INVESTIGATION OF THE FUNCTION OF MYOTUBULARIN THROUGH EXAMINATION OF PROTEIN-PROTEIN INTERACTIONS AND EXCLUSION OF MTMR1 AS A FREQUENT CAUSE OF X-LINKED MYOTUBULAR MYOPATHY
DISSERTATION
Presented in Partial Fulfillment of the Requirements for
the Degree Doctor of Philosophy in the Graduate
School of The Ohio State University
By
LaRae Meschelle Copley, B.S.
*****
The Ohio State University 2004
Dissertation Committee: Approved by Gail Herman, MD/PhD, Advisor
Arthur Burghes, PhD ______Charis Eng, MD/PhD Advisor Thomas Sferra, MD Molecular, Cellular & Developmental Biology
ABSTRACT
X-linked myotubular myopathy (MTM1, MIM# 310400) is a rare neuromuscular disorder presenting at birth with hypotonia, respiratory distress, and characteristic facies. Many patients do not survive past infancy. However, long-term survivors of MTM1 are known and have presented with additional medical problems later in life such as advanced bone age, mild spherocytosis, and peliosis hepatis.
In 1996, an MTM1 gene encoding a predicted protein called myotubularin was isolated.
Based on the structure of the gene, it was postulated that myotubularin would be a dual specificity phosphatase. Recently, researchers have discovered that myotubularin can act as a lipid phosphatase by dephosphorylating phosphatidylinositol-3-phosphate [PI(3)P], an intracellular membrane trafficking signal. Additionally, myotubularin is predicted to contain GRAM, PDZ and coiled coil domains which may be relevant for protein-protein interactions.
To further elucidate the function of myotubularin in vivo, we employed a yeast two- hybrid system to screen for proteins with which it interacts. We generated two binding domain fusion constructs, both containing the phosphatase domain of myotubularin. One
ii of the constructs harbors a C375S mutation at the active site of the enzyme and is known to abolish catalytic activity. Candidate interacting proteins were identified using a human fetal brain library (Clontech, Palo Alto, California). Our results yielded several clones, including the catalytic subunit of protein phosphatase 2A (PP2Ac), SCG10, and four novel ESTs. PP2A, a serine/threonine phosphatase, is involved in a variety of functions within the cell. SCG10 is a stathmin like protein implicated in microtubule depolymerization. Binding of the MTM1 protein with these candidates was determined to be specific in the yeast system, but confirmation of specific interactions in mammalian cells using coimmunoprecipitation and GST pulldown strategies was unsuccessful.
A second goal of this work was to develop antibodies to myotubularin. Regions of antigenicity were predicted in the myotubularin primary structure and corresponding peptides were generated and injected into rabbits. Upon screening and purifying sera, an antibody that specifically detects human myotubularin but not mouse myotubularin was generated as determined by immunoblotting.
To date, mutations in more than three hundred MTM1 families worldwide have been characterized, including many from our laboratory. However, no mutation within this gene has been found in approximately 20% of “MTM1” patients, including several males with an X-linked pattern of inheritance. MTMR1, a gene highly homologous to MTM1, is located 50KB telomeric to MTM1 on the X chromosome. We have screened 16 exons of MTMR1 in genomic DNA from 14 males with biopsy proven myotubular myopathy in whom we found no mutation in MTM1. This includes two males with X-linked
iii pedigrees and two with affected male siblings. No mutations were found in these patients.
These results suggest that mutations in MTMR1 are not a frequent cause of X-linked myotubular myopathy in males for whom no mutation is found in MTM1.
Finally, we evaluated changes in cell shape in the context of myotubularin overexpression. Recently, researchers discovered that myotubularin localizes to the plasma membrane and Rac1 induced membrane ruffles. We developed HeLa cell lines that stably express Flag-tagged myotubularin in a doxycycline inducible fashion using the tet-on system (Clontech). These cell lines, with variable expression levels of Flag-MTM1 protein, were used to quantitate cell spreading and membrane ruffling. We found a statistically significant decrease in spreading over 6 hours in cells with a high level of myotubularin overexpression compared to untreated cells. This effect was lessened as the level of Flag-MTM1 protein levels decreased. Cell lines with little to no inducible expression of Flag-MTM1 protein exhibited spreading levels similar to untreated cells.
This system was also used to quantitate membrane ruffling. A decrease in membrane ruffling was observed in cells expressing high levels of the myotubularin, but statistical significance could not be established with two independent observers.
iv
For Christopher
v
ACKNOWLEDGMENTS
I would like to thank my advisor, Dr. Gail Herman, for the opportunity to work in her lab.
Without her, none of this would have been possible. I have learned much from her. I
would also like to acknowledge my committee for their continuous assistance and encouragement. I especially thank Dr. Charis Eng for acting as my "Medical Scientist
Mentor" and for many discussions regarding my work, career, and life as a graduate
student. I also thank Dr. Arthur Burghes for being my "Basic Scientist Mentor" and
providing me with multiple suggestions for my project. I am also grateful for the
straightforwardness and enthusiasm for this project displayed by Dr. Thomas Sferra. I
extend my special thanks to MCDB’s Program Director, Dr. David Bisaro, for always
taking the time to discuss questions and concerns along the way.
I must also thank my lab mates, past and present, for making each day in the lab more
interesting. Without their support and encouragement, this entire process would have
been much less enjoyable. I am honored to have worked alongside and learned
enumerable lessons in science and politics from my friend, Dr. David Cunningham. I am
lucky to be leaving the lab having made lifelong friends with Marsha Lucas, Tessa Carrel
and Marlene Parker. Their abilities to listen, advise, and reassure amaze me. I thank Dr.
Hugo Caldas and Tara Grove for their support and helpful technical discussions. I am
vi grateful for the friendship and humor of Leon Humphries and Allison Parent. I am also
thankful to Fenglei Jiang for the encouraging discussions and willingness to be
"Independent Observer #2". Heartfelt thanks goes to David Zhao for much of my initial
training in molecular biology.
I cannot express how thankful I am to have had the support, love, and friendship of my
husband, Christopher Kelly. He has managed to grin and bear the late nights and
frustrating days, without ever losing sight of a goal that started out as mine, but ended as
ours. I am also very grateful to my parents, Larry and Davida Copley, who at every stage
of my education assured me that I could make it.
I also appreciate the assistance in predicting peptides from Dr. Pravin Kaumaya, the tet vectors from Dr. Reed Clark, and the GST vectors and protocols from Dr. Gregory
Taylor and Dr. Jack Dixon. I am thankful for the support received by way of the
Distinguished University, Molecular Life Sciences, and Bennett Fellowships from The
Ohio State University, and the General Mason Award from Children's Research Institute.
vii
VITA
September 7, 1974...... Born –Columbus, Ohio, USA
June, 1997………...... Bachelor of Science, Pharmacy The Ohio State University magna cum laude, with Honors and Distinction
1997-1999...... ……… Medical Scientist Fellow The Ohio State University.
