
IPTG(h)I - 1 2 - 1 2 p S p S p S p S

1711(0-
1101(0-

79kD — a... u...

BORD—
47kD —

351(0—

251(0—

Figure 3.5. Expression and solubility of GST-Roc-COR by IPTG induction
with coexpressed chaperones. GST-Roc-COR was coexpressed with either pGTf2 or
pGJKE8 chaperones in BL21 (DE3) cells by IPTG induction for the indicated times at
30°C. pGTf2 encodes 3 bacterial chaperones, groES, groEL and tig. pGKJ E8 encodes 5
bacterial chaperones, dnaK, dnaJ, dnaE, groES and groEL. Left panel, the induced level
of GST-Roc-COR with different chaperones at different induction time points was
assessed by Coomassie staining. Right panel, the solubility of GST-Roc-COR with
expressed chaperones was determined as described in Materials and Methods. P: pellet
fraction (insoluble), S: supernatant fraction (soluble). Equal cellular equivalents of

protein from each fraction were loaded

138

Table 3.1. Summary of the protein solubility of bacterial GST-Roc and GST-

Roc-COR variant proteins under various induced conditions.

 

 

 

 

 

 

 

 

 

 

 

construct mutation bacterial strain induction chaperones protein
solubility
no Rosetta (DE3) pLysS auto no soluble
A1 396T Rosetta (DE3) pLysS auto no soluble
R1 398L Rosetta (DE3) pLysS auto no soluble
Mfg-)2» T1 348N Rosetta (083) pLysS auto no soluble
R1441 G Rosetta (DE3) pLysS auto no soluble
R1 398L,
R1441 G Rosetta (DE3) pLysS auto no soluble
Rosetta (DE3) pLysS auto no insoluble
pG EX-5x-
1-Roc- no Rosetta (DE3) pLysS IPTG pGro insoluble
COR
pG-JKEB or .
BL21 (DEa) IPTG pG-Tf2 Insoluble

 

 

 

 

 

 

139

 

(A) (B)

 

 

>
> N _J I'-
> -' *- 9 *3 's a 8 8
$3 8 8 0 g") co ‘2 $2
(9 $3 2 N E E g m <
8 °‘ ‘5. '5 s 8 s 3 8
3 o 8 a: c: BSA .— 9 ‘F “7’ BSA
3 “F t? '- .— .1 .L m a. t; :,-,
17: 0": 5; t5 8 8 w <0 (119) 0 o o 0 (its)
0 (D o o ,—‘ r—r r3 '9‘ 1 2 I 1 f l r i F fl 1 2
171kD E1 E2 5152 5152 £152 8 £152 3 E1 E2 8 £152 3 E162
110kD-
791(0-
SOKD- on“
47kD-
H" -*-ooooo- 3:... “any“.
35kD- u . w
1

Figure 3.6. Purification of recombinant GST-Roc variant proteins induced
by auto-induction condition. GST, GST-Cdc42 612V and GST-Roe variant proteins
were induced by auto-induction media at 30°C for 16-18 h, and bacteria were lysed using
a French pressure cell. GST and GST-fusion proteins were incubated with glutathione
sepharose 4B beads at 4°C for 1.5 h followed by washing. Proteins attached to the
glutathione beads were eluted with appropriate buffer containing 0.5 mM DTT and 20
mM reduced glutathione. Proteins from beads (B) and each elution (E1, E2) were
assessed by Coomassie blue staining. GST and GST-Cdc42 G12V proteins were used as
controls. Equal cellular equivalents from each fraction were loaded. Protein
concentration was determined using a Biorad protein assay kit and re-confirmed with

Commassie blue staining as compared to the known amount of BSA.

140

Figure 3.7. The GTP-binding ability of Roc variant proteins. A, top, 5 ug purified
GST, GST-Roe variant proteins and GST-Cdc42 from E. Cali was incubated with GTP-
conjugated agarose beads. The level of GST-fusion proteins bound to GTP agarose beads
was assessed by Western blotting using the GST antibody. GST-Cdc42 protein was used
as a positive control. Bottom, the total amount of each GST-fusion protein used in the
GTP-binding experiments was examined by Western blotting with the GST antibody. B.
GTP-binding experiment was performed using total cellular lysates from HEK 293T cells
expressing various HA-Roc proteins. The level of HA-Roc variant proteins bound to the
GTP-conjugated agarose beads was accessed by Western blotting using a HA antibody.
The expression levels of various HA-Roc constructs and the loading control were
examined by immunoblotting using the antibodies against HA and actin, respectively.

WT: wild type.

141

(A)

 

GST- Roe
‘ a
Gav“ C? be
"’ ‘- "’ I

GST WB

 

GST fu . d u e- d ..
- sron
protein loading 4’ GST WB

1711(0—
1101(0-

79kD —

6010- u
47”). --- N ..-. Coomassie Blue
Staining
35kD—
25kD- " se. a
(B)
’5 K
HA-Roc: ' V5 95 4" ‘3“ 8*
GTP pull down ~—

"'""’ """" HA WB

u.-- HA WB

cellular lysates

”--“fl’ actinWB

142

conformationally active Roc variants, A1396T, bound to GTP resin inefficiently when
isolated from bacteria and cultured cells, suggesting that this mutation may not render
Roc conformationally active. These data suggest that Roc domain of LRRK2 is a GTP-
binding protein. Consistent with predictions, Arg at 1398 to Leu of Roc is able to bind

GTP, whereas replacement of Thr at 1348 to Asn fails to bind GTP.

