*. .5“
a.» “
«q
Ms

m.»
w»

~u»-.
.

L
age};
. ,-
Egg

m!
:A-u—u;

w
.
:29 on

Wax...
’ w... .“m
"“57“
u-
A“.
NM.
'NM
. ,Q'
3‘ — Mo“

 

 

_ -._

.lLr

THE$5

2004

l.’.“v'\ (';-'
. ‘/(,A Jl. '

 

 

_1_ any”. - ’5‘,

l—lDflHn l

Michigan State
University

 

 

This is to certify that the
dissertation entitled

REGULATION OF MIXED LINEAGE KINASE 3 BY
AUTOINHIBITION, PROTEOLYSIS AND BINDING PARTNERS

presented by

Hua Zhang

has been accepted towards fulfillment
of the requirements for the

Ph.D degree in Cell and Molecular Biology
Program

 

 

W;{ 62%

Major Professor’s Signature

(Zhuggqut. «EL/d} QZCX333

Date

MSU is an Affirmative Action/Equal Opportunity Institution

 

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.

 

DATE DUE DATE DUE DATE DUE

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

6/01 c:/ClRC/Date0ue.p65-p.15

REGULATION OF MIXED LINEAGE KINASE 3 BY

AUTOINHIBITION, PROTEOLYSIS AND BINDING PARTNERS

By

Hua Zhang

A DISSERTATION

Submitted to
Michigan State University
In partial fulfillment of the requirements
For the degree of

DOCTOR OF PHILOSOPHY

Cell and Molecular Biology Program

2003

ABSTRACT

REGULATION OF MIXED LIN EAGE KINASE 3 BY
AUTOINHIBITION, PROTEOLYSIS AND BINDING PARTNERS

By

Hua Zhang

Mixed lineage kinase 3 (MLK3) is a serine/threonine protein kinase widely
expressed in human tissues. MLK3 acts as a mitogen activated protein kinase (MAPK)
kinase kinase (MAPKKK) to phosphorylate and activate the dual specific kinases, MAPK
kinase 4 (MKK4) and MKK7, which, in turn, can phosphorylate and activate the c-Jun
NHz-terminal kinase (JNK) pathway. Overexpressed MLK3 also modestly activates the
p38 MAPK pathway.

In addition to its catalytic domain, MLK3 contains several domains, including an
amino-terminal Src Homology 3 (SH3) domain, a centrally located zipper region, a
Cdc42/Rac interactive binding (CRIB) motif and a Pro/Thr/Ser rich carboxyl terminus,
which may mediate protein-protein interactions to regulate its activity and/or signaling.
The work presented herein demonstrated that MLK3 is autoinhibited through an
interaction between its amino terminal SH3 domain and a sequence located between the
Zipper and CRIB motifs of MLK3. A single proline residue in the SH3-binding site is
required for this intramolecular autoinhibition. Mutation of either the SH3 domain or the
proline residue in the SH3—binding site increases MLK3 catalytic activity. The SH3

domain and the critical proline residue in the binding site are conserved in the closely

related family membérs, MLKl, MLKZ, MLK4 and the Drosophila homologue, Slipper,
suggesting a common autoinhibitory mechanism among these kinases.

MLK3 undergoes proteolysis to generate a stable carboxyl terminal fragment
(CTF) in vivo in an activity dependent fashion in HEK 293 cells and in a PMA —induced
fashion in Jurkat T Ag cells. The amino terminus of the CTF is Pro 252, suggesting that
the “MLK3 protease” cleaves between Gln 251 and Pro 252 within the kinase domain.
Site directed mutagenesis revealed that Leu 250 and Gln 251, but not Pro 252, are
required for this proteolytic event. Interestingly, the coronavirus main protease shares a
strikingly similar substrate specificity, where it cleaves immediately after Leu-Gln
sequences in its viral polyprotein substrates. Chapter III of this thesis presents data that
indicate mammalian cells contain a coronavirus protease-like activity. In addition, the
protease inhibitor, MG 132, blocks the PMA-induced proteolysis of MLK3.

Affinity purification coupled with mass spectrometry was employed to isolate and
identify MLK3 binding partners. In-gel trypsin digestion together with in-solution trypsin
digestion have revealed a total of 23 potential interacting proteins of MLK3. These
findings certainly will lead to new directions in the study of the regulations and biological
functions of MLK3.

In addition, the fimctional significance of the interaction between the molecular
chaperones Hsp90/Cdc37 and MLK3 was studied. Hsp90/Cdc37 was shown to bind to
the catalytic domain of MLK3 in a MLK3 activity-independent manner. Disruption of
Hsp90 function with geldanamycin decreases the protein level of MLK3 in MCF-7 cells.
Furthermore, Hsp90/Cdc37 regulate TNFoc-induced JNK activation at the MAPKKK

(MLK3) level, but not at MKK4/MKK7 or JNK level.

iv

T 0 my family

ACKNOWLEDGMENTS

I would like to thank my committee members Dr. Kathy Gallo, Dr. Susan Conrad,
Dr. John Wang, Dr. David Dewitt and Dr. Robert Hausinger for their guidance and
support. Especially I deeply appreciate my mentor, Dr. Gallo’s guidance, support and
encouragement throughout my graduate career. I would like to thank Wei Wu and Dr.
Watson for their collaboration on the mass spectrometry project. And I thank Weiqin
Chen and Dr. Walter Esselman for their collaboration on the MLK3 proteolysis project.

I am very thankful for the friendships with the members in Dr. Gallo’s lab during
my graduate training, especially Dr. Otis Vacratsis, Mary Chao, Karen Schatcher, Yan
Du, Geouyarh (Stancy) Liou, Ritesh Agrawal, and Dr.Barbara Bock. I thank them for

their valuable discussion and co-operation.

TABLE OF CONTENTS

Page
List of Tables ........................................................................................ xii
List of Figures ....................................................................................... xiii
Key to Abbreviations ............................................................................... xvi
I. Literature Review .................................................................................. 1
l. Mammalian protein kinase ............................................................. 1
2. Regulation of protein kinases .......................................................... 4
2.1. Regulation of protein kinases by autoinhibition ............................... 5
2.2. Regulation of protein kinases by intermolecular interactions ................ 6
2.3. Regulation of protein kinases by proteolysis ................................... 6
2.4. Regulation of protein kinases by heat shock proteins ........................ 8
3. Mitogen activated protein kinase pathways ........................................ 12
3.1. The ERK pathway ................................................................ 12
3.2. The JNK pathway ................................................................. 13
3.3. The p38 MAPK pathway ......................................................... 15
3.4. Scaffold proteins and specificity of MAPK pathways ....................... 16
4. Mixed lineage kinases ................................................................. 20
4.1. Family members of mixed lineage kinase ...................................... 20
4.2. Pathways regulated by MLKs ................................................... 22
4.2.1. The JNK pathway ........................................................ 22

vi

4.2.2. The p38 MAPK pathway ................................................. 22

4.2.3. The ERK pathway ........................................................ 23

4.2.4. The NF-KB pathway ..................................................... 23

4.3. The regulation of MLK3 ......................................................... 25
4.3.1. Regulation of MLK5 by extracellular signals ........................ 25

4.3.2. Regulation of MLK5 by phosphorylation ............................. 26
4.3.3. Regulation of MLK5 by oligomerization .............................. 27

4.3.4. Regulation of MLK5 by Rho GTPase ................................. 28

4.3.5. Regulation of MLK5 by scaffold proteins ............................ 29

4.4. The biological functions of MLK3 in mammalian cells ..................... 31
4.4.1. MLKs in neuronal apoptosis ............................................ 31

4.4.2. MLKs in cell transformation and cell morphology changes ....... 33

4.4.3. MLKs in cell cycle ....................................................... 33

4.5. The biological functions of MLK5 in Drosophila ............................ 34
4.6. MLK3 .............................................................................. 35

5. Objective of thesis ..................................................................... 38
6. Reference ............................................................................... 39
H. Autoinhibition of Mixed Lineage Kinase 3 Through Its SH3 Domain ................... 50
1. Abstract ................................................................................. 50
2. Introduction ............................................................................. 51
3. Materials and Methods ................................................................ 53
3.1 DNA Constructs and Mutagenesis .............................................. 53

vii

 

3.2 Expression and Purification of GST Fusion Proteins ........................ 55

3.3 Cell Culture, Transfections, and Lysis ......................................... 55

3.4 Immunoprecipitations and GST Pulldown Assays ........................... 55

3.5 Gel Electrophoresis and Western Blot Analysis .............................. 56

3.6 In vitro Kinase Assays ............................................................ 56
4. Results

4.1 Point mutations in the SH3 domain increases the in vitro kinase

activity of MLK3 .................................................................. 58
4.2 The SH3 domain of MLK3 interacts with MLK3 ............................. 58
4.3 Identification of the SH3 binding region within MLK3 ..................... 59
4.4 Point mutation in the SH3 binding site abolishes binding to the SH3
domain and increases MLK3 activity ........................................... 61
5. Discussion .............................................................................. 71
6. References .............................................................................. 77
III. Characterization of a Proteolytic Event of Mixed Lineage Kinase 3 .................... 81
1. Abstract ................................................................................. 81
2. Introduction ............................................................................. 83
3. Materials and Methods ................................................................ 86
3.1 . Reagents ............................................................................ 86
3.1. Construction of Mammalian Expression Vectors and Site-Directed
Mutagenesis ........................................................................ 86
3.2. Cell Culture, Transfections, and Lysis ......................................... 86
3.4 Immunoprecipitations ............................................................ 87
3.5 SDS-PAGE and Western Blot Analysis ....................................... 87

viii

5.

6.

3.6 Pulse Chase Analysis ............................................................. 88

3.7 Amino Terminal Sequencing .................................................... 88
3.8 MLK3 Kinase Domain Homology modeling ................................. 89
Results ................................................................................... 90

4.1.MLK3 undergoes activity dependent proteolysis in HEK 293 cells. . . . . ...90
4.2.Proteolysis occurs between amino acids 251-252 of MLK3 ................ 91
4.3.Gln 251 and Leu 250 are critical for MLK3 proteolysis ..................... 91

4.4. Inducible proteolysis of MLK3 by extracellular stimuli in Jurkat T
antigen cells ........................................................................ 92

4.5. Inducible proteolysis of MLK3 is independent of MLK3 activity in

Jurkat T antigen cells ............................................................. 93
4.6. Protease inhibitor MG 132 blocks the generation of CTF of MLK3 in

J urkat T Ag cells ................................................................... 93
4.7.The CTF is generated from newly synthesized MLK3 ...................... 94
4.8.ERK activation is required for MLK3 proteolysis ........................... 94
Discussion ............................................................................. 109
References ............................................................................ l 14

IV. Identification of Mixed Lineage Kinase 3 Binding Partners by Affinity Purification

and Mass Spectrometry ............................................................................ 117
1. Abstract ................................................................................ 117

2. Introduction ........................................................................... 119

3. Materials and Methods ............................................................... 123

3.1. Plasmid Constructs and Mutagenesis ....................................... 123

3.2. Construction of MCF-7 Inducible Cell Lines .............................. 123

ix

 

3.3. Purification of Flag-MLK3 Complexes .................................... 124

3.4. Nupage Gel Electrophoresis and Coomassie Staining .................... 125
3.5. In—gel Trypsin Digestion ...................................................... 125
3.6. In-solution Trypsin Digestion ................................................ 125
3.7. LC/MS/MS ..................................................................... 126
3.8. Transfection and Lysis of HEK 293 cells .................................. 127
3.9. Purification of GST Fusion Proteins and GST Pulldown Assays ....... 127
3.10. SDS~PAGE and Western Blot Analysis .................................... 128
4. Results ................................................................................. 129
4.1. Survey of endogenous levels of MLK3 .................................... 129
4.2. Inducible expression of MLK3 in MCF-7 cells ............................ 129
4.3. Purification of Flag-MLK3 complexes from MCF-7 cells ............... 129
4.4. Identification of putative MLK3 binding partners by in-gel
digestion and MS/MS ........................................................... 130
4.5. In-solution digestion and MS/MS ........................................... 131
4.6. Association of MLK3 with Clathrin Heavy Chain ........................ 132
5. Discussion ............................................................................. 147
6. References ............................................................................. 154
V. Regulation of Mixed Lineage Kinase 3 by Hsp90 and Cdc37 ........................... 158
1. Abstract ................................................................................ 158
2. Introduction ........................................................................... 159
3. Materials and Methods ............................................................... 161
3.1. Plasmid Constructs and Site Directed Mutagenesis ........................ 161

 

3.2. Cell lines and Transfections ................................................... 161

3.3. GST Pulldown Assays ......................................................... 162

3.4. Immunoprecipitations .......................................................... 162

3.5. In vitro Kinase Assays ......................................................... 163

3.6. SDS-PAGE and Western Blot Analysis ..................................... 163

4. Results ................................................................................. 165
4.1. Demonstration of the interaction between endogenous Hsp90/Cdc37

and F lag-MLK3 ................................................................. 165

4.2. Endogenous MLK3 is associated with Hsp90/Cdc37 ..................... 165

4.3. Hsp90/Cdc37 associates with the kinase domain of MLK3 ............... 165

4.4. MLK3 is associated with Hsp90/Cdc37 in an MLK3 activity
independent manner ............................................................ 166

4.5. Hsp90/Cdc37 regulates the protein level of MLK3 in MCF-7 cells. . ...l67

4.6. Hsp90/Cdc37 regulate TNFa-induced activation and signaling of

MLK3 in MCF-7 cells ......................................................... 168
5. Discussion ............................................................................. 176
6. References ............................................................................ 181
VI. Concluding Remarks ......................................................................... 184

xi

Table

List of Tables

Page
Chapter 3.

. Effect of protease inhibitors on PMA-induced generation of MLK3 CTF ......... 97
MAPK inhibitors ........................................................................... 97
Chapter 4.

. Previously identified mixed lineage kinases MLK1,2,3 binding partners. . . . . l 34
Proteins identified by in-gel trypsin digestion and LC/MS/MS .................... 135
Proteins identified by in-solution trypsin digestion and LC/MS/MS in three
independent experiments ................................................................ 136

xii

List of Figures

Figure Page
Chapter I.

. Domain structures of protein kinase C family members ............................ 10

. Autoinhibition and activation of tyrosine kinase Src ................................ 11

. The mammalian MAPK pathways ..................................................... 18

. Human MAPKs ........................................................................... 19

. Composite structure of human mixed lineage kinases .............................. 36

. Scaffold proteins that interact with MLKs ............................................ 37
Chapter H.

. Schematic of MLK3 ..................................................................... 63

. Effect of a point mutation in the SH3 domain of MLK3 on catalytic activity. . .64
. Assay for association of the SH3 domain of MLK3 with full length MLK3. . ..65
. Mapping of the SH3 binding region within MLK3 ................................. 66
. Effect of zipper mutations on the SH3 binding ...................................... 68

. Alignment of the SH3 binding region of MLK3 with MLKl and MLK2

and effect of MLK3 P469A on binding of SH3 domain ............................. 69
. Catalytic activity of MLK3 P469A .................................................... 70
Chapter IH.
. In vivo proteolysis of MLK3 in HEK 293 cells ........................................ 98
. MLK3 is cleaved within the kinase domain ............................................ 99

xiii

10.

11.

The proteolytic pattern of MLK3 variants ............................................. 100
Inducible proteolysis of MLK3 in Jurkat T Ag cells ................................. 101
The proteolytic pattern of transfected MLK3 variants in PMA-treated

Jurkat T Ag cells .......................................................................... 102
Effect of protease inhibitors on PMA-induced generation of the CTF

of MLK3 in Jurkat T Ag cells ........................................................... 103
Chemical structure of selected protease inhitors and comparison with the
sequence before the cleavage site in MLK3 ........................................... 104
Pulse chase experiments in HEK 293 cells ............................................. 105
Cycloheximide blocks PMA induced generation of the CTF of MLK3

in Jurkat T Ag cells ........................................................................ 106

Effect of MAPK pathway inhibitors on the PMA-induced generation

of the CTF of MLK3 in Jurkat T Ag cells ............................................. 107
Model for generation of CTF of MLK3 ................................................ 108
Chapter IV

. Expression of MLK3 in several human cell lines .................................... 138
Construction of inducible cell lines .................................................... 139
Inducible expression of Flag-tagged MLK3 .......................................... 140
Affinity purification of Flag MLK3 complexes ....................................... 141
Identification of band 1 as Clathrin heavy chain by LC/MS/MS ................... 142
Association of MLK3 with Clathrin heavy chain terminal domain ................. 145
Interactions of MLK3 variants with Clathrin heavy chain ........................... 146

xiv

Chapter V

. Association of Flag-MLK3 with Hsp90/Cdc37 ....................................... 169
. Endogenous MLK3 is associated with Hsp90 and Cdc37 ........................... 170
. Hsp90 and Cdc37 associate with the catalytic domain of MLK3 .................. 171
. Association of MLK3 variants with Hsp90/Cdc37 .................................... 172
. Geldanamycin (GA) decreases the endogenous level of MLK3 ................... 173
. GA abolishes TNFor-mediated activation of MLK3 and JNK ..................... 174

XV

 

Key to Abbreviations

 

3Cpm
AGC
ANT -2
AP50
ASKl
CaM
CAMK
CBP
CDK
CHC
CKl
CKI
CMGC
CRIB
DeltaCn
DLK
DMEM
ERK
ES

ESI
FKBP
FnIII
FRET
GA
GluR6
GST
HA
HPKl

3C-like proteinase

containing Protein Kinase A, Protein Kinase G, Protein Kinase C families
adenine nucleotide translocator-Z

Clathrin assembly protein complex 2 medium chain
apoptosis signal- regulating kinase 1

Calcium /Calmodulin

calcium/calmodulin dependent protein kinases
calmodulin-binding peptide

cyclin dependent protein kinases

clathn'n heavy chain

casein kinase 1

CDK inhibitors

containing CDK, MAPK, GSK3, CLK families
Cdc42/Rac interactive binding motif
1.0-normalized correlation score

dual leucine zipper protein kinase

Dulbecco’s Modified Eagle’s Medium
extracellular signal-regulated kinase
nanoelectrospray

electrospray ionization

FK binding proteins

fibronectin type III

fluorescence resonance energy transfer
geldanamycin

Kainate receptor glutamate receptor
glutathione S-transferase

hemagglutinin

hematopoietic progenitor protein kinase-1

 

xvi

 

Hsc70
HSP
IicBa

Ig

JIP

KIF3
LC
LZK
MALDI
MAPK
MAPKKK
MEK
MEKK
MKK
MLK
MLTK

Mpro

MS
NFAT4
NGF
NIMA
PAGE
PAK
PDGF

heat shock cognate 71 kD protein

heat shock protein

inhibitor of NF-KB alpha

islet brain

immunoglobulin

IKB kinase

JNK interacting protein

c-jun N-terminal kinase

kinesin superfamily motor KIF3

liquid chromatography

leucine zipper-bearing kinase

matrix assisted laser desorption lionization
mitogen activated protein kinase
MAPK kinase kinase
mitogen-activated protein/ERK kinase
MEK kinase

MAPK kinase

mixed lineage kinase

MLK-like mitogen-activated protein triple kinase
MEK partner 1

main proteinase

MLK-related kinase

mass spectrometry

nuclear factor of activated T cells 4
nerve growth factor

never in mitosis A

polyacrylamide gel electrophoresis
p21 activated protein kinase

platelet-derived growth factor

 

xvii

 

SH2
SH3
SPRK
STE
Syd
TAKl
TAP
TD
TEV
TGF
TK
TKL

TOF
Xcorr

ZAK

peptidyl-prolyl isomerase Pinl

protein kinase A

protein kinase C

plenty of SH3s

post-synaptic density protein 95
permeability transition pore
receptor guanylate cyclase

RNA interference
rank/preliminary score

sterile alpha motif
stress-activated protein kinases
severe acute respiratory syndrome
src-homology 2

src-homology 3

SH3 domain containing proline rich protein kinase
homologs of yeast Sterile 7, 11, 20 kinases
Sunday driver

TGFB activated protein kinase 1
tandem affinity purification
terminal domain

tobacco etch virus

transforming grth factor
tyrosine kinases

tyrosine kinase—like

tumor necrosis factor

time of Flight

cross-correlation score

leucine-zipper and sterile-alpha motif kinase

 

xviii

I. Literature Review

1. Mammalian protein kinases

Mammalian protein kinases catalyze the transfer of the y-phosphate of ATP to the
amino acid residues of protein substrates. Protein phosphorylation can be reversed
through the action of protein phosphatases. Mammalian protein kinases can be divided
mainly into three groups based on amino acid specificity. The tyrosine kinases
phosphorylate only Tyr residues within proteins, whereas the serine/threonine protein
kinases are specific for phosphorylation of Ser or Thr residues. The dual specificity
kinases are a small group of kinases capable of phosphorylating Tyr as well as Ser and
Thr in their protein substrates [1] [2]. In addition, histidine kinases are newly discovered
protein kinases in mammalian systems, which phosphorylate the 1 and/or 3 nitrogen
atoms in the histidine residues by forming a phosphoramidate bond. Histidine kinases are
better defined in bacteria and plants and the function of histidine phosphorylation in
mammals has remained obscure [3].

All eukaryotic protein kinases share a highly conserved catalytic domain, often
referred as the kinase domain. This kinase domain, of approximately 250-300 amino
acids, can be divided into 12 subdomains, which share conserved sequences among
protein kinases. The kinase domain folds into a two lobed structure. The smaller N-
terminal lobe consists of subdomains I-IV of the kinase. Its secondary structure is mainly
composed of antiparallel beta sheets with one conserved helix called the C—helix. The
primary function of the N-terminal lobe is to orient and bind to Mg2+ and ATP. The C-

terrninal lobe consists of subdomains VIA-XI. Its secondary structure is mainly alpha

helical. The C-terminal lobe functions mainly in protein substrate binding and in
initiating the phospho-transfer from ATP to the substrate. Subdomain V connects the two
lobes and plays an important role in recognition of the peptide substrate and in the
anchoring of the Mg2+ATP (reviewed in [4, 5]). Within the kinase domain, there is a so-
called “activation loop” which is located between two highly conserved sequence motifs
Asp-Phe-Gly of subdomain VII and Ala-Pro-Glu of subdomain VIII [6]. This activation
loop often contains site(s) which, upon phosphorylation, undergo significant
conformational changes to increase the catalytic activity of the protein kinases.

In addition to their catalytic domains, most kinases contain non-catalytic regions
or domains, which may control the subcellular location of protein kinases, mediate their
interaction with other proteins or molecules, affect their enzymatic activity, and influence
their stability for degradation by proteases. For example, Src homology 3 (SH3) domains
normally bind to proline-rich sequences to mediate protein-protein interactions. In
addition, a few protein kinases contain leucine zipper motifs. Leucine zipper motifs
mediate protein oligomerization by forming coiled coil structures, in which the leucine
residues spaced seven residues apart interact at the interface of opposing helices [7].
Several protein kinases contain a Cdc42—Rac interactive binding (CRIB) motif to which
small GTPase Cdc42 or Rac can bind [8].

The human genome encodes 518 protein kinases, making the protein kinases one
of the largest families of proteins. Based on sequence homology within the catalytic
domains, and assisted by information of domain arrangement outside of the catalytic
domain, mammalian protein kinases are classified into nine groups: AGC (containing

PKA, PK_(_}, PKQ families), CAMK (galcium/cglmodulin dependent protein kinases),

 

CMGC (containing CDK, MAPK, GSK3, QLK families), TK (tyrosine kinases), STE
(homologs of yeast firile 7, 11, 20 kinases), CKl (gasein kinase 1), TKL (tyrosine
kinase-like), RGC (receptor guanylate gyclase), and atypical kinases [9].

Protein kinases control many important physiological processes, such as energy
metabolism, cell proliferation, cell differentiation, and apoptosis. In addition, protein
kinases play critical roles in the operation of the nervous and immune systems and during

the development in cell-cell communication.

 

2. Regulation of protein kinases

Since virtually all physiological processes are regulated by reversible protein
phosphorylation, protein kinases should be precisely controlled. Most protein kinases are
normally kept in an inactive form, and in response to extracellular stimuli, such as growth
factors, hormones, or osmotic stress, certain protein kinases are activated and serve to
transduce signals from the outside to within the cell. A variety of mechanisms regulate
protein kinases, including phosphorylation and other post-translational modifications,
binding of activators or inhibitors, and changes in subcellular localization. Herein, two
well studied protein kinases, the serine/threonine protein kinase PKC (protein kinase C)
and the tyrosine protein kinase Src will be used as examples to discuss regulatory
mechanisms that are often utilized by protein kinases.

The PKCs are a family of protein kinases involved in modulating diverse cellular
responses, including immune responses, cell proliferation, and learning [10]. The
mammalian PKC family contains ten isozymes that cluster into three groups based on
their domain arrangements: conventional (or, y and the alternatively spliced [31 and [311),
novel (8, s, n/L, 0), and atypical (Q, 10») PKC isozymes (F ig.1). In addition to the kinase
domain in the carboxyl terminus, PKC contains domains, which regulate its subcellular
localization as well as catalytic activity. Conventional and novel PKC isoforms contain a
typical Cysteine-rich C1 domain, which binds to diacylglycerol/ phorbol esters. The C2
domain in conventional PKC, upon binding to Ca“, facilitates PKC translocation to the
plasma membrane [1 1].

The non-receptor tyrosine kinase c-Src, encoded by the c-src gene is involved in

many cellular processes, such as cell cycle, cell proliferation and migration [12]. Src, and

its family members, Hck and Fyn, share a similar domain arrangement. The amino
terminal SH4 domain of Src is critical for Src association with the plasma membrane. Src
also contains an SH3 domain and an SH2 domain, which are characterized as binding
region for proline—rich sequences or sequences containing phosphorylated tyrosine

residues respectively. The kinase domain of Src is located at its carboxyl terminus [13]

(Fig. 2).

2.] Regulation of protein kinases by autoinhibition

Many protein kinases are kept in an inactive state through autoinhibitory
intramolecular interactions. Some protein kinases contain pseudosubstrate domains that
bind the active site of the catalytic domain, blocking substrate access [14]. For example, a
pseudosubstrate autoinhibitory domain that precedes the Cl region of PKC, binds to the
kinase domain of PKC, inhibiting substrate binding and keeping PKC in an inactive state
[10].

Protein kinases can also be autoinhibited through intramolecular interactions
without the participation of the catalytic domain. One interesting example is provided by
the non-receptor tyrosine kinases of the Src family (Fig. 2), in which an autoinhibitory
intramolecular association involves both the SH2 domain and the SH3 domain of Src. In
the autoinhibited state of Src, the carboxyl terminal phosphorylated Tyr 527 interacts
with the SH2 domain while the SH3 domain of Src binds to a polyproline type H helix
sequence that is located between the SH2 domain and the kinase domain. These two
intramolecular interactions constrain the motion of the kinase domain and block the

binding of ATP and substrate, thus keeping Src in an inactive form [15, 16].

2.2. Regulation of protein kinases by intermolecular interactions

In the process of signal transduction, intermolecular interactions between small
molecules and protein play a critical role. For example, binding of diacylglycerol or
phorbol esters to the Cl domain of PKC disrupts the autoinhibitory interaction, and leads
to its recruitment to the plasma membrane and its activation [11].

Autoinhibition within some protein kinases can be released through binding to
another protein. For example, a proline-rich sequence in the Nef protein of the human
immunodeficiency virus-1 (HIV-1) can bind to the SH3 domain of Hck, a Src family
member, thus releasing Hck autoinhibition, and increasing Hck activity dramatically [l 7].

The activity of cyclin dependent protein kinases (CDKs) is regulated by binding
to either activating or inhibitory proteins. CDKs are nuclear serine-threonine kinases that
control the cell cycle in eukaryotes. CDKs are catalytically active only when bound to
cyclins. CDKs are negatively regulated through binding to CDK inhibitors (CKI) such as

p27Kipl or p16 [18, 19].

2.3. Regulation of protein kinases by proteolysis

Ubiquitin-mediated degradation is a highly conserved mechanism for selective
protein degradation in eukaryotes. Ubiquitin is a 76 amino acid protein of 8.6 kDa that is
covalently attached to proteins targeted for degradation. Three major enzymes are
required in ubiquitination. Ubiquitin is activated in an ATP-dependent step by a specific
activating enzyme (E1). The activated ubiquitin molecule is then transferred to a

ubiquitin-conjugating enzyme (E2). Then the ubiquitin protein ligase (E3) links the

ubiquitin through its carboxyl terminus to the s—amino group of a Lys residue in the
substrate protein. The 268 proteasome can recognize and degrade the ubiquitinated
proteins (reviewed in [20]). Many proteins involved in signal transduction and cell cycle
progression, including protein kinases, are subject to ubiquitin-mediated degradation.
For example, upon treatment of the renal epithelial cells with a diacylglycerol analogue,
PKCa undergoes ubiquitin-mediated degradation [21]. The non-receptor tyrosine kinases
Src and Blk, have also been reported to undergo ubiquitin mediated degradation [22]
[23]. Furthermore, this ubiquitination is activity dependent, as activated forms of Src are
less stable than wild type or kinase-inactive Src. The steady state level of Src can be up-
regulated by the proteasome inhibitors [22]. Interestingly, the activities of CDK3 are
indirectly regulated by ubiquitin-mediated degradation pathway. The protein levels of
cyclins and CKIs are controlled by the proteasome during the cell cycle, thus influencing
the activity of CDK3 along cell cycle [18, 19].

Many protein kinases are substrates for caspases, and several protein kinases
become activated after proteolytic cleavage by the caspases. Caspases are cysteine
proteases, which are activated during apoptosis and cleave before an aspartic acid within
the substrate [24-26]. For example, PKCe[27] and PKC6 [28] are activated by caspases-
mediated cleavage. In contrast, other protein kinases that control anti-apoptotic pathways,
are inactivated upon caspases cleavage [29].

Calpains are responsible for the degradation of some proteins in signal
transduction. Calpains are Ca2+-dependent intracellular cysteine proteases, widely
expressed in various mammalian tissues. There are two major forms of calpains: m-

calpain and p—calpain. Calpains show no strict sequence preferences. It has been

suggested that calpain-mediated cleavage may be dependent on protein structure [30].
The substrates of calpain include PKC isoforms [31]. Recent studies show that calpain
can be activated through ERK pathway [32, 33] and the precise mechanism of how this

activation occurs remains to be discovered.

2.4. Regulation of protein kinases by heat shock proteins

Heat shock proteins are molecular chaperones that assist in the folding of newly
synthesized proteins and maintain protein homeostasis in cells. Heat shock proteins are
constitutively expressed under normal growth conditions. Upon exposure to stress, such
as heat shock, the expression of some heat shock proteins is greatly enhanced [34].

Heat shock proteins, Hsp90 and Hsp70, are two of the most highly expressed molecular
chaperones. Hsp70 interacts with unfolded substrates and assists them in folding to
intermediate states, thus preventing irreversible aggregation. In contrast, Hsp90 can
prevent the aggregation of intennediately folded proteins, but lacks the ability to refold
denatured proteins [34].

Most of the Hsp90 client proteins are involved in signal transduction. Studies
indicate that some protein kinases interact with Hsp90, and the Hsp90 co-chaperone
Cdc37 is required for this interaction. Association with Hsp90/Cdc37 stabilizes the client
protein kinase and prevents its degradation by the proteasome. The ansamycin antibiotics,
such as geldanamycin (GA), occupy the ATP/ADP pocket of Hsp90, thus inhibiting
Hsp90 chaperone activity. Geldanamycin treatment destabilizes several protein kinases,
such as Src, Raf-1 and Akt [35] [36, 37], because geldanamycin treatment dissociates

dysfunctional Hsp90 from protein kinases and protein kinases undergo ubiquitin-

mediated degradation. However, some of the protein kinases, such as licB kinase (IKK),
although they have been reported to associate with Hsp90/Cdc3 7, are not destabilized by
geldanamycin [38].

Relatively less is known about how Hsp70 may regulate protein kinases. Recent
studies show that Hsp70 interacts with PKC. Hsp70 prolongs the lifetime of mature PKC
by stabilizing the protein and allowing re-phosphorylation of the kinase. Disruption of

Hsp70-PKC interaction prevents re-phosphorylation and targets PKC for down-regulation

[39].

pseudosubstgate
Conventional

a, B1, [311, v

 

 

novel C2
Novel

s,6,0,n/L

 

 

Atypical
C, l/A

 

 

Fig. 1. Domain structures of protein kinase C family members. The C-terminal kinase
domains are shown in solid bar. The regulatory region or domains are shown as

following: the pseudosubstrate (E ); the C1 domain ( ); the C2 domain (ED ).