1999-2000...... …… Distinguished University Fellow The Ohio State University
1999-2003...... Molecular Life Sciences Fellow
1999-2003...... Bennett Fellow
2000-2003…………………………………… Mason Fellow Children’s Research Institute
2003-2004……………………………………...Distinguished University Fellow The Ohio State University
PUBLICATIONS
Research Publications
1. Copley, LM., Zhao, WD., Kopacz K., Herman GE., Kioschis, P., Poustka, A., Taudien, S., Platzer, M. (2002) Exclusion of mutations in the MTMR1 gene as a frequent cause of X-linked myotubular myopathy. Am J Med Genet., 107(3):256-8.
viii
FIELDS OF STUDY
Major Field: Molecular, Cellular & Developmental Biology
ix
TABLE OF CONTENTS
P a g e
Abstract...... ii
Dedication...... v
Acknowledgments ...... vi
Vita ...... viii
List of Tables...... xiii
List of Figures ...... xiv
List of Abbreviations...... xvi
Chapters:
1. Introduction
1.1 Clinical Features of MTM1...... 1 1.2 Histopathology of MTM...... 5 1.3 Differential Diagnosis of MTM1...... 6 1.4 Muscle Differentiation and Development...... 7 1.5 Manifestations in Carrier Females ...... 8 1.6 Autosomal Forms of Myotubular Myopathies ...... 9 1.7 Isolation of MTM1, a Gene Mutated in X-Linked Myotubular Myopathy ...... 10 1.8 Protein Structure of Myotubularin ...... 11 1.9 Mutations in the Human MTM1 Gene...... 12 1.10 Genotype/Phenotype Correlations ...... 14 1.11 Substrate Identification for Myotubularin ...... 15 1.12 Myotubularin Expression During Myoblast Differentiation ...... 18 1.13 Identification of Other MTM Family Proteins ...... 19 1.14 Subcellular Localization of MTM1 ...... 25 1.15 Mouse Models for MTM1 ...... 27
x 2. Screening for Protein-Protein Interactions Involving Myotubularin
2.1 Introduction ...... 29 2.2 Materials and Methods ...... 35 2.2.1 General Reagents ...... 35
2.2.2 Yeast 2-Hybrid...... 35 2.2.2 A. Yeast Strains and General Yeast Protocols i. Media ...... 35 ii. Strains ...... 37 2.2.2 B. Cloning i. Generation of Bait Vectors for Use in Yeast 2-Hybrid Screenings...... 37 ii. Generation of pGBKT7-MTM1341-544 (C375S) by Site Directed Mutagenesis...... 40 2.2.2 C. Strain Development i. Generation of Bait Strains PJ692A- pGBKT7- MTM1341-544 and PJ692A- pGBKT7-MTM1341-544(C375S). . . .41 ii. Verification of Yeast Bait Strains...... 42 2.2.2 D. Yeast 2-Hybrid Screening i. Library Mating...... 44 ii. Isolation of Positive Clones...... 45
2.2.3 Coimmunoprecipitation Strategy...... 48 2.2.3 A. Cloning i. Cloning of pcDNA4His/Max TOPO-SCG10 and pcDNA4His/Max TOPO-PP2Ac...... 48 ii. Generation of pCMV-tag2B-hMTM1 and pTRE2pur-hMTM1...... 50 2.2.3 B. Generation of Stable Clones that Express Flag-MTM1 in a Tet-Inducible Fashion ...... 51 2.2.3 C. Immunoblotting and Antibodies ...... 52 2.2.3 D. Coimmunoprecipitation Studies...... 53
2.2.4 GST Pulldown Assays...... 55 2.2.4 A. Purification of GST-MTM1 Fusion Protein ...... 55 2.2.4 B. Generation of SCG10 and PP2Ac as Probes for GST Pulldowns ...... 57 2.2.4 C. Pulldown Assays ...... 59
2.3 Results ...... 60 2.3.1 Yeast 2-Hybrid Screenings...... 60 2.3.2 Coimmunoprecipitation Strategy to Confirm Candidate Protein-Protein Interactions...... 69 2.3.3 GST Strategy to Confirm Candidate Protein-Protein
xi Interactions...... 76
2.4 Conclusions and Discussion...... 81
3. Development and Characterization of an Antibody for Myotubularin
3.1 Introduction...... 87 3.2 Materials and Methods...... 90 3.2.1 Generation of Peptide Antibodies to Human Myotubularin 3.2.1 A. Peptide Prediction...... 90 3.2.1 B. Antibody Production...... 91 3.2.1 C. Antibody Purification...... 92 3.2.1 D. Peptide Competition...... 93 3.2.2 Generation of a Mouse MTM1 Mammalian Expression Construct. . . 93 3.3 Results...... 94 3.4 Conclusions and Discussion...... 106
4. Exclusion of Mutations in the MTMR1 Gene as a Frequent Cause of X-Linked Myotubular Myopathy
4.1 Introduction...... 108 4.2 Materials and Methods...... 109 4.2.1 Patient Genomic DNA and Exon Sequencing ...... 109 4.3 Results...... 111 4.4 Conclusions and Discussion...... 113
5. Effects of Myotubularin Overexpression on Cell Shape
5.1 Introduction...... 116 5.2 Materials and Methods...... 119 5.2.1 Cell Spreading Assays...... 119 5.2.2 Cell Ruffling...... 120 5.2.3 Rac1 Activation Assay...... 121 5.3 Results...... 123 5.4 Conclusions and Discussion...... 137
6. Future Directions ...... 140
List of References ...... 144 Appendix A:...... 155
xii
LIST OF TABLES
Table Page
1.1 Classification of MTM Family Proteins ...... 20
2.1 Amount of Dropout Supplement Required for Each Selection Media Used in Yeast 2-Hybrid ...... 36
2.2 Genotypes of Parent Strains Used in Yeast 2-Hybrid...... 37
2.3 Yeast Strains Employed in Yeast 2-Hybrid… ……………...... 61
2.4 Clones Obtained in Mutant and Wildtype Matings...... 66
2.5 Specific Yeast 2-Hybrid...... 68
2.6 Buffers Used in Coimmunoprecipitation Studies ...... 73
4.1 Primer Sequences for Amplification of MTMR1 Exons...... 111
4.2 Family History and Disease Severity in Patients Sequenced for MTMR1 Mutations ...... 113
A.1 General Information for Parent Vectors ...... 153
A.2 Genotypes and Sources of E. coli Strains...... 154
xiii
LIST OF FIGURES
Figure Page
2.1 Traditional Yeast 2-Hybrid Scheme...... … ...... 34
2.2 Whole Cell PCR Demonstrating the Presence of DNA-BD Vectors in PJ692A ...... 62
2.3 Verification of DNA-BD Strains Used for Library Matings...... 64
2.4 Scheme for Coimmunoprecipitation of Xpress/HisG-PP2Ac or Xpress/HisG-SCG10 with Flag-MTM1 ...... 72