Furthermore, the Roc domain containing the PD-associated substitution, R1441C,
bound GTP resin similarly to wild type Roc. This finding is consistent with our model

which shows that Argl44l is located far from the GTP binding site [25].

4.3 COR region affects the GTP binding of Roc

Because the Roc domain always exists in tandem with the COR region in proteins
of the ROCO family, they may be firnctionally interdependent. To date, the firnction of
COR regions in the ROCO proteins is a mystery. To decipher whether the presence of the
COR region regulates the ability of Roc to bind GTP, cellular lysates from HEK 293T
cells transiently expressing HA-tagged Roe-COR proteins with different mutations were
assessed in the GTP-binding assay. As shown in Figure 3.8, HA-Roc-COR R1398L,
predicted to be conformationally active, bound GTP conjugated resin slightly better than
the wild type HA—Roc-COR and its mutants including A1396T and T1348N. Surprisingly,
HA-Roc-COR T1348N, the predicted conformationally inactive variant which failed to
bind GTP-agarose in the context of HA-Roc, binds GTP-conjugated agarose as efficiently
as wild type HA-Roc-COR. This may indicate that within the Roe-COR context, the

T1348N variant either nonspecifically associates with GTP resin or the presence of the

143

HA-Roc-COR HA-Cdc42

 

f « VN r—fi
‘b
9 Q
43‘ 5'59 Q55 «5".» 0'0 «52“
GTP pull down HA we
.-
HA WB
cellular lysates
Ku- .- —— ' -- actinWB

 

Figure 3.8. The GTP-binding ability of Roe-COR variant proteins. Total
cellular lysates from HEK 293T cells expressing various HA-Roc-COR proteins and HA-
Cdc42 variants were used in GTP-binding experiments as described in Materials and
Methods. HA-Roc-COR variant proteins bound to the GTP-conjugated agarose beads
were detected by western blotting using a HA antibody. The expression levels of various
HA—Roc-COR constructs and the loading control were examined by immunoblotting
using the antibodies against HA and actin respectively. HA-Cdc42 G12V and HA-Cdc42

T17N were used as the positive and negative controls, respectively. WT: wild type.

144

COR region somehow alters the conformation of the Roc domain to facilitate GTP

binding despite the presence of the T1348N mutation. Direct comparison of the GTP

binding ability of HA-Roc and HA-Roc-COR proteins in parallel will be necessary to

firrther address whether COR regulates the GTP binding of Roc.

4.4 The GTP-binding specificity of Roc

The nucleotide specificity of Roc variants was tested in a competition assay. The
interaction of Roc variants with GTP resin was abolished upon addition of increasing
amount of flee GT P (Figure 3.9A). Moreover, addition of GTP, but not other nucleotides
including ATP or CTP led to release of Roc fiom GTP resin (Figure 3.93 and 9C),

suggesting that Roc preferentially binds GTP, but not other nucleotide triphosphates.

4.5 Hydrolysis of GTP by the Rec domain of LRRK2

To assess whether the Roc domain of LRRK2 possesses GTPase activity, an
enzymatic coupling method to detect free phosphate generated from GTP hydrolysis was
employed. Enzymatic conversion of free phosphate and 2-amino-6-mercapto—7-
methylpurine ribose (MESG) by purine nucleoside phosphorylase (PNP) produces 2-
amino-6—mercapto-7-methylpurine which absorbs at 360 run. As shown in Figure 3.10,
readings of absorbance at 360 run over time increased in the presence of purified GST-
Roc, suggesting that GST-Roe hydrolyzed GTP. Notably, no increase in absorbance was

observed using GST.

4.6 Interaction of Roe-COR region with the carboxyl terminal region of
LRRK2

145

(A)

 

 

 

GST-Roe: A1396T R1398L WT
[GTP] (mM) ‘ ' 1 l l 1
competition: ' 5 1° ' 5 1° " 5 1°

.. — — GST WB
GST-ROC. A1 396T R1398L WT
nucleotide Q Q Q Q Q Q Q Q Q

competltlon: ' (3‘ 4‘ 0‘ ' (3‘ 65 0" ' 0‘5“ 0‘

- _ __ - .- .. csr we
(3)

HA-ROC: A1 3961' R1 398L WT
nucleotide Q «Q «Q «Q «Q «Q «Q «Q .8
competition: - (3‘ Y’ O 0 V 0 ' 0 V‘ O

.— -- ~--- -- _— HAWB

Figure 3.9. GTP-binding specificity of Roc variant proteins. One microgram
purified GST-Roe variant proteins was incubated with GTP conjugated agarose beads for
90 min at 4°C followed by nucleotide competition with A, different concentrations of
exgenous GTP as indicated or B, 5 mM indicated nucleotides for 1 h. Levels of GST-
Roc variants retained in the GTP resin afier competition were evaluated by western
blotting using the GST antibody. C, cellular lysates from HEK 293T cells transiently
expressing HA-Roc variant proteins were used in the nucleotide competition assay and
analyzed as described above. Data shown is representative of three independent

experiments.