 

(Adapted from [11])

10

Extracellular signals

 

 

 

 

b
1’71"."iwmitr‘tflnmrwvrrrrrnnr".vrrnrr'r ."YTl'lTl'lTluv7T“*i'l‘lTi.‘[T'. wirnrr rrrr rrwr""r"r1f‘r “mama";
MJJAAJLLI“JLJ.14JL-V-JL‘J L41“ Allan“ ..‘LLL‘LLJ‘ - “PM- 55;; Will LLJL' ‘LL _1 Lu; T In
_ Proline rich
I sequence
Y Phosphotyrosine
527 Containing sequence

Inacfive

9 Phosphorylation of
‘— activation loop

$527 4— Dephosphorylation

Active

Fig. 2. Autoinhibition and activation of tyrosine kinase Src. Src is autoinhibited by
two intermolecular interactions. In response to extracellular signals, the activity of Src
increases by either releasing the autoinhibition through high affinity SH2 or SH3 ligands,
or by dephosphorylation of Tyr 527. PT K stands for protein tyrosine kinase. (Adapted

from [40])

11

3. Mitogen activated protein kinase pathways

Mitogen activated protein kinase (MAPK) pathways are composed of three
sequentially activated protein kinases, and the basic model is conserved from yeast to
humans (reviewed in [41—43]). MAPK pathways are activated in response to
extraordinarily diverse stimuli such as growth factors, cytokines, irradiation, and high
osmolarity (Fig. 3). The first kinase of the MAPK module is referred as a MAPK kinase
kinase (MAPKKK), which may be activated through phosphorylation by a MAPK kinase
kinase kinase (MAPKKKK) or through interaction with and, sometimes translocation by
an activated small GTPase such as Ras, Rae or Cdc42. An activated MAPKKK can then
phosphorylate and activate a dual specific MAPK kinase (MKK), which, in turn,
phosphorylates a specific Tyr and a Thr residue in the activation loop of the kinase
domain of a terminal MAPK. Phosphorylated activated MAPK have cytosolic substrates
but can also translocate to the nucleus, where it phosphorylates nuclear substrates
including transcription factors. There are three well-defined major mammalian MAPKs:
the extracellular signal-regulated kinase (ERK), ERK1 and ERK2; the c-jun N—terminal
kinase (JNK), JNKl, JNK2, and JNK3; and the p38 isoforms, p38a, p386, p38y, and p386
(Fig. 4). In addition, a fourth MAPK, ERKS has been recently described [44]. All
MAPKs are proline-directed kinases, phosphorylating their substrates on Ser or Thr

residues immediately precede a Pro residue [45].

3.1. The ERK pathway

12

In 1990, the first MAPK, erkI cDNA was isolated by Cobb’s lab [46]. ERK1 and
ERK2 (reviewed in [47]) are activated by a variety of extracellular signals, such as
growth factors, cytokines, virus infections, and lipopolysaccharides (LPS). MAPKKKs
that activate the ERK pathway include Raf-1 [48], A-Raf [49] and B-Raf [50]. The
activity of Raf can be elevated by the activated small GTPase Ras [51]. Additionally, the
MAPKKKs, MEKK1[52], MEKK2, MEKK3 [53], and Tpl2[54] have been reported to
activate ERK. In contrast to the diversity of MAPKKKS, there are only two MKKs that
activate ERK: mitogen-activated protein/ERK kinase 1(MEK1) and MEK2. Activated
MEK1/2 phosphorylates a Thr and Tyr in the activation loop of the catalytic domain of
ERK1/2 [55].

The ERK pathway is generally associated with proliferation and deregulation of
the components in the ERK pathway often leads to cellular transformation. Systematic
genome-wide screen from 15 cancer cell lines revealed Ras is mutated to an oncogenic
form in about 15% of human cancer. It was recently reported that 66% of malignant
melanomas contain mutations in the B-raf gene, which results in amino acid substitutions
in the kinase domain. A single substitution (V 599E) in the activation loop accounts for
80% of the mutations. Some of the mutated B-Raf proteins have enhanced in vitro kinase
activity and can transform NIH3T3 cells [56]. Activated MEK1/2 transforms fibroblasts
and produces tumors in nude mice [57], and activation of ERK2 transforms immortalized
cells [5 8]. Therefore, the ERK pathway has been targeted for anti-cancer therapies.

ERKl knockout animals are viable with a modest defect in T-cell development

[59], suggesting that there might be some fimctional compensation by ERK2.

13

 

3.2. The JN K pathway

INKS, also known as stress-activated protein kinases (SAPK), are a group of
MAPKs that are activated by cytokines and environmental stresses, such as heat shock,
ultraviolet (UV), osmotic stress (reviewed in [60, 61]). Ten INK isoforms, which
correspond to alternatively spliced isoforms derived from the Jnkl, JnkZ and Jnk3 genes,
exist. M1 and INK2 are ubiquitously expressed, while INK3 is selectively expressed in
the brain, heart, and testis [62]. Regulation of JNK at the MAPKKK level is very
complex and multiple MAPKKKs, at least upon ectopic expression, activate the JNK
pathway, including MEKK1-4 (reviewed in [63]), transforming growth factor (TGF) B
activated protein kinase 1 (TAKl) [64], apoptosis signal- regulating kinase 1 (ASKl)
[65], the seven mixed lineage kinases (MLKs) (reviewed in [66]). INK activation is
directly mediated by dual phosphorylation on Thr and Tyr residues by either MKK4 or
MKK7 [67]. In contrast to ERK, which has substrates both in the cytosol and nucleus, the
substrates of JNK are mainly transcription factors in the nucleus. INK phosphorylates the
transcriptional activation domains of several transcription factors including c-Iun, JunD,
Elk] and ATF2. For instance, c-Jun is a transcription factor with relatively short half-life
and is rapidly degraded by the ubiquitin-mediated degradation pathway. Activated JNK
phosphorylates c-Jun on Ser 63 and Ser 73 and these phosphorylation events stabilize c-
Jun and inhibit ubiquitination and degradation by the proteasome [68]. In contrast,
phosphorylation of nuclear factor of activated T cells 4 (NF AT4) by INK causes nuclear
exclusion of NFAT4, thus inhibiting NFAT transcriptional activity [69].

INK is involved in immune responses, development and apoptosis. Apoptosis is a

form of cell suicide that is triggered by a variety of stimuli and involves the activation of

14

members of the caspase family of cysteine proteases. During the process of apoptosis,
cytochrome c is released from mitochondria into the cytosol to activate caspase-9 which
subsequently activates the downstream caspases [70]. INK is required for UV-induced
apoptosis in primary murine embryonic fibroblasts. JnkIT’L-JnkZT/T mouse fibroblasts
have a defect in UV-induced cytochrome c release and subsequent mitochondrial death
signaling, thus protecting cells against UV-stimulated apoptosis [71]. The targeted
knockout of three INK genes (JnkI, Jnk2, Jnk3) individually results in viable mice that
have defects in immune responses and apoptosis [61, 72], suggesting that different JNK
isoforms may have overlapping functions. Indeed, Inkl"'/Jnk2"' double knockout
mutants die at mid-gestation, and have defects in neural-tube closure [73]. The Jnk3
knockout mice shows a reduction in seizure activity and decreased susceptibility to

hippocarnpal neuron apoptosis in response to excitotoxicity [74].

3.3. The p38 MAPK pathway

The p38 MAPKs were first identified as targets for drugs that inhibit TNFa-
mediated inflammatory responses [75]. Four p38 isoforms were identified: p3 8a, p385,
p3 87, and p3 86. The p38 MAPKs can be activated by cytokines, hormones, stresses such
as heat shock and osmotic shock [29]. Similar to ERK and INK, p38 is activated by dual
phosphorylation of Thr and Tyr in the activation loop of the kinase domain by MKK3
and MKK6. MAPKKKs that can activate the p38 pathway includes TAKl [76], ASKl
[65], and some MLKs [66].

The p38 pathway is involved in the regulation of cytokine production and in

cytokine-stimulated cell proliferation [77]. Inhibition of p38 by a specific inhibitor, SB

15

203580, blocks interleukin (IL)-2 and IL-7 induced T cell proliferation. In addition, p38
is also involved in apoptosis. Death receptors, such as Fas and TNF receptors, trigger the
activation of the p38 pathway [78]. Several studies have shown that different p38
isoforms have overlapping, but also distinct, physiological roles. In the human cervical
carcinoma cell line, HeLa, p380t induces apoptosis, whereas p385 promotes cell survival
[79]. Among the four p38 isoforms, p3 8a is the best studied and most widely expressed.
Disrupting the p38a gene in mice causes embryonic lethality [80, 81]. It is quite
interesting that the other p38 isoforms are unable to compensate the deletion of the p38a

gene, reiterating that the different p38 isoforms likely perform distinct biological

fimctions.

3.4. Scaffold proteins and specificity of MAPK pathways

MAPK signaling is complex at many levels. First, diverse extracellular signals
activate MAPK pathways. In addition, multiple protein kinases, particularly at the
MAPKKK level, have been implicated in the activation of a single MAPK pathway.
Finally, a single MAPKKK can often activates multiple MAPKs. All of these issues raise
the question as to how signaling specificity is achieved. This question can be answered, at
least partially, by considering protein expression patterns, the differential activation
mechanisms and selective interactions with MAPK scaffold proteins.

Although the catalytic domains of all these MAPKKKs are fairly similar, the
divergent sequences in the non-catalytic regions of these kinases may allow for their
differential activation in response to different upstream signals. Signal connectivity from

extracellular stimuli to MAPKKKs has been addressed by using RNA interference

16

(RNAi) in Drosophila Schneider S2 cells in Cobb’s lab [82]. In contrast to more than a
dozen MAPKKKs activating INK in mammalian cells, four such MAPKKKs are
expressed in Drosophila 82 cells. RNAi experiments revealed that maximal activation of
JNK by lipopolysaccharide requires TAKl. However, all four MAPKKKs (TAKl,
MLK2, ASKl, and MEKKl) are required to cause maximal INK activation by sorbitol.
One could imagine that the selectivity of signaling is more complex in mammalian
MAPK pathways due to the involvement of multiple MAPKKKs. These data indicate that
specific stimuli utilize different mechanisms to recruit different MAPKKKS to regulate
the JNK pathway [82]. The molecular details by which different upstream signals
regulate distinct MAPKKKs are still largely undefined.

Several scaffold proteins that co-ordinate MAPK signaling have been identified.
Scaffold proteins selectively bind to multiple components of the MAPK pathways and, in
some cases, may increase the local concentration of the signaling components, thus
enhancing the speed of signal transduction. For example, MEK partner 1(MP1) is a
scaffold protein involved in the activation of ERK pathway. MP1 binds MEKl and
ERKl , but not ERK2. Upon overexpression, MP1 enhances ERKl activation and
reporter gene expression driven by the transcription factor Elk-1 [83]. INK interacting
proteins (IIP) are another group of scaffold proteins involved in the JNK and p38

pathway (This content will be discussed in detail in the next section).

17

 

 

 

 

 

 

Raf, MEKKl-3, MEKK1-4, T AKl, ASK], MEKKZ
MKKK TplZ ASKl, TAKI, MLKS
MLKl-4, DLK

 

 

 

 

 

 

 

 

I I I I
M“ @@..

MAPK Esta/2 . p38 ERK5

395% $513 558 £58

 

 

 

 

 

 

 

 

 

Fig. 3. The mammalian MAPK pathways. In mammalian cells, there are four major
MAPK pathways: ERK, JNK, p38 and ERKS pathways. MAPKs are activated by dual

phosphorylation of Thr and Tyr in the activation loop.

18

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

MAPK Isoforms Molecular weight Primary accessnon
number
ERK] 43 kDa P27361
ERK
ERK2 41 kDa P28482
INKlal 46 kDa AAA36131
INKloLZ 55 kDa AAC50607
INKlBl 46 kDa AAC50610
JNKIBZ 55 kDa AAC50611
INKZal 46 kDa AAC50606
INK
INK2a2 55 kDa AAA56831
INKZBI 46 kDa HSU35002
INKZB2 55 kDa HSU35003
INKBOLI 46 kDa AAC50605
INK3012 55 kDa HSU34819
P38 or 41 kDa Q16539
P38 B 42 kDa Q15759
P38
P38 y 42 kDa P53778
P38 5 42 kDa 015264

 

 

 

 

 

 

 

 

Fig. 4. Human MAPKs. The molecular weight and primary accession number of

different MAPK isoforms are shown.

19

4. Mixed lineage kinases

Mixed lineage kinases (MLKs) are a family of serine/threonine protein kinases
(Fig. 5) that were first described in mammals. MLKs belong to the tyrosine kinase like
(TKL) group of protein kinases [9]. Their catalytic domains share sequence similarities
to both tyrosine and serine/threonine protein kinases, giving rise to the name mixed
lineage kinase. However, to date, only serine/threonine phosphorylation activity has been
documented in MLKs. Biochemical and genetic studies have revealed that MLKs
function as MAPKKKs to activate the JNK and p38 pathways. MLKs have been
implicated in neuronal apoptosis and are therapeutic targets for many neurodegenerative

diseases such as Parkinson’s disease.

4.1. Family members of mixed lineage kinases

Over the past several years, seven human MLK genes have been identified. Based
on the domain arrangement and sequence homology within the catalytic domains, MLKs
cluster into 3 subfamilies: the MLKs; the dual leucine zipper bearing kinases (DLKs);
and zipper sterile-oc-motif kinase (ZAK).

The MLK subfamily includes MLKI, MLK2, MLK3, and MLK4. MLK4 has two
splicing forms, designated MLK40t and MLK4B. MLKs all contain an amino terminal
SH3 domain followed by a catalytic domain, a centrally located zipper region and a
CRIB motif. The kinase domains of MLK5 share more than 75% sequence identity. The
carboxyl terminal regions of MLK5 are quite divergent except that they are all proline
rich. This diversity indicates that different MLKs might be regulated by distinct upstream

signals through this region. MLK3 mRNA is widely expressed in adult and fetal human

20

tissues [84] and protein is detected by Western blotting in a wide variety of hematopoietic
and epithelial cell lines (Gallo lab, unpublished data). MLKl mRNA is detected in
epithelial tumor cell lines of breast, colonic and esophageal origin [85]. Both the RNA
and protein levels of MLK1 are detected in an immature B-cell line, RIN-SAH cells, but
are absent in a more mature RIN-A12 cells, suggesting that MLK] is developmentally
regulated in B-cell development [86]. The mRNA for MLK2 is present in brain, testis,
and skeletal muscle and is expressed at a lower level in pancreas [87, 88]. There is
currently no published information on the MLK4 mRNA distribution. Of the four MLKs,
only MLK2 and MLK3 have been biochemically characterized.

The DLKs include DLK (also called zipper protein kinase (ZPK) [89]) and
leucine—zipper kinase (LZK). DLKs lack an SH3 domain and a CRIB motif, but do have
two small leucine-zipper motifs that are separated by a 31 amino acid spacer that follow
the catalytic domain. Analogous to the MLKs, DLKs have a proline—rich carboxyl
terminal region, whose function has not been well defined. The catalytic domains of DLK
and LZK are 87% identical. DLK mRNA is highly expressed in adult mouse brain [90],
and is detected primarily in the hippocampus, the cerebral cortex, and the Purkinje cell
layer of the cerebellum by in situ hybridization analysis of mouse brain sections [89].
LZK mRNA is widely expressed, with the highest expression in the pancreas [91]. There
is relatively little information on the protein level of DLK and LZK, presumably because
the endogenous protein level may be too low to be detected by Western blotting.

The sole member of the third subfamily is ZAK [92], which has two different
splicing forms, ZAKa and ZAKB. Other names for ZAK includes MLK-like mitogen-

activated protein triple kinase (MLTK) [93], and MLK-related kinase (MRK) [94]. In

21

addition to the catalytic domain, ZAKoc contains leucine zipper region and a sterile-0t
motif (SAM). Due to different mRNA splicing, ZAKB lacks the SAM motif. ZAK
mRNA is expressed widely in human tissues, with highest expression in heart and
skeletal muscle [93, 94].

MLKs are absent from yeast, but MLK homologues are found in Drosophila and
C. elegans [66]. Specifically, Drosophila have a DLK and MLK homologue and C.
elegans have a DLK and ZAK homologue. In addition, C. elegans also contains a

distantly related SH3 containing MLK, lacking the zipper region and the CRIB motif.

4.2. Pathways regulated by MLKs.
4.2.1. The JN K pathway.

Upon overexpression in mammalian cells, all MLKs tested, including MLK2 [95],
MLK3 [96], DLK [97], LZK [91], and ZAK [92] [93], activate the JNK pathway. MLK2
and MLK3 are capable of phosphorylating both MKK4 and MKK7 in an in vitro kinase
assay [96] [95]. However, DLK phosphorylates and activates recombinant MKK7, but
not MKK4, in vitro [97], suggesting that DLK specifically utilizes MKK7 to transduce
the signals and activate the JNK pathway in vivo. Overexpression of ZAK enhances both
the in vitro catalytic activity of MKK4 and MKK7 [93]. However, another study
indicated that co-expression of dominant negative MKK7, but not MKK4, with ZAK,
blocks ZAK mediated INK activation [98]. Further studies are necessary to clarify this

issue.

4.2.2. The p38 pathway.

22

MLK2 and MLK3, when overexpressed in COS cells, modestly activate the p38
kinase. [99] [96]. Irnmunoprecipitated p38 from cells co-transfected with DLK and p38
shows increased activity in an in vitro kinase assay [100]. In addition, ZAKB modestly

activates p38 upon overexpression [94].

In general, upon overexpression of MLK5, the fold activation of JNK is much
higher than the fold activation of p38, suggesting that MLKs may serve as a potent

activator of INK pathway, but it only partially activates the p38 pathway.

4.2.3. The ERK pathway.

When cotransfected with ERK2 in COS cells, MLK2 enhances ERK in vitro
kinase activity by 11 fold [99]. Our lab also observed that overexpression of MLK3 in
HEK (human embryonic kidney epithelial) 293 cells activates the ERK pathway
(unpublished data). However, another report shows that MLK3 fails to enhance ERKl
activity when both MLK3 and ERK are overexpressed in COS cells [96]. Overexpressed
ZAK activates the ERK pathway in COS cells [93, 94]. However, overexpression of DLK
in COS 7 cells and in NIH 3T3 cells fails to increase the in vitro kinase activity of ERK2
[100]. It is unclear whether the activation of ERK pathway by some of the MLKs is cell
type specific. The contradictory results from different labs using the same cell line make

this question more complex. Further investigations are necessary to clarify these issues.

4.2.4. The NF-KB pathway.

23

NF-KB is a transcription factor involved in immune, inflammatory, and anti-
apoptotic responses. NF-KB is normally sequestered in the cytosol by binding to the
inhibitor of NF-KB alpha (licBoz). Phosphorylation of IKBa induces its degradation,
allowing NF-KB to translocate to the nucleus, bind to its target promoters, and activate
transcription. Phosphorylation of 11(ch is mediated by the serine/threonine protein
kinases IKB kinase alpha (IKKa) and IKKB (reviewed in [101]). In Jurkat cells,
overexpressed MLK3 induces phosphorylation and activation of IKKa and IKKB [102].
A kinase-defective mutant of MLK3 strongly blocks the NF-KB dependent transcription
in response to T cell activation by CD3/CD28 antibodies.

LZK induces NF-KB activation weakly (1.4 fold) as judged by a reporter gene
assay. Cotransfection of LZK with IKKB increases IKKB in vitro catalytic activity.
Furthermore, LZK associates with the H<K complex through the kinase domain as shown
by a coimmunoprecipitation experiment [103]. In addition, ZAK was also reported to

activate the NF-KB pathway as judged by transcription reporter assay [92].

In summary, almost all the studies regarding the activation of MAPK pathways
and NF-KB pathway by MLKs were carried out using overexpressed systems in
mammalian cells. There are technical difficulties to obtain purified, fully active MLKs to
study their enzymology to determine the km and Km towards their putative substrates in
vitro. Many issue further complicate the in vivo situation, including phosphorylation and
activation states of MLK3, the subcellular localization of the MLK and its substrate, and

the possible signaling specificity derived from scaffold proteins.

24

4.3 The regulation of MLKs
4.3.1. Regulation of MLKs by extracellular signals

The extracellular signals that activate endogenous MLKs are not well defined.
Serum, platelet-derived grth factor (PDGF), anisomycin, PMA, CD3/CD28 antibodies,
but not TNFOL induce endogenous MLK3 mobility changes, suggesting of
phosphorylation of MLK3. In addition, the phosphopeptide map of in vivo [32F] labeled
endogenous MLK3 from PMA-treated Jurkat cells is virtually identical to that obtained
from MLK3 overexpressed with activated Cdc42 in HEK 293 cells, indicating that PMA
induces the same phosphorylation events on MLK3 (Gallo lab and Esselman lab,
unpublished data). However, it is unclear whether these changes correspond to changes in
MLK3 activity.

Recently, TNFOL has been shown to increase the activity of endogenous MLK3
about two fold, as judged by the ability of MLK3 to phosphorylate recombinant GST-
MKK4 in vitro [104]. In addition, the lipid C6—ceramide treatment in Jurkat T cells is
reported to increase endogenous MLK3 catalytic activity to about the same extent. The
ability of the MLKs inhibitor, CBP-11004 to attenuate ceramide-induced INK activation,
suggests that activation of JNK by ceramide is accomplished at least partially through
MLKs. The activity of MLK3 is also reported to increase in the presence of ceramide in
an in vitro kinase assay system, suggesting that MLK3 is a direct target of ceramide
[104]. However, it will be important for other researchers to confirm these reports.

Treatment of mammalian cell lines with the Ser/Thr phosphatase inhibitor,
okadaic acid, results in retarded mobility of DLK [105, 106] and induces JNK activation,

presumably through inhibition of phosphatases that negatively regulate DLK and /or INK

25

pathway [106, 107]. However, it is unclear whether okadaic acid may affect the activity
of DLK.

Afier transient expression in COS-7 cells, ZAK has relatively high basal activity.
Treatment with osmotic shock, such as sorbitol, KCl or NaCl, increases the activity of
ZAK. However, the activity of ZAK remains unchanged upon treatment with fetal calf
serum, TGF-B, anisomycin, or UV, which are also upstream signals that activate INK
pathway [93].

Taken together, it seems that the activation of different subfamilies of MLKs
requires distinct upstream stimuli. This specificity may be mainly due to the different
domain arrangement and sequence diversity in the carboxyl terminal regions among the

subfamily members.

4.3.2. Regulation of MLKs by phosphorylation

Phosphorylation of protein kinases may rapidly and reversibly change the
activities of kinases, regulate its association with other molecules, and affect its
subcellular localization. Activation loop phosphorylation is a prominent regulatory
mechanism of protein kinases. Mutation of Thr 277 and Ser 281 to Glu (to mimic the
phosphorylated Ser and Thr) in the activation loop region in MLK3 results in fillly
functional kinase, whereas the MLK3 T277A or 8281A mutant has low activity,
suggesting that these two residues may serve as the phosphorylation sites to regulate
MLK3 activity [108]. Since the Thr and Ser residues in the activation loop are highly
conserved, phosphorylation in the activation loop is very likely to be a regulatory

mechanisms for all the MLKs.

26

Using mass spectrometry and phosphopeptide mapping techniques, our lab
identified eleven in viva phosphorylation sites of MLK3 overexpressed with activated
Cdc42 in HEK 293 cells [109]. Seven of the identified phosphorylation sites are followed
immediately by a proline residue, suggesting that MLK3 is a target for proline-directed
kinases. Since MLK3 activates INK, ERK and p38 to some extent, one interesting idea is
that MLK3 may be phosphorylated by MAPKs in a feedback mechanism. The precise
mechanism is under investigation in our lab. In addition, It has been shown that
overexpressed MLK2 can be phosphorylated in its carboxyl terminal region by JNK in
vitro [110]. However, how phosphorylation by INK or other MAPKs regulates MLK2 or
MLK3 is still unclear. It is necessary to generate antibodies directly against specific

phosphorylation sites to more carefully probe their function.

4.3.3. Regulation of MLKs by oligomerization

All seven MLK family members contain some type of leucine zipper. MLK3 is
reported to form disulfide-linked dimers upon overexpression in HEK 293 cells. The
leucine zipper region in MLK3 is necessary and sufficient for MLK3 oligomerization
[111]. Using site directed mutagenesis to deliberately introduce a helix disrupting proline
mutation in the zipper region, work in our lab has shown that the zipper mediated
oligomerization of MLK3 is required for proper phosphorylation of its substrate MKK4
and subsequent activation of JNK pathway [112].

Leucine zipper mediated DLK dimerization is necessary for DLK activity and
subsequent INK activation [113]. Using chimeric DLK—FK-binding protein (FKBP)

constructs, artificially induced dimerization of DLK monomers is sufficient to induce

27

DLK phosphorylation and INK activation[106]. LZK forms dimers/oligomers and
deletion of a zipper motif of LZK prevents activation on INK [114]. The regulation of
ZAK by its zipper region has not been determined.

Zipper-mediated oligomerization is very likely to be a common regulatory
mechanism to modulate the activity of all MLKs. Oligomerization may trigger the trans-
autophosphorylation of MLKs, leading to MLK activation and subsequent INK

activation.

4.3.4. Regulation of MLKs by Rho GTPase

The mammalian Rho GTPase family contains seven distinct proteins: Rho (A, B,
and C isoforms), Rac (1 and 2 isoforms), Cdc42 (Cdc42Hs and G25K isoforms), RhoD,
RhoG, RhoE, and TClO [115]. Rho GTPases cycle between active, GTP-bound states and
inactive, GDP-bound states, acting as molecular switches to control many cellular
processes, such as cytoskeletal organization, gene transcription, and proliferation [116].
Constitutively active forms of Rae and Cdc42 have been shown to activate the JNK and
p38 pathways [117]. Only MLK subfamily members (MLKl-4) contain a CRIB motif.
The CRIB motif is a 14-16 amino acid sequence with eight consensus residues that
mediates binding with activated form of Cdc42/Rae [8]. The CRIB motif is necessary but
not sufficient for MLK3’s interaction with Cdc42 [118]. Coexpression of MLK3 with
activated Cdc42 increases MLK3 catalytic activity and potentiates MLK3-mediated JNK
activation, and this is accompanied by a change in the in viva phosphorylation pattern of
MLK3 [118, 119]. Attempts to activate MLK3 by Cdc42 in an in vitro system using

purified, recombinant proteins have so far failed, suggesting that an additional component

28

‘i:

or the cellular environment may be required for MLK3’s activation [1 18]. Recent work in
our lab has shown that prenylated activated Cdc42 induces translocation of MLK3 to a
plasma-membrane enriched fraction (Gallo lab, manuscript in preparation).

Activated Cdc42 seems to promote the oligomerization of MLK3, suggesting that
the activation mechanism by Cdc42 involves oligomerization of MLK3 [111] [106].
Although most studies on the regulation of MLK5 by small GTPases have been focused

on Cdc42, activated Rac also binds to and potentiates MLK3 activity [119].

4.3.5. Regulation of MLKs by scaffold proteins

JIPs are scaffold proteins that function in the JNK and p38 pathways. J 1P1 was
first identified in a two-hybrid screen for INK-interacting proteins [120]. Subsequently it
was shown that JIPl binds MLKs, but not other MAPKKKs; MKK7, but not MKK4; and
multiple INK isoforms [121]. To date, three JIPs have been identified. While IIPl is
ubiquitously expressed in human tissues, IIP2 is predominantly expressed in brain [122].
IIP3 is structurally distinct from JIPl/2 and is highly expressed in brain, heart and lung
[123].

Several MLKs, including MLK2, MLK3, DLK and LZK have been shown to
interact with IIPl [121, 122] (Fig. 6). It has been suggested that JIPs facilitate the JNK
signaling by assembling the kinase components in the signaling pathway. This idea is
supported by the finding that coexpression of JIPs with MLK3 potentiates INK activation
[121-123]. Indeed, recent work with yeast MAPK scaffolds supports the idea that
scaffolds may simply function to increase the local concentration of signaling

components [124]. However, later studies indicate that the regulation of the JNK

29

pathway by JIPs may be more complex. These data suggest that I 1P1 bound DLK is
monomeric, unphosphorylated and catalytically inactive. INK promotes the dimerization,
phosphorylation and activation of IIP-associated DLK [106].

Compared with I 1P1 , JIP2 binds INK more weakly [122], suggesting that I IP2
may have other binding partners in cells. Indeed, I 1P2 has been shown to act as a scaffold
protein in the p38 pathway. JIP2 associates with MLK3, MKK3 and p38. In addition,
JIP2 associates with Tiaml and Ras-GRF, guanine nucleotide exchange factors for Rae.
Expression of JIP2 potentiates Tiaml or Ras-GRF 1 induced p38 activation. Expression of
Tiaml or Ras-GRFI also enhances association of JIP2 with MLK3, MKK3 and p38
[125].

Although the INK binding region is conserved, JIP3 is structurally distinct from
IIP1/2. The only MAPKKK reported to interact with I 1P3 is MLK3. JIP3 interacts with
MKK7 and multiple INK isoforms to facilitate INK signaling [123].

Genetic studies revealed that the Drosophila homologue of I 1P3, Sunday dn'ver
(Syd), is involved in kinesin-dependent axonal transport [126]. In addition, IIPs have
been shown to bind several other proteins that have diverse biological functions,
including the kinesin light chain. JIPs may associate with kinesin motor proteins and
serve as the loading dock for INK signaling components moving along microtubules
[127].

In addition to the IIPs, a protein named POSH (plenty of SH3s) also acts as a
scaffold for the JNK pathway to trigger neuronal apoptosis. POSH interacts directly with
activated Racl, MLKs (MLK1-3 and DLK), and indirectly with MKK4/7 and INKS,

presumably through another protein. POSH overexpression induces neuronal apoptosis

30

and this can be suppressed by dominant negative forms of MLKs, MKK4/7 and c-Jun,
and by the MLK inhibitor, CBP-1347 [128].

Post-synaptic density protein 95 (PSD-95) is a scaffold protein that contains three
PDZ domains, a SH3 domain, and an enzyrnatically inactive guanylate kinase domain
[129]. The PDZ domain of PSD-95 binds to and clusters ion channels, including the
kainate receptor glutamate receptor 6 (GluR6), in the postsynaptic density membrane of
neurons [130]. Expression of GluR6 activates the JNK pathway and induces neuronal
apoptosis. The SH3 domain of PSD-95 binds to MLK2 and MLK3. Co-expression of
kinase defective MLK2 or MLK3 significantly inhibits GluR6-mediated INK activation
and neuronal toxicity [131], suggesting that the PSD-95 acts as a scaffold protein to
mediate the signaling from ion channels to MLKs induced INK activation and neuronal

apoptosis.

4.4. The biological functions of MLKs in mammalian cells.
4.4.1. MLKs in neuronal apoptosis.

Nerve grth factor (N GF) deprivation-induced neuronal apoptosis requires both
the activity of the small GTPase Cdc42 and the activation of JNK pathway [132].
Recently, several studies in neuronal systems indicate that MLKs are involved in
neuronal apoptosis induced by NGF deprivation. Overexpression of MLK family
members, including MLKl-3 and DLK, effectively induces apoptosis in neuronal-like
PC12 cells and sympathetic neurons. NGF deprivation-induced apoptosis can be blocked

by expression of catalytically inactive forms of MLKs [133, 134].

31

The presence of polyglutamine-expanded huntingtin is a hallmark in the neuronal
cells of Huntington’s disease patients. The SH3 domain of MLK2 has been shown to
interact with normal huntingtin protein, but not polyglutamine-expanded huntingtin in
transfection experiments. Expression of catalytically inactive MLK2 attenuates neuronal
apoptosis induced by the polyglutamine-expanded huntingtin, suggesting that MLK2 may
be involved in polyglutamine-expanded huntingtin-mediated neuronal apoptosis [135].

The INK pathway has become a target for drugs against neurodegenerative
diseases. Cephalon has developed inhibitors that block INK activation and neuronal
apoptosis in cells and animal models. In vitro and in viva studies indicate that the direct
targets of these drugs are not INK, but MLKs. K252a [136] and its derivative, CBP-1347
[137], act as competitors of ATP to inhibit the kinase activity of recombinant MLKs in
vitro. CBP-1347 blocks INKl activation induced by MLKs (MLKl-3, DLK, LZK), but
not MEKKl. In addition, K252a and CBP-1347 also activate Akt and ERK in a MLK-
independent manner, suggesting that the neuroprotective and neurotrophic effects of
K252a and CEP1347 may involve modulation of several neurotrophic signaling pathways
[136].

Overexpression of gplZO, the major coat protein of the HIV-1, induces apoptosis
in primary hippocampal neurons, which is blocked by CBP-1347. In addition,
overexpression of MLK3 in hippocampal pyramidal neurons potentiates gplZOIIIB-
induced neuronal apoptosis, whereas expression of a catalytically inactive MLK3 blocks
the apoptosis induced by gp120IIIB. These results indicate that MLKs may mediate
gpl2OIIIB-induced neuronal death, suggesting potential clinical utility of CBP-1347 in

inhibiting AIDS-related dementia [138].