2.5 PP2Ac Does Not Coimmunoprecipitate with Myotubularin Under Low Salt/Triton Lysis Conditions ...... 74
2.6 SCG10 Does Not Coimmunoprecipitate with Myotubularin Under RIPA Lysis Conditions ...... 75
2.7 GST-MTM1 Protein Probed with PP2Ac and SCG10 Made Via in vitro Transcription/Translation ...... 77
2.8 GST-MTM1(C375S) Protein Probed with Xpress/HisG-SCG10 Transfected into HeLa ...... … ...... 79
2.9 GST-MTM1 and GST-MTM1(C375S) Proteins Probed with Xpress/HisG-SCG10 and Xpress/HisG-PP2Ac Transfected into HeLa...... 80
2.10 Comparison of a Similar Domain in ERα and α4 to an MTM1 Amino Acid Sequence Contained in the Baits Used for Yeast 2-Hybrid...... … ...... 84
xiv 3.1 Alignment of MTM1 with Other Members of the Gene Family in the Peptides Used for Antibody Production...... 97
3.2 Immunoblots Using Unpurified Sera ...... … ...... 99
3.3 Immunoblots Using Purified Sera ...... 101
3.4 Competition of CRI-3 with the MTM1 574-591 Peptide ...... 103
3.5 Immunoprecipitation of Flag-MTM1 Protein with Purified CRI-3 Serum ...... 103
3.6 Purified CRI-3 Serum Detects Human Myotubularin, but Not Mouse Myotubularin by Immunoblotting ...... 105
4.1 Pedigrees of Two MTM Families Studied That Are Consistent with an X-Linked Pattern of Inheritance ...... 112
4.2 Schematic Representation of Splice Forms of MTMR1 Described in the Literature ...... 115
5.1 Cell Spreading in Three Independent Clones Over Three Independent Experiments ...... 124
5.2 Cell Spreading in Clone 83 Over a Range of Dosages ...... 127
5.3 Cell Spreading in Clone 16 Over a Range of Dosages ...... 131
5.4 Comparison of Percent Ruffling in Three Independent HeLa Tet-on Clones Expressing Flag-MTM1 Protein ...... 134
5.5 Clone 83 Cells Demonstrate Fewer Ruffles Qualitatively When Treated with Doxycycline to Induce Flag-MTM1 Expression ...... 135
5.6 Percent Ruffling as Noted by Two Independent Observers in Three Experiments ...... 136
5.7 RacI Activation Assay Performed on Clone 83 Cellular Lysates After Treatment with Doxycycline ...... 137
xv
LIST OF ABBREVIATIONS
α alpha
AIDS Acquired Immunodeficiency Syndrome aa amino acid(s) amp ampicillin
β beta
BCS Bovine Calf Serum
BODIPY 4,4-difluoro-5,7-dimethyl-4-bora-3a,4a-diaza-s-indacene-3-propionic acid
BSA Bovine Serum Albumin
°C degrees Celsius
CFA Complete Freund’s Adjuvant
CMV cytomegalovirus
DAPI 4’,6-diamidino-2-phenylindole
DDO Double Dropout Media
DMEM Dulbecco’s Modification of Eagle’s Media
DMF N,N-dimethylformamide
DMSO dimethylsulfoxide
DNA deoxyribonucleic acid
xvi dNTP deoxynucleoside triphosphate
Dox doxycycline
EEA1 early endosomal antigen 1
EGF epidermal growth factor
EST expressed tag sequence
FBS Fetal Bovine Serum
g gram(s)
GOI gene of interest
GDP guanosine diphosphate
GST glutathione-S-transferase
GTP guanosine triphosphate
HRP horseradish peroxidase
hr hour(s)
IFA Incomplete Freund’s Adjuvant
IPTG isopropyl-β-D-1-thiogalactopyranoside
I.U. International units
kan kanamycin
L liter(s)
LB Luria Broth
LPA lysophosphatidic acid
µ micro
M moles per liter min minute(s) xvii mL milliliter(s)
MLB Magnesium lysis buffer mM millimolar mRNA messenger ribonucleic acid
MTM1 X-linked myotubular myopathy mol mole(s)
NEB New England Biolabs
Neo Neomycin
Ni-NTA nickel- N-[5-amin-1-carboxypentyl]aminodiacetic acid
PBS Phosphate Buffered Saline
PBD Pak-1 Binding Domain
PCR polymerase chain reaction
PDGF platelet derived growth factor
PP2Ac catalytic subunit of protein phosphatase 2A
Pen Penicillin
Puro Puromycin
QDO Quadruple Dropout Media rpm rotations per minute
RFLP Restriction Fragment Length Polymorphisms
RNA ribonucleic acid
RT-PCR Reverse Transcriptase Polymerase Chain Reaction
SCG10 Superior Cervical Ganglion 10 sec second(s)
xviii
SD Synthetic Dropout Media
Strep Streptomycin
TAE Tris acetate electrophoresis buffer tRNA Transfer ribonucleic acid
U Unit(s)
UTR untranslated region x-gal 5-bromo, 4-chloro-3-indolyl β-D-galactopyranoside
Zeo Zeocin
xix
CHAPTER 1
INTRODUCTION
X-linked myotubular myopathy (MTM1; MIM # 310400) is a rare neuromuscular
disorder in males presenting at birth with hypotonia, respiratory distress and
characteristic facies. The gene mutated in the majority of myotubular myopathy cases,
MTM1, was isolated in 1996 2 and codes for the protein myotubularin. This introduction
will cover the clinical features of myotubular myopathy, the isolation and structure of the
MTM1 gene, other genes in the MTM family, and possible mechanisms of action of
myotubularin in cells.
1.1 Clinical Features of MTM1:
Classically, the clinical features of MTM1 manifest at birth and include severe hypotonia,
absent deep tendon reflexes, and respiratory distress, often requiring ventilator support3,6-
8. In one study, hypotonia was seen in 100% of cases, and intubation, presumably due to
weakness in the diaphragm and intercostal muscles, was required at birth in 80% 3.
Neonatal respiratory distress is the most concerning feature at birth, and is associated
1 with a high risk of mortality. Other perinatal features have been described in MTM1 patients as well; polyhydraminos and decreased fetal movements were documented in
45% and 58% of cases in one clinical study 3. Prematurity, defined as less than 36 weeks
gestation, was seen in 30.5% of cases 8. At birth, in addition to hypotonia and respiratory
distress, areflexia (62%) and cryptorchidism (55%) have been reported 3,8. One study of
55 MTM1 patients found that the average initial hospitalization after birth was 90 days
and 14.5% died before discharge 3.