146

 

 

 

 

 

 

? t
0 so 100 150 zoo
Time (see)

Figure 3.10. The GTP hydrolysis ability of Roc. The GTPase activity of Roc
was assessed using an Enzchek phosphate assay kit. One micrograrn purified GST-Roe
was used in the reaction as described in Materials and Methods. 0.5 mM GTP was
added to initate GTP hydrolysis reaction at 30°C. Readings of absorbance at 360 nm in

every 10 seconds were plotted. Purified GST was used as a negative control.

147

In addition to the protein kinase domain, LRRK2 also possesses several predicted
protein-protein interaction domains. Emerging genetic studies have indicated that
mutations in these potential protein interacting regions result in familial Parkinson’s
disease, suggesting a pivotal role for LRRK2 protein interactions in the pathogenesis of
PD. In familial PD, three different mutations impact Arg 1441 resulting in substitutions
with Cys, Gly or His. Since three distinct amino acid substitutions at Arg 1441 result in
pathogenesis, we have hypothesized that, rather than altering some catalytic protein of
LRRK2, they might disrupt protein interactions to cause Parkinson’s disease. This idea is
supported by molecular modeling of Roc which predicts that Arg 1441 resides on the Roc

surface distant from the catalytic site where protein interactions might occur. It is

 

'T’i: -- .

conceivable that PD mutations in the Roe-COR region might affect its interaction with

other parts of LRRK2, impacting LRRK2 kinase activity and mediating pathogenesis.

Many multi-domain protein kinases adopt conformations that involve domain-
domain interactions within the same protein kinase. To test whether the Roe-COR
domain interacts with other regions of LRRK2, expression vectors for the NHz—terminal
or COOH-terminal halves of LRRK2 were cotransfected with Roe-COR in HEK 293T
cells and co-immunoprecipitation experiments were performed. The N-terminal half of
LRRK2 includes the NHz-terminal through the m repeat domain (N -Ankyrin), and
the C-terminal half of LRRK2 encompasses the _Roc domain, COR region, _lginase domain
and ED 40 repeats through the end (RCKWE). RCKWE, but not N-Ankyrin, interacted
with the Roc-COR domain (Figure 3.11A). The central region of LRRK2 containing the

LRR, Roe, _COR and kinase domains (LRCK), but lacking the C-terminal WD40 domain,

148

Figure 3.11. Interaction of Roe-COR with different regions of LRRK2.
Different LRRK2 truncation proteins were transiently coexpressed with or without Roc-
COR in HEK293T cells for 20-24 h. The interaction of different LRRK2 truncation
variants including (A) the N- and C-terminal halves as N-Ank and RCKWE respectively,
(B) the region consist of LRR-Roc-COR-Kinase, (C) WD 40 repeats of LRRK2 with
Roc-COR was assessed by co-immunoprecipitation assay. Levels of expressed proteins
and of endogenous actin as a loading control, were detected by Western blotting with the
indicated antibodies. N-Ank: N-tenninal half of LRRK2 up to Ankyrin, RCKWE: Roc-
COR-kinase-WD40 to the end of LRRK2, LRCK: LRR-Roc-COR-kinase, WD: WD 40

repeats, R-C: Roe-COR. Data shown is representative of three independent experiments.

149

(A) (B)
V5-RC

e a- 5’
~6ng
o 0
she-3e:

HA-LRRKZ: -
IP: HA V5 WB
. . '0 _- HA we
_ -
" "" “ V5 WB

_———-——--- actinWB

(C) e (0
HA: 49 99- - e9 e9 -
V5: - - RC we RC we -
f : ' «lgG
V5 WB
IP: HA
- .-
HA WB
I c- O

 

150

V5-RC

ofoci‘fl 602$

HA-LRRKZ: - ,3- a.

" V5WB
..- -- HAWB

IP: HA

—-

u“ HAWB
n "'V vswe

-- can... aCtin WB

«5‘2.

HA: - qS’Q- - - Q-C’Qs’
V5: - - -RCWERCWE
- -

HAWB

. .-

-

V5WB

. -

\ ____._.— actin WB

failed to interact with Roe-COR in the co-immunoprecipitation assay (Figure 3.113),
suggesting that the region containing WD-40 domain through the end (WE) of LRRK2 is
required for interaction with Roc-COR. To test whether the WE region is sufficient to
interact with Roc-COR, HEK 293T cells transiently expressing WE and Roc-COR of
LRRK2 were used in the co-immunoprecipitation assays. No WE region was detected in
the Roc-COR immunoprecipitates (Figure 3.11C). Taken together, these data suggest
that the RCKWE interacts with Roe-COR and the WE region is necessary, but not

sufficient, for this interaction.