32

4.4.2. MLKs in cellular transformation, cell morphology changes and cancer.

Overexpression of wild-type MLK3 leads to morphological transformation of NIH
3T3 fibroblasts, whereas overexpression of catalytically inactive MLK3 fails to do so.
The MEK inhibitor PD 098059 partially reverts the MLK3-transformed phenotype [139].
However, another recent report showed that MLK3 inhibits Rac-mediated cellular
transformation [140]. Recently, comprehensive screening of 35 colorectal cancer cell
lines for mutations of genes encoding all the TK and TKL protein kinases in the human
genome was performed. Mutations in MLK4, but not other MLKs, were detected in
several cancer cell lines [141]. The identified mutations in the kinase domain of MLK4
occur in highly conserved residues (H261, G291, A293, and W296), suggesting that these
mutations may render MLK4 catalytically inactive. Therefore, it is conceivable that
MLK4 act as a tumor suppressor gene.

Overexpression of ZAKa, but not ZAKB, in Swiss 3T3 (mouse embryonic
fibroblast) cells and 10T1/2 (mouse embryonic fibroblast) cells, causes disruption of actin
stress fibers and cell shrinkage. The kinase defective mutant of ZAKa does not cause this
phenomenon [93] [98]. This suggests that ZAK may play a role in the regulation of actin

organization.

4.4.3. MLKs in cell cycle.
Using a bioinformatics approach, MLK3 was identified as a NIMA (never in
mitosis A)-related protein kinase. NIMA is a Ser/Thr kinase involved in cell cycle

regulation. Loss of NIMA function causes cell arrest in G2 phase [142]. During the G2/M

33

phase of synchronized HeLa cells, MLK3 is localized to the centrosome and there is a
retarded mobility of endogenous MLK3 due to hyperphosphorylation, which correlates
with an increase of MLK3 catalytic activity. However, INK activity remains unchanged
during the cell cycle, suggesting that MLK3 may function in a INK independent pathway
during cell cycle progression. In addition, overexpresed MLK3 causes profound
disruption of cytoplasmic microtubules and a nuclear distortion, suggesting that
endogenous MLK3 may perform a similar function during G2/M [143].

Interestingly, expression of ZAK increases the G2/M cell population [94, 98],
suggesting that ZAK may be involved in the regulation of the G2 checkpoint control.
Furthermore, catalytically inactive ZAK attenuates the y-radiation induced cell cycle

arrest [94]. Taken together, these data suggest a role for MLKs in cell cycle regulation.

4.5. The biological functions of MLKs in Drosophila.

The Drosophila melanogaster homologue of MLK1 -4, Slipper (Slpr), is involved
in INK mediated dorsal closure during embryogenesis. Dorsal closure is a process of cell
sheet movement, in which the dorsal ectoderrn moves from a lateral position to the dorsal
midline, thus enclosing the embryo in a continuous protective epidermis [144]. Dorsal
closure involves a pathway that requires small GTPase dRacl, the MKK7 homologue
Hep, the INK homologue Bsk and the transcription factor dIun and dFos [145]. Genetic
studies revealed that the Slpr mutant embryos have a dorsal open cuticle phenotype and
fail to maintain the dramatic cell shape changes within the dorsal epithelium. Genetic
epistasis tests indicate that Slpr functions at the downstream of GTPase dRac and

upstream of Hep and Bsk to regulate dorsal closure [146].

34

4.6. MLK3

MLK3, also called SH3 domain-containing proline—rich protein kinase (SPRK) or
protein tyrosine kinase 1 (PTK-l) is a serine/threonine kinase with a predicted molecular
weight of 93 kDa [84, 147, 148]. MLK3 contains a unique NHz-terminal glycine-rich
region followed by an SH3 domain and a kinase catalytic domain. Following the kinase
domain, there are two closely spaced leucine/isoleucine zipper motifs followed by a
CRIB motif and a stretch of basic amino acids. The final 220 amino acids of MLK3 are
rich in proline (24%), serine (12%) and threonine (13%) [84].

Our lab has been investigating the molecular mechanisms that regulate MLKs
activities using MLK3 as a paradigm. While several laboratories have focused on the
downstream effects of MLK3 expression in different cell types, less is known about the
molecular mechanisms regulating its activity and its intracellular signaling pathways.
Many aspects of MLK3 regulation are likely to hold for other family members that share
homologous domains. Understanding the molecular mechanisms that regulate MLK3
activity should provide insight into the regulation of protein kinases and aid in the design

of isofonn-specific MLK inhibitors that may be useful therapeutics in the future.

35

SH3 lGnase LZ CRIB

r’“ w"! II—w l
I.

P—'

MLK1 L

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

.m .__._. J .1 LI -J
1066
MLK2 LILIT‘TJJ 14E 1
Gly 954
MLK MLK3 Ll"::.: __. ‘1 T C: l
847
MLK4a F f 1:: ‘:*: :1": “‘1 I: 1
570
LMLK4B L l 1:“ :1“ 3 IE: 1
r. 1036
DLK I :;___ __ ' *1} :1 “j
DLK 359
__LZK I r 7 l [l ‘l
r— SAM 966
ZAKa 1:7?" "" 1T1 1
ZAK 800
3’“ LL ll 1
455

Fig. 5. Composite structure of human mixed lineage kinases. The relative position of
SH3 domain, leucine zipper (LZ), CRIB motif, and sterile a motif (SAM) was shown.

(Adapted from [66])

36

 

Fig. 6. Scaffold proteins that interact with MLKs. IIPl/3 specifically bind to
MLKs, MKK7, and INK to coordinate JNK pathway. JIPZ binds to MLKs, MKK3
and p38. POSH interacts with activated Rae and MLKs directly. Interaction of
MKK4/7 and INKS with POSH may be mediated by an additional protein,

designated here as ‘X’.

37

5. Objective of Thesis

Protein kinases are involved in many physiological processes, such as
proliferation, differentiation and apoptosis. Therefore, their activities are tightly
regulated.

Our lab has been focused on the molecular mechanisms regulating MLK3. In
addition to its catalytic domain, MLK3 contains several domains, which are predicted to
be critical for regulating the activity and signaling of MLK3. Previous work in the lab has
explored the importance of the zipper region and CRIB motif. However, relatively little is
known about the function of the SH3 domain and the carboxyl terminal proline rich
region. In addition, the extracellular signals that affect MLK3 activity have not been well
defined, and the proteins that interact with MLK3 are largely unknown. This thesis
describes the examination of various aspects of mechanisms regulating MLK3.

Chapter 11 presents investigation of the function of the SH3 domain of MLK3 in
regulating MLK3 activity. Chapter III describes an observed proteolytic event of MLK3.
The study examines the cleavage site of MLK3, the specificity of the proteolysis, as well
as the protease that is involved in the proteolysis. Chapter IV details the identification of
MLK3 binding partners by affinity purification and mass spectrometry. Chapter V
describes the investigation of the regulation of MLK3 by heat shock protein
Hsp90/Cdc37. The study examines whether Hsp90/Cdc37 affect MLK3 protein level, as

well as MLK3 mediated INK signaling.

38

6. References:

10.

11.

12.

13.

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

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

Steeg, P.S., et al., Histidine kinases and histidine phaspharylatedprateins in
mammalian cell biology, signal transduction and cancer. Cancer Lett, 2003.
190(1): p. 1-12.

Hanks, SK. and T. Hunter, Protein kinases 6. The eukaryotic protein kinase
superfamily: kinase (catalytic) domain structure and classification. Faseb I , 1995.
9(8): p. 576-96.

Johnson, L.N., et al., The Eleventh Datta Lecture. The structural basis for
substrate recognition and control by protein kinases. FEBS Lett, 1998. 430(1-2):
p. 1-11.

Bossemeyer, D., Protein kinases--structure and function. FEBS Lett, 1995.
369(1): p. 57-61.

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.

Burbelo, P.D., D. Drechsel, and A. Hall, A conserved binding motif defines
numerous candidate target prateinsfar bath Cdc42 and Rac GTPases. J Biol
Chem, 1995. 270(49): p. 29071-4.

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

Newton, A.C., Protein kinase C: structure, function, and regulation. J Biol Chem,
1995. 270(48): p. 28495-8.

Newton, A.C., Protein kinase C: structural and spatial regulation by
phosphorylation, cafactars, and macromolecular interactions. Chem Rev, 2001.
101(8): p. 2353-64.

Tatosyan, AG. and O.A. Mizenina, Kinases of the Src family: structure and
functions. Biochemistry (Mosc), 2000. 65(1): p. 49-58.

Xu, W., S.C. Harrison, and M.J. Eck, Three-dimensional structure of the tyrosine
kinase c-Src. Nature, 1997. 385(6617): p. 595-602.

39

14.

15.

l6.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

28.

Soderling, T.R., Protein kinases. Regulation by autoinhibitory domains. J Biol
Chem, 1990. 265(4): p. 1823-6.

Hubbard, S.R., Src autoinhibition: let us count the ways. Nat Struct Biol, 1999.
6(8): p. 711-4.

Williams, I.C., R.K. Wierenga, and M. Saraste, Insights into Src kinase functions:
structural comparisons. Trends Biochem Sci, 1998. 23(5): p. 179-84.

Moarefi, I., et al., Activation of the Src-family tyrosine kinase Hck by SH3 domain
displacement. Nature, 1997. 385(6617): p. 650-3.

Olashaw, N. and W.I. Pledger, Paradigms of growth control: relation to Cdk
activation. Sci STKE, 2002. 2002(134): p. RE7.

Steinman, R.A., Cell cycle regulators and hematopoiesis. Oncogene, 2002.
21(21): p. 3403-13.

Hershko, A. and A. Ciechanover, The ubiquitin system. 1998.

Lee, H.W., et al., Ubiquitinatian ofpratein kinase C-alpha and degradation by
the proteasome. J Biol Chem, 1996. 271(35): p. 20973-6.

Harris, K.F., et al., Ubiquitin-mediated degradation of active Src tyrosine kinase.
Proc Natl Acad Sci U S A, 1999. 96(24): p. 13738-43.

Oda, H., S. Kumar, and RM. Howley, Regulation of the Src family tyrosine
kinase Blk through E6AP—mediated ubiquitination. Proc Natl Acad Sci U S A,
1999. 96(17): p. 9557-62.

Eamshaw, W.C., L.M. Martins, and SH. Kaufrnann, Mammalian caspases:
structure, activation, substrates, and functions during apoptosis. Annu Rev
Biochem, 1999. 68: p. 383-424.

Wolf, BB. and DR. Green, Suicidal tendencies: apaptatic cell death by caspase
family prateinases. J Biol Chem, 1999. 274(29): p. 20049-52.

Shi, Y., Mechanisms of caspase activation and inhibition during apoptosis. Mol
Cell, 2002. 9(3): p. 459-70.

Basu, A., et al., Proteolytic activation of protein kinase C-epsilan by caspase-
mediated processing and transduction afantiapaptatic signals. I Biol Chem,
2002. 277(44): p. 41850-6.

Emoto, Y., et al., Proteolytic activation ofpratein kinase C delta by an ICE-like
protease in apaptatic cells. Embo I, 1995. 14(24): p. 6148-56.

40

29.

30.

31.

32.

33.

34.

35.

36.

37.

38.

39.

40.

41.

Widmann, C., S. Gibson, and G.L. Johnson, Caspase-dependent cleavage of
signaling proteins during apoptosis. A turn-off mechanism for anti-apaptatic
signals. I Biol Chem, 1998. 273(12): p. 7141-7.

Sorimachi, H., S. Ishiura, and K. Suzuki, Structure and physiological function of
calpains. Biochem J, 1997. 328(Pt 3): p. 721-32.

Wang, K.K., Calpain and caspase: can you tell the diflerencea by kevin K. W.
WangVal. 23, pp. 20-26. Trends Neurosci, 2000. 23(2): p. 59.

Glading, A., et al., Epidermal growth factor receptor activation of calpain is
required for fibroblast motility and occurs via an ERK/MAP kinase signaling
pathway. J Biol Chem, 2000. 275(4): p. 2390-8.

Glading, A., et al., Membrane proximal ERK signaling is required for M-calpain
activation downstream of epidermal growth factor receptor signaling. J Biol
Chem, 2001. 276(26): p. 23341-8.

Nollen, EA. and RI. Morimoto, Chaperoning signaling pathways: molecular
chaperones as stress-sensing 'heat shock' proteins. I Cell Sci, 2002. 115(Pt 14): p.
2809-16.

An, W.G., T.W. Schulte, and L.M. Neckers, The heat shock protein 90 antagonist
geldanamycin alters chaperone association with p210bcr-abl and v-src proteins
before their degradation by the proteasome. Cell Growth Differ, 2000. 11(7): p.
355-60.

Schulte, T.W., et al., Disruption of the Raf-1-Hsp90 molecular complex results in
destabilization of Raf-1 and loss of Raf-I -Ras association. J Biol Chem, 1995.
270(41): p. 24585-8.

Basso, A.D., et al., Aktfarms an intracellular complex with heat shock protein 90
(Hsp90) and Cdc3 7 and is destabilized by inhibitors of Hsp90 function. J Biol
Chem, 2002. 277(42): p. 39858-66.

Chen, G., P. Cao, and D.V. Goeddel, INF-induced recruitment and activation of
the IKK complex require Cdc3 7 and Hsp90. Mol Cell, 2002. 9(2): p. 401-10.

Gao, T. and AC. Newton, The turn motif is a phosphorylation switch that
regulates the binding of Hsp 70 to protein kinase C. J Biol Chem, 2002. 277(35):
p. 3 1 5 85-92.

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

Widmann, C., et al., Mitogen-activated protein kinase: conservation of a three-
kinase module from yeast to human. Physiol Rev, 1999. 79(1): p. 143-80.

41

42.

43.

44.

45.

46.

47.

48.

49.

50.

51.

52.

53.

54.

55.

Johnson, G.L. and R. Lapadat, Mitogen-activated protein kinase pathways
mediated by ERK, JNK, and p38 protein kinases. Science, 2002. 298(5600): p.
191 1-2.

Chang, L. and M. Karin, Mammalian MAP kinase signalling cascades. Nature,
2001. 410(6824): p. 37-40.

Zhou, G., Z.Q. Bao, and IE. Dixon, Components of a new human protein kinase
signal transduction pathway. J Biol Chem, 1995. 270(21): p. 12665-9.

Su, B. and M. Karin, Mitogen-activated protein kinase cascades and regulation of
gene expression. Curr Opin Immunol, 1996. 8(3): p. 402-11.

Boulton, T.G., et al., An insulin-stimulated protein kinase similar to yeast kinases
involved in cell cycle control. Science, 1990. 249(4964): p. 64-7.

Cobb, M.H., MAP kinase pathways. Prog Biophys Mol Biol, 1999. 71(3-4): p.
479-500.

Kyriakis, I.M., et al., Raf-1 activates MAP kinase-kinase. Nature, 1992.
358(6385): p. 417-21.

Wu, X., et al., Selective activation of MEK I but not MEK2 by A-Raf from
epidermal growth factor-stimulated Hela cells. J Biol Chem, 1996. 271(6): p.
3265-71.

Vaillancourt, R.R., A.M. Gardner, and G.L. Johnson, B-Raf-dependent regulation
of the MEK-I/mitogen-activated protein kinase pathway in PC12 cells and
regulation by cyclic AMP. Mol Cell Biol, 1994. 14(10): p. 6522-30.

Vojtek, A.B., S.M. Hollenberg, and IA. Cooper, Mammalian Ras interacts
directly with the serine/threonine kinase Raf Cell, 1993. 74(1): p. 205-14.

Lange-Carter, C.A., et al., A divergence in the MAP kinase regulatory network
defined by MEK kinase and Raf. Science, 1993. 260(5106): p. 315-9.

Blank, I.L., et al., Molecular cloning of mitogen-activated protein/ERK kinase
kinases (MEKK) 2 and 3. Regulation of sequential phosphorylation pathways
involving mitogen-activatedpratein kinase and c-Jun kinase. J Biol Chem, 1996.
271(10): p. 5361-8.

Ceci, I.D., et al., T pl-2 is an oncogenic kinase that is activated by carboxy-
terminal truncation. Genes Dev, 1997. 11(6): p. 688-700.

Zheng, CE. and KL. Guan, Cloning and characterization of two distinct human

extracellular signal- regulated kinase activator kinases, MEK] and MEK2. J Biol
Chem, 1993. 268(15): p. 11435-9.

42

56.

57.

58.

59.

60.

61.

62.

63.

64.

65.

66.

67.

68.

69.

70.

Davies, H., et al., Mutations of the BRAF gene in human cancer. Nature, 2002.
417(6892): p. 949-54.

Mansour, S.J., et al., T ransfarmatian of mammalian cells by constitutively active
MAP kinase kinase. Science, 1994. 265(5174): p. 966-70.

Robinson, M.J., et al., A constitutively active and nuclear form of the MAP kinase
ERK2 is suflicientfor neurite outgrowth and cell transformation. Curr Biol, 1998.
8(21): p. 1141-50.

Pages, G., et al., Defective thymocyte maturation in p44 MAP kinase (Erk 1)
knockout mice. Science, 1999. 286(5443): p. 1374-7.

Ip, Y.T. and R]. Davis, Signal transduction by the c-Jun N-terminal kinase

(JNK)--from inflammation to development. Curr Opin Cell Biol, 1998. 10(2): p.
205-19.

Weston, CR. and R]. Davis, The flVK signal transduction pathway. Curr Opin
Genet Dev, 2002. 12(1): p. 14-21.

Gupta, S., et al., Selective interaction of JNK protein kinase isoforms with
transcription factors. Embo I, 1996. 15(11): p. 2760-70.

Schlesinger, T.K., et al., The TAO of MEKK. Front Biosci, 1998. 3: p. D1 181-6.

Yamaguchi, K., et al., Identification of a member of the MAPKKK family as a
potential mediator of T GF -beta signal transduction. Science, 1995. 270(5244): p.
2008-11.

Ichijo, H., et al., Induction of apoptosis by ASK I , a mammalian MAPKKK that
activates SAPK/JNK and p38 signaling pathways. Science, 1997. 275(5296): p.
90-4.

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.

Whitrnarsh, AI and RI. Davis, T ranscriptian factor AP-I regulation by mitogen-
activated protein kinase signal transduction pathways. J Mol Med, 1996. 74(10):
p. 589-607.

Fuchs, S.Y., et al., Phosphorylation-dependent targeting of c-Jun ubiquitination
by Jun N— kinase. Oncogene, 1996. 13(7): p. 1531-5.

Chow, C.W., et al., Nuclear accumulation of NFA T 4 opposed by the JNK signal
transduction pathway. Science, 1997. 278(5343): p. 1638-41.

Martinou, I.C., S. Desagher, and B. Antonsson, Cytochrome c release from
mitochondria: all or nothing. Nat Cell Biol, 2000. 2(3): p. E41-3.

43

71.

72.

73.

74.

75.

76.

77.

78.

79.

80.

81.

82.

83.

84.

85.

Tournier, C., et al., Requirement of JNK for stress-induced activation of the
cytochrome c- mediated death pathway. Science, 2000. 288(5467): p. 870-4.

Davis, R.I., Signal transduction by the JNK group of MAP kinases. Cell, 2000.
103(2): p. 239-52.

Sabapathy, K., et al., Defective neural tube morphogenesis and altered apoptosis
in the absence of both JNK] and JNKZ. Mech Dev, 1999. 89(1-2): p. 115-24.

Yang, D.D., et al., Absence of excitotoxicity-induced apoptosis in the
hippocampus of mice lacking the Jnk3 gene. Nature, 1997. 389(6653): p. 865-70.

Lee, I.C., et al., A protein kinase involved in the regulation of inflammatory
cytokine biosynthesis. Nature, 1994. 372(6508): p. 739-46.

Moriguchi, T., et al., A novel kinase cascade mediated by mitogen-activated
protein kinase kinase 6 and MKK3. J Biol Chem, 1996. 271(23): p. 13675-9.

Crawley, I.B., et al., T cell proliferation in response to interleukins 2 and 7
requires p38MAP kinase activation. J Biol Chem, 1997. 272(23): p. 15023-7.

Brenner, B., et al., Fas- or ceramide-induced apoptosis is mediated by a RacI-
regulated activation of Jun N-terminal kinase/p38 kinases and GADD153. J Biol
Chem, 1997. 272(35): p. 22173-81.

Nemoto, S., et al., Induction of apoptosis by S8202190 through inhibition of
p38beta mitogen-activated protein kinase. J Biol Chem, 1998. 273(26): p. 16415-
20.

Tamura, K., et al., Requirement for p38alpha in erythropaietin expression: a role
for stress kinases in erythropoiesis. Cell, 2000. 102(2): p. 221-31.

Adams, R.H., et al., Essential role of p38alpha MAP kinase in placental but not
embryonic cardiovascular development. Mol Cell, 2000. 6(1): p. 109-l6.

Chen, W., M.A. White, and M.H. Cobb, Stimulus-specific requirements for MAP3
kinases in activating the JNK pathway. J Biol Chem, 2002. 277(51): p. 49105-10.

Schaeffer, H.J., et al., MP1: a MEK binding partner that enhances enzymatic
activation of the MAP kinase cascade. Science, 1998. 281(5383): p. 1668-71.

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

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

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.

44

86.

87.

88.

89.

90.

91.

92.

93.

94.

95.

96.

97.

DeAizpurua, H.I., et al., Expression of mixed lineage kinase-1 in pancreatic beta-
cell lines at different stages of maturation and during embryonic pancreas
development. J Biol Chem, 1997. 272(26): p. 16364-73.

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

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

Blouin, R., et al., Cell-specific expression of the ZPK gene in adult mouse tissues.
DNA Cell Biol, 1996. 15(8): p. 631-42.

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

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 kinase from human brain. I Biol
Chem, 1997. 272(45): p. 28622-9.

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.

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.

Gross, E.A., 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.

Cuenda, A. and D.S. Dorow, Differential 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.

Rana, A., et al., The mixed lineage kinase SPRKphosphorylates 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 MKK 7 and not MKK4
as substrate. I Biol Chem, 1999. 274(15): p. 10195-202.

45

98.

99.

100.

101.

102.

103.

104.

105.

106.

107.

108.

109.

110.

Yang, J.I., 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.

Nagata, K., et al., The MAP kinase kinase kinase MLK2 co-lacalizes with
activated JNK along microtubules and associates with kinesin superfamily motor
KIF3. Embo I, 1998. 17(1): p. 149-58.

Fan, G., et al., Dual leucine zipper-bearing kinase (DLK) activates p46SAPK and
p38mapk but not ERK2. J Biol Chem, 1996. 271(40): p. 24788-93.

Israel, A., The [K 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.

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.

Mata, M., et al., Characterization of dual leucine zipper-bearing kinase, a mixed
lineage kinase present in synaptic terminals whose phosphorylation state is
regulated by membrane depolarization via calcineurin. J Biol Chem, 1996.
271(28): p. 16888-96.

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

Barancik, M., P. Htun, and W. Schaper, Okadaic acid and anisomycin are
protective and stimulate the SAPK/JNK pathway. I Cardiovasc Pharrnacol, 1999.
34(2): p. 182-90.

Leung, I.W. and N. Lassam, 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.

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.

Phelan, D.R., et al., Activated JNKphosphorylates the c-terminal domain of

MLK2 that is required for MLK2-induced apoptosis. J Biol Chem, 2001. 276(14):
p. 10801-10.

46

111.

112.

113.

114.

115.

116.

117.

118.

119.

120.

121.

122.

123.

Leung, I.W. and N. Lassam, 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.

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 JNK pathway. Monomeric SPRK
L410P does not catalyze the activating phosphorylation of Thr258 of murine
MIT OGEN-A C T I VA T ED protein kinase kinase 4. J Biol Chem, 2000. 275(36): p.
27893-900.

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

required for complex formation and stress-activatedpratein kinase activation. J
Biol Chem, 2000. 275(10): p. 7273-9.

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

Mackay, DJ. and A. Hall, Rho GT Pases. J Biol Chem, 1998. 273(33): p. 20685-8.

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

Coso, O.A., et al., The small GT P-binding proteins RacI and Cdc42 regulate the
activity of the JNK/SAPK signaling pathway. Cell, 1995. 81(7): p. 1137-46.

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.

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
ralefor mixed lineage kinase 3/protein-tyrosine kinase 1, a novel member of the
mixed lineage kinase family. I Biol Chem, 1996. 271(44): p. 27225-8.

Dickens, M., et al., A cytoplasmic inhibitor of the JNK signal transduction
pathway. Science, 1997. 277(5326): p. 693-6.

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

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

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

47

124.

125.

126.

127.

128.

129.

130.

131.

132.

133.

134.

135.

136.

137.

Park, S.H., A. Zarrinpar, and WA. Lim, Rewiring MAP kinase pathways using
alternative scaflold assembly mechanisms. Science, 2003. 299(5609): p. 1061-4.

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

Bowman, A.B., et al., Kinesin-dependent axonal transport is mediated by the
sunday driver (S YD) protein. Cell, 2000. 103(4): p. 583-94.

Verhey, K.I., et al., Cargo of kinesin identified as JIP scaffolding proteins and
associated signaling molecules. I Cell Biol, 2001. 152(5): p. 959-70.

Xu, Z., N.V. Kukekov, and LA. Greene, POSH acts as a scafloldfar a

multiprotein complex that mediates JNK activation in apoptosis. Embo I, 2003.
22(2): p. 252-61.

Kim, B, et al., Heteramultimerization and NMDA receptor-clustering activity of

Chapsyn- 110, a member of the PSD-95 family of proteins. Neuron, 1996. 17(1):
p. 103-13.

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.

Mielke, K. and T. Herdegen, JNK and p38 stresskinases--degenerative efi’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.

Xu, 2., et al., The MLK family mediates c-Jun N-terminal kinase activation in
neuronal apoptosis. Mol Cell Biol, 2001. 21(14): p. 4713-24.

Liu, Y.F., D. Dorow, and J. Marshall, Activation of MLK2-mediated signaling

cascades by palyglutamine- expanded huntingtin. J Biol Chem, 2000. 275(25): p.
19035-40.

Roux, P.P., et al., K252a and CEP1347 are neuroprotective compounds that
inhibit mixed-lineage kinase-3 and induce activation of Akt and ERK. I Biol
Chem, 2002. 277(51): p. 49473-80.

Maroney, A.C., et al., Cep-1347 (KT 75 I 5), a semisynthetic inhibitor of the mixed
lineage kinase family. J Biol Chem, 2001. 276(27): p. 25302-8.

48

138.

139.

140.

141.

142.

143.

144.

145.

146.

147.

148.

Bodner, A., et al., Mixed lineage kinase 3 mediates gpIZOIIIB-induced
neurotaxicity. J Neurochem, 2002. 82(6): p. 1424-34.

Hartkamp, I., I . Troppmair, and UK. Rapp, The JNK/SAPK activator mixed
lineage kinase 3 (MLK3) transforms NIH 3T 3 cells in a MEK-dependent fashion.
Cancer Res, 1999. 59(9): p. 2195-202.

Lambert, I.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.

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

Osmani, A.H., S.L. McGuire, and SA. Osmani, Parallel activation of the NIMA
and p34cdc2 cell cycle-regulated protein kinases is required to initiate mitosis in
A. nidulans. Cell, 1991. 67(2): p. 283-91.

Swenson, K.I., K.E. Winkler, and AR. Means, A New Identity far 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.

Knust, E., Drosophila morphogenesis: movements behind the edge. Curr Biol,
1997. 7(9): p. R558-61.

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

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.

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.

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

49

II. Autoinhibition of Mixed Lineage Kinase 3 Through Its SH3 Domain

1. Abstract

Mixed lineage kinase 3 (MLK3) is a serine/threonine protein kinase that functions
as a mitogen-activated protein kinase kinase kinase to activate the c-Iun NHz-terminal
kinase pathway. MLK3 has also been implicated as an IKB kinase kinase in the
activation of NF-KB. Amino-terminal to its catalytic domain, MLK3 contains a src-
homology 3 (SH3) domain. SH3 domains harbor three highly conserved aromatic amino
acids that are important for ligand binding. In this study, we mutated one of these
corresponding residues within MLK3 to deliberately disrupt the function of its SH3
domain. This SH3-defective mutant of MLK3 exhibited increased catalytic activity
compared to wildtype MLK3 suggesting that the SH3 domain negatively regulates MLK3
activity. We report herein that the SH3 domain of MLK3 interacts with full length
MLK3, and we have mapped the site of interaction to a region between the zipper and the
Cdc42/Rac Interactive Binding (CRIB) motif. Interestingly, the SH3 binding region
contains not a proline-rich sequence but, rather, a single proline residue. Mutation of this
sole proline abrogates SH3 binding and increases MLK3 catalytic activity. Taken
together, these data demonstrate that MLK3 is autoinhibited through its SH3 domain.
The critical proline residue in the SH3 binding site of MLK3 is conserved in the closely
related family members, MLKl and MLK2, suggesting a common autoinhibitory
mechanism among these kinases. Our study has revealed the first example of SH3
domain-mediated autoinhibition of a serine/threonine kinase and provides insight into the

regulation of the mixed lineage family of protein kinases.

50

 

2. Introduction

Virtually all physiological processes are regulated by reversible protein
phosphorylation. Therefore protein kinases and phosphatases should be highly regulated
enzymes. Physical interactions with other signaling molecules can modulate protein
kinase activity by changing subcellular location, autophosphorylation state, or
conformation of a protein kinase. In the absence of intermolecular associations, many
kinases are kept in an inhibited state that is maintained by intramolecular interactions.

Mixed lineage kinase 3 (MLK3)1, also called src homology 3 (SH3) domain-
containing proline-rich kinase (SPRK)[1], is a serine/threonine kinase that functions as a
mitogen-activated protein kinase (MAPK) kinase kinase (MKKK) to phosphorylate and
activate the dual specific kinases, MAPK kinase 4 (MKK4) [2]and MKK7 [3] which, in
turn, can phosphorylate and activate c-Iun NHz-terminal kinase (JNK). Recent evidence
suggests that IKB kinase (1 (IKKa) and IKKB are also substrates of MLK3, and that
MLK3 acts as an IKK kinase to activate the NF-KB pathway in response to T cell
receptor costimulation [4].

MLK3 contains several domains that are predicted to mediate protein-protein
interactions including a src-homology 3 (SH3) domain, a leucine zipper, and a
Cdc42/Rac Interactive Binding (CRIB) motif (Fig. 1). The small GTPase Cdc42, in its
activated state, binds to MLK3 through the centrally located CRIB motif, and increases
MLK3 catalytic activity [5-7]. This activation is accompanied by a change in the
phosphopeptide map pattern of in viva labeled MLK3, suggesting that Cdc42 binding
induces activating phosphorylation event(s) on MLK3 [7]. A leucine zipper that resides

NHz-terminal to the CRIB motif is necessary for MLK3 homo-oligomerization [8].

51

Work in our lab has shown that zipper-mediated oligomerization is not required for
MLK3 activation by Cdc42, but instead is critical for proper interaction and
phosphorylation of a downstream target, MKK4, and subsequent INK activation [9].

The NHz-terminal portion of MLK3 contains an SH3 domain. SH3 domains are
independently folding modules of about 60 amino acids. Although different SH3
domains have distinct ligand preferences [10], the consensus binding site is composed of
a short proline-rich sequence, oflen Pro-Xxx-Xxx-Pro preceded or followed by basic
amino acids [1 1-14]. The SH3 domain of MLK3 has been shown to interact with a
proline-rich region in hematopoietic progenitor protein kinase-1 (HPK1)[15], but an
effect on MLK3 activity has not been demonstrated.

Since MLK3 contains a large proline-rich COOH-terminal region (Fig. 1), we
were interested in the possibility that the SH3 domain of MLK3 might bind to this region
in an autoregulatory fashion. In the work presented here, we show that the SH3 domain
of MLK3 does indeed play a critical role in autoregulation. However, unexpectedly, the
SH3 domain binds to a region between the zipper and CRIB motifs that lacks a classical
SH3 binding site. Point mutations, either in the SH3 domain or in the newly identified
SH3 binding site of MLK3, which disrupt this interaction, result in increased MLK3
activity. The data presented here indicate that the catalytic activity of MLK3 is
negatively regulated by its SH3 domain. This is the first demonstration of SH3-mediated

autoinhibition of a serine/threonine kinase.

52

3. Materials and Methods

3.1 DNA Constructs and Mutagenesis

Construction of the cytomegalovirus-based expression vectors carrying the
cDNAs for wildtype MLK3 (pRK5-NFlag.mlk3), MLK3 L410P, and MLK3 A430-486
has been described elsewhere [7, 9]. A series of truncation variants of MLK3 were
generated by 15 cycles of amplification of pRK5-NFlag.mlk3 using the polymerase chain
reaction (PCR) with Pfic polymerase (Invitrogen). The same 5’ oligonucleotide was used
for the construction of a series of variants with successive COOH-terrninal deletions: 5’-
GCATTAGCTAGCACCATGGACTACAAG-3’.