Males with MTM1 often have a characteristic facies at birth. Typically, they have a high
forehead and long face with an arched palate. Midface hypoplasia and weak facial
muscles are seen as well. MTM1 patients have a large head at birth, with head
circumference exceeding the 90th percentile in 61% of patients 3. Additionally, MTM1
patients are long at birth, with 69% of patients exceeding the 90th percentile 3.
Overall, MTM1 has a poor prognosis, with mortality reported as high as 46% within the
first 18 months of life 8. Considering all of the 116 patients in this study, those that died had a median age of death at 2.5 months, with a range between 0-12 years. Death frequently resulted from respiratory failure.
Since the discovery of the MTM1 gene, the phenotypic spectrum associated with X- linked myotubular myopathy due to a mutation in the MTM1 locus has been expanded3,7,8. Two studies by Herman et al and McEntagart et al described survivors
2 and allowed for the classification of MTM1 patients as severely, moderately, or mildly
affected 3,8. Patients with a severe phenotype have long term ventilator dependence 3 and
seem to be the most common type of patient seen in clinical studies (79% of total patients
studied in8). The average age that these pateints are able to sit unassisted is 30 months,
indicating significant delay in attainment of gross motor milestones. The study by
McEntagart et al 8 documented that only 50% of their severely affected patients were able
to sit unassisted. None of the severely affected patients attained the ability to walk. The
severe respiratory phenotype was associated with a low perinatal evaluation score (PES)
which was based on the presence or absence of polyhydraminos, fetal movement, deep
tendon reflexes, joint contractures, complicated delivery, and intubation at birth.
Significantly higher PES scores were seen in patients with milder respiratory symptoms.
This would suggest that the severity of the disease at birth may predict the severity of the
disease later in life, but the investigators caution that the variability of this observation is
too wide to recommend its use clinically8. Patients with a moderate phenotype
represented 14% of survivors (>1year old) in a study by Herman, et al and 6% of all
patients in the study by McEntagart, et al. These patients attained motor milestones
faster; the majority of patients were able to sit up and 45% achieved ambulation8.
Herman et al found that mild to moderate patients were able to sit unassisted at an average age of 9.5 months and able to walk at 24 months, indicating a delay in motor milestones that was less severe than that seen in the classic severe patients. These patients are ventilator dependent for only part of the day, when healthy. The mildly affected patients represented 15-20% of the survivors3,8. These patients are ventilator independent
3 and are able to walk. It should be noted that both studies reported normal intelligence in the majority of these MTM1 patients3,8, although delays in speech occurred in 60% of
patients generally due to dysarthria and ventilator use8.
Long term survivors with MTM1 often develop additional clinical features consistent with a myopathy. These include scoliosis, which usually develops by age 5 (60%), feeding problems requiring gastrostomy (88%), myopia (45%), and ophthalmoplegia
(30%) 3. There have also been a few reports in the literature of adult patients with mild
presentations of MTM1 9 including a 67 year old fertile man who is the oldest known
MTM1 patient 10. Interestingly, unanticipated complications have been described in long
term survivors of MTM1, challenging the idea that MTM1 is strictly a myopathy 3. The numbers of patients presented are small for each complication, but represent a first attempt to describe complications in long term survivors. Six patients have been reported with transient or persistent increases in liver enzymes, 4 with gallstones, 4 with pyloric stenosis and 1 with kidney stones 3. Three patients developed peliosis hepatis, a rare and
serious liver pathology characterized by several blood filled cystic cavities in the liver3,10.
This condition usually is associated with AIDS and wasting diseases, and is very rare in childhood. Two patients have been described on autopsy as having a subcapsular hematoma of the liver 3,11. Two patients had documented spherocytosis as detected by
osmotic fragility testing. Two patients had a vitamin K responsive coagulopathy 3. Five patients have been described with hydrocephalus, although it is unclear if the hydrocephalus was secondary to perinatal complications 7. With these complications
4 documented in some patients, Herman et al suggested yearly blood counts, abdominal ultrasonograghy and liver function tests along with a thorough examination of clotting factors before attempting surgery 3.
1.2 Histopathology of MTM1
Diagnosis of MTM1, when suspected clinically, is further evaluated by muscle
biopsy3,7,12. Histopathologically, large numbers of centrally located nuclei are surrounded
by a region lacking myofibrils, as indicated by myosin ATPase staining 13,14. Muscle cells with centrally located nuclei are often referred to as myotubes (Section 1.3). The periphery of the affected muscle fibers appears to have normal myofibrils, while the central “clearing” contains increased glycogen and mitochondria as indicated by increased PAS and oxidative enzyme staining, respectively 6,13,15. One study found that
muscle biopsies from MTM1 patients contained widely spaced central nuclei much like
the placement of nuclei in biopsies taken from 8-15 week gestation fetuses. This was in
direct contrast to muscle biopsies taken from term infants that demonstrated peripherally
placed nuclei 14. Type I fiber hypotrophy has been noted and affected muscle fibers have
a smaller diameter than unaffected fibers. The ratio of muscle fibers appearing affected or
as myotubes compared to healthy muscle fibers may vary from patient to patient 6,15. In
one study of four male MTM1 patients, the percentage of muscle fibers demonstrating centrally located nuclei varied from 2-40% 13. Strikingly, biopsies from different muscles
of the same patient showed varied percentages of centrally located nuclei, necessitating in
5 one case a repeat biopsy to confirm the clinical suspicion of MTM113. Certainly, while muscle biopsy is common for confirming clinically suspected MTM1, it is not clear that biopsy of one muscle will authenticate the diagnosis in every case or what minimum percentage of affected myofibers indicates MTM1.
1.3 Differential Diagnosis of MTM1
The most common disease that must be considered as a differential diagnosis for MTM1 is congenital myotonic dystrophy 16,17. This disease is caused by an expansion of the CTG triplet repeat in the 3’UTR of the DMPK gene, with the size of the repeat increasing through generations and resulting in more severe clinical phenotypes. Large CTG amplifications are associated with the congenital form and almost exclusively seen in the affected offspring of affected women (OMIM #160900). This CTG triplet repeat can be detected molecularly to rule out congenital myotonic dystrophy. Patients presenting at birth with this autosomal dominant disorder display hypotonia and respiratory distress;
25% of patients fail to survive past 18 months of age and of the survivors, 75% have reduced cognitive ability requiring special education. By examination of muscle biopsy, central nuclei and type I hypotrophy are seen along with ringed fibers and sarcoplasmic masses.