4.7 Effect of PD mutations of Roe-COR on the protein interaction of Roc-
COR with RCKWE of LRRK2

To dissect which region within Roe-COR is responsible for its interaction with
RCKWE of LRRK2 and how PD mutations within the Roe-COR region affect this
association, Roc and Roc-COR domains with or without PD mutations were assessed in
the co-immunoprecipitation experiments. As shown in Figure 3.12, more RCKWE is
detected in the Roc-COR immunoprecipitates than in the Roc immunoprecipitates,

suggesting the association of RCKWE with Roc is weaker than its association with Roc-

COR.

The impact of PD-associated mutations on interactions among LRRK domains
was assessed in co-immunoprecipitation experiments. Introduction of the PD-associated
mutations in Roc, Roc R1441C and Roc R1441G, dramatically reduced the association
with RCKWE (Figure 3.13A). Consistently, substitution of Arg 1441 to Cys and/or Gly

in the Roc-COR region also diminished the association of Roc-COR with RCKWE

151

HA-RCKWE
$0 ’0 ‘
- ,3}? :‘S 0““ (K350 $80
- HA WE

an.“ «:96

Myc we

K ---~. q-lgG

IP: Myc

 

Myc we

..- -- a— actin WB

Figure 3.12. Association of Roc with Roc-COR-kinase—WD 40 through the

end of LRRK2. Expression vector consists of either Roc or Roc-COR was

cotransfected with Roc-COR-kinase-WD 40 to the end of LRRK2 construct in HEK

293T cells for 24 h. The presence of RCKWE in the Roc or Roe-COR

immunoprecipitates was evaluated by Western blotting using the HA antibody. Levels of

expressed proteins and of endogenous actin as a loading control, were detected by

Western blotting with the indicated antibodies.

152

Figure 3.13. Effect of PD mutations found in the Roe and COR domains on
their interaction with the region of Roc-COR-kinase—WD 40 through the end
of LRRK2. Roc or Roe-COR variant proteins harboring PD mutations were
coexpressed with HA-RCKWE in HEK 293T cells. The presence of RCKWE in the Roc
or Roc-COR immunoprecipitates was assessed by immunoblotting using the HA antibody.
Clarified lysates were analyzed by Western blotting with appropriate antibodies for the

levels of expressed proteins and of endogenous actin as a loading control. WT: wild type,

RC: R1441C, RG: R1441G, YC: Y1699C.

153

(A)

 

HA-RCKWE: - + - +
Myc-Roe: - - wt RC RG wt RC RG
"" * HA WB
IP: Myc

- -- ----- Myc we

"" “I.-- HA WB

-----—--—--- actinWB

(B)

HA-RCKWE: - + - *-
r'___'fi r——'—\
Myc-Roc-COR: - - wt RC RG YC wt RC RG YC
r
". I HA WB (short exposure)
IP: Myc HA WB (long exposure)

M Myc WB

 

__ an--- HAWB

154

(Figure 13B) although the effect was less dramatic than for Roc. Interestingly, Roe-COR
Y1699C which harbors a PD-associated mutation within the COR region more efficiently

complexed with RCKWE of LRRK2 than did its wild type counterpart.

155

5 Discussion

Small GTPases are key signal transducing proteins regulating diverse cellular
processes through interactions with key effector proteins, including protein kinases.
Dysregulation of GTPases, or their regulators, is commonly associated with human
disease, especially cancer. LRRK2 is a large protein containing a putative protein kinase
and GTPase which, when mutated, can cause Parkinson’s disease. In the present study,
we provide evidence that the Roc domain of LRRK2, expressed as a GST fusion in
bacteria, functions as a bonafide GTPase. However, solubility issues and variability of
some preparations of some LRRK2 domains and variants produced in E. coli suggest that
alternative expression systems, such as insect cells, may prove useful for kinetic and

structural analyses.

In order to determine whether the GTPase activity and/or nucleotide binding
properties of the Roc domain might influence LRRK2 activity and/or pathogenesis, we
sought to generate Roc variants with different GTP binding/hydrolysis properties, based
on known properties of variants of small GTPases of the Ras superfamily. As expected,
the predicted conformationally inactive variant, T1348N, failed to bind GTP. Two
predicted conformationally active variants, R1398L and A1396T, were also generated.
The R1398L variant bound GTP as predicted whereas the A1396T exhibited variable
GTP binding properties depending on its context. These data are consistent with the
notion that the R1398L mutation renders Roc conformationally active. GTPase assays

will be needed to fully characterize this Roc variant.