The following oligonucleotides were used as 3’ primers:

pRK5-NF1ag.mlk3 1-635: 5’-CGTTAAGGATCCTCAAGAGCTGCTACCGCG-3’;
pRK5 -NFlag.mlk3 1 -598: 5 ’-CGTTAAGGATCCTCAGGATGAGTCATCTGA-3 ’;
pRK5 -NFlag.mlk3 1 -5 29: 5 ’-CGTTAAGGATCCTCACCGGGGAAAGGTGGG-3 ’;
pRK5 -NFlag.mlk3 1-485: 5 ’-CGTTAAGGATCCTCACGCCCGGAGCTTGCT-3 ’;
pRK5-NFlag.mlk3 1-386: 5’-CGTTAAGGATCCTCATTCCCGTAGGACCTG-3’.
The amplified mlk3 fragments were subcloned in-frarne with the Flag coding sequence
into pRK5-NFlag.mlk3 using NheI and BamHI.

For the construction of the hemagglutinin (HA)- tagged MLK3 variants, the
following oligonucleotides were used in the PCR reactions with pRK5-NFlag.mlk3 as the
template:
pCGN-HA.mlk3 1-114: 5’-CGTTAGTCTAGAATGGAGCCCTTGAAGAG-3’ and 5’-

GCATTAGGATCCTC AGAAGCTGGC C ACCTC GC -3 ’

53

 

pCGN-HA.mlk3 115-399: 5’-CGTTAGTCTAGACAGGAGCTGCGGCTGG-3’ and 5’-
GCATTAGGATCCTCACCAGCCTTCCTGCATGG-3’

pCGN-HA.mlk3 400-591: 5’-CGTTAGTCTAGAAAGCGCGAGATCCAGGG-3’ and
5’-GCATTAGGATCCTCAGTACCATGTGGCTTCG-3’

pCGN-HA.mlk3 592-847: 5’-CGAATGTCTAGACTGGATTCAGATGACTC-3’ and 5’-
GCATTAGGATCCTCAAGGCCCCGCTTCCGGC-3 ’

The amplified mlk3 fragments were subcloned in-frame with the HA coding sequence
into the pCGN mammalian expression vector [16] using XbaI and BamHI .

Variants of MLK3 containing point mutations were constructed using the Quick
Change site-directed mutagenesis method (Stratagene) with pRK5-NFlag.mlk3 as the
template DNA using Pfx polymerase (Invitrogen) and 15 cycles of amplification. The
following oligonucleotides, and their reverse complements, were used as mutagenesis
primers:
pRK5-NFlag.mlk3 Y52A: 5’-GGACAGCCCTGTTCGACGCCGAGCCCAGTGG-3’
pRK5-NFlag.mlk3 P469A: 5 ’-GTGGACCGCGAGCGAGCGCACGTGC-3 ’

The presence of the appropriate mutations was confirmed by DNA sequencing (MSU
DNA Sequencing Facility).

A Glutathione S-transferase (GST) expression vector (pGEM-2T-SH3) carrying
the cDNA encoding the SH3 domain of MLK3 (amino acids 43-104) was constructed by
PCR-mediated amplification of the corresponding coding sequence from pRK5-mlk3
followed by subcloning into the pGEM-2T vector. The Y52A mutation was introduced
by site-directed mutagenesis of pGEM—2T-SH3 using the same oligonucleotides that were

used to construct pRK5-NFlag.mlk3 Y52A.

54

3.2 Expression and Purification of GST Fusion Proteins

GST and GST fusion proteins (GST-SH3 and GST-SH3 Y52A) were expressed in
E. coli and purified using glutathione Sepharose 4B, according to the manufacturer’s
protocol (Amersham Phamacia Biotech). Eluted fractions containing the GST fUSlOI'l
protein, as determined by SDS-polyacrylamide gel electrophoresis (PAGE) followed by
Coomassie Blue staining, were pooled and concentrated to about 1 mg/ml using a

Centriprep concentrator (Amicon).

3.3 Cell Culture, Transfections, and Lysis

Human fetal kidney 293 cells were cultured on 100-mm dishes and transfected
using the calcium phosphate method as previously described [7]. Cells were harvested 16
hours after transfection, and lysed by the addition of 1 ml of lysis buffer (50 mM HEPES
(pH 7.5), 150 mM NaCl, 1.5 mM MgC12, 2 mM EGTA, 1% Triton X-100, 10% glycerol,
10 mM sodium fluoride, 1 mM Na4PPi, 100 pM B-glycerophosphate, 1 mM Na3VO4, 2
mM phenylmethylsulfonyl fluoride, and 0.15 units/ml aprotinin) as described previously

[7].

3.4 Immunoprecipitations and GST Pulldown Assays

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 antibody (Sigma) directed against the Flag epitope (0.45 rig/pl slurry). For
the GST-pulldown experiment, the GST fusion proteins were preincubated with

glutathione Sepharose 4B resin for 30 min. Clarified lysate (400 p1) was incubated with

55

20 pl of antibody-bound Protein A-agarose or with 20 pl of glutathione Sepharose 4B
resin prebound with 5 pg GST fusion protein for 90 min at 4 °C. Irnmunoprecipitates and
GST-pulldowns were washed with HNTG buffer (20 mM HEPES (pH 7.5), 150 mM
NaCl, 0.1% Triton-X-lOO, and 10% glycerol). Irnmunoprecipitates used for kinase
assays were washed three times with HNTG buffer containing 1 M LiCl, three times with
HNT G buffer, and twice with kinase assay buffer (50 mM Tris-HCl (pH 7.5), 100 mM

NaCl, 1 mM MnClz, 10 mM MgC12, 0.1 mM Na3VO4).

3.5 Gel Electrophoresis and Western Blot Analysis

Proteins from lysates and immunoprecipitates were resolved by SDS-PAGE
according to Laemmli [17] and transferred to nitrocellulose membranes. The membranes
were immunoblotted using MLK3 antiserum (1 pg/ml), M2 Flag monoclonal antibody (9
pg/ml) or HA antibody (BAbCO) (Spg/ml), followed by the appropriate horseradish
peroxidase-conjugated secondary antibody (Invitrogen). Western blots were developed
by the chemiluminescence method. Multiple exposures of the Western blots were
developed, and densitometry (NIH Image) of unsaturated films was used to determine

relative expression levels.

3.6 In vitro Kinase Assays

Kinase assays were performed in 20 pl of kinase assay buffer containing 50 pM
ATP and 5 pCi [y-32P]-ATP (3000 Ci/mmol) (NEN Life Science Products), 10 pg of
mixed histones (Roche), and the reactions were carried out for either 15 min or 30 min at

room temperature as previously described [7]. Following the kinase assay, proteins were

56

separated by SDS-PAGE. Gels were rinsed in phosphate-buffered saline, dried, and the
incorporation of radioactivity into the kinase or substrates was determined by

PhosphorImaging (Molecular Dynamics).

57

4. Results

4.1 Point Mutation in the SH3 Domain Increases the In Vitro Kinase Activity of
MLK3

Sequence and structural analyses of SH3 domains reveal three highly conserved
aromatic amino acids that participate in ligand binding [18, 19]. Replacement of any one
of these conserved residues with alanine in the SH3 domain of the adaptor protein Sem-
5/Grb2 abolishes binding to its proline-rich targets [18]. The corresponding conserved
residues in the SH3 domain of MLK3 are Tyr 52, Trp 83, and Tyr 99. To deliberately
disrupt the function of the SH3 domain of MLK3, site-directed mutagenesis was used to
substitute the tyrosine residue at position 52 with an alanine residue. The catalytic
activity of this SH3 mutant, MLK3 Y52A, was compared with that of wildtype MLK3.

Cells transiently expressing MLK3 or MLK3 Y52A were lysed, the MLK3
variants were immunoprecipitated, and in vitro kinase assays were performed using
histones as an exogenous substrate. Data from a representative experiment are shown in
Fig. 2. Based on three independent experiments, MLK3 Y52A exhibits a two-fold
increase in autophosphorylation activity and a 2.5-fold increase in histone
phosphorylation activity when compared with wildtype MLK3. These data suggest that

the catalytic activity of MLK3 is negatively regulated by its functional SH3 domain.

4.2 The SH3 Domain of MLK3 Interacts with MLK3
The finding that a disruptive mutation in its SH3 domain increases MLK3
activity, coupled with the fact that MLK3 contains nine potential SH3 binding sequences

(Pro-Xxx-Xxx-Pro) in its COOH terminal 220 amino acids, led us to investigate whether

58

the SH3 domain may bind to MLK3 itself. The SH3 domain of MLK3 (amino acids 43-
104) was expressed as a fusion protein with GST in E. coli and purified. The ability of
MLK3 to interact with its SH3 domain in an intermolecular fashion was assessed in GST
pulldown experiments. As shown in Fig. 3, both wildtype MLK3 and MLK3 Y52A
associate with GST-SH3 but not with GST. The observation that wildtype MLK3
associates with GST-SH3 to a lesser extent than does MLK3 Y52A suggests that an SH3-
mediated intramolecular interaction within wildtype MLK3 competes with the

intermolecular binding of GST-SH3.

4.3 Identification of the SH3 Binding Region Within MLK3

Variants with progressive deletions from the COOH-terminus of MLK3 were
constructed with an NHz-terminal Flag tag, expressed, and tested for their ability to
associate with GST-SH3. MLK3 1-635 and MLK3 1-598 lack the Pro/Ser/Thr-rich
region; MLK3 1-529 ends 23 amino acids after the CRIB motif; MLK3 1-485 terminates
just after the zipper motif; and MLK3 1-3 86 contains the NHz-terminus through the
kinase domain. All variants expressed at similar levels in 293 cells (Fig. 4A, bottom
panel). In order to create a more appropriate negative control for the GST pulldown
experiment, the Tyr residue in GST-SH3 that corresponds to Tyr 52 in MLK3 was
replaced with Ala, thus giving rise to GST-SH3 Y52A. Both fusion proteins were
expressed in E. coli, purified, and used in GST pulldown assays to test their binding to
the F lag-tagged MLK3 variants. Surprisingly, as shown in Fig. 4A, deletion of the
Pro/Ser/Thr-rich COOH-terminal region does not prevent association with GST-SH3. In

fact, of all the truncation variants tested, only MLK3 1-3 86 fails to detectably bind to

59

GST-SH3. This suggests that the SH3 binding site is within amino acids 387-485 of
MLK3. However, it is also conceivable that the SH3 binding site may be present but
masked in MLK3 1-386. In any case, one can conservatively conclude that a binding site
for MLK3’s SH3 domain is present in amino acids 1-485 of MLK3. All MLK3 variants
fail to associate with GST-SH3 Y52A, indicating that introduction of this point mutation
successfully blocks the binding ability of the SH3 domain.

To further define the SH3 binding site within MLK3, individual or tandem
domains of MLK3 were constructed as NHz-terminal HA-tagged variants, expressed, and
analyzed for their capacity to interact with the SH3 domain of MLK3. MLK3 1-114
contains the NHz-terminal glycine-rich region and the SH3 domain; MLK3 115-399
contains the kinase domain; MLK3 400-591 includes the zipper region, the CRIB motif
and an adjacent stretch of basic amino acids; and MLK3 592-847 contains the COOH-
terrninal Pro/Ser/Thr-rich region. All variants expressed at similar levels although MLK3
592-847 apparently undergoes some proteolytic degradation (Fig. 43, bottom panel).
Data from GST pulldown assays show that, of these variants, only MLK3 400-591 retains
the ability to interact with GST-SH3 (Fig. 4B). As expected, none of the variants
associates with GST-SH3 Y52A. Taking together all of the results from the GST
pulldown experiments, we conclude that the binding site for MLK3’s SH3 domain is
within amino acids 400-485 of MLK3.

To confirm that the SH3 binding region lies within MLK3 400-485, we took
advantage of a previously constructed deletion mutant, MLK3 A430-486, which lacks the
COOH-terminal half of the zipper and the flanking stretch of basic amino acids [7]. As

shown in Fig. 5, MLK3 A430-486 fails to associate with GST-SH3. These data suggest

60

that the critical SH3 binding determinants lie within amino acids 430-485 of MLK3.
However, an alternative explanation is that disruption of zipper-mediated oligomerization
prevents access to the SH3 binding site. We recently reported that a leucine zipper point
mutant, MLK3 L410P, behaves as a monomer [9]. Our results show that GST-SH3
associates with both MLK3 L410P and wildtype MLK3 to approximately the same extent

(Fig. 5).

4.4 Point Mutation in the SH3 Binding Site Abolishes Binding to the SH3 Domain
and Increases MLK3 Activity

An SH3 binding region has been mapped to amino acids 400-485 of MLK3. The
predicted leucine zipper which comprises amino acids 400-463 of MLK3 is devoid of
proline residues, and GST-SH3 binding does not require zipper-mediated homo-
oligomerization of MLK3 (Fig. 5), suggesting that the zipper domain should not contain
the SH3 binding region. Amino acids 463-485 of MLK3 are therefore predicted to be
crucial for the binding of the SH3 domain. As shown in Fig. 6A, no typical SH3 binding
motif of Pro-Xxx-Xxx-Pro is found in this region of MLK3, although a single proline
residue is found at position 469 and four contiguous arginine residues (472-475) are
present. Substitution of all four arginine residues with neutral glutamine residues
reduces, but does not abolish, binding of MLK3 to GST-SH3 (data not shown). The sole
proline residue in this region was substituted with an alanine residue by site-directed
mutagenesis and the ability of this mutant, MLK3 P469A, to associate with MLK3’s SH3

domain was tested in a GST-pulldown assay. As shown in Fig. 6B, MLK3 P469A does

61

not detectably associate with GST-SH3, indicating that Pro 469 is critical for SH3
binding.

A point mutation in the SH3 domain increases MLK3 activity (Fig. 2),
presumably because the mutation disrupts an autoinhibitory interaction. It has now been
determined that Pro 469 is critical for binding to MLK3’s SH3 domain. A reasonable
prediction, therefore, is that mutation of Pro 469 should likewise disrupt an autoinhibitory
interaction and increase MLK3 activity. Indeed, as measured in an in vitro kinase assay
(Fig. 7), MLK3 P469A has about four-fold higher autophosphorylation and histone
phosphorylation activity than does wildtype MLK3. These data argue that MLK3 is
autoinhibited by a physical interaction between its SH3 domain and a SH3 binding site

that is located between its zipper and CRIB motif.

62

 

 

 

 

1 43 103 115 384 400 463 492 506 632 847
C

Gly SH3 Kinase Zipper If Pro/SerlThr-rich
B

 

 

 

 

 

 

 

 

 

Fig. 1. Schematic of MLK3. The numbers in the diagram represent amino acid number.

The glycine-rich region (amino acid 1-42) is denoted by Gly. CRIB stands for
“Cdc42/Rac interactive binding” motif.

63

autoradiogram
<— MLK3

 

-- <-
. histones

1P .1 - Flag-MLK3 we
Autophosphorylation Histone phosphorylation
3 4
2
2
1
0 0
MLK3: WT Y52A MLK3: WT Y52A

Fig. 2. Effect of a point mutation in the SH3 domain of MLK3 on catalytic activity.
Cells were transfected with expression vectors containing the cDNAs for Flag-tagged
wildtype (WT) MLK3 or MLK3 Y52A. A minus sign indicates that a control empty
vector was transfected. The MLK3 variants were isolated from cellular lysates by
immunoprecipitation (1P) using the Flag antibody, and kinase activity was assessed in
vitro as described under “Experimental Procedures”. A, in vitro kinase assay of MLK3
and MLK3 Y52A using histones as a substrate. The top panel shows an autoradiogram
with bands corresponding to MLK3 autophosphorylation and histone phosphorylation
indicated by arrows. A Western blot (WB) of the immunoprecipitated MLK3 variants
using the Flag antibody is shown in the bottom panel. B and C, the mean : standard
errror for fold increase in MLK3 autophosphorylation and histone phosphorylation from
three independent experiments is shown. Data were quantitated by phosphorimaging and
normalized to MLK3 expression levels as described under “Experimental Procedures”.

64

g 5
MLK3: E g . E g .
GST pulldown I “i O. Flag-MLK3 we

.u- ' 1‘ . ‘ :L '0
J
1

GST pulldown W W “W— ... _ ~ Coomassie

 

 

staining
L J \ J
GST-SH3 GST
3‘.
MLK3: 5 g .
- ‘ Flag-MLK3 WB

Fig. 3. Assay for association of the SH3 domain of MLK3 with full length MLK3.

Cells were transfected with expression vectors containing the cDNAs for MLK3 variants.

Cellular lysates expressing the indicated MLK3 variants were incubated with glutathione
Sepharose 4B resin to which purified GST-SH3 or GST had been prebound. T op panel,

the presence or absence of bound MLK3 variants was assessed by Western blotting with

the Flag antibody. Middle panel, equal loading of GST-SH3 or GST on the glutathione

Sepharose 4B resin was confirmed by Coomassie staining. Bottom panel, the expression

levels of MLK3 variants were assessed by Western blotting of cellular lysates with the

Flag antibody. The data shown is representative of four independent experiments.

65

MB
ww
03
my.
mm.
T9
Sh
GF

GST-SH3]
Coomassie
staining

1'1.-

man

Y

  

' w

can...

Flag-MLK3 WB

  

nae...

66

mum. w

«an;

mac...

<Nm>

 

MLK3: E

3' MLK3:1-114 115-399 400-591 592-847 -

f—\ m
Eli-"#1 ‘ at“ GST pulldown]
5"," ' :fi’giir “’11“ HA-MLK3 we

I . fl ..
.Q at." ‘H I! .'
cafe 4.2.-95?

as... """‘"""‘:".'f"": GST-SH3]

 

“' Coomassie
WYWYWYWYWY staining
3 s. N

:5 n "a $
MLK3: ‘7 2 8 g

\- I- V In -

”' HA-MLK3WB

in“? _‘

r. ..

  

.33 "-’ in.

Fig. 4. Mapping of the SH3 binding region within MLK3. Cells were transfected
with expression vectors containing the cDNAs for MLK3 truncation variants. The
numbers in the Figure represent amino acid number in MLK3. A, B, cellular lysates
expressing the indicated MLK3 truncation variants were incubated with glutathione
Sepharose 4B resin to which purified GST-SH3 (W) or GST-SH3 Y52A (Y) had been
prebound. T op panels, the presence or absence of bound MLK3 variants was assessed by
Western blotting with the Flag antibody (A) or the HA antibody (B). Middle panels,
equal loading of GST-SH3 or GST-SH3 Y52A on the glutathione Sepharose 4B resin
was confirmed by Coomassie staining. Bottom panels, the expression of MLK3
truncation variants was assessed by Western blotting of cellular lysates with the Flag
antibody (A) or the HA antibody (B). The data shown is representative of at least four

independent experiments.

67

MLK3: WT L410P 13430-486 -

 

GST pulldown]

 

MLK3 WB
. -; '. _‘ ' Coomassie
-< -<' .< staining
> > >
CD
a i
‘- M
MLK3: E 3 :5 -
can. mm

Fig. 5. Effect of zipper mutations on the SH3 binding. Cells were transfected with
expression vectors containing the cDNAs for MLK3, MLK3 L410P, and MLK3 A430-
486. Cellular lysates expressing the indicated MLK3 variants were incubated with
glutathione Sepharose 4B resin to which purified GST-SH3 or GST-SH3 Y52A had been
prebound. T op panel, the presence or absence of bound MLK3 variants was assessed by
Western blotting with the MLK3 antibody. Middle panel, equal loading of GST-SH3 or
GST—SH3 Y52A on the glutathione Sepharose 4B resin was confirmed by Coomassie
staining. Bottom panel, the expression levels of MLK3 variants were assessed by
Western blotting of cellular lysates with MLK3 antibody. The data shown is

representative of three independent experiments.

68

A
MLK3 4 6 3 - QVDRERPHVRRRRGTFKRSKL . RA- 4 8 5

MLKl 4 3 8 - QLCQEKPRVKKRKGKFRKSRL . KL - 4 6 O
MLK2 4 4 4 - QLSQEKPRVRKRKGNFKRSRLLKL - 4 6 7

MLK3; WT Y52A P469A -

N f

 

 

\ GST pulldown]
MLK3 WB

a“ hlll ii“ GST-SH3!

h 3,. . Coomassie
lb

.<

UI

to

>

‘i .< staining
g a.

N

>

 

MLK3 WB

 

Fig. 6. Alignment of the SH3 binding region of MLK3 with MLK] and MLK2 and
effect of MLK3 P469A on binding of SH3 domain. A, alignment of the SH3 binding
region of MLK3 with MLK family members, MLKl and MLK2. Amino acid numbers
are indicated to the left and right of each sequence. The sole proline residue and the four
contiguous basic residues are shown in bold. B, cells were transfected with expression
vectors containing the cDNAs for the indicated MLK3 variants. Cellular lysates
expressing the MLK3 variants were incubated with glutathione Sepharose 4B resin to
which purified GST-SH3 or GST-SH3 Y52A had been prebound. Top panel, the
presence or absence of bound MLK3 variants was assessed by Western blotting with
MLK3 antibody. Middle panel, equal loading of GST-SH3 or GST-SH3 Y52A on the
glutathione Sepharose 4B resin was confirmed by Coomassie staining. Bottom panel, the
expression levels of MLK3 variants were assessed by Western blotting of cellular lysates

with MLK3 antibody. The data shown is representative of four independent experiments.

69

 

 

     

 

A. < 3‘.
E a s -
MLK3: >- n. _ autoradiogram
.. -‘ <—MLK3
”‘ "' “ 4—
histones
‘—
'P 5.. MLKs we
Autophosphorylation Histone phosphorylation
fi 51
4i 4T
2 2 _
o 0 ‘
MLK3: WT Y52A P469A MLK3: WT Y52A P469A

Fig. 7. Catalytic activity of MLK3 P469A. A, in vitro kinase assay of MLK3 variants
using histone as a substrate. The top panel shows an autoradiogram with bands
corresponding to MLK3 autophosphorylation and histone phosphorylation indicated by
arrows. A Western blot of the immunoprecipitated samples using a Flag antibody is
shown in the bottom panel. B and C, the mean j; standard error for fold increase in
MLK3 autophosphorylation and histone phosphorylation from three independent
experiments is shown. Data was quantitated by phosphorimaging and normalized to

MLK3 expression levels as described under “Experimental Procedures”.

70

5. Discussion

Protein phosphorylation is a ubiquitous regulatory event in biology. Many
pathological processes are associated with dysregulation of protein kinase activity. It is
not surprising, therefore, that nature has evolved and coordinated multiple mechanisms to
exquisitely control the activity of protein kinases. Many protein kinases are maintained
in low activity or inactive states through intermolecular or intramolecular association
with inhibitory molecules or domains. These inhibitory interactions may be disrupted in
response to appropriate stimuli, allowing the protein kinase to adopt an active
conformation.

SH3 domains were first identified as targeting domains that through binding to
proline-rich sequences mediate intermolecular associations among signaling molecules
[13, 14]. Extensive structural information has revealed three aromatic residues conserved
in SH3 domains that participate in binding to proline-rich ligands which usually adopt
left-handed polyproline type H helices [18, 19]. MLK3 contains an NHz-terminal SH3
domain. Notably, proline residues comprise 24% of the COOH-terminal 220 amino
acids. Our finding that the mutation of one of the conserved aromatic residues in MLK3,
Tyr 52, to Ala renders MLK3 more active (Fig. 2), suggested to us that MLK3’s SH3
domain might bind to MLK3 itself to negatively regulate its activity. Pulldown
experiments showed that full-length MLK3 associates with the SH3 domain of MLK3
fused to GST (Fig. 3). However, the extent of association of MLK3 Y52A with the GST-
SH3 fusion protein is much greater than that of wildtype MLK3 (Fig. 3), suggesting that
the functional SH3 domain of wildtype MLK3 competes with exogenous GST-SH3.
These results argue that an intramolecular association involving the SH3 domain of

wildtype MLK3 negatively regulates MLK3 activity.

71

Despite the fact that the COOH-terminal Pro/Ser/Thr—rich region of MLK3
contains nine Pro—Xxx-Xxx-Pro sequences, deletion of this region has no effect on
association of MLK3 with GST-SH3 (Fig. 4A). Instead, a series of mapping experiments
has identified a sequence located between the zipper and CRIB motifs that is important
for the interaction of MLK3 with its SH3 domain (Fig. 4B). Unexpectedly, this sequence
is not proline-rich but, rather, contains a single proline residue at position 469. Mutation
of this sole proline residue to alanine in MLK3 abolishes binding to GST-SH3. In accord
with the supposition that MLK3 is autoinhibited by an SH3-mediated intramolecular
interaction, replacement of Pro 469 with Ala increases the catalytic activity of MLK3.

While MLK3 represents the first demonstrated example of SH3-mediated
autoinhibition of a serine/threonine kinase, the well-studied tyrosine kinases of the Src
family offer some interesting parallels. Crystal structure and biochemical studies have
revealed an autoinhibitory intramolecular association that involves both the SH2 domain
and the SH3 domain of Src [20—22]. A COOH-terminal phosphorylated Tyr provides a
ligand for the SH2 domain, and the SH3 domain of Src binds to its so-called “SH2-kinase
linker” region, a sequence located between the SH2 and kinase domains. These
intramolecular interactions constrain the movement of the two lobes of the kinase domain
and therefore, block the binding of ATP, keeping Src in an inactive form (reviewed in
[23-25]). Interestingly, the SH3 binding sequence in Src, and also in its relative Fyn,
harbors only a single proline residue, analogous to what we have discovered in MLK3,
even though the preferred ligand of Src’s SH3 domain was found by phage display

methods to be the proline-rich sequence Len-Xxx-Xxx-Arg-Pro-Leu-Pro-Xxx-Pro [10].

72

One might imagine that the SH3 binding sequence in the “zipper-CRIB linker”
region of MLK3 would provide a suboptimal ligand for MLK3’s SH3 domain. The high
effective concentration afforded in intramolecular associations in all likelihood allows
this otherwise weak interaction to occur. Indeed, evolutionary selection may have given
rise to a relatively low affinity intramolecular ligand for MLK3’s SH3 domain so that it
might be outcompeted by the presentation, in response to an appropriate physiological
signal, of a high affinity proline-rich ligand on another signaling molecule. For instance,
binding of a proline rich sequence in the Nef protein of human immunodeficiency virus-l
to the SH3 domain of Hck, a Src family member expressed primarily in myeloid cells,
overcomes autoinhibition, dramatically increasing the activity of Hck [26, 27].

Signaling molecules that overcome SH3-mediated autoinhibition of MLK3 have
yet to be identified. One potential candidate is the MKKK kinase HPKI that contains
proline-rich sequences through which it interacts with the SH3 domain of MLK3 [15] and
that, like MLK3, can activate both the JNK pathway and the NF-KB pathway [28, 29].
Furthermore, HPKl phosphorylates kinase-inactive MLK3 in an in vitro kinase assay
[15]. However, no effect of HPKl on MLK3 activity has been observed in our hands or
reported in the literature.

The JNK pathway scaffold proteins, JNK interacting protein 1 (J IPl), J IPZ and
JIPB, have been shown to bind to mixed lineage kinases including MLK3 [30—32].
Cotransfection of JIPl , J 1P2 or J [PB with MLK3 in COS-7 cells enhances the activation
of JNK by MLK3 [31, 32]. This may suggest that binding of JIP increases MLK3
activity. It is conceivable that J [P binding might relieve SH3-mediated autoinhibition of

MLK3. However, the mixed lineage kinase family member, dual leucine zipper kinase

73

(DLK), which lacks an SH3 domain, also binds to J [PI and J 1P2 [33], and recent data
from Holzman’s lab suggests that JIP-associated DLK is catalytically inactive [34].
Further studies are necessary to clarify these issues.

We previously reported that activated Cdc42 binds to MLK3 in a CRIB-
dependent manner and results in a change in the in vivo phosphorylation pattern of MLK3
as judged by phosphopeptide mapping [7]. The close proximity of the newly identified
SH3 binding sequence of MLK3 to its CRIB motif may suggest interplay between
Cdc42-mediated activation and SH3-mediated autoinhibition of MLK3. The precise
mechanistic relationship is likely to be complex and is currently under investigation in
our laboratory.

The mixed lineage family of serine/threonine kinases are so-named for the
sequence similarity in their catalytic domains to both tyrosine and serine/threonine
kinases. The major function ascribed to this group of protein kinases is activation of the
JNK pathway, specifically by acting as MKKKs to phosphorylate and activate MKK4 or
MKK7. MLK] [35], MLK2 [36, 37], and MLK3 share the same general domain
arrangement, which includes an NHz-terminal SH3 domain and centrally located zipper
and CRIB motifs. However, the sequences in their COOH-termini diverge. The more
distantly related mixed lineage kinases include leucine zipper-bearing kinase (LZK) [3 8],
DLK/ZPK/MUK [39-41], leucine-zipper and sterile-alpha motif kinase (ZAK) [42], and
MLK-like mitogen-activated protein triple kinase (MLTK) [43]. These kinases lack both
SH3 domains and CRIB motifs. In this report we have identified a single proline residue
in MLK3 that is required for binding and autoinhibition through MLK3’s SH3 domain.

Interestingly, this proline residue is conserved in the closely related family members

74

MLK] and MLK2 (Fig. 6A), but not in the more distantly related mixed lineage kinases.
Thus this newly discovered mechanism of SH3-mediated autoinhibition of MLK3 is

predicted to apply to the regulation of MLK1 and MLK2 as well.

75

 

Footnotes

 

Aclmowledgements- We are grateful to Dr. Ronald William Henry (Michigan State
University) for providing the pCGN vector.

76

6. References

10.

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

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

Rana, A., et al., The mixed lineage kinase SPRKphosphorylates 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.

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.

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.

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.

Leung, I.W. and N. Lassam, 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.

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 JNK pathway. Monomeric SPRK
L410P does not catalyze the activating phosphorylation of Thr258 of murine

MI T OGEN-A C T 1 VA T ED protein kinase kinase 4. J Biol Chem, 2000. 275(36): p.
27893-900.

Sparks, A.B., et al., Distinct ligand preferences of Src homology 3 domains from
Src, Yes, Abl, Cortactin, p53bp2, PLCgamma, Crk, and GrbZ. Proc Natl Acad Sci
U S A, 1996. 93(4): p. 1540-4.

77

11.

12.

13.

14.

15.

l6.

l7.

18.

19.

20.

21.

22.

23.

24.

25.

Ren, R., et al., Identification of a ten-amino acid proline-rich SH3 binding site.
Science, 1993. 259(5098): p. 1157-61.

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

Feller, S.M., et al., SH2 and SH3 domains as molecular adhesives: the
interactions of Crk and Abl. Trends Biochem Sci, 1994. 19(11): p. 453-8.

Mayer, B]. and M.J. Eck, SH3 domains. Minding your p’s and q’s. Curr Biol,
1995. 5(4): p. 364-7.

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

Tanaka, M. and W. Herr, Diflerential transcriptional activation by Oct-1 and Oct-

2: interdependent activation domains induce Oct-2 phosphorylation. Cell, 1990.
60(3): p. 375-86.

Laemmli, UK, Cleavage of structural proteins during the assembly of the head
of bacteriophage T4. Nature, 1970. 227(259): p. 680-5.

Lim, WA. and RM. Richards, Critical residues in an SH3 domain from Sem-5
suggest a mechanism for proline-rich peptide recognition. Nat Struct Biol, 1994.
1(4): p. 221-5.

Weng, Z., et al., Structure-function analysis of SH3 domains: SH3 binding
specificity altered by single amino acid substitutions. Mol Cell Biol, 1995. 15(10):
p. 5627-34.

Williams, J .C., et al., The 2.35 A crystal structure of the inactivated form of

chicken Src: a dynamic molecule with multiple regulatory interactions. J Mol
Biol, 1997. 274(5): p. 757-75.

Xu, W., S.C. Harrison, and M.J. Eck, Three-dimensional structure of the tyrosine
kinase c-Src. Nature, 1997. 385(6617): p. 595-602.

Sicheri, F., I. Moarefi, and J. Kuriyan, Crystal structure of the Src family tyrosine
kinase Hck. Nature, 1997. 385(6617): p. 602-9.

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

Williams, J .C., R.K. Wierenga, and M. Saraste, Insights into Src kinase functions:
structural comparisons. Trends Biochem Sci, 1998. 23(5): p. 179-84.

Hubbard, S.R., Src autoinhibition: let us count the ways. Nat Struct Biol, 1999.
6(8): p. 711-4.

78

26.

27.

28.

29.

30.

31.

32.

33.

34.

35.

36.

37.

38.

Moarefi, I., et al., Activation of the Src-family tyrosine kinase Hck by SH3 domain
displacement. Nature, 1997. 385(6617): p. 650-3.

Briggs, S.D., et al., SH3-mediated Hck tyrosine kinase activation and fibroblast
transformation by the Nef protein of H1 V-I . J Biol Chem, 1997. 272(29): p.
1 7899-902.

Hu, M.C., et al., Human HPKI, a novel human hematopoietic progenitor kinase
that activates the JNK/SAPK kinase cascade. Genes Dev, 1996. 10(18): p. 2251-
64.

Hu, M.C., et al., Hematopoietic progenitor kinase-1 (HPK 1) stress response
signaling pathway activates IkappaB kinases (IKK-alpha/beta) and IKK-beta is a
developmentally regulated protein kinase. Oncogene, 1999. 18(40): p. 5514-24.