6
1.4 Muscle Differentiation and Myotubular Myopathy
Human muscle development is a progressive process usually completed during
gestation18. Several genes play a role in myogenesis including Pax3 in myogenic
precursors, MyoD and Myf5 in committed myoblasts, and myogenin in myoblast
differentiation19-22. Mature muscle cells begin as myogenic precursors which develop into
myoblasts, round cells lacking myofibrils and containing centrally located nuclei. As
differentiation proceeds, the myoblasts elongate and fuse forming multinucleate cells
with increased levels of myofibrils. At this “myotube” stage, cells still have centrally
located nuclei. As these myotubes develop into mature fibers, myofibrils increase in
number, filling the center of the cell, and nuclei are repositioned to the periphery of the
cell. One hypothesis concerning the pathogenesis of MTM1 is that it represents an arrest
in muscle cell differentiation 14,18. The similarity of the muscle fibers found in MTM1 to fetal myotubes was first noted in 1966 by Spiro et al 18 in a 12 year old boy with a
clinical course consistent with MTM1. It was not until 1990 with the work of Sarnat 14, that muscle biopsies from four 8-15 week fetuses and four term infants receiving biopsies to rule out congenital myopathies were compared to those from four infants with biopsy proven MTM1. Sarnat found that in many ways, muscle tissue from MTM1 patients did resemble fetal muscle: nuclei were centrally located along with mitochondria and glycogen. Sarnat also detected the strong expression of vimentin and desmin in MTM1 muscle as determined by immunohistochemistry. Vimentin is an intermediate filament
7 seen in myoblasts and myotubes, but not present in mature striated muscle. Desmin is an intermediate filament that increases after myoblast fusion and can only be detected at
very low levels in mature muscle. The strong staining pattern of vimentin in MTM1
muscle was demonstrated again in 5 patients, but the tissues were not tested for desmin
staining 13. To date, there has been no large scale study evaluating the desmin, vimentin
or other skeletal muscle marker staining patterns in MTM1 patients. While this
hypothesis of an arrest in differentiation still persists, it is controversial. This is because
affected fibers in MTM1 patients do not exactly mimic myotubes.
1.5 Manifestations in Carrier Females:
While MTM1 is generally associated with males due to its X-linked inheritance, clinical
manifestations in female carriers have been described 23-26. Phenotypes described in
female carriers have ranged from mild weakness and easy fatigability 24, to overt and
disabling muscle disease 23. Carrier females may have a high incidence of miscarriage
and stillborn males 8,27. Muscle biopsies from carriers may or may not reveal the classic histopathology associated with MTM1 7. In one study of 12 obligate female carriers, only
6 demonstrated centrally located nuclei on muscle biopsy 6. Many of the female carriers
described as symptomatic did not have muscle symptoms in the neonatal period, but
muscle symptoms developed in childhood that progressed with age 25,26,28. Multiple
investigators have suggested that skewed X-inactivation may account for the appearance
of a phenotype in heterozygous female carriers 23,28, but these case reports have been
8 anecdotal. One study 24 examined DNA from females from 18 MTM1 families and found a tendency for severely affected females to have skewed X-inactivation; however, these results were not statistically significant.
1.6 Autosomal Forms of Myotubular Myopathies
Both autosomal dominant (MIM# 160150) and autosomal recessive (MIM# 255200) forms of “myotubular myopathy” have been described in the literature, but relatively few families have been documented 6,27. The autosomal “myotubular myopathies” are sometimes referred to as centronuclear myopathies. The histopathology seen in muscle biopsy with MTM1 is similar to that observed in autosomal forms. In general, patients with autosomal dominant disease have generalized muscle weakness, usually involving the proximal and facial muscles, that develops in late childhood and early adulthood. The course of the disease is milder than MTM1; however, the muscle symptoms are slowly progressive 6. In one study, researchers found that all X-linked cases of myotubular myopathy presented by birth, whereas 41% of autosomal dominant cases presented by age 9 and 27% between the ages of 10 and 19 years 27. Twenty-nine percent of autosomal dominant patients died between the third to fifth decade of life, while 92% of the X-linked patients succumbed within 1 year 27. Patients with autosomal recessive disease present in infancy, childhood, or early adulthood. The severity of autosomal recessive disease appears intermediate between classic X-linked MTM1 and the milder
9 autosomal dominant form 6. Currently, no defective genes have been identified in the
autosomal myopathies, presumably due to the scarcity of families available for linkage analysis.
1.7 Isolation of MTM1, a Gene Mutated in X-Linked Myotubular Myopathy
The X-linked inheritance pattern for MTM1 was suggested by evaluating families with
more than one affected male related through maternal lineages. Linkage to Xq28 was
established using RFLPs 29-31. The critical region was refined to 5 megabases by studying
a female myotubular myopathy patient with skewed X-inactivation and a cytogenetically
detectable heterozygous deletion in this region 32. Careful study of this patient, along with
linkage analysis and the development of highly informative microsatellite markers,
narrowed the region to 600Kb 33. Further refinement of the region to 430 Kb was possible
with the identification of two male patients with MTM1 and ambiguous genitalia 34. This
additional finding was hypothesized to result from a contiguous deletion of the MTM1
locus and one or more additional neighboring genes 35. As more microsatellite markers
and more informative families became available, the region was narrowed to 280 Kb
between the markers DXS1282 and DXS455, and YAC and cosmid contigs spanning the
region were constructed 2,36. Exon trapping, exon prediction in silico and cDNA selection
were used to identify transcribed sequences within these contigs. One gene discovered in
this region, MTM1, was assembled by screening multiple cDNA libraries and selected
10 cDNA clones. Eight exons were identified initially, five of which were sequenced for
mutations in MTM1 patients, revealing frameshifts or missense mutations that were
conserved in C. elegans and S. cerevisiae homologues of MTM1. This conservation
suggested that the residues mutated in the MTM1 patients are important for protein
function. Conservation of the protein in yeast suggested that myotubularin may play an evolutionarily conserved role in cells; this was a surprise for a gene responsible for a seemingly muscle specific disease.
The final structure of MTM1 (Genbank # NT 011726) is a 15 exon gene yielding a 603 amino acid protein. The additional exons were described after screening cDNA libraries and ESTs in the dBEST database 2. The gene covers approximately 100Kb of genomic
DNA and is made up of type 0 exons except between exons 3 and 4. When exons are described as being type 0, it means that the exon ends and begins with a complete, in frame codon; thus, if a type 0 exon is deleted, the predicted protein downstream of the
deletion would still be in frame. Transcription of MTM1 results in a ubiquitously
expressed 3.9 kB transcript by Northern analysis with an alternate 2.4 kB transcript found
in muscle and testes, representing an alternate polyadenylation signal 2.