156

Introduction of the PD-associated mutation, R1441C, had no effect on the GTP
binding of Roc, consistent with our hypothesis that this PD mutation disrupts protein
interactions rather than influences the Roc GTP binding and/or GTPase activity. It is
thus conceivable that the Roc domain of LRRK2 interacts with other cellular proteins and
that, in Parkinson’s disease, these Roc-mediated interactions are disrupted leading to
pathogenesis. Since LRRK2 is such a large multidomain protein, it seemed plausible that
the R00 domain might interact with another domain(s) within LRRK2 and that PD- F
associated mutations might impact these domain-domain interactions. Indeed
coimmunoprecipitation experiments indicate that the tandem Roc-COR domains are able

to interact with a carboxyl terminal region of LRRK2 encompassing Roc-COR, kinase

 

domain and WD 40 repeats though the end. Thus LRRK2 may be regulated through
intra- or inter-molecular protein interactions involving its Roc-COR region. Furthermore,
pathogenic mutations within the Roe-COR region impacted these interactions.

R1441C/G of Roc domain slightly attenuated Roe-COR interaction with the RCKWE of

LRRK2, whereas Y1699C located in COR region dramatically enhanced this interaction.

LRRK2 is a member of the ROCO family of proteins. ROCO proteins, which
contain both Roc and COR domains, are found in prokaryotes, plants and metazoa.
Considerable domain diversity exist outside of the Roe and COR regions of ROCO
proteins. Very few ROCO proteins have been studied in any detail, and their biological
functions are largely unknown. In particular, no biochemical function has been described
for any COR domain. Results fiom our GTP-binding study have demonstrated that the
efficiency of GTP binding of Roc variants was influenced when it was in conjunction

with COR, giving a possibility that COR may regulate the GTP binding ability of Roc. A

157

direct comparison of the GTP binding ability of Roe and Roc-COR will be crucial for
addressing this hypothesis. In addition, the studies presented here mapping the region
within Roc-COR responsible for its interaction with RCKWE, indicate that Roc interacts
with RCKWE much weakly than Roc-COR. These results may suggest that this
interaction may primarily mediated by COR. Future experiments assessing the ability of
COR to interact with RCKWE in the co-immunoprecipitation assay will provide key

clues for understanding the role of COR in LRRK2 regulation.

Recently, the GTPase activity of Roc has also been demonstrated by other
research groups using different experimental contexts. It has been demonstrated that
expressed full length human LRRK2 hydrolyzed GTP in a slow rate as judged by the thin
layer chromatography assay [14][29]. Disruption of the nucleotide binding site in the
Roc domain of full length LRRK2 by substitution of Lys 1347 to Ala failed to hydrolyze
GTP [14]. These data suggest that the Roc domain in the fill length LRRK2 construct is
a GTPase. Introduction of the PD-associated mutation, R1441C, in full length LRRK2
reduced the GTP hydrolysis rate [14, 29]. In these studies, the low expression level of
full length LRRK2 may be the cause of the low GTPase acitivity. In addition, unequal
amount of the full length LRRK2 containing Roc mutations used in GTPase assay makes
quantitaion more difficult. Li et a1. has reported that the purified Roc domain of murine
LRRK2 from bacteria possessed a GTPase activity [15], consistent with my finding
presented in this Chapter. The Roc T1348N of murine LRRK2 which failed to bind GTP
did not hydrolyze GTP [15]. The GTPase activity was moderately reduced in Roc
R1441C and R1441G of murine LRRK2 [15] which is consistent with the result in the

full length LRRK2 study. However, a recent study from other lab has demonstrated that

158

bacterially expressed GST-Roc R1441C has same GTP hydrolysis rate as wild type Roc
[29]. Altogether, these data suggest the Roc domain is able to hydrolyze GTP. However,
the effect of PD-associated mutation, R1441C/G, on GTPase activity of Roc remains
unconclusive. In this Chapter, I also demonstrated that the Roc domain of human
LRRK2 functions as a bonafide GTPase. The intrinsic GTPase activity of Roc is
relatively low in vitro which is consistent with previous studies of Ras and Rho subfamily
GTPases as well as Roc. An insect cell expression system and the mammalian
expression system stimulated by stresses are commonly used to evaluate the ability of
GTP hydrolysis of Ras and Rho subfamily GTPases since these activated GTPases
usually undergo post-translational modifications including prenylation [40] with a
farnesyl [41] or geranylgeranyl [42] group which renders them hydrophobic and be
targeted to cellular membranes. Loss of appropriate cellular structure(s) such as lipids
during purification procedure may result in the low GTPase activity of Roc. Based on
this, the insect cell expression vectors of His-tagged Roc variants and Roe-COR are
under construction in the lab. Moreover, the GTP hydrolysis ability of purified Roc
variants as well as Roe-COR from insect cells will be tested and compared in parallel
with that from bacteria. Roc is very distinct from Ras and Rho subfamilies and no
potential prenylation sites are predicted in the Roc-COR region. However, previously, we
also proposed that the LR and WD40 domains have high net positive charge located on
the surface, allowing LRRK2 possibly bind phospholipids [25]. It suggests that the LRR

and/or WD40 domain might be required for higher Roc GTPase activity.