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 scaflold
proteins. Mol Cell Biol, 1999. 19(10): p. 7245-54.

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.

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

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

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

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 MS T, a novel (putative)
serine/threonine kinase with SH3 domain. Oncogene, 1995. 10(7): p. 1447-51.

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

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

79

39.

40.

41.

42.

43.

Hirai, S., et al., Activation of the flVK pathway by distantly related protein
kinases, MEKK and MUK. Oncogene, 1996. 12(3): p. 641-50.

Holzman, L.B., S.E. Merritt, and G. Fan, Identification, molecular cloning, and
characterization of dual leucine zipper bearing kinase. A novel serine/threonine
protein kinase that defines a second subfamily of mixed lineage kinases. J Biol
Chem, 1994. 269(49): p. 30808-17.

Reddy, U.R. and D. Pleasure, Cloning of a novel putative protein kinase having a
leucine zipper domain from human brain. Biochem Biophys Res Commun, 1994.
202(1): p. 613-20.

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.

Gotoh, 1.1., M. Adachi, and E. Nishida, Identification and characterization of a
novel MAP kinase kinase kinase, MLT K. J Biol Chem, 2000. 19: p. 19.

80

III. Characterization of a Proteolytic Event of Mixed Lineage Kinase 3

1. Abstract

Mixed lineage kinase 3 (MLK3) is a mitogen-activated protein kinase kinase
kinase (MAPKKK) that activates MAPK pathways, including the JNK and p38
pathways. When MLK3 is transiently expressed in HEK 293 cells, MLK3 undergoes
proteolysis to generate a stable carboxyl terminal fragment (CTF). The site of proteolysis
has been deduced to be Gln 251-Pro 252, which is located within the kinase domain of
MLK3. Based on the predicted structure of the kinase domain of MLK3 from homology
modeling, the cleavage site is in a solvent accessible region. Site-directed mutagenesis
studies revealed that Gln 251 and Leu 250 are required for this proteolytic event of
MLK3. Interestingly, the substrates of coronavirus main protease share the same
sequence determinants, with a required sequence Leu-Gln just amino terminus to the
cleavage site. Our data suggest that a protease with substrate specificity similar to that of
the coronavirus protease exists in mammalian cells. It is also conceivable that MLK3 may
be a cellular target of the coronavirus protease during viral infection.

Overexpressed MLK3 undergoes phorbol 12-myristate 13-acetate (PMA) (or
TNFa, or anti-CD3 antibody) induced proteolysis in Jurkat T antigen cells. The ability of
various protease inhibitors to block PMA-induced proteolysis of MLK3 was examined.
Among the many commercially available protease inhibitors tested, only MG 132 blocks
the generation of CTF, presumably through competitive inhibition by mimicking the

sequence at the cleavage site and occupying the catalytic core of the protease.

81

In addition, pulse chase experiments suggest that MLK3 undergoes proteolysis
shortly after synthesis and the protein synthesis inhibitor cycloheximide efficiently blocks
PMA-induced proteolysis of MLK3 in Jurkat T Ag cells.

MLK3 undergoes proteolysis in an activity dependent manner in HEK 293 cells,
whereas the catalytic activity of MLK3 is not required for PMA induced generation of the
CTF in J urkat T Ag cells. We hypothesized that activation of a particular pathway or
several pathways by MLK3 (293 cells) or by PMA (J urkat T Ag cells) is required for
MLK3 proteolysis. Using specific inhibitors that target the ERK, p38 or JNK pathways
respectively, our data suggest that ERK activation may be required for the generation of

the CTF.

82

2. Introduction

Protein kinases and other molecules involved in signaling transduction are often
regulated by proteolysis. According to enzyme nomenclature, proteases are divided into
five subclasses based on their catalytic mechanisms, including serine proteases, cysteine
proteases, aspartic proteases, metalloproteases, and a subclass of proteases for which the
catalytic mechanism is yet unknown [1]. The amino acid residues at the cleavage sites of
the substrates are designated as --'P3, P2, Pl-Pl', P2', P3'---, where the cleavage site is
denoted as P1- P1’ [2]. Several proteases, which have been implicated in regulating
signaling pathways, are described in details in the following.

Ubiquitin-mediated degradation is a highly conserved mechanism for selective
protein degradation in eukaryotes. Ubiquitin is an 8.6 kDa protein that is covalently
attached to proteins targeted for degradation. There are three major enzymes that are
involved in ubiquitination. Ubiquitin is activated in an ATP-dependent step by a specific
activating enzyme (E1). The activated ubiquitin molecule is then transferred to a
ubiquitin-conjugating enzyme (E2). Then the ubiquitin protein ligase (E3) links the
ubiquitin through its carboxyl terminus to an e-amino group of a Lys residue in the
substrate protein. The 26S proteasome then recognizes and degrades the ubiquitinated
proteins. The 26S proteasome is a large multisubunit complex, consisting of a central
catalytic 208 proteasome and a 198 regulatory complex. The tertiary structure of the 20S
proteasome is highly conserved from yeast to mammals and consists of a stack of four
rings, each containing seven subunits, a7B7B7a7 (reviewed in [3-5]). The 208 proteasome
is a novel “threonine” protease, in which the Thr residue in the 0 subunit acts as

nucleophile to attack the carbonyl carbon of the substrate [6]. Many proteins involved in

83

signal transduction and cell cycle progression, such as protein kinase Ca, p53, cyclins,
CDK inhibitors and transcription factors, are degraded through the ubiquitin-mediated
degradation pathway. Lactacystin, a natural metabolite of Streptomyces, upon entering
the cells, undergoes a spontaneous conversion (lactonization) to the clasto-lactacystin B-
lactone, which specifically inhibits the activity of proteasome by covalently modifying
the active site Thr residue [7].

Caspases are cysteine proteases involved in apoptosis. Caspases are synthesized
as inactive zymogens and become activated through proteolytic processing and self-
association of two procaspases in the apoptotic cascade. The P1 position in the substrate
of the caspases is restricted to Asp. Caspase substrates include cytoskeletal proteins,
nuclear proteins, cell cycle regulators, and protein kinases. In contrast to the ubiquitin-
mediated degradation pathway, which generally degrades its substrates to inactive, small
fragments, caspases cleavage can either render a protein kinase inactive, or constitutively
active by removal of autoinhibitory domains. Interestingly, upon cleavage by caspases,
some protein kinases, such as protein kinase C6, further contribute to the apoptotic
response (reviewed in [8]). Thus caspase-mediated proteolysis can set in motion a
positive feedback loop.

The calpains constitute another family of proteases that regulate signaling
pathways. Calpains are Ca2+-dependent intracellular cysteine proteases, widely expressed
in various mammalian tissues. There are two major forms of calpains: m-calpain and u-
calpain. Calpains show no strict sequence preferences. Instead, it has been suggested

that calpain may recognize and cleave specific tertiary conformations of proteins.

84

Calpain substrates include cytoskeletal proteins such as actin, protein kinases such as
protein kinase C, nuclear transcription factors, and caspases (reviewed in [9]).

The secretases are a family of aspartyl proteases that exist as complexes in the
transmembrane space. y-secretase is an intramembrane proteolytic complex that is
involved in the proteolysis of amyloid precursor protein. The proteolytic product,
amyloid B-protein, is a major constituent of the abundant neuritic plaques which are the
hallmark of Alzheimer’s disease [10]. The catalytic core of y-secretase is presenilin. In
addition to amyloid precursor protein, many substrates of presenilin, including the cell
surface receptor Notch [11] and the receptor tyrosine kinase ErbB4 [12], localize
predominantly at the plasma membrane [13]. Presenilin inhibitors are being developed
by many pharmaceutical companies as possible therapeutics for Alzheimer’s disease. A
commercially available hydroxyethylene dipeptide isostere, L-685, 458, functions as a

transition state analogue and potently inhibit the activity of presenilin [l4].

Herein, we report that MLK3 undergoes proteolysis to generate a stable carboxyl

terminal fragment in vivo and describe the characterization of this proteolytic event.

85

 

3. Materials and Methods
3.1. Reagents

Lactacystin, MG132, MDL 28170, leupeptin, L-685, 45 8, ZVAD-FMK, E64d, U
0126, SP 600125, and SB 203580 were purchased from Calbiochem. Cycloheximide,
pepstatin A, chloroquine, and phorbol 12-myristate 13-acetate (PMA) were purchased

from Sigma.

3.2. Construction of Mammalian Expression Vectors and Site-Directed Mutagenesis

Construction of the cytomegalovirus-based expression vectors carrying the
cDNAs for wild type MLK3 (pRK-mlk3) and for the kinase defective mutant of MLK3
(pRK- mlk3 K144A) and monomeric form of MLK3 (pRK mlk3 L410P) has been
described elsewhere [15, 16]. The expression plasmid encoding amino-terminal Flag
epitope-tagged constitutively active variant (pRK5-NFlag.Cdc42 V" '2) of Cdc42 was
kindly provided by Avi Ashkenazi (Genentech, Inc.). The wildtype MLK3 expression
vector was used as a template to create mutants (mlk3 L250A, L250V, Q251A, Q251E,
Q251L, Q251N, P252A, P252L, and P252D) using the Quick Change Site-directed
Mutagenesis method (Stratagene). The presence of the desired mutation was confirmed
by automated DNA sequencing in the Genomics Technology Support Facility at

Michigan State University.

3.3. Cell Culture, Transfections, and Lysis

Human embryonic kidney (HEK) 293 cells were cultured in Ham’s F12:low

glucose Dulbecco’s modified Eagle’s media (1:1) (Invitrogen) supplemented with 8%

86

fetal bovine serum (Invitrogen), 2 mM L-glutamine, and penicillin/streptomycin
(Invitrogen). J urkat T antigen cells were cultured in RPMI 1640 (Invitrogen)
supplemented with 8% fetal bovine serum, 2 mM L-glutamine, 50 11M 2-
mercaptoethanol, and penicillin/streptomycin. HEK 293 cells were transfected with
plasmid DNA using the calcium phosphate method as previously described . Jurkat T
antigen cells were transfected by electroporation (250 V, 950 uF) using a Bio-Rad Gene
Pulser electroporator. Cells were harvested, washed with ice cold PBS and lysed for 5
min on ice by the addition of 1 ml of lysis buffer (50 mM HEPES (pH7.5), 150 mM
NaCl, 1.5 mM MgC12, 2 mM EGTA, 1% Triton X-100, 10% glycerol, 10 mM sodium
fluoride, 1 mM NaaPPi, 100 uM B-glycerophosphate, 1 mM Na3VO4, 2 mM
phenylmethylsulfonyl fluoride, and 0.15 units/m1 aprotinin). To disrupt cells under

denaturing conditions, 0.2% SDS and/or 8 M urea were added to the lysis buffer.

3.4. Immunoprecipitations

MLK3 antibody (0.4 [lg/pl slurry) was prebound to protein A-agarose beads for
0.5 h at room temperature. Clarified lysate (300 pl) was incubated with 20 pl of antibody
bound Protein A-agarose for 90 min at 4 °C. Irnmunoprecipitates were washed with
HNTG buffer (20 mM HEPES (pH 7.5), 150 mM NaCl, 0.1% Triton-X-lOO, 10%

glycerol).

3.5. SDS-PAGE and Western Blot Analysis

Lysates and immunoprecipitates of proteins were resolved by SDS-PAGE.

Proteins were transferred to nitrocellulose membranes and immunoblotted using MLK3

87

 

antibody (7.7ug/ml), followed by the appropriate horseradish peroxidase-conj ugated

secondary antibody (Bio-rad). Western blots were developed by chemiluminescence.

3.6. Pulse Chase Analysis

16 h post transfection, HEK 293 cells were washed twice with PBS, and the
culture medium was replaced with methionine/cysteine fi'ee high glucose Dulbecco’s
modified Eagle’s medium (Invitrogen) containing 10% dialysed fetal bovine serum
(Invitrogen) and penicillin /streptomycin. After 1 h incubation, cells were pulse-labeled
by the addition of 150 uCi/ml [3 5 S] methionine /cysteine to the culture medium for an
additional 1 h. Cells were washed twice with PBS and then either lysed immediately or
incubated in complete medium containing 100 fold excess of methionine (3 mg/ml) for
the indicated times prior to cell lysis. MLK3 was immunoprecipitated from the cleared
lysates with anti-MLK3 antibody as described above and subjected to SDS-PAGE.
Proteins were transferred to nitrocellulose membrane and the relative amounts of

radiolabel in full length and CTF of MLK3 were measured by phosphorimaging.

3.7. Amino Terminal Sequencing

HEK 293 cells (in 16 X 150 mm tissue culture plates) transiently expressing
MLK3 were disrupted in lysis buffer and immunoprecipitated with 77 pg of MLK3
antibody (0.4 pg/ul slurry) for 1.5 h as described above. Immunoprecipitates were
washed six times with HNTG buffer and resolved by SDS/PAGE. Proteins were
transferred to a nitrocellulose membrane and stained with Coomassie Blue. The

membrane slice corresponding to the MLK3 CTF was excised from the membrane and

88

T"

washed several times with water. The amino terminus of the fragment was sequenced by
Edman degradation in the Genomics Technology Support Facility at the Michigan State

University.

3.8. MLK3 Kinase Domain Homology Modeling
A three dimensional model of MLK3’s catalytic domain was constructed with the
program Homology based onthe kinase domain of Src. This work was done by Ritesh

Agrawal with the assistance from Dr. Leslie Kuhn at Michigan State University.

89

4. Results

4.1. MLK3 undergoes activity dependent proteolysis in HEK 293 cells.

Afier transient expression of MLK3 in HEK 293 cells followed by Western blot
analysis, a band corresponding to full length MLK3 as well as an immune-reactive band
of about 66 kDa were observed (Fig. 1A). Since the MLK3 antibody was raised against
the carboxyl terminal eight amino acids of MLK3, this suggests that MLK3 undergoes
proteolysis to generate a stable carboxyl terminal fragment (CTF). We refer to generation
of the CTF as the “proteolytic event”. When we used an antibody against the amino
terminal SH3 domain of MLK3, we were unable to detect a corresponding 27 kDa amino
terminal fragment of MLK3 (data not shown). It is possible that the amino terminal
fragment is rapidly degraded after it is produced.

The proteolytic event of MLK3 may occur in vivo or may occur after cell lysis.
However, when urea and SDS, in addition to several protease inhibitors, were included in
the cell lysis buffer, the CTF was still observed (Fig. 1A). This is consistent with the idea
that the proteolytic event occurs in vivo.

In our lab, several variants of MLK3 have been constructed to study its function.
Interestingly, when MLK3 mutants with diminished catalytic activity were expressed in
HEK 293 cells, little or no CTF was observed (Fig. 1B). For instance, expression of the
kinase-defective variant MLK3 K144A fails to generate the CTF. Perhaps, the most
interesting mutant is MLK3 L410P, which has a helix disrupting point mutation in its
zipper domain and fails to oligomerize [16]. When transfected alone in HEK 293 cells
(lane 5), its catalytic activity is low and no CTF is observed. When cotransfected with

constitutively active Cdc42, the activity of MLK3 L410P increases as judged by

90

autophosphorylation and histone phosphorylation in an in vitro kinase assay [16], and the
CTF is observed (lane 6). This is consistent with the idea that MLK3 activity is required

for proteolysis in HEK 293 cells.

4.2. Proteolysis occurs between amino acids 251-252 of MLK3

In order to determine the cleavage site within MLK3, the CTF was isolated and
subjected to amino terminal sequencing by Edman degradation. No amino acid was
detected in the first cycle, but the following cycles revealed the correct sequence of
MLK3 from amino acids 253-273 (Fig. 2A). This is consistent with the idea that the
proteolytic event occurs between Gln 251 and Pro 252 of MLK3. It is unclear why the
proline was not detected in the first cycle. Perhaps this proline is modified either in vivo
or in vitro under the conditions of isolation. The CTF contains an incomplete kinase
domain and thus, this proteolytic event should render MLK3 inactive.

Among the solved structures of protein kinase domains, the kinase domain of Src
shares the closest similarity with that of MLK3 with 28% identity . A three dimensional
model of the kinase domain of MLK3 was generated based on the crystal structure of Src
using Homology modeling [18]. As indicated by the arrow in Fig 28, Gln 251 and Pro
252 are located in the loop region between the amino terminal lobe and carboxyl terminal
lobe of the kinase domain, thus this bond is predicted to be a solvent accessible region.
This model suggests that the site of proteolysis in MLK3 should be available to an

“MLK3 protease”.

4.3. Gln 251 and Leu 250 are critical for MLK3 proteolysis.

91

Since MLK3 is cleaved between Gln 251-Pro 252, one might hypothesize that
Gln 251 and Pro 252 are important for the recognition by the “MLK3 protease”.
Therefore, a series of MLK3 variants containing amino acid substitutions at Gln 251 or
Pro 252, were constructed and transiently expressed in HEK 293 cells. As shown in Fig.
3A, upon substitution of Gln 251 with Ala, Glu, Leu or Asn of MLK3, no CTF was
generated. In contrast, substitution of Pro 252 of MLK3 with Ala, Leu, or Asp resulted in
the generation of the CTF. These data suggest that the “MLK3 protease” may specifically
recognize Gln 251 at the P1 cleavage site. However, Pro 252 at P1 ’ is not essential for
recognition. In addition, the Leu 250 was mutated to Ala and Val respectively and the
effect on MLK3 proteolysis was tested. As shown in Fig. 3B, L250A and L250V do not
generate CTF, suggesting that the Leu 250 at the P2 position is also critical for

recognition by the MLK3 protease.

4.4. Inducible proteolysis of MLK3 by extracellular stimuli in Jurkat T antigen (Ag)
cells.

The CTF of MLK3 is not present when MLK3 is transiently expressed in J urkat T
Ag cells. However, Weiqin Chen (Esselman lab) observed the CTF of MLK3 upon
treatment of cells with PMA. We have extended this initial observation to demonstrate
the appearance of the CTF as soon as 1 h after PMA treatment, and the accumulation of
the CTF in a time dependent manner. In addition, Stancy Liou (Gallo lab) has observed
that anti-CD3 antibody, which clusters and activates CD3 receptors, as well as TNFa

treatment results in the generation of the CTF (Fig 4).

92

‘—

 

4.5. Inducible proteolysis of MLK3 is independent of MLK3 activity in Jurkat T Ag
cells

Expression of the kinase defective variant MLK3 K144A or the monomeric
variant MLK3 L410P in Jurkat T Ag cells results in the generation of the CTF in
response to PMA treatment (Fig. 5). This suggests that generation of the CTF does not
require an active conformation of MLK3, but rather requires signaling events that can
emanate from PMA, CD3 antibody, or TNFa. MLK3 Q251A, L250A and L250V do not
produce the CTF upon PMA treatment, consistent with the idea that Gln 251 and Leu 250

are retained as the sequence determinants for the MLK3 protease.

4.6. Protease inhibitor MG 132 blocks the generation of CTF of MLK3 in Jurkat T
Ag cells.

The inducible proteolytic event of MLK3 in Jurkat T Ag cells provides a model
system to investigate the attributes/identity of the MLK3 protease. The ability of several
protease inhibitors to block PMA-induced generation of the CTF was tested in J urkat T
Ag cells (Table 1). Treatment of MLK3-transfected J urkat T Ag cells with a general
aspartyl protease inhibitor (pepstatin A), general cysteine protease inhibitors leupeptin
and E64d, a caspases inhibitor (ZVAD-FMK), a calpain inhibitor (MDL 28170), and a
lysosomal protease inhibitor (chloroquine) all failed to block the PMA-induced
generation of the CTF. Surprisingly, only MG132, a peptide aldehyde that is marketed as
a proteasome inhibitor [19], is able to block the PMA-induced MLK3 proteolytic event.
To determine whether generation of the CTF reflects a proteasome-mediated event, the

more specific inhibitor Lactacystin was used, but failed to block this proteolytic event. In

93

addition, MG 132 has been reported to inhibit secretases [20, 21]. However, a specific
inhibitor of y-secretase, L685, 45 8, failed to prevent PMA-induced generation of the CTF
(Fig. 6). These data suggest that MG 132 is targeting a protease other than the

proteasome or y-secretase to inhibit MLK3 proteolysis.

4.7. The CTF is generated from newly synthesized MLK3

To demonstrate a precursor: product relationship between full length MLK3 and
the CTF, pulse chase experiments in HEK 293 cells were performed. Afier transiently
expressing wildtype MLK3 and the cleavage site mutant MLK3 Q251A, HEK 293 cells
were pulse-labeled with [3SS]-Met/Cys for l h and chased for up to 6 h. Fig. 8 shows
representative results from four independent experiments. Surprisingly, the CTF is
apparent at the end of the 1 h pulse period, yet during the chase period, the labeled CTF
did not accumulate. This suggests that the MLK3 proteolysis occurs rapidly after MLK3
synthesis, and the proteolysis does not occur or occurs very slowly afterwards.

Furthermore, when J urkat T Ag cells transiently expressing MLK3 were treated
with the protein synthesis inhibitor cycloheximide along with PMA for 3 h, the CTF was
not detected (Fig. 9). Taken together, these two sets of experiments are consistent with

the idea that only newly synthesized MLK3 is a substrate for the “MLK3 protease”.

4.8. ERK activation is likely required for MLK3 proteolysis
MLK3 undergoes proteolysis to generate the CTF in an activity dependent
manner in HEK 293 cells. Only active MLK3 variants produce the CTF, whereas

catalytically inactive variants do not. The occurrence of MLK3 proteolytic event may

94

depend on the conformation of MLK3 and the MLK3 protease might specifically
recognize the active conformation of MLK3. Another possibility is that proteolysis of
MLK3 may occur through a signaling pathway that is activated by MLK3 or other
factors. The data obtained from the J urkat T Ag cells support the second hypothesis.
When MLK3 is overexpressed in J urkat T Ag cells, its catalytic activity is low compared
with that of MLK3 overexpressed in HEK 293 cells. PMA induces MLK3 proteolysis,
but does not change MLK3 catalytic activity (data not shown, from Weiqin Chen).
Furthermore, PMA also induces the proteolysis of catalytically inactive mutants of
MLK3. It is reasonable to hypothesize that proteolysis of MLK3 may require a pathway
that can be activated by overexpressed MLK3 in HEK 293 cells or induced by PMA in
J urkat T Ag cells.

If this hypothesis is true, a reasonable prediction, therefore, is that PMA should
trigger proteolysis of the kinase defective mutant MLK3 K144A in HEK 293 cells.
Indeed, we observed the corresponding CTF of MLK3 K144A after PMA treatment (data
not shown).

It has been reported that PMA activates the ERK, the JNK and the p38 pathways
[22]. MLK3 has been reported to activate both the JNK and p38 pathways [23].
Unpublished data from our lab has shown that upon overexpression, MLK3 activates the
ERK pathway in HEK 293 cells (data not shown). It is possible that activation of one or
more of the MAPK pathways is required for the proteolytic event of MLK3. To test this
hypothesis, the ability of specific inhibitors of ERK, JNK or p38 pathways to block
PMA-induced MLK3 proteolysis was assessed in J urkat T Ag cells. J urkat T Ag cells

transiently expressing MLK3 were pretreated with the MEK inhibitor U0126 [24], the

95

“Ola;

JNK inhibitor SP 60012 [25], the p38 inhibitor SB 203580 [26] (Table 2), or a
combination of them for 0.5 h prior to treatment with PMA for 3 h. As shown in Fig. 10,
the MEK inhibitor U0126 alone partially blocks the generation of the CTF. Neither the
JNK inhibitor nor the p38 inhibitor alone was effective in blocking the MLK3
proteolysis. Generation of the CTF is potently inhibited by using all three inhibitors.
These results suggest that ERK activation is likely required for MLK3 proteolytic event.
The p38 and/or JNK activation may partially contribute to the proteolytic event in MLK3.
Finally, we cannot rule out the possibility that another MAPK independent pathway,
activated by MLK3 in HEK 293 cells, or by PMA in Jurkat T Ag cells, might be critical

for generation of the CTF.

96

Table 1. Effect of protease inhibitors on PMA-induced generation of MLK3 CTF.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Concentration
Inhibitor Targeted protease CTF detected
used
Pepstatin A aspartyl protease 20 11M Yes
Leupeptin cysteine protease 100 pM Yes
E64d cysteine protease 20 uM Yes
ZVAD-FMK caspases 200 nM Yes
MDL 28170 [27] calpain 5 uM Yes
Chloroquine [28] lysosomal protease 100 nM Yes
Lactacystin [7] proteasome 5 uM Yes
L-685, 458 [14] y-secretase 100 nM Yes
M6132 [19-21] pmteasome 10 “M No
y-secretase
Table 2. MAPK inhibitors.
Inhibitor Targeted kinase Targeted pathway Concentration
U0126 MEK1,2 ERK 10 11M
SP600125 JNK1,2,3 JNK 20 pM
SB203580 p38 M APK p38 6 uM

 

 

 

 

 

 

97

Lysrs buffer
+02% SDS
+8 M urea
+02% SDS
8 M urea

*1
t“-

MLK3—' . 3 2. ..I E .5

MLK3 we
CTF—~> ' '-

MLK3 WT K144A L410P vector
(fl t—\ r n r \

 

Cdc42V12 - + - + - + - +
MLK3 —> W! MLK3 WB
CTF ——> QQ -

Fig. 1. In vivo proteolysis of MLK3 in HEK 293 cells. HEK 293 cells were transfected
with expression vectors containing the cDNA for MLK3 in the presence or absence of
cDNA for activated Cdc42. Cells were lysed using lysis buffer containing the protease
inhibitors aprotinin and PMSF along with EGTA. Strong denaturants were added to the
lysis buffer as indicated. A Western blot (WB) is shown with full length MLK3 and the
CTF of MLK3 indicated by arrows. A. MLK3 proteolysis pattern in different lysis
conditions. B. Generation of CTF of MLK3 variants.

98

A.

 

 

 

o
lGIy‘ SH3] ‘ Kinase j ‘Zipper 3‘
W
Pl-Pl’

| NNILLL‘ _ 1 J1’IESDDMEHKTLKITDFGLAREWHETTQM

V

(aa 251—252)

 

Fig. 2. MLK3 is cleaved within the kinase domain. A. MLK3 is cleaved between Gln

251 and Pro 252. The amino acids determined by Edman degradation are underlined. B.

MLK3 kinase domain model based on Homology modeling. The arrow indicates the

cleavage site Gln 251-Pro 252.

99

 

:5 93 .—.' E :5. a a
L0 to LO LO LO LO LO
MLK3: E 8 3 8 3 E E E
MLK3 + '--fl-"‘-”‘
MLK3WB
CTF -> C ’1... $364
< >
8 8
MLK3: g <3 3
Full length MLK3 +dfli-
«w w- - MLK3WB

CTF—>03

Fig. 3. The proteolytic pattern of MLK3 variants. HEK 293 cells were transfected with
expression vectors containing the cDNAs as indicated above each lane. Afier 20 h, cells
were lysed and Western blotting using the MLK3 antibody was performed. Bands
corresponding to full length MLK3 and the CTF are indicated.

100

00.51 2 4 h

PMA . . _-

MLK3 WB

 

0025a51 3 h

-
TNFa - -l .. MLK3WB

---.‘
cos Ab .. .... MLK3 wa

rflm

Fig. 4. Inducible proteolysis of MLK3 in Jurkat T Ag cells. Jurkat T Ag cells were
transfected with expression vectors containing the cDNA Forty hour post-transfection,
cells were treated with 50 ng/ml PMA, 10 nM of TNFa, or 0.5 pg/ml CD3 antibody for
the indicated times and cells were lysed. MLK3 and CTF were detected by Western
blotting using the MLK3 antibody. (TNFa and CD3 antibody data from Stancy Liou,
Gallo lab)

101

PMA - + - +
-i‘ u-

MLK3 WB
Q ‘
k—J ;___/

MLK3: K144A L410P

PMA - + - + — + - +
H MLK3WB
U
k j k j\ I

 

\_)
MLK3: wt L250A L250V 0251A

Fig. 5. The proteolytic pattern of transfected MLK3 variants in PMA-treated J urkat
T Ag cells. J urkat T Ag cells were transfected by electroporation with expression vectors
containing the cDNAs for the indicated MLK3 variants. After 40 h, the tranfected J urkat
T Ag cells were treated with solvent control (DMSO) or 50 ng/ml PMA for 2 h and cells
were lysed. The presence or the absence of the CTF was detected by Western blotting
using the MLK3 antibody. A. proteolytic pattern of catalytically inactive MLK3 and

monomeric MLK3. B. proteolytic pattern of cleavage site MLK3 variants.

102

+ MG-132
+ Lactacystin
+ L685, 458

PMA: - +

“In---
- MLK3WB

i- is“

Fig. 6. Effect of protease inhibitors on PMA-induced generation of the CTF of
MLK3 in Jurkat T Ag cells. J urkat T Ag cells were transfected with expression vectors
containing the cDNAs for MLK3. After 40 h, cells were pretreated with vehicle control
(DMSO), MG132 (10 nM), lactacystin (SnM), or L685, 458 (100 nM) for 0.5 h, then
treated with 50 ng/ml PMA for an additional 2 h. Cells were lysed and Western blotting
was performed using the MLK3 antibody.

103

un MG-132
a 6.5 ..3 ~

Carbobenzoxy-Leu-Leu-Leu-CHO

Cleavage site in MLK3: Leu-Leu-Leu-Glnfro

Lactacystin

 

, I ~. ' ; J L-685,458

Fig. 7. Chemical structure of selected protease inhitors and comparison with the

sequence before the cleavage site in MLK3. A, the chemical structure of MG-132 in

comparison with the cleavage site within MLK3. The arrow indicates the cleavage site.

B, the chemical structure of lactacystin. C, the chemical structure of L-685, 458.

104

MLK3WJ MLK 21A
h \ (_3_Q_§_\

Chase time: 0 2 4 6 O 2 4 6 (h) autoradiogram
H but $13 I“
I t“ 9"
I . - .

 

‘ . 1:: 'I .

’b I'm-.4 II .
H 5,4 64 i 4:

, . .. . , . <—CTI=
LJ 9.1 t "t ’ ‘a‘d .

.4 b» 50 fl‘ .- _ .-
IP .. MLK3WB
I». ‘3 El H

a ‘ I. at

Fig. 8. Pulse chase experiments in HEK 293 cells. Cells were transfected with MLK3
wildtype and MLK3 Q251A cDNA. After 16 h post-transfection, cells were incubated
with methionine-cysteine-free media for 1 hour, then supplemented with 150 pCi of [358]
Met/Cys label mix in the same media. After 1 hour, cells were placed in fresh media
containing 300-fold excess of unlabeled methionine. Cells were harvested and lysed at
various time points. Cell lysates were immunoprecipitated with the MLK3 antibody and
immunoblotted MLK3 was resolved by SDS-PAGE and transferred to a PVDF
membrane. The radiolabeled protein was detected by phosphorimaging and the total
immunoprecipitated MLK3 was assessed by Western blot. The top panel shows the
radiolabeled MLK3, and the bottom panel shows levels of immunoprecipitated MLK3

variants.

105

cux --+-
PMA - + + -

he

MLK3 WB

Fig. 9. Cycloheximide blocks PMA-induced generation of the CTF of MLK3 in

J urkat T Ag cells. J urkat T Ag cells were transfected with expression vectors containing
the cDNAs for MLK3. After 40 h, cells were treated with 10 ng/ml PMA in the presence
or absence of 10 ug/ml cycloheximide (CHX) for 3 h. Cells were lysed and Western
blotting using the MLK3 antibody was performed.

106

 

 

 

 

U 1026 - - + - - + + - +
SP 600125 - - - + - + _ + +
88 203580 - - - - + - + + +

PMA - + + + + + + + +

 

 

 

 

 

 

 

 

 

 

 

MLK3 WB

Fig. 10. Effect of MAPK pathway inhibitors on the PMA-induced generation of the
CTF of MLK3 in J urkat T Ag cells. J urkat T Ag cells were transfected with expression
vectors containing the cDNAs for MLK3. After 40 h, cells were pretreated with vehicle
control (DMSO), 10 M U0126 (MEK inhibitor), 20 nM SP 600125 (JNK inhibitor),
and/or 6 uM SB 203580 (p38 inhibitor) as indicated for 0.5 h prior to treatment with 50
ng/ml PMA for 2 h. Cells were lysed and Western blotting using the MLK3 antibody was

performed.

107

 

HEK 293 cells Jurkat T Ag cells
MLK3... PMA

\, l

l\ /l

ERK/‘2

l
l

MLK3 protease

F ig. 11. Model for generation of CTF of MLK3. The CTF of MLK3 may be generated
through activation of one or more pathways by MLK3 (in HEK 293 cells) or PMA,
TNFor and anti-CD3 antibody (in J urkat T Ag cells). ERK pathway is likely required for
the generation of the CTF.