1.8 Protein Structure of Myotubularin
The predicted amino acid sequence of MTM1 contains a Cys-X5-Arg domain in exon 11
which is found at the active site of protein tyrosine phosphatases 2. The entire domain
11 spans amino acids 273-471. Further analyses demonstrated that the likely substrates for
MTM1 are phospholipids (See Section 1.10). Cui et al proposed that MTM1 also possessed a SET Interacting Domain (SID) in exons 12 and 13 (amino acids 436-486), based on the interaction of a fragment of MTM1 with a SET domain in a yeast 2-hybrid assay37. SET domains are present in nuclear proteins involved in transcriptional silencing and chromatin aggregation. However, the physiological consequences of this motif are unknown at this time (See Section 1.12) 37-39. Additional domains potentially important for protein-protein interactions were also described. They include the coiled coil (aa 553-
580) and PDZ-binding site domains at the C-terminal of MTM1 whose functions remain unclear 40,41 and a Rac Induced localization Domain (RID) between amino acids 161-272 that is required for the subcellular localization of MTM1 to membrane ruffles 39 (See section 1.13). The final domain found in MTM1 is the GRAM domain (aa 29-160); this domain is named from the other proteins in which it is found, including glucosyltransferases, Rab-like GTPase activators (and myotubularins) and has unknown significance 42.
1.9 Mutations in the Human MTM1 Gene
Prior to the identification of the MTM1 gene, patients without severe neonatal onset of disease and a clear X-linked pedigree were generally believed to have an autosomal form of myotubular myopathy 43,44. Since the discovery of the gene, some patients with milder phenotypes have been determined to have mutations in MTM1. To date, 328 families
12 have been described with 193 different disease causing mutations in the MTM1
gene10,12,44-48. The mutations described occur throughout the gene, but exons 4, 12, 3, 8,
9, and 11 are the most frequently affected in descending order when exon length is taken
into consideration 44. Of the patients described to date, 29.5% have missense mutations,
20.7% have nonsense mutations, 23.1% have small insertions or deletions, 19.8% have
splice site point mutations, and 7% have large deletions 10,12,44-48. Approximately 20% of
patients with suspected MTM1 do not have an identifiable mutation in the MTM1 gene12.
This may be due to a mutation in the promoter or other non-coding parts of the gene, misdiagnosis or failure to detect a mutation with the protocol used. Eighty-six percent of mothers with affected male offspring carry the mutation themselves 10. This carrier rate is
higher than the predicted rate of 66% for a lethal X-linked disease and makes genetic
counseling particularly important. Germline mosaicism has been described in five cases,
further complicating the genetic counseling for this disease 24,49. Prenatal diagnosis by
sequencing DNA can now be performed in familial cases or to rule out gonadal
mosaicism 44,50.
Most missense mutations in MTM1 occur between exons 8 and 12 which encompass the
catalytic active site (exon 11) and the putative Set Interacting Domain (SID; exons 12,
13) 44 (See Section 1.12). In one large study that reported 58 patients with 37 different
missense or single amino acid deletion mutations, all except 6 of the affected residues
were conserved in mouse MTM1, human MTMR1, and the Drosophila orthologue
dMTMH1. In the other six mutations, the affected amino acid was functionally conserved
13 in human MTM1, human MTMR1 and dMTMH1 44, meaning that the residues differ but
have similar physical and chemical properties. In the same study, 43 of the mutations that
were in CpG hypermutable sites and could be explained by a C to T transition from
deamination of a methylated cytosine. Of these 43 mutations, 84% were recurrent.
1.10 Genotype/Phenotype Correlations
Clearly a spectrum of phenotypes is seen in patients with a mutation in MTM1. This is
more complex than one might imagine, since multiple affected individuals within the same family and unrelated individuals sharing the same mutation may have different severities of disease 8,44,45. Although the precise reasons for this variability are unclear, it
is possible that different modifier loci are interacting in the disease, possibly explaining
the variability in phenotype seen in patients. An attempt to find such correlations was
undertaken in one large study involving 116 patients 8. Researchers found that patients
with non-truncating mutations, defined as missense and in frame small deletions or
insertions, were more likely to have a mild phenotype as judged by ventilator
dependence. Patients with truncating mutations, defined as frameshift, nonsense or large
deletion, showed a lower survival rate and increased ventilator dependence than patients
with non-truncating mutations. Clearly, establishing genotype/phenotype correlations is
important so that reproductive decisions and medical care can be recommended
considering prognosis.
14 The most recent study of mutations in MTM1 patients, encompassing all published cases,
revealed that five different mutations account for 20% of the cases of MTM1 10. The most
common mutation, c.1261-10A>G, results in the generation of an alternate splice site and the insertion of three amino acids (FIQ) at the beginning of exon 12. This mutation accounts for 7.3% of the reported MTM1 cases and is always associated with a severe phenotype. The second most common mutation, R241C, results in an amino acid change in exon 9, accounts for 4% of MTM1 cases, and is associated with a milder phenotype in approximately 60% of the patients where data are available. The third most common
MTM1 mutation, accounting for 4% of cases, is c141-144 del AGAA which results in a
frameshift in exon 4 and is always associated with a severe phenotype. The fourth and
fifth most common mutations seen, R37X and R421Q, account for 2.8% and 2.5% of
MTM1 cases, respectively. Overall, only 5% of reported patients with a large deletion,
frameshift or nonsense mutation have mild disease, but all missense mutations affecting
the SID or catalytic active site of MTM1 were severe 44.
1.11 Substrate Identification for Myotubularin
Myotubularin was originally predicted to be a protein tyrosine phosphatase (PTPase)
based on its Cys-X5-Arg active site motif and can dephosporylate both serine and tyrosine containing peptides in vitro 2,37. However, a significant advance in
understanding the molecular nature of myotubularin came with the work of Taylor and
Dixon who noted that the in vitro phosphatase activity of myotubularin with the artificial
15 PTPase substrate para-nitrophenylphosphate (pNPP) was very inefficient and that the
active site of myotubularin bore some similarity to the active site of Sac1p in S.
cerevisiae, a known phosphoinositide phosphatase51. They subsequently tested
myotubularin for in vitro enzyme activity against a variety of phosphoinositides,
including phosphotidylinositol-3-phosphate (PI3P). Myotubularin was found to
hydrolyze PI3P to phosphotidylinositol (PI) with much greater efficiency (Km=43uM)
than hydrolysis of pNPP (Km=21mM). These data suggested that PI3P may be the
physiological target for myotubularin. PI3P is thought to be involved in vesicular
trafficking and has been localized to early endosomes 52. To address PI3P metabolism
physiologically, cells were labeled with 3H-inositol and transfected with either wildtype
myotubularin or MTM1 (C375S) protein, a catalytically inactive mutant. Using HPLC, the authors demonstrated that PI3P levels decreased with overexpression of myotubularin and increased with overexpression of the inactive MTM (C375S) protein compared to untransfected cells. These results suggested that PI3P is a physiological substrate of
myotubularin. Taylor and Dixon also suggested that MTM1 may result from reduced ability to dephosporylate PI3P since recombinant mutants of myotubularin associated
with severe disease demonstrated no activity towards PI3P.