Emerging biochemical studies from last couple years have mainly focused on

several particular pathogenic mutations of LRRK2 elevated its catalytic activity in vitro

159

resulted in morphological degeneration of neurons. However, whether mutations within
Roe-COR region of LRRK2 directly impact LRRK2 kinase activity in vitro responsible
for neuronal cell death remains obscure. In our study, the Roc domain of LRRK2 not
only functioned as a real GTPase but also competently interacted with the RCKWE
region of LRRK2, suggesting that the Roc domain involved in regulation of LRRK2 itself
through an intra or inter molecular interaction. Interestingly, this interaction was
attenuated by the PD mutations of Roc, R1441C/G, whereas it was increased by the PD
mutation of COR, Y1699C, suggesting that the alteration of the LRRK2 itself domain-
domain interactions may contribute to the pathogenesis of PD. These data presented here
provide important information to obtain a novel mechanism of LRRK2 regulation and
open a possibility that the dysregulation of LRRK2 protein itself interaction mediated by

Roc-COR domain may lead to Parkinson’s disease.

160

6

fl
0

10.

11.

12.

References

F ahn, 8., Description of Parkinson 's disease as a clinical syndrome. Ann N Y Acad
Sci, 2003. 991: p. 1-14.

Cookson, M.R., The biochemistry of Parkinson 's disease. Annu Rev Biochem, 2005.
74: p. 29-52.

Sherer, T.B., et al., An in vitro model of Parkinson's disease: linking mitochondrial

impairment to altered alpha-synuclein metabolism and oxidative damage. J
Neurosci, 2002. 22(16): p. 7006-15.

Moore, D.J., et al., Molecular pathophysiology of Parkinson 's disease. Annu Rev
Neurosci, 2005. 28: p. 57-87.

Paisan-Ruiz, C., et al., Cloning of the gene containing mutations that cause PARK8-
linked Parkinson's disease. Neuron, 2004. 44(4): p. 595-600.

Zirnprich, A., et al., Mutations in LRRK2 cause autosomal-dominant parkinsanism
with pleomorphic pathology. Neuron, 2004. 44(4): p. 601-7.

Berg, D., et al., Type and frequency of mutations in the LRRK2 gene in familial and
sporadic Parkinson ’3 disease *. Brain, 2005. 128(Pt 12): p. 3000-11.

Khan, N.L., et al., Mutations in the gene LRRK2 encoding dardarin (PARK8) cause
familial Parkinson's disease: clinical, pathological, olfactory and functional
imaging and genetic data. Brain, 2005. 128(Pt 12): p. 2786-96.

Mata, I.F., et al., erk2 pathogenic substitutions in Parkinson's disease.
Neurogenetics, 2005. 6(4): p. 171-7.

Gloeckner, C.J., et al., The Parkinson disease causing LRRK2 mutation 12020T is
associated with increased kinase activity. Hum Mol Genet, 2006. 15(2): p. 22332.

Greggio, B., et al., Kinase activity is required for the toxic eflects of mutant
LRRK2/dardarin. Neurobiol Dis, 2006. 23(2): p. 329-41.

Ito, G., et al., G]? binding is essential to the protein kinase activity of LRRK2, a

causative gene product for familial Parkinson '3 disease. Biochemistry, 2007.
46(5): p. 1380-8.

161

13.

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

J aleel, M., et al., LRRK2 phosphorylates moesin at threonine-558: characterization
of how Parkinson 's disease mutants aflect kinase activity. Biochem J, 2007.
405(2): p. 307-17.

Lewis, P.A., et al., The R1441 C mutation of LRRK2 disrupts GT P hydrolysis.
Biochem Biophys Res Commun, 2007. 357(3): p. 668-71.

Li, X., et al., Leucine-rich repeat kinase 2 (LRRK2)/PARK8 possesses GTPase
activity that is altered in familial Parkinson's disease R1441C/G mutants. J
Neurochem, 2007. 103(1): p. 238-47.

Smith, W.W., et al., Kinase activity of mutant LRRK2 mediates neuronal toxicity.
Nat Neurosci, 2006. 9(10): p. 1231-3.

West, A.B., et al., Parkinson 's disease-associated mutations in leucine-rich repeat
kinase 2 augment kinase activity. Proc Natl Acad Sci U S A, 2005. 102(46): p.
16842-7.

West, A.B., et al., Parkinson 's disease-associated mutations in LRRK2 link

enhanced GT P-binding and kinase activities to neuronal toxicity. Hum Mol Genet,
2007. 16(2): p. 223-32.

MacLeod, D., et al., The familial Parkinsonism gene LRRK2 regulates neurite
process morphology. Neuron, 2006. 52(4): p. 587-93.

Bosgraaf, L. and P.J. Van Haastert, Roc, a Ros/GTPase domain in complex proteins.

Biochim Biophys Acta, 2003. 1643(1-3): p. 5-10.

Marshall, C.J., eta1., Signal transduction by p21ras. Int J Cancer Suppl, 1989. 4: p.
29-3 1 .

Lim, L., et al., Regulation of phosphorylation pathways by p21 GT Pases. The p21
Ras-related Rho subfamily and its role in phosphorylation signalling pathways.
Eur J Biochem, 1996. 242(2): p. 171-85.

Schiller, M.R., Coupling receptor tyrosine kinases to Rho GTPases--GEFs what's
the link Cell Signal, 2006. 18(11): p. 1834-43.