108

5. Discussion

We have characterized a proteolytic event of MLK3. MLK3 undergoes
proteolysis to generate a stable CTF of 66 kDa. Based on our observations in both HEK
293 cells and Jurkat T Ag cells, we hypothesize that proteolysis of MLK3 may occur by a
feedback mechanism. Therefore, specific inhibitors targeting the ERK, JNK and p38
pathways were used separately or in combination to test whether they may block
proteolysis. Inhibitor studies suggest that ERK activation may be required for this
proteolytic event as an ERK inhibitor diminished MLK3 proteolysis. However, p38
activation and JNK activation may contribute to the generation of the CTF of MLK3 as
well. Since PMA is a potent activator of all three MAPK pathways, we cannot rule out
the possibility of cross talk among the different pathways. It might be useful to test these
MAPK inhibitors in TNFor-triggered proteolysis to find out whether JNK or p38
pathways actually play a role in this MLK3 proteolytic event.

Data presented thus far is derived from transiently overexpressed MLK3. We
have not reproducibly observed the production of CTF from endogenous MLK3. It could
be due to the fact that endogenous MLK3 is present at relatively low levels, making it
technically difficult to detect the CTF using the immunoblotting method. It is also
conceivable that we have not identified an appropriate signal to trigger endogenous
MLK3 proteolysis. The strict sequence specificity of the MLK3 proteolytic event largely
rules out the possibility that the proteolysis is just an artifact of MLK3 overexpression. It
is conceivable that, under physiological condition, ERK activation promotes cell
proliferation, and may trigger the activation or expression of the MLK3 protease, which

cleaves and inactivates MLK3, thus blocking MLK3-mediated apoptosis.

109

Pulse chase experiments in HEK 293 cells support the idea that proteolysis occurs
in newly synthesized MLK3. Cycloheximide, a protein synthesis inhibitor, blocks PMA
induced proteolysis, further strengthening this idea. We are unaware of reports of the
proteases that act on newly synthesized proteins, except that the 26S proteasome located
in the cytosol degrades misfolded or unassembled newly synthesized proteins [29]. It is
unclear why the MLK3 protease selectively chooses newly synthesized MLK3.
Interestingly, we observe that the 5’ untranslated region (UTR) in mlk3 gene is required
for this proteolytic event, since an MLK3 construct lacking the 5’ UTR does not generate
the CTF (data not shown). It is conceivable that the MLK3 protease may physically
interact with the translational machinery near the 5’ UTR.

Western blotting using an antibody raised against the amino terminal SH3 domain
of MLK3 fails to detect the predicted 27 kDa amino terminal fragment of MLK3,
suggesting that the amino terminal fragment may be unstable. The existence of the stable
CTF suggests that CTF may have some biological function. Since the CTF does not
contain a complete kinase domain, it should render MLK3 catalytically inactive.
However, co-immunoprecipitation experiments using the CTF and full length MLK3
reveals that the CTF can associate with MLK3, presumably through its zipper region
(data not shown). It has been shown that zipper-mediated homo-oligomerization is
required for MLK3 activation and JNK signaling [16, 30]. This leads to the possibility
that the CTF may modulate MLK3 catalytic activity through heterodimerization. In
addition, CTF contains the CRIB motif responsible for association with small GTPase
Rae and Cdc42. Thus, although speculative, it may act as dominant negative effector to

regulate GTPase signaling.

110

Site-directed mutagenesis studies indicate that the Leu 250 and Gln 251 at the P2
and P1 sites are critical for generation of the CTF. Among the four MLK family members
MLKl-4, the Leu and Gln are only conserved in MLKl and MLK3, suggesting that it
may be a unique regulatory mechanism for these two kinases. Extensive searches of the
literature have failed to reveal a mammalian protease that cleaves after a Gln residue.
However, a group of related coronavirus proteases does display a requirement for Gln at
the Pl site. Coronaviruses are positive-strand RNA viruses with exceptionally large
genomes (27 to 31 kb) and one strain of human coronavirus, 229E, causes upper
respiratory tract illness in humans, in particular, the common cold [31]. The coronavirus
main proteinase (Mpm), also called 3C-like proteinase (3Cp’°), is a cysteine protease [32]

that cleaves protein substrate at Leu-Gln l(Ser, Ala, Gly) sequences (the cleavage site is

denoted by l). Mpro functions in proteolysis and maturation of the coronavirus replicase

pro

polyproteins. Cellular substrates for M have not been identified [33]. The newly
identified severe acute respiratory syndrome (SARS) Mpro also requires a Gln residue at
the P1 site, with some flexibility at the P2 site [34]. Our data indicate that a coronavirus
protease-like activity exist in mammalian cells. There are a couple of interesting thoughts
regarding this finding. First, several labs have developed inhibitors against the Mpro of
SARS for therapeutical application. It would be intriguing to test whether those inhibitors
actually block the MLK3 proteolytic event. It is also conceivable that MLK3 may be a
cellular target for M‘”r0 upon viral infection. Based on the literature [35] and the simple
BLAST search performed by us, there are no mammalian homologues of Mp”. The

MLK3 protease may share little primary sequence homology with Mp”, but shares

sufficient structural similarity to retain the same recognition site specificity.

111

The ability of several commercially available protease inhibitors to block the
PMA-induced proteolytic event of MLK3 was tested. General cysteine or aspartyl
protease inhibitors failed to block the generation of the CTF. However, one protease
inhibitor, MG-132 is able to block the PMA-induced proteolysis. In the literature, MG-
132 has been reported to inhibit the activity of proteasome and secretases. However, a
specific inhibitor of proteasome, lactacystin, and a specific inhibitor of y-secretase, L-
685, 458 are unable to block the proteolysis. These findings suggest that MG-132 blocks
the proteolytic event of MLK3, by inhibiting a protease other than the proteasome or the
y-secretase. The structure of MG-l 32 is carbobenzoxy-Leu-Leu-Leu-CHO. The sequence
before the cleavage site of MLK3 is Leu-Leu-Leu-Gln (Fig. 7). It is very likely that MG-
132 acts as a competitive inhibitor that binds to the catalytic core of the MLK3 protease,
thus preventing MLK3 binding and subsequent cleavage by the protease. To gain fiirther
support for this idea, the peptide Ac-lle-Leu-Leu-Leu-Gln-CO-NHZ is being synthesized
and its ability to block this proteolytic event of MLK3 will be tested. If this proves
successful, it may be possible to use the peptide inhibitor or another such compound as an
affinity tag to isolate the MLK3 protease from cellular lysates so that the identity of the

protease can be determined by mass spectrometry.

112

 

Footnotes

 

Acknowledgements:
1. The kinase domain model of MLK3 was done by Ritesh Agrawal.
2. The PMA induced MLK3 proteolysis was first observed by Weiqin Chen in Dr.
Esselman’s lab.
3. TNFOI and CD3 antibody induced MLK3 proteolysis was observed by Geou-yarh
(Stancy) Liou.

113

 

6. Reference:

10.

11.

12.

13.

14.

Keil, B., Specificity of proteolysis. 1992: Berlin ; New York : Springer-Verlag.

Schechter, I. and A. Berger, 0n the size of the active site in proteases. I. Papain.
Biochem Biophys Res Commun, 1967. 27(2): p. 157-62.

Seemuller, E., et al., Proteasome from T hermoplasma acidophilum: a threonine
protease. Science, 1995. 268(5210): p. 579-82.

Groll, M., et al., Structure of 205 proteasome from yeast at 2.4 A resolution.
Nature, 1997. 386(6624): p. 463-71.

Unno, M., et al., The structure of the mammalian 20S proteasome at 2. 75 A
resolution. Structure (Camb), 2002. 10(5): p. 609-18.

Hershko, A. and A. Ciechanover, The ubiquitin system. 1998.

Dick, L.R., et al., Mechanistic studies on the inactivation of the proteasome by
lactacystin in cultured cells. J Biol Chem, 1997. 272(1): p. 182-8.

Eamshaw, W.C., L.M. Martins, and SH. Kaufrnann, Mammalian caspases:
structure, activation, substrates, and functions during apoptosis. Annu Rev
Biochem, 1999. 68: p. 383-424.

Sorimachi, H., S. Ishiura, and K. Suzuki, Structure and physiological function of
calpains. Biochem J, 1997. 328(Pt 3): p. 721-32.

Marks, N. and M.J. Berg, APP processing enzymes (secretases) as therapeutic
targets: insights from the use of transgenics ( T gs) and transfected cells.
Neurochem Res, 2003. 28(7): p. 1049-62.

Fortini, M.E., Gamma-secretase-mediatedproteolysis in cell-surface-receptor
signalling. Nat Rev Mol Cell Biol, 2002. 3(9): p. 673-84.

Lee, H.J., et al., Presenilin-dependent gamma-secretase-like intramembrane
cleavage of ErbB4. J Biol Chem, 2002. 277(8): p. 6318-23.

Ray, W.J., et al., Cell surface presenilin-I participates in the gamma-secretase-
like proteolysis of Notch. J Biol Chem, 1999. 274(51): p. 36801-7.

Shearman, M.S., eta1., L-685,458, an aspartyl protease transition state mimic, is

a potent inhibitor of amyloid beta-protein precursor gamma-secretase activity.
Biochemistry, 2000. 39(30): p. 8698-704.

114

 

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

27.

Bock, 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.

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 JNK pathway. Monomeric SPRK
L410P does not catalyze the activating phosphorylation of Thr258 of murine
MIT OGEN-A C T 1 VA T ED protein kinase kinase 4. J Biol Chem, 2000. 275(36): p.
27893-900.

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

Xu, W., S.C. Harrison, and M.J. Eck, Three-dimensional structure of the tyrosine
kinase c-Src. Nature, 1997. 385(6617): p. 595-602.

Rock, K.L., et al., Inhibitors of the proteasome block the degradation of most cell

proteins and the generation of peptides presented on MHC class I molecules.
Cell, 1994. 78(5): p. 761-71.

Klafki, H., et al., The carboxyl termini of beta-amyloid peptides 1-40 and 1-42 are
generated by distinct gamma-secretase activities. J Biol Chem, 1996. 271(45): p.
28655-9.

Steinhilb, M.L., R.S. Turner, and J.lR. Gaut, The protease inhibitor, M0132,
blocks maturation of the amyloid precursor protein Swedish mutant preventing
cleavage by beta-Secretase. J Biol Chem, 2001. 276(6): p. 4476-84.

Widmann, C., et al., Mitogen-activated protein kinase: conservation of a three-
kinase module from yeast to human. Physiol Rev, 1999. 79(1): p. 143-80.

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.

DeSilva, D.R., et al., Inhibition of mitogen-activated protein kinase kinase blocks

T cell proliferation but does not induce or prevent anergy. J Immunol, 1998.
160(9): p. 4175-81.

Bennett, B.L., et al., SP600125, an anthrapyrazolone inhibitor of Jun N—terminal
kinase. Proc Natl Acad Sci U S A, 2001. 98(24): p. 13681-6.

Davies, S.P., et al., Specificity and mechanism of action of some commonly used
protein kinase inhibitors. Biochem J, 2000. 351(Pt 1): p. 95-105.

Rami, A., D. F erger, and J. Krieglstein, Blockade of calpain proteolytic activity

rescues neurons from glutamate excitotoxicity. Neurosci Res, 1997. 27(1): p. 93-
7.

115

28.

29.

30.

31.

32.

33.

34.

35.

Ohkuma, S. and B. Poole, Fluorescence probe measurement of the intralysosomal
pH in living cells and the perturbation of pH by various agents. Proc Natl Acad
Sci U S A, 1978. 75(7): p. 3327-31.

Wickner, 8., MR. Maurizi, and S. Gottesman, Posttranslational quality control:
folding, refolding, and degrading proteins. Science, 1999. 286(5446): p. 1888-93.

Leung, I.W. and N. Lassam, 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.

Myint, S., The coronavirdae. 1995: Plenum, New York.

Ziebuhr, J ., G. Heusipp, and 8G. Siddell, Biosynthesis, purification, and
characterization of the human coronavirus 229E 3C-like proteinase. J Virol,
1997. 71(5): p. 3992-7.

Hegyi, A. and J. Ziebuhr, Conservation of substrate specificities among
coronavirus main proteases. J Gen Virol, 2002. 83(Pt 3): p. 595-9.

Anand, K., et al., Coronavirus main proteinase (3 CLpro) structure: basis for
design of anti-SARS drugs. Science, 2003. 300(5626): p. 1763-7.

Anand, K., et al., Structure of coronavirus main proteinase reveals combination

of a chymotrypsin fold with an extra alpha-helical domain. Embo J, 2002. 21(13):
p. 3213-24.

116

IV. Identification of Mixed Lineage Kinase 3 Binding Partners by

Affinity Purification and Mass Spectrometry

1.Abstract

Mixed-lineage kinase 3 (MLK3) is a mitogen-activated protein kinase kinase
kinase (MAPKKK) that activates MAPK pathways, including the JNK and p38
pathways. Western blotting of cellular lysates from human cell lines revealed that MLK3
is present at high levels in breast cancer cell lines. Thus the human breast cancer cell
line, MCF-7, was used as a source for the identification of proteins that interact with
MLK3. MCF-7 cells were engineered to inducibly express Flag-tagged MLK3. Flag-
MLK3 protein complexes were affinity purified from MCF-7 cellular lysates using anti-
Flag M2 affinity resin. Flag-MLK3 and associated proteins were resolved on SDS-
polyacrylamide gels and stained with Coomassie Blue. Bands representing potential
MLK3 binding partners were identified by trypsin digestion followed by liquid
chromatography/ nanoelectrospray tandem mass spectrometry (LC/MS/MS). Ten proteins
were identified using this strategy, including heat shock proteins Hsp90 a,B, Hsc70,
Hsp70, and Hsp90 co-chaperone Cdc37; Clathrin heavy chain, Clathrin assembly protein
complex 2 medium chain; tubulin, adenine nucleotide translocator-Z, and ribosomal
protein S3. The strategy described in this chapter should be applicable to the
identification of binding partners of other proteins expressed in mammalian cells.

As a second approach, in-solution digestion coupled with LC/MS/MS was also
performed to identify MLK3 binding partners. This method not only confirmed the

identity of most binding partners discovered by in-gel digestion, but also identified

117

thirteen additional potential MLK3 binding partners. These findings provide new avenues
to explore that may help us further understand the regulation and biological function of

MLK3.

118

 

2. Introduction

Cells respond to a variety of extracellular stimuli such as growth factors, osmotic
stress, hormones, and nutritional deprivation. Many of these responses are mediated by
cell surface receptors, and activation of these receptors triggers cellular responses, such
as proliferation, differentiation, or apoptosis. In these signaling cascades, physical
protein-protein interactions are often critical in transducing the signal. The activity of a
protein kinase can be modulated by physical interactions with other signaling molecules.
For instance, binding of an activator or inhibitor may directly impact the catalytic activity
of protein kinases. Phosphorylation by other kinase may alter the properties of a protein
kinase. Conversely, protein kinases can interact with and phosphorylate downstream
substrate proteins. In addition, protein interactions may dictate the subcellular location of
a particular protein kinase. Finally, chaperone proteins can influence the stability of
protein kinases.

MLK3 is a serine/threonine kinase that functions as a mitogen activated protein
kinase (MAPK) kinase kinase (MAPKKK) to activate the JNK and p38 pathways [1-4].
Though more controversial, it has also been published that MLK3 is capable of
phosphorylating IKK to positively regulate the NFIcB pathway in response to T cell
costimulation [5]. In addition to its catalytic domain, MLK3 contains an SH3 domain, a
leucine zipper, a CRIB motif and a carboxyl terminal Pro/Ser/Thr-rich region. These
regions may mediate interactions with other proteins, which either regulate MLK3
activity, or impact MLK3 signaling.

MLK3 belongs to the mixed lineage kinase subfamily of protein kinases. The four

MLK members, MLK1-4, all contain an SH3 domain, a kinase domain, a zipper region

119

and a CRIB motif. Of the four MLK subfamily members, only MLK3 and MLK2 have
been biochemically characterized. Using a variety of approaches, several proteins have
been shown to interact with MLK2 and/or MLK3 as summarized in Table 1. However,
the mechanisms regulating MLK3 activity are still not well understood, and the
biological functions of MLK3 are not well defined. The identification of additional,
physiologically relevant MLK3 binding should shed light on the biological roles of
MLK3.

Affinity purification coupled with mass spectrometry has emerged as a popular
tool for isolating protein complexes and identifying their components. Two groups have
successfully applied a single step purification using the anti-Flag M2 affinity resin to fish
out protein complexes [6, 7].

Mass spectrometry (MS) has become a critical component in the study of
biological systems. It can be applied to many aspects of biological research, such as the
study of protein conformational changes by hydrogen/deuterium exchange, the
identification of posttranslational modifications, and the identification of unknown
proteins. In addition, methods are being developed to use mass spectrometry to determine
relative levels of specific cellular proteins in different biological/pathological states.

All mass spectrometers are composed of three basic components: an ionization
source wherein molecules are ionized and vaporized, a mass analyzer where the ions are
separated based on the mass to charge ratios, and a detector. Matrix assisted laser
desorption lionization (MALDI) and electrospray ionization (ESI) are the two most
commonly used ionization methods applied to proteins and peptides. MALDI uses a laser

pulse to induce the formation of singly charged peptide ions [8], and is often coupled to a

120

Time of F light (TOF) mass analyzer for peptide mass mapping or analysis of
posttranslational modification, including phosphorylation.

In ESI, molecules form a beam of ions in the presence of a strong electrical field
by first forming a spray of droplets, which quickly form smaller droplets upon solvent
evaporation. This process continues until individual charged peptide ions are formed [9].

E51 is often coupled with a tandem MS (MS/MS) instrument to derive sequence
information of ionized peptides. In an MS/MS experiment, a precursor ion is mass-
selected by the first mass analyzer and focused into a collision region in which the
precursor ion is induced to fragment thus yielding product ions. Product ions are mass
analyzed by a second mass analyzer. There are two commonly used tandem MS
instruments. The triple quadruple mass analyzer is arranged in series in space, whereas
the ion trap mass analyzer is arranged in series in time [9].

Several programs exist to analyze the raw MS/MS data of peptides, including
PROF OUND, MASCOT, and SEQUEST. Using the SEQUEST algorithm, acquired
fragmentation spectra of peptides are correlated with the predicted amino acid sequences
in protein databases. The proteins can be identified by the resulting list of peptide
sequences [10]. Several parameters are used in SEQUEST to indicate the accuracy of the
prediction. The cross-correlation score (XCorr) measures the similarity between the
mass-to-charge ratios for the fragment ions predicted from amino acid sequences
obtained from the database and the fragment ions observed in the tandem mass spectrum.
A peptide with an XCorr above 2.5 is generally considered reliable. The DeltaCn (1 .0-
normalized correlation score) indicates the normalized numerical difference between the

top and second highest XCorr in the database. RSp(Rank/preliminary score) indicates a

121

preliminary ranking based on the number of matched ion peaks. A subjective
combination of these scores, as well as other factors such as the charge of the precursor
ion, and the number of peptides that map to a given protein, is typically used to evaluate
the accuracy of each prediction [11]. Finally, further biochemical analysis must be
performed to validate these findings.

Using affinity purification coupled with electrospray/MS/MS, more than twenty
potential MLK3 binding partners were identified, suggesting MLK3 forms complexes

with other proteins in vivo.

122

3. Materials and Methods

3.1. Plasmid Constructs and Mutagenesis

Construction of the cytomegalovirus-based expression vectors carrying the
cDNAs for wild type MLK3 (pRK5-NFlag.mlk3), and MLK3 truncation variants (pCGN-
HA mlk3 1-114, 115-399, 400-591, 592-847) have been described elsewhere [12]. The
expression vector encoding pGEX-2T GST-Clathrin 1-5 79 was kindly provided by Dr.
James H. Keen at Thomas Jefferson University (Philadelphia, PA). The expression vector
encoding wildtype MLK3 was used as a template to create mlk3 K144R using the Quick
Change Site-directed Mutagenesis method (Stratagene). The presence of the desired
mutation was confirmed by automated DNA sequencing in the Genomics Technology

Support Facility at Michigan State University.

3.2. Construction of MCF-7 Inducible Cell Lines

MCF-7 stable cell line expressing the transcription factor vector pLzNz-
RHSBH/ZF 3 (ARIAD) was kindly provided by Dr. Susan Conrad. The target gene vector
pLH-lel-PL mlk3 was constructed by polymerase chain reaction-mediated amplification
of the corresponding coding sequence from pRK5 -NF1ag.mlk3 followed by subcloning
into the target gene vector at the Hind III and C1a 1 sites. The pLH-lel-PL mlk3 plasmid
was transfected into the stable MCF-7 cell line using the calcium phosphate method as
described elsewhere [13]. Stable clones were selected under neomycin (4O rig/ml)
(Invitrogen) and hygromycin (50 rig/ml) (Invitrogen) containing Dulbecco’s Modified
Eagle’s Medium (DMEM) (5% fetal bovine serum) (Invitrogen) for about 2 weeks.

Individual colonies that formed after about 4 weeks were picked and cultured separately.

123

AP21967 was added to induce the expression of MLK3 and the expression level was
confirmed by resolving cellular lysates on SDS-PAGE followed by immunoblotting using
Flag antibody. Individual clones inducibly expressing MLK3 were maintained under
neomycin (40 rig/ml) and hygromycin (10 rig/ml) containing DMEM (5% fetal bovine

serum).

3.3. Purification of Flag-MLK3 Complexes

MCF-7/ iFlag-MLK3 (inducibly expressing F lag-MLK3) cells (1X 108) in eight
15 cm dishes were cultured in the presence of 100 nM AP21967 for 20 h to induce Flag-
MLK3 expression. Equal amounts of cells were cultured in the absence of the AP21967
as a negative control. The cells were washed once with PBS and lysed in lysis buffer (50
mM Tris HCl, pH 7.4, 150 mM NaCl, lmM EDTA, 1% Triton X-100) containing a
cocktail of protease inhibitors (Sigma). The lysates were pooled, centrifuged at 14,000
rpm for 15 min, and the supernatants passed through a 0.22 M filter to remove any
remaining cells and particulates. Cleared cellular lysates were then incubated with 400 u]
of anti-F lag M2 agarose gel (Sigma) (previously equilibrated in TBS (SOmM Tris HCl,
pH 7.4, 150 mM NaCl) buffer) for 90 min with gentle mixing at 4°C. The resin was
collected by centrifugation at 1000x g for 5 min and was washed with TBS buffer three
times to remove non-specific binding proteins. Flag-MLK3 complexes were eluted from
the resin by incubation for 60 min at 4°C with 400 111 TBS containing 300 ng/ul 3X Flag
peptide (Sigma). The proteins collected in the supernatant were then precipitated by ice-

cold acetone (3:1 acetone: supernatant).

124

3.4. Nupage Gel Electrophoresis and Coomassie Staining

After acetone precipitation, the Flag-MLK3 complexes were dissolved in 30 pl of
water and Nupage Lithium Dodecyl Sulfate (LDS) sample buffer (4X) (Invitrogen) and
50 mM DTT was added. Protein mixtures were denatured at 70°C for 10 min and then
separated on Nupage 4-12 % Bis-Tris Gels (Invitrogen). Gels were stained with

SimplyBlue SafeStain (Coomassie G—250) (Invitrogen).

3.5. In-gel Trypsin Digestion

The protein bands were excised from the Coomassie stained gel and chopped into
1 mm3 pieces. Gel pieces were transferred to siliconized 1.5 m1 Eppendorf tubes (Dot
Scientific), and proteins were reduced with 10 mM dithiotheritol (DTT) by incubating at
56°C for 30 min, and then alkylated with 55 mM iodoacetamide in the dark at room
temperature for 20 min. The gel pieces were dehydrated with 100% acetonitrile and dried
using a Speed-Vac. Gel pieces were covered with sequencing grade trypsin (13 ng/pl)
(Promega) and incubated at 37°C overnight. Tryptic peptides were extracted with 60%
acetonitrile/1% trifluoroacetic acid and were completely dried using a Speed-Vac. The

dried sample was reconstituted by adding 8 pl of 1% formic acid.

3.6. In-solution Trypsin Digestion
After acetone precipitation, proteins were dissolved in 15 pl buffer containing 6M
urea, 50 mM Tris pH 8, 2 mM DTT and incubated at 60°C for 45 min to reduce and

denature the proteins. Sequencing grade trypsin (100 ng in 85 pl of 50 mM ammonium

125

 

 

bicarbonate) was added to digest the protein mixture at 37°C overnight. Tryptic peptides
were concentrated to approximately 20 pl by Speed-Vac evaporation.
3.7. LC/MS/MS

A portion of the tryptic peptides (typically 6 pl out of 20 pl sample) was loaded
onto a peptide capillary trap (Michrom 004/25108/32) through the auto sampler of the
CapLC system (Waters Co., Milford, MA). An auxiliary pump was used to wash the trap
with 1% aqueous formic acid at a flow rate of 20 pl/min for 5 min. At the end of the first
5 min wash, an automated valve on the CapLC system was switched allowing the
gradient mixture of solvent A and B (Solvent A: 1% formic acid in water; Solvent B: 1%
formic acid in acetonitrile) to flow through the trap and onto the nano-spray column
(PicoFrit column from NewObjective, PFC7515-AQ--5, 15 pm tip id, 75 um column id, 5
cm of C18 packing resin before the fiit at the tip). The chromatography was run
isocratically for 5 min at 5% B/95%A and then was increased to 60% B in a linear
gradient over 25 minutes. For in solution digests, the chromatography was run
isocratically for 10 min at 5% B and then increased in a linear gradient to 60% B over
140 minutes. The flow rate from the PicoFrit column tip was 0.1 til/min. The flow from
the pump was 9 pl/min, allowing the gradient to be ramped up or down smoothly. A
micro T-splitter (Upchurch Scientific, Oak Harbor, WA) between the CapLC pumps for
solvent A/B mixture and the capillary trap interfaced the two flow rates. The flow from
the PicoFrit column tip was directly sprayed into a Finnigan LCQ-Deca mass
spectrometer (Thenno Finnigan, San Jose, CA). The source voltage was set at 2.7 kV
and the heated capillary temperature was set at 200°C. The quadruple ion trap mass

spectrometer scanned from m/z 200 to m/z 2000 at a rate of 1.5 sec/scan. Data-dependent

126

 

scan methods were used. The dynamic exclusion scan was set so that MS/MS would be
performed on the three highest peaks over the set threshold. The same m/z detected over
5 min was scanned only three times. Zoom scan was not applied in order to perform more
MS/MS scans. The calibration was accomplished by infusion of l pmol/pl of angiotensin
peptide at the same flow rate (0.1 III/min). The SEQUEST software (version 2,
TurboSequest) (Thermo F innigan, San Jose, CA) was used to extract the MS/MS files
from the raw data file, and to match the peptides in the human Fasta protein database.
The precursor ion mass tolerance was set at 2.5 Da. A smaller deviation between the
experimental mass and the calculated mass results in a higher final score. Scoring of all
the candidate peptides was based on the matching of the observed b, y ions in an MS/MS

spectrum with the calculated counterparts from the candidate sequences.

3.8. Transfection and Lysis of HEK 293 cells
Human embryonic kidney (HEK) 293 cells were cultured on lOO-mm dishes and
transfected using the calcium phosphate method as previously described [13]. Cells were

harvested 16 h after transfection, and lysed as described previously[12].

3.9. Purification of GST Fusion Proteins and GST Pulldown Assays

GST and GST-Clathrin heavy chain (1-579) were expressed in DE3 E. coli and
purified on glutathione-Sepharose 43, according to the manufacturer's protocol
(Amersham Phamacia Biotech). Clarified lysate (400 pl) was incubated with 20 pl of

glutathione-Sepharose 4B resin prebound with 5 pg of GST fusion protein for 90 min at

127

 

4 °C. GST pull-downs were washed twice with HNTG buffer (20 mM HEPES (pH 7.5),

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

3.10. SDS-PAGE and Western Blot Analysis

Lysates and GST pulldowns were resolved by SDS-PAGE. Proteins were
transferred to nitrocellulose membranes and immunoblotted using the Flag antibody (4
pg/ml), actin mouse monoclonal antibody (Sigma), or phospho-c-Jun (KM-l) mouse
monoclonal antibody (2 pg/ml) (Santa Cruz), followed by the appropriate horseradish
peroxidase-conj ugated secondary antibody (Bio-Rad). Western blots were developed by

chemiluminescence.

128

4. Results

4.]. Survey of endogenous levels of MLK3.

Total cellular lysates from a variety of human cell lines, HEK 293 cells (human
embryonic kidney epithelial cells), Jurkat T lymphocyte cells, MCF-7 cells (human breast
carcinoma), A549 (human lung carcinoma), and HeLa (human cervix adenocarcinoma)
were surveyed for their levels of endogenous MLK3. As shown in Fig. 1, of the cell lines
tested, MLK3 is present at highest levels in MCF-7 cells, suggesting that MCF-7 cells

may be a good source of physiologically relevant binding partners of MLK3.

4.2. Inducible expression of MLK3 in MCF-7 cells

To establish a reliable system for affinity purification of MLK3 complexes, MCF-
7 cells were engineered to inducibly express Flag-tagged MLK3 using the Ariad system
(Fig. 2). As shown in Fig. 3, Flag-MLK3 was inducibly expressed upon the addition of
the dimerizer AP21967. Since MLK3 activates the JNK pathway, the phosphorylation of
the transcription factor c-Jun was used as readout of MLK3 activity in vivo. The
phosphorylation of c-Jun dramatically increased upon induced expression of Flag-MLK3,

suggesting that inducibly expressed Flag-MLK3 is catalytically active.

4.3. Purification of Flag-MLK3 complexes from MCF-7 cells

Inducibly expressed Flag-MLK3 in MCF-7/iFlag-MLK3 cells was
immunoprecipitated with anti-Flag M2 affinity gel as described in “Materials and
Methods”. Flag-MLK3 complexes were eluted by addition of 3X Flag peptide and were

analyzed by SDS-PAGE and Coomassie staining. Uninduced MCF-7/iFlag-MLK3 cells

129

were used as a negative control. As shown in Fig. 4, in addition to the band representing
Flag-MLK3, there are multiple bands representing proteins that co-purified with Flag-
MLK3, but were absent in the fraction purified from control cells. Protein bands were

assigned numbers as indicated in Fig. 4.

4.4. Identification of putative MLK3 binding partners by in-gel trypsin digestion
and MS/MS

The bands of interest were excised from the gel and trypsin digestion was
performed as described in “Materials and Methods”. After digestion, tryptic peptides
were analyzed directly by LC/MS/MS to obtain sequence information. Analysis of band 1
is shown as an example in Fig. 5. Tryptic peptides were separated based on
hydrophobicity by reverse phase I-IPLC, which greatly reduces the complexity of the
tryptic peptide mixture, and eluting peptides were ionized by electrospray under a high
voltage field. Ionized peptides enter the ion trap sequentially to generate mass spectra.
The mass spectrum taken at 25.36 min is shown as an example in Fig. 5B. In this
spectrum, there are three peaks with high intensity. The three corresponding peptides all
undergo collision-induced dissociation and produce MS/MS. The MS/MS of peptide peak
m/z=617.21 and peak m/z=668.13 are shown in Fig. 5C and 5D. The SEQUEST
algorithm was used to analyze the raw MS/MS data and to match the peptide sequences
to proteins in human Fasta protein database. Based on the sequence information obtained
by MS/MS, these two peptides with a m/z =617.21 or 668.13, respectively, match

Clathrin heavy chain in the protein database. For protein band 1, fourteen peptide

130

sequences determined by MS/MS match the Clathrin heavy chain in the human protein
database. Therefore, protein band 1 represents Clathrin heavy chain.

Using in-gel digestion followed by LC/MS/MS, ten proteins were identified as
potential MLK3 binding partners (Table 2). Three of the bands (band 2, 3 and 6) excised
from the gel contained heat shock proteins. These proteins are heat shock protein 90
(Hsp90) a, B and its co-chaperone Cdc37, heat shock 70 kD protein 1 (Hsp70) and heat
shock cognate 71 kD protein (Hsc70). Two bands (band 1 and 4) excised from the gel
represent Clathrin heavy chain (CHC) and Clathrin assembly protein complex 2 medium
chain (AP2-p2). In addition, (It-tubulin, adenine nucleotide translocator-2 (ANT2), and

ribosomal protein S3 were identified in Flag-MLK3 complexes.

4.5. In-solution digestion and MS/MS

In-gel digestion requires separation of protein samples in advance and the
catalytic activity of trypsin is reduced in gel. Yates et al. have successfully performed in-
solution trypsin digestion coupled with multidimensional liquid chromatography and
MS/MS to identify large numbers of proteins in complex mixtures [14] [15]. Since
MLK3 complexes are relatively simple after affinity purification, we utilized an in-
solution digestion coupled with one-dimensional HPLC to separate tryptic peptides
derived from Flag-MLK3 complexes. Three independent experiments were performed.
As shown in Table 3, except for AP2-p2, all other proteins that were identified by in-gel
digestion were detected in these experiments. Although a-tubulin was not identified, 13-
tubulin was detected. This suggests that MLK3 may associate with microtubules which

are composed of (1/13 tubulin heterodimers [16]. Thirteen additional potential MLK3

131

 

 

binding partners were also identified. However, since each of these proteins was
identified in only one experiment, they might represent false positives rather than bona
fide binding partners. Further biochemical analysis is required to analyze and confirm

their interactions with MLK3.