In 2002, Mandel et al demonstrated a lack of PTPase activity for myotubularin in vivo in
S. pombe 53. The researchers also noted that S. pombe null for MTM1 accumulated PI3P
by thin layer chromatography. This phenotype was rescued when human myotubularin
16 was overexpressed, but not by human MTM1 (C375S) or human MTM1 (D278A)
proteins, catalytically inactive and substrate trapping mutations, respectively.
The discovery that myotubularin is a lipid phosphatase renewed speculation about its
function in vivo. Researchers from the Barr laboratory evaluated the substrate specificity
of MTM1 54. Recombinant MTM1 was produced and tested in vitro against a panel of
phospholipids. These experiments demonstrated that myotubularin could dephosphorylate
PI(3,5)P2, yielding PI5P, as well as PI3P. To address this new substrate specificity in vivo, myotubularin was overexpressed in S. cerevisiae labeled with 3H-inositol, and
analyzed by HPLC. PI5P, the proposed product of PI(3,5)P2 dephosphorylation, is not
normally present at detectable levels in yeast, but could easily be detected with
myotubularin overexpression; additionally, increasing the levels of PI(3,5)P2 in yeast by
osmotic shock caused an increase in PI5P levels with myotubularin overexpression. No activity was seen with the MTM (C375S) mutant protein.
Enzyme kinetics were performed using recombinant myotubularin, yet typical Michaelis-
Menton kinetics were not observed, suggesting that allosteric mechanisms were present54.
Researchers found that the in vitro activity of myotubularin towards PI(3,5)P2 and PI3P
was actually stimulated by the addition of PI5P, a proposed product of the enzyme.
Additionally, myotubularin activity increased as the concentration of enzyme in the
reaction was increased; this suggested that oligomerization of the MTM1 enzyme itself
may contribute to enzyme activity. To visualize these hypothetical oligomers, electron
17 microscopy was employed on an in vitro reaction with MTM1, PI3P and PI5P.
Microscopy revealed that in the presence of substrate, the enzyme organized into a
heptameric ring structure approximately 12.5 nm in diameter. Collectively, these data revealed that MTM1 is a complex enzyme with allosteric regulation by PI5P and
oligomerization contributing to its activity in vitro. The physiological significance from
these in vitro data is yet to be addressed.
1.12 Myotubularin Expression During Myoblast Differentiation
Since the muscle histopathology associated with myotubular myopathy has been
hypothesized to resemble fetal muscle, questions arose about its expression during
muscle differentiation. To begin to address these questions, Kim et al monitored the RNA
levels of MTM1 by Northern analysis in differentiating C2C12 cells 55. C2C12 cells are
mouse myoblasts that can be induced to differentiate into myotubes with the withdrawal
of fetal bovine sera and the addition of horse sera that has lower mitogen content. As the
culture differentiated into myotubes, as indicated by the expression of myosin heavy
chain and visual inspection, the levels of MTM1 RNA increased.
Variability in myotubularin levels during differentiation were also studied by Laporte et al 39. Upon immunoprecipitation of endogenous protein from various adult mouse tissues,
they found that two isoforms of myotubularin were present in skeletal muscle and heart.
18 Upon differentiation of C2C12 cells, the form exhibiting slower migration increased; they
hypothesize that this may represent a post-translationally modified form of myotubularin.
1.13 Identification of Other MTM Family Proteins
Starting with the sequence for MTM1, the public databases (Genbank) were analyzed for
related genes 4. This initial analysis yielded two related proteins, MTM Related 1 and 2
(MTMR1 and MTMR2) that possessed 75% and 65% amino acid identity to MTM1, respectively 2. Since this initial discovery, a total of 13 human MTM-related genes have been described and categorized into six subgroups based on amino acid homology56-59
(Table1). Each of these six categories contain orthologues from C. elegans and D. melanogaster and there are yeast orthologues in the active phosphatase categories (1-
3)56,60.
MTMR1 and MTMR2 have been classified in the same category as MTM1 because they
all contain the GRAM domain, the RID domain, an active phosphatase domain, a coiled coil domain, and a PDZ binding site. MTMR1 was mapped to Xq28 just 20 kb telomeric to MTM1 4. Multiple splice forms have been described5. MTMR1 also possesses activity
towards PI3P, although 30 and 100x less than MTM1 and MTMR2, respectively61 . This gene will be discussed further in Chapter 4.
19
x IN(1,3)P2
.
x x x 56-59 PI(3,5)P
x x x x x x x ? Substrate Specificity PI3P
x x FYVE
g
x x x x x PDZ Domain Bindin
x x PH
x x x x x x x x x x x x x x Coil Coiled
x x x x x x x x x x x x x x SID
x x x x x x Inactive
x x x x x x x x on Amino Acid Structure as Described in references Phosphatase Active
x x x x x x x x x x x x x x RID
x x x x x x x x x x x x x x GRAM
x x DENN
MTM1 Assn. Disease CMT4B1 CMT4B2
SBF1 SBF2 DSP1 DSP2 Name 3-PAP FYVE- FYVE- Alternate LIP-STYX FLJ20313
MTM Family Domains are present from N-terminus to C-terminus. Table 1.1: Classification of MTM Family Proteins Based Category 1 MTM1 MTMR1 MTMR2 Category 2 MTMR3 MTMR4 Category 3 MTMR6 MTMR7 MTMR8 Category 4 MTMR5 MTMR13 Category 5 MTMR9 Category 6 MTMR10 MTMR11 MTMR12
20 MTMR2, initially discovered by searching for ESTs in the public databases, was mapped
to 11q22 62. Mutations in MTMR2 have been described in patients with Charcot Marie
Tooth disease Type 4B1 (CMT4B1), a peripheral demyelinating neuropathy characterized by focally folded myelin within neural sheaths 62,63. Recombinant MTMR2
dephosphorylates PI3P with comparable efficiency to myotubularin in vitro55.
Recombinant mouse MTMR2 protein, which exhibits 97% identity to the human protein,
64 dephosphorylated PI(3,5)P2 four times more efficiently than PI3P .
Multiple CMT4B1-associated MTMR2 mutations when tested in vitro using the mouse
homologue have been demonstrated to lack phosphatase activity on both phospholipid
substrates 64. A protein-protein interaction has been discovered between MTMR2 and
MTMR5 (SBF1), which is a catalytically inactive member of the myotubularin family
(see below) 65. When this interaction was recapitulated with recombinant proteins in
vitro, the enzymatic activity of MTMR2 was increased 4.6 and 3.4 times with PI3P and
65 PI(3,5)P2, suggesting that MTMR5 cooperates with MTMR2 . Domain mapping revealed that the coiled coil domain in MTMR2 protein is necessary for its interaction with MTMR5 protein. Additionally, the subcellular localization of MTMR2 with
MTMR5 is lost when the coiled coil domain is deleted from MTMR2, suggesting that this interaction is necessary to localize MTMR2 to the proper subcellular compartment 65.