Manning, G., et al., The protein kinase complement of the human genome. Science,
2002. 298(5600): p. 1912-34.

Mata, I.F., et al., LRRK2 in Parkinson's disease: protein domains and fimctional
insights. Trends Neurosci, 2006. 29(5): p. 286-93.

162

26.

27.

28.

29.

30.

31.

32.

33.

34.

35.

36.

37.

38.

Bock, B.C., et al., Cdc42-induced activation of the mixed-lineage kinase SPRK in
vivo. Requirement of the Cdc42/Rac interactive binding motif and changes in
phosphorylation. J Biol Chem, 2000. 275(19): p. 14231-41.

Du, Y., et al., Cdc42 induces activation loop phosphorylation and membrane
targeting of mixed lineage kinase 3. J Biol Chem, 2005. 280(52): p. 42984-93.

Xu, Z., et al., The MLK family mediates c-Jun N-terminal kinase activation in
neuronal apoptosis. Mol Cell Biol, 2001. 21(14): p. 4713-24.

Guo, L., et al., The Parkinson 's disease-associated protein, leucine-rich repeat
lcinase 2 (LRRK2), is an authentic GT Pase that stimulates kinase activity. Exp
Cell Res, 2007. 313(16): p. 3658-70.

Korr, D., et al., LRRK] protein kinase activity is stimulated upon binding of GT P to
its Roc domain. Cell Signal, 2006. 18(6): p. 910-20.

Gibbs, J .B., et al., Intrinsic GT Pase activity distinguishes normal and oncogenic ras
p21 molecules. Proc Natl Acad Sci U S A, 1984. 81(18): p. 5704-8.

Manne, V., E. Bekesi, and HF. Kung, Ha-ras proteins exhibit GT Pase activity:

point mutations that activate Ha-ras gene products result in decreased GT Pase
activity. Proc Natl Acad Sci U S A, 1985. 82(2): p. 376-80.

McGrath, J .P., et al., Comparative biochemical properties of normal and activated
human ras p21 protein. Nature, 1984. 310(5979): p. 644-9.

Sweet, R.W., et al., The product of ras is a GT Pase and the T24 oncogenic mutant
is deficient in this activity. Nature, 1984. 311(5983): p. 273-5.

Der, C.J., T. Finkel, and GM. Cooper, Biological and biochemical properties of
human rasH genes mutated at codon 61 . Cell, 1986. 44(1): p. 167-76.

Temeles, G.L., et al., Yeast and mammalian ras proteins have conserved
biochemical properties. Nature, 1985. 3 13(6004): p. 700-3.

Farnsworth, CL. and LA. Feig, Dominant inhibitory mutations in the Mg(2+)-
binding site of RasH prevent its activation by GT P. Mol Cell Biol, 1991. 11(10): p.
4822-9.

F eig, LA. and GM. Cooper, Inhibition of NIH 3 T 3 cell proliferation by a mutant
ras protein with preferential aflinity for GDP. Mol Cell Biol, 1988. 8(8): p. 3235-
43.

163

39. Nishihara, K., et al., Chaperone coexpression plasmids: dijferential and synergistic
roles of DnaK-DnaJ-GrpE and GroEL-GroES in assisting folding of an allergen
of Japanese cedar pollen, Cryj2, in Escherichia coli. Appl Environ Microbiol,
1998. 64(5): p. 1694-9.

40. Silvius, J .R., Mechanisms of Ras protein targeting in mammalian cells. J Membr
Biol, 2002. 190(2): p. 83-92.

41. Seabra, M.C., Membrane association and targeting of prenylated Ras-like GT Pases.
Cell Signal, 1998. 10(3): p. 167-72.

42. Leung, K.F., R. Baron, and MC. Seabra, Thematic review series: lipid
posttranslational modifications. geranylgeranylatian of Rab GT Pases. J Lipid Res,
2006. 47(3): p. 467-75.

164

IV Concluding Remarks

This work presented in this thesis is comprised of two major parts. The first part
describes work evaluating the interaction of the protein kinase MLK3 with the
mitochondrial ATP/ADP translocase ANT2 and the investigation of how this interaction
is regulated. The second part of this thesis is focused on a Parkinson’s associated protein
kinase, LRRK2. Specifically the Roc domain of LRRK2 has been shown to function as
a bonafide GTPase. Previously we proposed that the Roc domain of LRRK2 may
activate its own protein kinase domain. Furthermore we predicted that PD-associated
mutations of Arg 1441within the Roc domain should interfere with protein interactions
rather than influence the GTPase activity of Roc since the mutated residue is far fi'om the
GTPase catalytic site and multiple, diverse amino acid substitutions (Cys,His, and Gly)
found in PD patients result in the same disease phenotype. In addition, an intra- or inter-
molecular interaction of LRRK2 mediated by the Roe-COR tandem domains is also

demonstrated.