4.6. Association of MLK3 with clathrin heavy chain

Band 1 was identified as clathrin heavy chain (CHC). Clathrin is a main
component of the clathrin-coated vesicles involved in endocytosis. Clathrin is a complex
of trimeric heavy chains, each of which is associated with a regulatory light chain. The
amino terminus of CHC consists a globular terminal domain (TD), which has been shown
to interact with a variety of proteins that function in the clathrin mediated endocytosis
[17] [18]. We were interested to test whether MLK3 interacts with CHC through its TD.
The TD of CHC (amino acids 1-579) was expressed as a fusion protein with GST in E.
coli and purified with glutathione Sepharose 4B beads. The ability of MLK3 to interact
with GST-CHC TD was assessed in GST pulldown experiments. As shown in Fig. 6,
MLK3 associates with GST-CHC TD but not with GST. A kinase inactive form of
MLK3, K144R, retains the ability to bind to GST-CHC TD, suggesting the kinase
activity is not required for this interaction.

A series of previously constructed amino terminal HA-tagged MLK3 variants,
which contains individual or tandem domains of MLK3, were used to define the clathrin
binding region within MLK3. These variants were expressed and analyzed for their
capacity to interact with the TD domain of CHC. Data from GST pull down assays show
that, of these variants, only MLK3 592-847 retains the ability to interact with GST-CHC

TD (Fig. 7). Thus we conclude that the binding site for clathrin TD is within amino acids

132

592-847 of MLK3. Interestingly, MLK3 592-847 apparently undergoes some proteolytic
degradation and the amino terminal proteolytic fragment associates with GST-CHC TD,
suggesting that clathrin binding region is not within the extreme carboxyl terminus of
MLK3.

Some proteins that function in clathrin-mediated endocytosis, such as the adaptor
proteins (AP), arrestin and epsinl, interact with the terminal domains of CHC through a
“clathrin box” motif with a consensus sequence of “LLpL-” (where p and — denote a
polar and a negatively charged residue respectively). This clathrin box motif is critical for
binding to the TD domain of CHC [l 8]. In the carboxyl terminus of MLK3, there is a
“clathrin box” like sequence “LLDVG” (aa7l3-717). However, mutation of one of the
five amino acid residue, Asp 715 to Ala in MLK3 did not abolish the binding ability of
MLK3 with clathrin heavy chain as judged by the GST pulldown assay (data not shown).

The nature and the consequences of this interaction are under investigation.

133

 

 

TABLE 1. Previously identified binding partners of MLK1, MLK2, and MLK 3

 

Binding partners

Reports on

Binding region in
MLKI ,2,3

Effect of interaction

 

14-3-3 e [3]

MLK2

No detectable effect on MLK2 induced
activation of JNK

 

Akt [19]

MLK3

511-847

Akt phosphorylates MLK3 S674 to
negatively regulate MLK3 activity and
prevent MLK3 induced cell death

 

Clathrin heavy chain
[201

MLK2

MLK2 931-935
Clathrin box motif

Overexpression of MLK2 may
negatively influence clathrin mediated
endocytosis

 

Dynamin [21]

MLK2

SH3 domain

MLK2 SH3 domain synergistically
stimulates dynamin activity with PtdSer

 

Hippocalcin [3]

MLK2

No detectable effect on MLK2 induced
activation of JNK

 

HPKI [22]

MLK3

SH3 domain

HPKl phosphorylates kinase inactive
MLK3 and but no report showing it
could alter MLK3 kinase activity

 

Huntingtin [23]

MLK2

SH3 domain

MLK2 associates with normal
huntingtin, but not polyglutamine-
expanded Huntington. Normal
huntington may negatively regulate
MLK2 activity.

 

JIP] [24,25]

MLK2, 3

J IPl enhances MLK3-mediated JNK
activation

 

JIP2/1B2 [4, 25];

MLK2, 3

MLK3 1-204

J 1P2 enhances MLK3-mediated JNK
activation

Tiaml enhances association of JIP2 with
MLK3 and JIP2 enhances Tiaml
activation of p38.

 

JIP3 [26]

MLK3

JIP3 enhances MLK3 mediated JNK
activation (J 1P3 is reported to not
associate with MLK2)

 

KIF3[3]

MLK2, 3

MLK2 497-953
MLK3 348-847

No detectable effect of MLK2 induced
activation of JNK. MLK2 co-localize
with JNK1/2 to punctate structures along
microtubules.

 

NeuroD [27]

MLK2

MLK2 phosphorylates NeuroD and
stimulate its transcripflion activity.

 

Pinl [28]

MLK3

Pinl may preferentially bind to
hyperphosphorylated MLK3

 

POSH [29]

MLK1,2,3

POSH act as a scaffold protein for Racl
and MLKs to facilitate JNK mediated
apoptosis.

 

13313-95 [30]

MLK2,3

The SH3 domain of PSD-95 associates
with MLK2, 3 during GluR6 mediated
JNK activation.

 

 

Rae/Cdc42 [3, 31]

 

MLK2, 3

 

CRIB

 

Enhances MLK3 autophosphorylation
and substrate phosphorylation

 

134

 

TABLE 2. Proteins identified by in-gel trypsin digestion and LC/MS/MS

The table lists the proteins, their abbreviations, GI (Gene Identifier) numbers in NCBI

database and molecular weight. Identified peptides and sequence coverage from one out

of three sets of experiments are indicated.

 

 

Protein Protein name Abbreviat- GI M.W. Identified Sequence
band ions (kDa) peptides Coverage
1 Clathrin heavy chain CHC 4758012 191.6 14 10.8%

2 Heat shock protein 90 a Hsp90a 123678 84.7 12 15.4%
Heat shock protein 90 [3 Hsp9013 6680307 83.3 9 12%

3 Heat shock 70 kD protein 1 Hsp70 123648 70.0 10 18.9%
Heat shock cognate 71 kDa Hsc70 462325 70.9 2 2.9%
protein

4 Clathrin assembly protein AP2-p2 1 13332 49.7 3 6.7%
complex 2 medium chain /AP50

5 a-tubulin 20455316 50.0 2 7.3%

6 P50 “‘37 Cdc37 21542000 44.5 1 4.5%

7 Adenine nucleotide ANT2 4502099 32.9 7 23.5%
translocator-2

8 Ribosomal protein S3 S3 417794 26.7 9 39.9%

 

135

TABLE 3. Proteins identified by in-solution trypsin digestion and LCfMS/MS in

three independent experiments.

 

 

Gene M.W. Identified Detected in

Identified proteins Identifier (kDa) Peptides Seq cov. Exp.
MLK3 GI:4505195 92.6 28 48.3% 1,2,3
Hsp9001* GI: 123678 84.6 10 17.9% 1,2,3
Hsp9OB* GI:11277141 84.8 7 14.0% 1,2
Hsc70* GI:5729877 70.9 9 34.4% 1,2
Hsp70* GI:2119712 69.9 5 14.1% 1,2
CDC37* GI:5901922 44.4 5 19.0% 3
Ribosomal protein S3* GI:13640373 26.7 3 18.5% 3
ANT2* GI:4502099 32.9 4 18.5% 1,3
Clathrin heavy chain* GI:4758012 191.6 2 3.0% 1
B-tubulin“ GI: 135471 49.7 2 9.2% 1,3
Clathrin assembly protein GI:20178274
complex 2 a-A large chain 107.6 2 6.3% 1
Ribosomal protein 85 GI:4506729 22.8 1 13.0% 1
Sorting nexin 12 GI:24418866 18.9 1 12.8% 2
Voltage-dependent N-type
calcium channel u-IB subunit GI:1705854 262.5 1 0.9% 2
Methylosome subunit t pICln GI:1708393 26.2 1 10.4% 3
Fibronectin receptor beta
subunit GI:124963 88.5 1 2.4% 3
Calcineurin-binding protein
Cabin 1 GI:6685261 246.4 1 1.8% 3
Transcription factor SL1 GI:2136304 95.3 1 3.9% 3
Glycoprotein gCquP GI:730772 31.4 1 9.1% 3
Son of sevenless protein
homolog 2 (SOS-2) GI:6175038 153.0 1 1.9% 3
SRPKl a protein kinase
(SRPKI) GI: 14252988 92.4 1 4.8% 3
Interferon-induced 35 kDa
protein (IFP 35) GI:19861692 31.5 1 13.3% 3
Death receptor 5, TRAIL
receptor-2 GI: 17380321 47.9 1 8.3% 3

 

136

Table 3 (cont’d)

 

Also detected in negative

 

 

control:
Titin GI:213628O 299.2 30 3.0% 1,2,3
Fission yeast Skbl protein
homolog GI:5174683 72.7 6 21.7% 2,3
Eukaryotic translation
initiation factor 48 GI:4503533 69.2 3 9.0% 1,2,3
sttathionine beta-synthase GI:543959 60.5 1 5.6% 3
Notes:

1. The asterisk indicates proteins that were also identified by in-gel digestion and

LC/MS/MS.

This table includes the identified potential MLK3 binding partners that contain at
least one peptide with a “XCorr” (cross-correlation score) above 2.5.

For the proteins that were identified in more than one experiment, the sequence

coverage from one set of experiments was listed.

137

 

MLK3 we *1 3

Jurkat MCF-7 HeLa 293 A549

Fig. 1. Expression of MLK3 in several human cell lines. Cellular lysates containing
equal amounts of total protein (20 pg) from the indicated cell lines were resolved by
SDS-PAGE and immunoblotted for MLK3.

138

 

 

' 5‘9" . Activation
Drug i -‘ 1 domain
binding , .2"
domain '

AP21967

£156“
Drug 1 _ x

, . _ —§ 1 .
binding domain 1‘ . 3"

Heterodimerizer

binding 110212111: i 1‘ mm
\VA _2\\ \\5\\'ma\ DNA
Flag-MLK3

 

Fig. 2. Construction of inducible cell lines. The cDNA encoding F lag-tagged mlk3 was
inserted into pLH-Z1 21-PL vector and transfected into an MCF-7 cell line that stably
expresses the transcription factor vector pL2N2-RHS3H/ZF 3. Stable clones were selected
using neomycin (40 pg/ml) and hygromycin (50 pg/ml) containing media. Clones were
isolated and tested for AP21967-induced expression of Flag-MLK3. (Figure adapted
from ARIAD)

139

AP21967 - +

h Flag MLK3 we

“ P-c-jun we

 

 

Actin WB

Fig. 3. Inducible expression of Flag tagged MLK3. A selected clone from the MCF-7
cells which inducibly expresses Flag-MLK3 was cultured either in the presence or in the
absence of AP21967 (100 nM) for 20 h. Cells were lysed and cellular lysates were
subjected to Western blot analysis. Top panel, the presence of Flag—MLK3 in the total
cellular lysates was assessed by immunoblotting using the Flag antibody. Middle panel,
the level of phosphorylated c-Jun in the total cellular lysates was assessed by
immunoblotting using an antibody directed against phospho-c-Jun. Bottom panel, equal
loading of the cellular lysates was confirmed by immunoblotting using an antibody

directed against actin.

140

- +

AP21967

  

31 kD ; *3... .

  

‘l‘ . - “33's!

. ..
s
..x
393

Fig. 4. Affinity purification of Flag MLK3 complexes. MCF-7 cells inducibly
expressing Flag-MLK3 were cultured either in the presence or in the absence of AP21967
for 20 h. Cleared cellular lysates were immunoprecipitated with anti-Flag M2 agarose.
Immunoprecipitates were washed and Flag-MLK3 complexes were eluted with 3X Flag
peptide. After acetone precipitation, Flag-MLK3 and associated proteins were resolved
by SDS-PAGE on a 4-12% acrylamide gradient gel and stained with Coomassie Blue.
Molecular mass markers are indicated on the left in kilodaltons. The lines with band
numbers indicate the positions of the proteins that were excised, digested with trypsin and

analyzed by mass spectrometry.

141

 

 

 

 

 

 

 

A
0)
8
ca 100 p
‘0
g 75%“
.0
<01) 50%
3% 25%“
'65 0%*—-—~"+— , . . - ,
0‘ o 10 20 30 4o 50 60
Time
B
668.13
100% l
E“
m
g 80% —
C
60% _ 617.21
40% n
20% *
0V0 MW
400 700 mlz1000

142

 

 

 

 

 

 

 

 

 

 

 

 

 

C MD
Wit 1;
Intensity ‘°°%‘ ya ’8 v9 v8 V7 y6 vs V4 V3
75% * b2
50% .‘
mi
0% i
0 300 600 900 1200 1500 mlz
b4 b5 b6 b8 b9
u v-Dttigii-HL-R
100°/ Y"
. ° ye vs
Intensity l y,
75% .
b4
50% ~
25% j
i i =
0% illLlli iii“ i
0 300 900 1200 1500
mlz

143

Fig. 5. Identification of band 1 as clathrin heavy chain by LC/MS/MS. A,
Reconstructed chromatogram of an LC/MS/MS run of a tryptic digest from an excised
gel band. x axis represents time; y axis represents the intensity of the base peak in the
mass spectrum acquired at that time. B, mass spectrum taken at 25.36 min (scan number:
810). There are at least three peaks representing three different peptides in this spectrum.
MS/MS was performed on each of them. The two MS/MS spectra that show adequate
fragmentation are plotted in C and D. C, MS/MS spectrum of the doubly charged tryptic
peptide IYIDSNNNPER (m/z 668.13). A nearly complete series of b ions and a nearly
complete series of y ions can be identified. D, MS/MS spectrum of another doubly
charged peptide VDKLDASESLR (m/z 617.21). The observed y and b fragment ions

corresponding to the indicated peptide sequence are shown.

144

MLK3 WT K144R -

 

 

{ W I \ f \
..I a.“ Flag WB
GST P“"d°W” Ki .-.H_....é ...773 GST-CHC TD

 

Flag WB

Fig. 6. Association of MLK3 with clathrin heavy chain terminal domain. HEK 293
Cells were transfected with expression vectors containing the cDNAs for MLK3 and
MLK3 K144R. Cellular lysates expressing the indicated MLK3 variants were incubated
with glutathione-Sepharose 4B resin to which purified GST-CHC TD or GST had been
bound. Top panel, the presence or absence of bound MLK3 variants was assessed by
Western blotting using the Flag antibody. Middle two panels, equal loading of GST-
Clathrin TD on the glutathione-Sepharose 4B resin was confirmed by Coomassie
staining. Bottom panel, the expression levels of MLK3 variants were assessed by Western
blotting of cellular lysates using the Flag antibody. The data shown are representative of

three independent experiments.

145

MLK3: 1-114 115-399 400-591 592-847
r 1 r 1 r \ (——\

 

“ .
-< HAWB
GST Pulldown ‘3” w w GST-CHC TD
_ .... _, __ GST
MLK3: E g g: g _
'l'
“'2
Input HAWB
'-

Fig. 7. Interactions of MLK3 variants with clathrin heavy chain. Afier transfection,
GST pulldown assays were performed as described in the “Material and Methods”. Top
panel, the presence or absence of bound MLK3 variants was assessed by Western
blotting using the HA antibody. Middle two panels, equal loading of GST-CHC TD on
the glutathione-Sepharose 4B resin was confirmed by Coomassie staining. Bottom panel,
the expression levels of MLK3 variants were assessed by Western blotting of cellular
lysates using the HA antibody. The data shown are representative of three independent

experiments.

146

 

5. Discussion

In the signal transduction cascades, many signals are transduced through protein-

 

protein interactions. The data presented here document how the affinity purification
coupled with mass spectrometry can be used to isolate interacting proteins. Based on its
generic nature, this method can be widely applicable to the identification and
characterization of interacting proteins in mammalian cells.

The anti-Flag M2 antibody covalently linked to the agarose was used in the
isolation of Flag-MLK3 complexes. Addition of the 3X Flag peptide to the affinity
agarose results in quantitative elution of MLK3 (data not shown), which not only reduces
contamination with light and heavy chains of the antibody, but also allows the eluted
proteins to be precipitated with acetone. Acetone precipitation significantly reduces the
volume of the sample, making it feasible to load large amounts of total proteins onto the
polyacrylamide gels.

In this study, tandem affinity purification (TAP) was not applied. The TAP
methodology is a protein tag-based affinity purification technique, which was originally
developed in yeast [32, 33]. One TAP tag consists of two immunoglobulin-binding
domains of protein A, followed by a cleavage site for the tobacco etch virus (TEV)
protease, and the calmodulin-binding peptide (CBP). Bait protein and its interacting
proteins in the cellular extract are first recovered by affinity selection on an IgG agarose.

After washing, the TEV protease is added to release the bound complexes. The eluate is

 

incubated with calmodulin-coated beads in the presence of calcium. After washing, the

bound complexes are released with EGTA. This two-step purification greatly reduces

147

non-specific binding, but also requires high-affinity binding among the complex proteins
and is most suitable for stable protein complexes, such as translational machinery or
transcription complexes. However, MLK3 is a protein kinase involved in signal
transduction and its interactions with other proteins may be weak or transient. The
affinity purification and elution procedure applied in this study are relatively gentle,
which enhances the capture of MLK3 complexes.

Electrospray coupled with ion trap mass spectrometer was used in the
identification of potential MLK3 binding partners. In contrast to MALDI, electrospray
ionizes peptide samples directly from solution; thus peptides mixture can be resolved by
capillary HPLC, which greatly reduces the complexity of the peptide mixtures. Ion trap
mass spectrometry is capable of sequencing the peptides ions by collision-induced
dissociation and this significantly facilitates the identification of unknown proteins. For
example, the peptide sequence with the highest XCorr score obtained from in-gel
digestion of band 6 matches a peptide sequence in Cdc37 in the human Fasta protein
database. Therefore, Cdc37 was assigned as an MLK3 binding partner.

We noticed that the identified MLK3 binding partners are relatively abundant
proteins in cell using the in-gel digestion approach. This could be due to the fact that only
Coomassie stainable bands were analyzed. Certainly, ES/MS/MS is capable of
identifying less abundant proteins. In-solution digestion coupled with one-dimensional
reverse phase C18 HPLC was attempted three times to identify additional MLK3 binding
partners. In these experiments, thirteen proteins were identified in the MLK3 complexes
(Table 3), including some signaling proteins, such as the guanine nucleotide exchange

factor SOSZ, SRPKla protein kinase, interferon-induced 35 kDa protein, and death

148

 

receptor 5. This suggests that MLK3 may form complexes with those signaling proteins
to mediate signal transduction. However, the in-solution digestion experiment is unable
to provide the molecular weight information of the unknown proteins. In addition, most
proteins, including some proteins already identified by in-gel digestion, were not detected
consistently in the three experiments. Irnmunoblotting of gels of resolved MLK3
complexes with specific antibodies against each potential candidate is necessary to
examine their presence in the MLK3 complexes. Due to the uncertainty of their
interaction with MLK3, the various proteins that have been identified using the in-
solution digestion will not be discussed in details.

Several additional approaches to increase the sensitivity of the identification of
MLK3 binding partners can be applied in future studies. For example, a cross-linking
reagent can be used to capture the transient interacting proteins of MLK3. Cellular
fractionation may reduce the complexity of MLK3 complexes. Additionally, in the in-
solution digestion followed by LC/MS/MS experiment, the peptides mixture may
overwhelm the resolution capacity of the single-dimensional chromatography system.
Orthogonal two-dimensional separation methods (cation exchange chromatography
followed by reverse-phase chromatography) have been reported to dramatically improve
the resolution of complex mixtures compared with any one-dimensional separation [14].
In the future, in-solution digestion coupled with two-dimensional chromatography may
be a practical way to identify low abundance MLK3 binding partners.

Clathrin heavy chain was identified as a potential MLK3 binding partner by mass
spectrometry and its interaction with MLK3 was confirmed by GST pulldown assay

using a fusion protein of clathrin heavy chain terminal domain with GST. In addition, the

149

 

 

carboxyl terminus of MLK3 (amino acids 592-847), which contains one potential clathrin
box motif, is capable of interacting with CHC. However, mutation of one of the five
conserved amino acids in this clathrin box motif, Asp 715 to Ala, does not abolish the
binding to GST-CHC TD, suggesting that the aspartic acid residue is not critical for the
interaction between MLK3 and CHC. The clathrin binding region within MLK3 needs to
be firrther defined. Previous studies showed that MLK2 interacts with clathrin heavy
chain through the extreme carboxyl terminal clathrin box motif within MLK2 [20].
However, interestingly, sequence alignment of MLK1 , 2, 3 and 4 indicates that the
clathrin box motif identified in MLK2 is conserved in the extreme carboxyl terminus of
MLK1 and MLK4, but not in MLK3. This suggests that MLK3 has a distinct clathrin
binding motif.

In addition to clathrin, there are large numbers of other proteins that participate in
the process of endocytosis. In this study, three other proteins that are involved in the
clathrin mediated endocytosis were identified as MLK3 binding partners. AP2-p2 and
Hsc70 were identified in the in-gel digestion. AP2-a, together with clathrin heavy chain
was discovered in the first set of in-solution digestion experiments. AP2 are
heterotetramen'c protein complexes of subunits a (100-110 kDa), [32 (105 kDa), u2 (50
kDa) and 0'2 (20 kDa). AP2 is the second most abundant protein in coated vesicles afier
clathrin, which links the clathrin shell to the membrane through interaction of its u2 and

a subunits with membrane proteins and lipids [34]. In vivo studies have shown that the

u2 subunit is phosphorylated and this phosphorylation is essential for clathrin-mediated
endocytosis [35]. Adaptor-associated kinase 1 (AAKl) has been reported to

phosphorylate u2 in vitro [36]. However, the identity of the “u2 kinase” in vivo has not

150

 

 

been defined. It is conceivable that MLK3 may phosphorylate AP2-p2. Hsc70 is involved
in the disassembly of clathrin lattices and also contributes to adaptor uncoating [3 7]. The
uncoating of clathrin coated vesicles is inhibited in vivo by overexpressing a dominant
negative mutant of Hsc70 [38]. However, since Hsc70 is a constitutive member of the
heat shock protein family, we cannot rule out the possibility that the sole role of Hsc70 is
to fold and maintain the conformation of MLK3. Further studies are necessary to
elucidate the potential biological function of MLK3 on clathrin mediated endocytosis.

In addition to Hsc70, Hsp70 and Hsp90 and its cochaperone Cdc37, were
identified as MLK3 binding partners. Heat shock proteins play critical roles in protein
folding and maturation. They also control protein homeostasis by regulating protein
stability [39]. Hsp70 has been reported to associate with the unphosphorylated variant of
PKC [311, thus prolonging the lifetime of mature PKC [40]. Whether phosphorylation of
MLK3 regulates its association with Hsp70 has not yet been determined. The interaction
of Hsp90/Cdc37 with MLK3 will be discussed extensively in the next chapter.

ANT2 was identified as a potential MLK3 binding partner. There are three ANT
isoforms, ANTI , ANT 2 and ANT3. ANTl is expressed mainly in heart and skeletal
muscle and ANT3 is ubiquitously expressed in all differentiated tissue [41]. Interestingly,
ANT 2 has relatively low expression level in most mature tissues [41, 42], with the
highest expression levels detected in myoblasts and tumor cells. ANT is synthesized in
the cytosol before export to the inner membrane of mitochondria where it catalyzes the
ADP/ATP transport. In addition, mounting evidence suggests that ANT is involved in
apoptosis, by controlling the opening of the permeability transition pore (PTPC)

(reviewed in [43]). It has been reported that some protein kinases, such as PKC6 [44],

151

. ._-

and Raf-1 [45] can translocate to mitochondria to modulate the mitochondrial mediated
apoptosis. Further experiments are needed to understand the association of ANT2 with
MLK3 and the potential role of this interaction in regulating apoptosis through the
mitochondrial membrane permeabilization.

In summary, affinity purification coupled with mass spectrometry has been used
to identify potential MLK3 binding partners. Confirmation of these interactions and
further analysis of the consequences of these interactions will be the subject of future

studies.

152

Footnotes

 

Acknowledgements:

1.

The LC/MS/MS experiments were performed in collaboration with Wei Wu in
Dr. Jack Watson’s lab.

Endogenous MLK3 expression levels (Fig. 1) were screened by Yan Du.

The MCF-7 cell line inducibly expressing MLK3 was constructed by Sarah
Santos (Dr. Conrad Lab) and Yan Du.

The GST-CHC (1-579) construct (pGEX-2T GST-clathrin 1-579) was kindly

provided by Dr. James H. Keen in Thomas Jefferson University.

153

6. References

10.

11.

12.

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

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

Nagata, K., et al., The MAP kinase kinase kinase MLK2 co—localizes with
activated JNK along microtubules and associates with kinesin superfamily motor
KIF3. Embo J, 1998. 17(1): p. 149-58.

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

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.

Westermarck, J ., et al., The DEW/H-box RNA helicase RHII/Gu is a co-factor
for c-Jun-activated transcription. Embo J, 2002. 21(3): p. 451-60.

Boudeau, J ., et al., Heat-shock protein 90 and Cdc3 7 interact with LKBI and
regulate its stability. Biochem J, 2003. 370(Pt 3): p. 849-57.

Hillenkamp, F., et al., Matrix-assisted laser desorption/ionization mass
spectrometry of biopolymers. Anal Chem, 1991. 63(24): p. 1193A-1203A.

Griffiths, W.J., et al., Electrospray and tandem mass spectrometry in
biochemistry. Biochem J, 2001. 355(Pt 3): p. 545-61.

Eng J K, M.A., Yates, JR 111, An approach to correlate tandem mass spectral data
of peptides with amino acid sequences in a protein database. Journal of the
American Society for Mass Spectrometry, 1994. 5(11): p. 976-989.

J immy Eng, J .Y., SEQUEST web homepage.
http://fields.sc1ipps.edu/seguest/indexlitml.

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

154

13.

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

25.

26.

Bock, 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.

Link, A.J., et al., Direct analysis of protein complexes using mass spectrometry.
Nat Biotechnol, 1999. 17(7): p. 676-82.

Washbum, M.P., D. Wolters, and J .R. Yates, 3rd, Large-scale analysis of the

yeast proteome by multidimensional protein identification technology. Nat
Biotechnol, 2001. 19(3): p. 242-7.

Nogales, E., Structural insight into microtubule function. Annu Rev Biophys
Biomol Struct, 2001. 30: p. 397-420.

Dell'Angelica, E.C., et al., Association of the AP-3 adaptor complex with clathrin.
Science, 1998. 280(5362): p. 431-4.

Kirchhausen, T., Clathrin. Annu Rev Biochem, 2000. 69: p. 699-727.

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

Akbarzadeh, S., et al., Mixed lineage kinase 2 interacts with clathrin and
influences clathrin- coated vesicle traflicking. J Biol Chem, 2002. 277(39): p.
36280-7.

Rasmussen, R.K., et al., Mixed-lineage kinase 2-SH3 domain binds dynamin and
greatly enhances activation of GT Pase by phospholipid. Biochem J, 1998. 335(Pt
1): p. 119-24.

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

Liu, Y.F., D. Dorow, and J. Marshall, Activation of MLK2-mediated signaling
cascades by polyglutamine- expanded huntingtin. J Biol Chem, 2000. 275(25): p.
19035-40.

Whitmarsh, 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 scafl'old
proteins. Mol Cell Biol, 1999. 19(10): p. 7245-54.

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.

155

27.

28.

29.

30.

31.

32.

33.

34.

35.

36.

37.

38.

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.

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 Centrosomes
and Influences Microtubule Organization. Mol Biol Cell, 2003. 14(1): p. 156-72.

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

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.

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.

Rigaut, G., et al., A generic protein purification method for protein complex
characterization and proteome exploration. Nat Biotechnol, 1999. 17(10): p.
1030-2.

Puig, 0., et al., The tandem affinity purification (TAP) method: a general
procedure of protein complex purification. Methods, 2001. 24(3): p. 218-29.

Takei, K. and V. Haucke, Clathrin-mediated endocytosis: membrane factors pull
the trigger. Trends Cell Biol, 2001. 11(9): p. 385-91.

Olusanya, 0., et al., Phosphorylation of threonine 156 of the mu2 subunit of the
AP2 complex is essential for endocytosis in vitro and in vivo. Curr Biol, 2001.
11(11): p. 896-900.

Conner, SD. and S.L. Schmid, Identification of an adaptor-associated kinase,
AAKI, as a regulator of clathrin-mediated endocytosis. J Cell Biol, 2002. 156(5):
p. 921-9.

Wakeham, D.E., et al., Molecular structures of proteins involved in vesicle coat
formation. Traffic, 2000. 1(5): p. 393-8.

Newmyer, S.L. and S.L. Schmid, Dominant-interfering Hsc70 mutants disrupt

multiple stages of the clathrin-coated vesicle cycle in vivo. J Cell Biol, 2001.
152(3): p. 607-20.

156

39.

40.

41.

42.

43.

45.

Nollen, B.A. and RI. Morimoto, Chaperoning signaling pathways: molecular
chaperones as stress-sensing ’heat shock' proteins. J Cell Sci, 2002. 115(Pt 14): p.
2809-16.

Gao, T. and AC. Newton, The turn motif is a phosphorylation switch that
regulates the binding of Hsp 70 to protein kinase C. J Biol Chem, 2002. 277(35):
p. 3 1 585-92.

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

Lunardi, J ., et al., The multiple ADP/A T P translocase genes are differentially
expressed during human muscle development. J Biol Chem, 1992. 267(22): p.
1 5267-70.

Belzacq, A.S., et al., The adenine nucleotide translocator in apoptosis. Biochimie,
2002. 84(2-3): p. 167-76.

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.

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

157

 

 

V. Regulation of Mixed Lineage Kinase 3 by Hsp90 and Cdc37

1. Abstract

Mixed lineage kinase 3 (MLK3) is a mitogen-activated protein kinase (MAPK)
kinase kinase (MAPKKK) that activates MAPK pathways, including the JNK and p38
pathways. In Chapter IV, affinity purification coupled with LC/MS/MS revealed that
components of Flag-MLK3 complexes include heat shock protein 90 a, B (Hsp90), and
its co-chaperone Cdc37. The association of endogenous MLK3 with Hsp90/Cdc37 was
confirmed by co-immunoprecipitation and immunoblotting. Domain mapping
experiments demonstrated that Hsp90/Cdc37 binds to the catalytic domain of MLK3. The
specific inhibitor for Hsp90, geldanamycin, has been shown to modulate the stability of
steroid hormone receptors and certain protein kinases including Raf and Src. Upon
treatment of MCF-7 cells with geldanamycin, MLK3 levels decrease dramatically.

TNFa treatment increases the catalytic activity of endogenous MLK3 and JNK.

Pretreatment of MCF-7 cells with geldanamycin abolishes TNFa-induced activation of

MLK3 and JNK. Taken together, the data presented in this chapter demonstrate a role for

molecular chaperones Hsp90/Cdc37 in MLK3 signaling.

158

 

 

2. Introduction

MLK3 is a serine/threonine protein kinase that acts as a MAPKKK to
phosphorylate and activate the dual specific kinases, MKK4 [1] and MKK7 [2] which, in
turn, can phosphorylate and activate JNK pathway. Overexpression of MLK3 also
modestly activates the p38 MAPK pathway [3, 4].

JNKs and p38 MAPKs are activated by diverse extracellular signals, including
cytokines, osmotic stress, UV and heat shock [5]. However, there are more than a dozen
MAPKKKs demonstrated to activate the JNK and p38 pathways. The stimuli that can
specifically activate endogenous MLK3 are not well characterized. Recently, it has been
shown that treatment of J urkat T lymphocytes with tumor necrosis factor-a (TNFa)
activates endogenous MLK3 as judged by the ability of MLK3 to phosphorylate
recombinant MKK4 in vitro [6].

In addition to its catalytic domain, MLK3 contains several domains, which may
mediate protein-protein interactions to regulate its activity and/or signaling. The small
GTPases, Cdc42 and Rac, in their activated states, bind to MLK3 in a Cdc42/Rac
interactive binding (CRIB) motif-dependent manner and increase MLK3 catalytic activity
[7]. Work described in Chapter H shows that MLK3 is autoinhibited through an
interaction between its amino terminal SH3 domain and a sequence located between the
zipper and CRIB motifs of MLK3 [8].

Hsp90 is a heat shock protein highly conserved in all eukaryotic cells. Hsp90
functions as a molecular chaperone to prevent the aggregation and to maintain the
homeostasis of client proteins. The client proteins of Hsp90 include steroid hormone

receptors and protein kinases (reviewed in [9, 10]). The co-chaperone Cdc37 is required

159

 

tut-Ii. .

for the interaction between Hsp90 and it client protein kinases. The identified clients of
Hsp90/Cdc37 include several protein kinases involved in proliferation, cell cycle control,
and tumorigenesis, including Src [11, 12], Raf [13, 14], and CDK4 [15].