Our lab has focused on understanding MLK3 function and regulation by
identifying its interacting proteins in human breast cancer MCF-7 cell line. MLK3 is
expressed at a high level in MCF-7 cells. Using the immunoaffinity purification coupled
with mass spectrometry, numerous potential MLK3 interacting proteins have been
identified in the lab and ANT is one of them. Data in Chapter II demonstrates that
expressed MLK3 interacts with ANT 2 using the co-immunoprecipitation technique.
Furthermore, the interaction of expressed MLK3 with ANT is specific for the ANT 2
isoform. The kinase domain of MLK3 is sufficient for the MLK3-ANT2 interaction.

The activation status of MLK3 regulates its interaction with ANT 2, with higher activity

165

 

forms of MLK3 complexing more efficiently with ANT 2. A portion of expressed MLK3
present in or associated with the mitochondria is capable of coimmunoprecitating with
ANT 2, implicating that the MLK3-ANT 2 interaction occurs at the mitochondria.
Coexpression of MLK3 with ANT2 synergistically increases total cellular ATP levels,
suggesting that the elevation of ATP levels is primarily a consequence of the MLK3-

ANT2 interaction.

The biological function of the elevated cellular ATP levels from the MLK3-
ANT2 interaction requires further investigation. Based on unpublished work in our lab,
transient knockdown of MLK3 reduces cell proliferation induced by growth factors and
steroid hormones, estrogen and progesterone, in MCF—7 cells. It is conceivable that the
elevated cellular ATP levels from the MLK3-ANT2 coexpression may contribute to
MCF-7 cell proliferation. Knockdown of ANT 2 with or without MLK3 in MCF—7 cells
under estrogen and/or progesterone stimulation will be necessary to further test whether
the MLK3-ANT 2 interaction contributes to cancer cell proliferation. Total cellular ATP
levels are increased upon the coexpression of MLK3 and ANT 2, important questions
remain. The source of the increased ATP, whether glycolysis or oxidative
phosphorylation, is as yet unknown. The detailed mechanism by which cellular ATP
levels increase upon MLK3/ANT 2 coexpression remains to be determined. Based on
ANT ’s classical firnction in ATP/ADP exchange, one would not necessarily predict that
ANT 2 should affect total level of ATP. Intriguingly, we obtain the peptides
corresponding to the mitochondrial proteins of oxidative phosphorylation including ATP
synthase in our mass spectrometry data. It will be important to determine whether these

proteins of the oxidative phosphorylation are also present in the MLK3-ANT2 complex.

166

Such future studies, in conjunction with this work, may provide a novel mechanism for

the regulation of cellular energetics in cancer cells.

Chapter III has described the investigation and characterization of the novel Roc
GTPase domain of LRRK2. Conditions for expression and purification of soluble GST-
Roc variant proteins were optimized. Unfortunately, GST-Roc-COR largely remained
insoluble under numerous conditions that were tested. The purified GST-Roe from
bacterial expression system possesses a weak intrinsic GTPase activity, consistent with
the concurrent published reports. Activated GTPases are commonly post-translationally
modified by prenylation. These modifications render the activated GTPases hydrophobic
and allow the activated GTPases to be targeted to cellular membranes. LRRK2 lacks a
classical motif for posttranslational prenylation, but it is conceivable that the absence of
appropriate cellular structure(s) such as lipids during expression or purification processes
may result in low GTPase activity. For instance, Cdc42 and Rac, display much higher
catalytic activities when expressed and purified from insect cells as compared with
bacterially expressed proteins. Therefore insect cell expression vectors encoding His-
tagged Roc variants are under construction in the lab. The assessment of the GTPase
activity of the purified Roc from insect cells in parallel with the purified Roc from

bacteria will give insight into the cause of the low GTPase activity of GST-Roe.

The second portion of Chapter HI reveals that Roe-COR can interact in trans with
a region of LRRK2 containing Roe, COR, kinase, WD40 through the end (RCKWE).
Roc weakly interacts with RCKWE, suggesting that COR contributes to the interaction of
Roc-COR with RCKWE. Notably, R1441C/G, the PD associated mutation in the Roc

domain, slightly reduces the Roe-COR mediated domain interaction whereas Y1699C,

167

the PD associated mutation in the COR domain, enhances this interaction. It is possible
that, in the context of full length LRRK2, the Roc-COR domain is captured in an
inhibitory conformation. Introduction of the PD associated mutation, Y1699C, in this
domain disrupts its inhibition, allowing Roe-COR Y1699C efficiently interacts with the
RCKWE region. The interaction between Roe and COR will be a key to the proposed
hypothesis and future experiments will test this interaction using co-immunoprecipitation
assays. These findings provide knowledge to establish a new mechanism for explaining

how these amino acid substitutions result in PD.

In conclusion, the research findings in this thesis contribute to the understanding
of the new role of MLK3 in regulation of cellular energetics through its mitochondrial
interacting protein, ANT 2. These studies shown herein also support our previous
hypothesis of the LRRK2 regulation and provide knowledge to develop a novel
mechanism of Roe-COR regulates LRRK2 molecular interaction that might be involved

in PD pathogenesis.

168

   

u([lllljlljlfllllljjljll111119111