As detailed in Chapter IV, using affinity purification coupled with mass
spectrometry, heat shock protein Hsp90 a,B and its co-chaperone Cdc37 were identified
as binding partners of MLK3 in breast cancer cell line MCF-7 cells. In this chapter, the
functional importance of the association of molecular chaperones Hsp90/Cdc37 on the

stability and signaling of MLK3 is presented and discussed.

160

 

3. Materials and Methods

3.1. Plasmid Constructs and Site Directed Mutagenesis

Construction of the cytomegalovirus-based expression vectors carrying the
cDNAs for wild type (pRK5-NFlag.mlk3) and inactive (pRK5-NFlag.mlk3 K144A) has
been described elsewhere [7]. The mammalian expression vector pGST-Cdc37 was
kindly provided by Dr. Nicholas Grammatikakis in Tufts University School of Medicine.

Construction of pCGN-HA.mlk3 115-399 was described in Chapter II. For the
construction of the additional hemagglutinin (HA)-tagged MLK3 variants, the following
oligonucleotides were used in the polymerase chain reactions with pRK5-NFlag.mlk3 as
the template;

pCGN-HA.mlk3 1-399, 5'-CGTTAGTCTAGAATGGAGCCCTTGAAGAG-3'
and 5'- GCATTAGGATCCTCACCAGCCTTCCTGCATGG-3';

pCGN-HA.mlk3 400-847, 5'-CGTTAGTCTAGAAAGCGCGAGATCCAGGG-3'
and 5'-GCATTAGGATCCTCAAGGCCCCGCTTCCGGC-3'.

The amplified mlk3 fragments were subcloned in-frame with the HA coding

sequence into the pCGN mammalian expression vector using XbaI and BamHI.

3.2. Cell lines and Transfections

MCF-7 cells were maintained in high glucose Dulbecco's modified Eagle's media
(1:1) (Invitrogen) supplemented with 8% fetal bovine serum (Invitrogen), 2 mM
glutamine, and penicillin/streptomycin (Invitrogen). MCF-7 cells were plated at a density

of 1 X 106 cells/ ml and after one day, were pretreated for 24 h with 5 uM of

geldanamycin or with DMSO. Cells were then treated with TNFa for the indicated

161

 

 

periods of time, harvested, and lysed as described previously[8]. A portion of the pelleted
cells was boiled with SDS-loading buffer and used for immunoblotting of the phospho-c
Jun level.

Human fetal kidney 293 cells were maintained in Ham's F-12/low glucose
Dulbecco's modified Eagle's media (1:1) (Invitrogen) supplemented with 8% fetal bovine
serum (Invitrogen), 2 mM glutamine, and penicillin/streptomycin (Invitrogen). HEK
293 cells were transfected using the calcium phosphate method as previously described

[7]. Cells were harvested 16 h afier transfection, and lysed as described previously[8].

3.3. GST Pulldown Assays

Mammalian constructs (pGST or pGST-cdc3 7) encoding GST or GST-Cdc37
respectively were cotransfected with expression vectors encoding MLK3 truncation
variants in HEK 293 cells. Cells were lysed 16 h after transfection and the clarified lysate
(400 [41) was incubated with 20 ul of glutathione-Sepharose 4B resin (Amersham
Phamacia Biotech) for 90 min at 4 °C. The pelleted resin was washed twice with HNTG
buffer (20 mM HEPES (pH 7.5), 150 mM NaCl, 0.1% Triton X-100, and 10% glycerol),

and bound proteins were resolved by SDS-PAGE.

3.4. Immunoprecipitations

Antibodies against the proteins of interest were prebound to protein A-agarose
beads for 30 min at room temperature: MLK3 antibody (0.25 ug/ul slurry), GST antibody
(0.25 pig/pl slurry). Clarified lysates (400 '41- 600 pl) were incubated with 20 p1 of
antibody-bound Protein A-agarose for 90 min at 4 °C. For detection of interactions

between endogenous proteins, 2 ml of clarified lysates from MCF-7 cells were incubated

162

 

 

with 50 p1 of antibody-bound Protein A-agarose. Immunoprecipitates were washed with
HNTG buffer. Immunoprecipitates used for kinase assays and endogenous associations
were washed twice with lysis buffer, and twice with kinase assay buffer (50 mM Tris-

HCl (pH 7.5), 100 mM NaCl, 1 mM MnClz, 10 mM MgC12, 0.1 mM Na3VO4).

3.5. In vitro Kinase Assays

Immunoprecipitated MLK3 was incubated in 20 pl of kinase assay buffer
containing 10 nM ATP and 5 nCi (y-32P)-ATP (3000 Ci/mmol) (NEN Life Science
Products), 6 pg of recombinant GST-MKK7 K165A for 30 min at room temperature.
Following the kinase assay, proteins were separated by SDS-PAGE, and transferred to a
nitrocellulose membrane, and the incorporation of radioactivity into the GST-MKK7 was
determined by Phosphorlrnaging (Molecular Dynamics). The membrane was Western
blotted with MLK3 antibody and GST antibody sequentially to detect the MLK3 and

GST-MKK7 level.

3.6. SDS-PAGE and Western Blot Analysis

Lysates and GST-pulldowns of proteins were resolved by SDS-PAGE. Proteins
were transferred to nitrocellulose membranes and immunoblotted using anti-MLK3 rabbit
polyclonal antibody (4ug/ml) [7], anti-actin mouse monoclonal antibody (Sigma), HA
mouse monoclonal antibody (5 ug/ml)(BAbCO), phospho-c-Jun (KM-1) mouse
monoclonal antibody (0.2 [lg/ml), MKK4 rabbit polyclonal antibody (0.2 jig/ml), MKK7
goat polyclonal antibody (0.4 pig/ml), JNK rabbit polyclonal antibody (0.2 jig/ml), Hsp90
rabbit polyclonal antibody (0.4 pig/ml), Cdc37 mouse monoclonal antibody (0.8 rig/ml)

GST mouse monoclonal antibody (0.2 jig/ml) (all of the above antibodies were purchased

163

 

 

 

from Santa Cruz, unless indicated), followed by the appropriate horseradish peroxidase-
conjugated secondary antibody (Bio-rad). Western blots were developed by

chemiluminescence.

164

 

4. Results

4.1. Demonstration of the interaction between endogenous Hsp90/Cdc37 and Flag-
MLK3.

In Chapter IV of this thesis, affinity purification of Flag-MLK3 complexes from
MCF-7 cells coupled with mass spectrometry identified Hsp90/Cdc37 as potential MLK3
binding partners. To verify the interactions between Hsp90/Cdc37 and MLK3, Flag-
MLK3 complexes were isolated from MCF-7 cells inducibly expressing Flag-MLK3 and
subjected to Western blot analysis using antibodies directly against Hsp90 and Cdc37. As
shown in Fig. 1, Hsp90 and Cdc37 were coimmunoprecipitated when MLK3 expression
was induced. Hsp90 and Cch7 could not be detected in the un-induced control,

confirming that the interaction is specific for Flag-MLK3.

4.2. Endogenous MLK3 is associated with Hsp90/Cdc37.

To exclude the possibility that the association of MLK3 with Hsp90/Cdc37 is due
to an artifact of Flag-MLK3 overexpression in these cells, endogenous MLK3 was
immunoprecipitated from MCF-7 cell lysates. As shown in Fig. 2, endogenous
Hsp90/Cdc37 coimmunoprecipitated with endogenous MLK3 as judged by
immunoblotting. We also observed that only a very small proportion of the total cellular
pool of Hsp90/Cdc37 is associated with MLK3 (data not shown). This observation is not
surprising as the Hsp90 and Cdc37 are relatively abundant proteins that associate with

many other cellular protein kinases.

4.3. Hsp90/Cdc37 associates with the kinase domain of MLK3.

165

 

 

 

Previous studies have revealed that Hsp90/Cdc37 binds to client protein kinases
IKB kinase (IKK)[16] and Raf [17] through their catalytic domains. To map the region in
MLK3 that interacts with Hsp90/Cdc37, two NHz-terminal HA-tagged MLK3 truncation
variants were constructed. MLK3 1-399 contains the glycine rich region, SH3 domain
and the kinase domain, whereas MLK3 400-847 includes the zipper region, CRIB motif
and the proline rich carboxyl terminus. These two variants, together with a previously
constructed variant MLK3 115-399 (which contains the kinase domain only) were used to
map the region of MLK3 that interacts with Hsp90/Cdc37. MLK3 variants were either
expressed together with GST-tagged Cdc37 (GST-Cdc37) or GST, and GST pulldown
experiments were performed to pull down GST-Cdc37, endogenous Hsp90 and
associated MLK3 variants. All variants expressed at similar levels as shown in Fig. 3B.
Data from GST pulldown assays show that, of these variants, MLK3 1-399 and MLK3
115-399 are able to interact with GST-Cdc37 and endogenous Hsp90 (Fig. 3A). As
expected, none of the variants associates with GST. Based on these experiments, MLK3
115-399 is sufficient to interact with GST-Cdc37. Therefore, we conclude that MLK3

interacts with Hsp90/Cdc37 through its kinase domain.

4.4. MLK3 is associated with Hsp90/Cdc37 in an MLK3 activity independent
manner.

We were interested to test whether the activation state of MLK3 affects its
interaction with Hsp90/Cdc37. We selected three forms of MLK3 that reflect different
activation states. MLK3 K144A is catalytically inactive. Previous work from our lab has

shown that, when wildtype MLK3 is overexpressed alone, it has basal catalytic activity

166

 

 

 

[18]; when it is coexpressed with activated Cdc42, the activity of MLK3 increases about
4 folds [7]. After transient expression of MLK3 variants with or without activated Cdc42,
MLK3 variants were immunoprecipitated and the presence of endogenous Hsp90/Cdc37
was assessed by Western blotting using appropriate antibodies. As shown in Fig. 4, equal
amount of Hsp90/Cd037 immunoprecipitates with the MLK3 variants under these
conditions. These data are consistent with the idea that MLK3 activity does not influence

its association with Hsp90/Cdc37.

4.5. Hsp90/Cdc37 regulates the protein level of MLK3 in MCF—7 cells.

The ansamycin antibiotics, such as geldanamycin (GA), inhibit Hsp90 chaperone
activity by occupying its ATP/ADP pocket [19, 20]. Upon GA treatment, the protein
level of several protein kinases, such as Raf [l4] and Akt [21], decreases dramatically.
However, GA does not affect the protein level of IKK although IKK proteins have been
shown to interact with Hsp90/Cdc37 [16]. We tested whether GA alters the expression
level of endogenous MLK3. MCF-7 cells were treated with increasing concentration of
GA for 24 h, and levels of endogenous MLK3 were assessed by Western blotting. We
observed a loss in the cellular levels of MLK3. GA (1 M) is sufficient to trigger the loss
of MLK3 in MCF-7 cells (Fig. 5A). In addition, GA treatment causes a decline in the
level of MLK3 in a time dependent manner (Fig. 5B). On average, levels of MLK3 were
reduced by more than 80% at 24 h. These results indicate that Hsp90/Cdc37 regulate the

protein levels of MLK3 in MCF-7 cells.

 

167

4.6. Hsp90/Cdc37 regulate TNFa-induced activation and signaling of MLK3 in
MCF-7 cells.

TNFa treatment has been shown to increase MLK3 catalytic activity in J urkat T
cells [6]. We examined whether TNFor increases MLK3 activity in MCF-7 cells. After 30
min treatment with TNFoc, endogenous MLK3 was immunoprecipitated and its activity
was assessed by an in vitro kinase assay using recombinant GST-MKK7 as a substrate.
The basal activity of endogenous MLK3 is low. Upon TNFa treatment, the catalytic
activity of MLK3 increases about 2.5 fold (Fig. 6A). Consistent with the activation of
endogenous MLK3, JNK activity in cells, as measured by the phosphorylation of
transcription factor c-Jun in the cell lysates, was greatly enhanced with parallel kinetics
(Fig. 6C). This suggests that TNFoc-induced JNK activation is through the activation of
MLK3. We then examined the effect of GA on TNFOL signaling to MLK3 and JNK.
MCF-7 cells were pretreated with or without GA for 24 h, and cells were treated with
TNFa for 30 min. As shown in Fig. 6, GA pretreatment completely abolished the TNFa
induced activation of MLK3 although there is about 20% of residual MLK3 protein in the
cellular lysates. In addition, pretreatment with GA not only lowered the basal level of c-
Jun phosphorylation in MCF-7 cells, but also abolished the TNFa-induced JNK
activation. Since the TNFOL induced JNK activation requires the three modules of the
MAPK cascades, we also examined the protein level of MKK4, MKK7, and JNK upon
GA treatment. The levels of MKK4, MKK7 and JNK were not affected by GA. Taken
together, these results support the idea that Hsp90/Cdc37 regulate TNFa induced JNK

activation through controlling the protein level of MAPKKKS, including MLK3.

168

AP21967 - _+_

Hsp90 we ‘

‘— IgG
<— Cdc37

Cdc37 WB

 

Flag MLK3 we

 

Fig.1. Association of Flag-MLK3 with Hsp90/Cdc37. Affinity purified Flag-MLK3
complexes from MCF7 cells inducibly expressing Flag-MLK3 were separated by SDS-
PAGE. The presence of endogenous Hsp90, Cdc37 and Flag-MLK3 was assessed by
immunoblotting with Hsp90, Cdc3 7, and Flag antibody respectively. In addition to
Cdc3 7, the presence of the anti-F lag M2 antibody (IgG) heavy chain is indicated by

arrow in the middle panel.

169

 

 

Cdc37 WB

f Hsp90 we

MLK3 WB

 

d1 10111103

Z
r
E
E

Fig. 2. Endogenous MLK3 is associated with Hsp90 and Cdc37. MLK3 was
immunoprecipitated (IP) from MCF-7 cell lysates using an MLK3 antibody prebound to
protein A agarose beads. A control immunoprecipitation was performed in parallel using

a GST antibody. The presence of Cdc37, Hsp90, and MLK3 was assessed by Western

blotting using the indicated antibodies.

170

 

1-399 115-399 400-847
MLK3 { h f \f \

HA MLK3
.-
GST
pulldown
- - - Hsp90
“ "‘ . GST-Cdc37
“Q I- III : Q GST

1-399 115-399 -
MLK3 I \f V400 8471

Input HA MLK3

cell lysates

 

rafi

Fig. 3. Hsp90 and Cdc37 associate with the catalytic domain of MLK3. HEK 293

cells were transfected with expression vectors containing the cDNA for MLK3 truncation
variants and GST—Cdc37 (or GST). The numbers in the figure represent amino acid
numbers in MLK3. Cellular lysates expressing the indicated MLK3 truncation variants
and GST-Cdc37 (or GST) were incubated with glutathione-Sepharose 4B resin. A, top
two panels, the presence or absence of bound MLK3 variants and Hsp90 was assessed by
Western blotting with the HA antibody and Hsp90 antibody respectively. Bottom two
panels, equal binding of GST-Cdc37 or GST on the glutathione-Sepharose 4B resin was
confirmed by Western blotting with GST antibody. B, the expression of MLK3 variants
in cellular lysates probed with the HA antibody. The data shown are representative of

three independent experiments.

Cdc42V12 - + - -

 

 

MLK3 wt wt K144A -
(.- fl - CdC37WB
MLK31P "'" 2 a Hsp90 we
L ! ~~ MLK3 we

F ig. 4. Association of MLK3 variants with Hsp90/Cdc37. HEK 293 cells were
transfected with expression vectors containing the cDNA for MLK3 wildtype (wt),

K144A in the presence or absence of Cdc42VIZ

. MLK3 was immunoprecipitated from
cellular lysates using MLK3 antibody preincubated with protein A agarose beads. The
presence or absence of Cdc3 7, Hsp90 and MLK3 was assessed by Western blotting using

the indicated antibodies.

172

 

 

GA(uM) o 0.01 0.1 1 5 10

‘ T

529-5. 2a..-- ...‘ci MLK3

- --"- actin

Time(h) 0 2 4 8 12 24
u..."”' “"' “ " MLK3

. --~ “ ‘ actin

Fig. 5. Geldanamycin (GA) decreases the endogenous level of MLK3. MCF-7 cells
were treated with various concentrations of GA for 24 h (A) or with 5 11M of GA for the
indicated times (B). The endogenous level of MLK3 was assessed by Western blotting of
cellular lysates using the MLK3 antibody. Western blotting for endogenous actin was

used to confirm that equal cell equivalents were loaded.

173

 

TNFa 0 30 0 30 (min)
GA + + (24h)

5... fi. g .N . _. ®-MKK7
a. - , - GST-MKK7

{-0 E: MLK3 IP

”“4

 

 

 

 

 

 

 

 

.E
33693 1 1
c x
5 = 1 ~
.2 o -
Time 0 30 (min)

174

TNFa 0 30 0 30 (min)
GA - - + 4» (24h)
.- -. fl ‘ MLK3
.‘_ __ ____ «1...... MKK4
.. .. - E :1 MKK7

Fig. 6. GA abolishes TNFa-mediated activation of MLK3 and JNK. MCF-7 cells
were pretreated with GA or the vehicle DMSO for 24 h, then treated with TNFa for the
indicated time, and lysed. A, in vitro kinase assay of MLK3 using GST-MKK7 K165A as
a substrate. MLK3 was isolated from cellular lysates by immunoprecipitation using an
MLK3 antibody, and kinase activity was assessed in vitro using recombinant GST-
MKK7 K165A as a substrate. Top panel shows a phosphorimage with bands
corresponding to MKK7 phosphorylation. A Western blot of the immunoprecipitated
MLK3 is shown in the bottom panel. The equal loading of GST-MKK7 substrate in the
kinase assay was confirmed by immunoblotting using an antibody against GST as shown
in the middle panel. B, the mean 1 SE. for fold increase in GST-MKK7 K165A
phosphorylation from three independent experiments is shown. C, equal amounts of
protein from cellular lysates were resolved by SDS-PAGE and Western blotted using the

indicated antibodies.

175

 

5. Discussion

Using affinity purification coupled with mass spectrometry, we detected MLK3 in
association with Hsp90/Cd037 and as well other proteins as described in Chapter IV.
Herein, we confirm that endogenous MLK3 is associated with Hsp90/Cdc37 in MCF-7
cells. Our finding that Hsp90/Cdc37 interact with MLK3 suggests that Hsp90/Cdc37 play
an important role for the physiological function of MLK3.

Several protein kinases have been reported to associate with Hsp90 and Cdc37. In
the few cases where domain interactions have been mapped, Hsp90/Cdc37 associates
with the catalytic domains of its client protein kinases. In accord with these studies, we
also find that the catalytic domain of MLK3 is sufficient for association with recombinant
Cdc37. In addition to MLK3, other MLKs exist [22]. Given their highly homologous
catalytic domains, it seems likely that Hsp90/Cdc37 may associate with and regulate the
other MLKs as well. The lack of commercially available antibodies that recognize the
other MLKs makes it difficult to test this prediction.

Despite the consistent findings that the catalytic domain of protein kinases
mediates binding to Hsp90/Cdc3 7, whether Hsp90 or Cdc3 7, or both, directly interact
with client protein kinases is less clear. Coomassie stained gels of Flag-MLK3 complexes
suggest that Hsp90 and Cdc37 associate with MLK3 at equal molar ratios. However, our
experiments were not able to assess whether both Hsp90 and Cdc37 directly associate
with MLK3. According to Grammatikakis et al., both Cdc37 and Hsp90 interact directly
with recombinant protein kinase Raf-1 in vitro and the association between Hsp90 and
Raf-1 was enhanced in the presence of Cdc37 [23]. Similar results were obtained from
Goeddel’s lab when studying Hsp90/Cdc37 association with IKK. Recombinant Cdc37

binds to IKK weakly and Hsp90 itself binds poorly to IKK. However, Cdc37 interacts

176

more strongly with IKK in the presence of recombinant Hsp90 [16]. It is likely that
Hsp9/Cdc37 bind in a cooperative fashion to MLK3, but this issue deserves further study.

Our data suggest that MLK3 associates with Hsp90/Cdc37 in an activity
independent manner, as both a kinase defective variant of MLK3, and a Cdc42 activated
form of MLK3, associate with Hsp90/Cdc37 equally well. The MAPKKK Raf-l
constitutively associates with Hsp90/Cdc37 as well. The Raf-Hsp90-Cdc37 complex was
not only observed in serum deprived cells, but also upon serum or phorbol ester
stimulation conditions, whereas the small GTPase Ras recruits the Raf-Hsp90-Cdc37
complex to the plasma membrane [24]. These findings suggest that Hsp90/Cdc37
interaction with Raf is also independent of kinase activity. In contrast, Hsp90/Cdc37 only
associates with tyrosine kinase v-Src during its biosynthesis. Once v-Src is inserted into
the plasma membrane by virtue of myristoylation, it no longer associated with Hsp90
[11].

An important role of Hsp90 is to stabilize its client proteins and maintain their
correctly folded states, thus preventing their degradation through the proteasome. The
stability of many, but not all, client protein kinases, such as Src and Raf-l, are regulated
by Hsp90. Treatment of cells with geldanamycin, a specific inhibitor of Hsp90, decreases
the protein level of MLK3. In the previous chapter, we have shown that MLK3 is present
at high levels in MCF-7 cells. Interestingly, many cancer cell lines, including MCF-7,
have high levels of Hsp90 [9, 25]. Conceivably, it is these high levels of Hsp90 that are
responsible for the high levels of MLK3 in breast cancer cells. Disruption of the function
of Hsp90 does not affect the level of many protein kinases, such as MKK4, MKK7 and

JNK studied here. It is likely that Hsp90 does not associate with these kinases. However,

177

 

 

IKK has been detected to form a complex with Hsp90/Cdc3 7, yet geldanamycin does not
affect the protein level of IKK [ 1 6].

The extracellular stimuli that activate MLK3 have not been well defined.
Recently, it has been reported that TNFOL activates endogenous MLK3, leading to JNK
activation in J urkat T cells. In the study presented here, we have shown that TNFa
activates endogenous MLK3 in MCF-7 cells. In addition, the phosphorylation of c-Jun is
greatly enhanced in response to TNFO. with similar kinetics. Pretreatment of the MCF-7
cells with GA abolishes TNFa-induced MLK3 activation, in parallel with the loss of
MLK3 in cells. This data indicates that geldanamycin inhibits MLK3 activity primarily
by decreasing MLK3 stability. In addition to MLK3, another MAPKKK, apoptosis
signal-regulating kinase 1 (ASKl), has been shown to mediate TNFa induced JNK
activation [26]. Therefore, we could not rule out the possibility that GA may inhibit the
protein level of ASK1 , and therefore indirectly its activity as well. Our data indicate that
Hsp90/Cdc37 regulates TNFa-induced JNK signaling at the MAPKKK level, but not at
the MAPKK (MKK4/MKK7) or MAPK (JNK) level. Our findings share some interesting
parallels with the Hsp90/Cdc37 regulation of the ERK pathway. Hsp90/Cdc37 controls
the stability of Raf-l, but not MEK] , MEK2 or ERK1/2. Destabilization of Raf-l by
geldanamycin abolishes PMA induced ERK activation [14]. Taken together, it is

consistent with the idea that Hsp90/Cdc37 regulates MAPK pathways at the MAPKKK
level.

Most of the previously identified Hsp90/Cdc37 clients are involved in cell
proliferation, cell cycle control and tumorigenesis. Therefore, Hsp90 has been targeted

for cancer therapy. The geldanamycin analogue, l7-AAG is currently in phase I clinical

178

 

d_‘_ '

trials for many types of cancer. Although the biological function of MLK3 in MCF-7
cells is not defined so far, MLK3 has been implicated in neuronal apoptosis and activates
JNK pathway, which mediates apoptosis in many cell types. Our findings indicate that
the Hsp90/Cdc37 not only regulates the protein kinases that promote proliferation and/or
tumori genesis, but it appears that anti-proliferative /proapoptotic protein kinases are also
clients for Hsp90/Cd037. This is supported by the recent report that LKBl, a
serine/threonine protein kinase that act as tumor suppressor, associates with
Hsp90/Cdc37 [27]. This finding, taken together with the data presented in this chapter,
indicate that the use of Hsp90 inhibitor in cancer therapy should be approached with

caution.

 

179

Footnotes

 

Acknowledgements:

1.

We are grateful to Dr. Nicholas Grammatikakis (Tufts University School of
Medicine) for providing the mammalian GST-Cdc37 and GST vectors

The bacteria expression vector of GST-MKK7 was kindly provided by Dr. Roger
J. Davis at University of Massachusetts Medical School. Geou-yarh (Stancy) Liou
engineered the GST-MKK7 K165A mutant and purified the corresponding

recombinant protein.

180

6. References:

10.

11.

12.

Rana, A., et al., The mixed lineage kinase SPRKphosphorylates 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 MKK 7 and not MKK4
as substrate. J Biol Chem, 1999. 274(15): p. 10195-202.

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.

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

Johnson, G.L. and R. Lapadat, Mitogen-activated protein kinase pathways
mediated by ERK, JNK, and p38 protein kinases. Science, 2002. 298(5600): p.
1911-2.

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.

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.

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

Nollen, B.A. and RI Morimoto, Chaperoning signaling pathways: molecular
chaperones as stress-sensing ’heat shock' proteins. J Cell Sci, 2002. 115(Pt 14): p.
2809-16.

Jolly, C. and RI. Morimoto, Role of the heat shock response and molecular
chaperones in oncogenesis and cell death. J Natl Cancer Inst, 2000. 92(19): p.
1564-72.

Xu, Y. and S. Lindquist, Heat-shock protein hsp90 governs the activity of pp60v-
src kinase. Proc Natl Acad Sci U S A, 1993. 90(15): p. 7074-8.

Whitesell, L., et al., Inhibition of heat shock protein HSP90-pp60v-src
heteroprotein complex formation by benzoquinone ansamycins: essential role for

181

13.

14.

15.

16.

17.

18.

19.

20.

21.

22.

23.

24.

stress proteins in oncogenic transformation. Proc Natl Acad Sci U S A, 1994.
91(18): p. 8324-8.

Stancato, L.F., et al., Raf exists in a native heterocomplex with hsp90 and p50 that
can be reconstituted in a cell-free system. J Biol Chem, 1993. 268(29): p. 21711-
6.

Schulte, T.W., et al., Disruption of the Raf-1-Hsp90 molecular complex results in
destabilization of Raf-1 and loss of Raf-1 -Ras association. J Biol Chem, 1995.
270(41): p. 24585-8.

Dai, K., R. Kobayashi, and D. Beach, Physical interaction of mammalian CDC3 7
with CDK4. J Biol Chem, 1996. 271(36): p. 22030-4.

Chen, G., P. Cao, and D.V. Goeddel, TNF-induced recruitment and activation of
the IKK complex require Cdc3 7 and Hsp90. Mol Cell, 2002. 9(2): p. 401-10.

Silverstein, A.M., et al., p5 0(cdc3 7) binds directly to the catalytic domain of Raf
as well as to a site on hsp90 that is topologically adjacent to the tetratricopeptide
repeat binding site. J Biol Chem, 1998. 273(32): p. 20090-5.

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

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

Stebbins, C.E., et al., Crystal structure of an Hsp90-geldanamycin complex:
targeting of a protein chaperone by an antitumor agent. Cell, 1997. 89(2): p. 239-
50.

Roe, S.M., et al., Structural basis for inhibition of the Hsp90 molecular

chaperone by the antitumor antibiotics radicicol and geldanamycin. J Med Chem,
1999. 42(2): p. 260-6.

Basso, A.D., et al., Akt forms an intracellular complex with heat shock protein 90
(Hsp90) and Cdc3 7 and is destabilized by inhibitors of Hsp90 function. J Biol
Chem, 2002. 277(42): p. 39858-66.

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.

Grammatikakis, N., et al., p5 0(cdc3 7) acting in concert with Hsp90 is required
for Raf-1 function. Mol Cell Biol, 1999. 19(3): p. 1661-72.

Wartmann, M. and R.J. Davis, The native structure of the activated Raf protein

kinase is a membrane-bound multi-subunit complex. J Biol Chem, 1994. 269(9):
p. 6695-701.

182

25.

26.

27.

Vercoutter-Edouart, A.S., et al., Proteomic detection of changes in protein
synthesis induced by fibroblast growth factor-2 in MCF-7 human breast cancer
cells. Exp Cell Res, 2001. 262(1): p. 59-68.

Ichijo, H., et al., Induction of apoptosis by ASK 1 , a mammalian MAPKKK that
activates SAPK/JNK and p38 signaling pathways. Science, 1997. 275(5296): p.
90-4.

Boudeau, J ., et al., Heat-shock protein 90 and Cdc3 7 interact with LKBI and
regulate its stability. Biochem J, 2003. 370(Pt 3): p. 849-57.

183

VI. Concluding Remarks

This thesis describes work aimed at the elucidation of the molecular mechanisms
that regulate MLK3. In particular, the regulatory role of SH3 domain of MLK3 was
examined, and a proteolytic event of MLK3 was characterized. In addition, twenty-three
potential binding partners of MLK3 were identified by affinity purification coupled with
mass spectrometry.

MLK3 contains an amino terminal SH3 domain and a large proline-rich carboxyl-
terminal region. The finding that mutation of the SH3 domain increases MLK3 activity
led to the original hypothesis that the SH3 domain of MLK3 might bind to this proline
rich region in an autoregulatory fashion. However, unexpectedly, the SH3 domain does
not bind to the extreme carboxyl terminal region, but rather binds to a region between the
zipper and CRIB motifs that lacks a classical SH3 binding site. Point mutations, either in
the SH3 domain or in the newly identified SH3 binding site of MLK3, which disrupt this
interaction, result in increased MLK3 activity. There are some important implications
regarding this work. This is the first finding of a SH3 domain mediated autoinhibition in
a serine/threonine kinase. Second, the proline residue in the SH3 binding site is
conserved in other MLK family members including MLK1,2,4 and the Drosophila
homolog Slipper. Thus this mechanism of SH3 mediated autoinhibition of MLK3 is
predicted to apply to the regulation of other MLKs as well. Third, there is no classical
proline rich sequence in the binding region. One might imagine that this sequence would
provide a suboptimal ligand for MLK3’s SH3 domain so this autoinhibition could be

released by the presentation of a high affinity ligand in response to an upstream signal.

184

 

This discovery also refined our working hypothesis of how small GTPases
Cdc42/Rac regulate MLK3 activity. The close proximity of the SH3 binding site and the
CRIB motif imply that the Cdc42 mediated activation of MLK3 should disrupt the
autoinhibition. Together with the published data, it seems that Cdc42/Rae activation of
MLK3 is a multi-step process which includes release of the SH3 mediated autoinhibition,
promotion of zipper-mediated oligomerization, induction of hyperphosphorylation on
MLK3 and presumably translocation of MLK3 to plasma membrane. The precise
mechanism is under investigation by Karen Schatcher in our lab using fluorescence
resonance energy transfer (FRET) techniques.

Chapter III described the characterization of a proteolytic event of MLK3. This
proteolytic event holds several unique features. First, MLK3 undergoes proteolysis to
generate a stable carboxyl terminal fragment (CTF) and the cleavage site has been
mapped to Gln 251-Pro 252 within the kinase domain. Thus, this proteolysis would
render MLK3 inactive. Second, evidence is presented that suggests the proteolysis
occurs shortly after MLK3 is synthesized. Third, MLK3 proteolysis appears to occur by a
feedback mechanism that requires the activation of ERK pathway. The CTF seems
relatively stable, suggesting that it might have some as yet unclarified biological function.
The protease responsible for generation of the CTF of MLK3 has not yet been identified.
MG 132 inhibition of the proteolysis has prompted us to consider fishing out the protease
by affinity purification. Stancy Liou in our lab is currently working on the project to
identify the MLK3 protease. We were unable to observe the generation of the CTF
consistently and reliable in endogenous MLK3, which may be due to the low level of

endogenous MLK3. It is also possible that we have not been able to find the appropriate

185

 

upstream signal to trigger endogenous MLK3 proteolysis. In order to show the
physiological significance of the MLK3 proteolysis, it will be important to establish
conditions under which the endogenous event occurs.

Using affinity purification coupled with mass spectrometry, more than twenty
potential MLK3 interacting proteins were identified. This certainly provides new
directions for the study of the regulation as well as the biological functions of MLK3.
The identification of clathrin and associated adaptor proteins in MLK3 complexes
suggests that MLK3 may be involved in clathrin-mediated endocytosis. This is an area
that requires further study. In addition, it seems that the stability and activity of MLK3
may be regulated by multiple molecular chaperones including Hsp70, Hsc 70 and
Hsp90/Cd037. The function of Hsp90/Cdc37 on MLK3 signaling has been examined in
detail. It may be necessary to characterize the role of Hsp70 and Hsc70 on MLK3
signaling in the future. Finally, many of MLK3 binding partners have not been
reproducibly identified. Verification of these interactions by coimmunoprecipitation and
immunoblotting using specific antibodies is required.

In conclusion, the research findings presented in this thesis constitute important
contributions to the understanding of the molecular mechanisms regulating MLK3 and of
the biological function of MLK3. In addition, this work aids in the overall understanding

of the regulation of the protein kinases.

186

     

i11111111111111]111]i11111111