The second category of the MTM1 gene family is represented by MTMR3 (FYVE-DPS1) and MTMR4 (FYVE-DSP2). These proteins are similar to MTM1 in that they possess the
21 GRAM, RID, phosphatase, SID and coiled coil domains, but they lack the PDZ-binding
domain and contain a C-terminal FYVE domain 57. The physiological significance of the
FYVE domain in these proteins is unknown; however FYVE containing proteins bind
PI3P and an FYVE domain probe can detect PI3P 52,66,67. MTMR3 is an efficient
phosphatase for PI3P and PI(3,5)P both in vitro with recombinant enzyme and in
yeast61,68. With the identification of MTMR3, came the discovery of MTMR4 from the same laboratory, by database mining and RACE analysis 69. Expression of MTMR4 was
examined in multiple human tissues by RT-PCR. Exogenously expressed MTMR4
protein was found to have a perinuclear subcellular localization by immunofluorescence
and efficiently dephosphorylated PI3P but at a rate approximately 20x lower than
myotubularin. Further study of proteins from this subcategory is needed to determine the
physiological roles they play.
The third group of MTM related proteins contains MTMR6, MTMR7, and MTMR8.
These proteins all possess the GRAM, RID, phosphatase, SID, and coiled coil domains 57.
The major difference in this subcategory of MTM family members is the absence of the
C-terminal PDZ binding site motif. MTMR6 has been described as interacting with
MTMR9 as discussed below 70. MTMR6 has also been shown to dephosphorylate PI3P in
vitro, but at an activity 30-100x lower than that of MTM1 and MTMR2 proteins 61.
MTMR7 was recently described 70. It is predicted to have various splice forms deleting the coiled coil domain. The significance of any alternate splicing is unknown. By
22 Northern analysis, transcripts were seen in several tissues, yet Western analysis indicated
that this protein was expressed predominantly in brain tissue. Analysis of substrate
specificity in vitro with recombinant MTMR7 revealed that it dephosphorylated PI3P
with an activity similar to MTMR6 protein. Surprisingly, MTMR7 protein also
dephosphorylated inositol(1,3)P2 , a non-lipid substrate, with activity 10x higher than that
against PI3P. MTMR7 is hypothesized to be enriched in Golgi or endosomes based on its
colocalization by immunofluorescence with γ-adaptin positive granules 70.
The fourth category of MTM family members includes MTMR5 (sbf1) and MTMR13
(sbf2) 57. These two proteins resemble myotubularin in that they possess the GRAM,
RID, phosphatase, SID, coiled coil, and PDZ binding site domains. However, although
they contain a phosphatase domain, the key amino acids necessary for activity have been
altered, rendering them catalytically inactive (dead phosphatases). Secondly, they contain
a C-terminal extension with a DENN (Differentially Expressed in Neoplastic versus
Normal cells) domain 57. Thirdly, they possess a pleckstrin homology (PH) domain,
which has been shown to bind PI3P 66.
The better studied of the proteins, MTMR5 was initially identified from a yeast 2-hybrid
experiment screening for proteins that interact with SET domains 37. SET domains are
found in proteins affecting chromatin silencing such as mammalian Hrx, Drosophila
Su(var), and Enhancer of zeste. It was the identification of MTMR5, its homology to
myotubularin, and its SET interacting Domain (SID) that suggested the possibility of a
23 SID in myotubularin. From this initial study, MTMR5 was postulated to prohibit
differentiation in C2C12 myoblasts, localize to the nucleus, and transform fibroblasts.
Three years later, the protein originally described as MTMR5 was found to be
incomplete38. MTMR5 was subsequently found to contain an N-terminal DENN motif
and a C-terminal PH domain. The complete protein showed very different characteristics
than originally reported in the literature38; it was no longer capable of transforming NIH
3T3 cells, was localized in the cytoplasm, and actually inhibited growth of NIH 3T3 cells. Taken together, these data suggest that the original protein-protein interactions described with SET containing proteins are unlikely to be physiologically relevant 38.
MTMR13 has recently been described as one of the genes mutated in Charcot Marie
Tooth disease 4B2 (CMT4B2) with early onset glaucoma 59. Two consanguineous
families were identified with nonsense mutations in MTMR13 that were predicted to
truncate the protein upstream of the inactive phosphatase and SID motifs.
MTMR9 (LIP-STYX) is the only member of the fifth group of MTM related proteins.
MTMR9 was actually identified as a binding partner for MTMR7 70. Mapping the
relevant residues for the interaction revealed that the coiled coil domain was sufficient for
MTMR9 protein to interact with MTMR7 protein in vitro. Interestingly, cooperation was
seen between the MTMR7 and MTMR9 proteins; the activity of MTMR7 protein towards
inositol (1,3)P2 was increased in vitro in the presence of MTMR9 protein compared to
24 MTMR7 alone. Using GST-MTMR9 and coimmunoprecipitation, Mochizuki and
Majerus determined that MTMR9 also interacted with MTMR6 70.
The sixth and final category of the MTM gene family contains MTMR10, MTMR11 and
MTMR12 (3-PAP) 57. These proteins lack the PDZ-binding site and are catalytically
inactive. MTMR10 is a hypothetical protein that currently has only been assessed a
Genbank accession number (FLJ20313). Little work has been done on MTMR11.
MTMR12 protein was originally isolated from rat brain and named 3-PAP 71. When it was isolated, MTMR12 protein was found to have a binding partner visible after
coimmunoprecipitation from brain lysates; this binding partner was identified by mass
spectroscopy to be MTM1 72 and was confirmed via coimmunoprecipitation from K562
cell lysates. This interaction will be further discussed in Chapter 2.
1.14 Subcellular Localization of Myotubularin
Endogenous myotubularin has not been detected directly in cells by immunofluorescence.
Detection of myotubularin from whole cell lysates typically requires
immunoprecipitation first 73, presumably due to very low levels in cells. Subcellular
localization has been performed using tagged MTM constructs that overexpress the
resultant protein 39,61. The first study, by Cui et al 37 suggested that myotubularin was
localized in the nucleus, while multiple subsequent studies support a cytoplasmic
localization 39,61.
25
One study attempted to correlate PI3P pools with subcellular localization of myotubularin61. PI3P was stained with a biotinylated 2xFYVE probe and was visualized in a punctate pattern representing endosomes, as expected. Upon overexpression of GFP-
MTM1 the punctate PI3P signal was abolished. From these data, it was concluded that myotubularin was able to act upon PI3P in endosomes. Loss of punctate endosomal PI3P staining was not demonstrated upon overexpression of MTMR2, suggesting that dephosphorylation of PI3P in endosomes is specific to myotubularin.
Subcellular localization has been further investigated by the Mandel Laboratory39.
Overexpression of tagged myotubularin demonstrated localization in a dense cytoplasmic network and filopodia in Cos cells. This pattern of localization was also apparent in both undifferentiated and differentiated C2C12 cells. However, colocalization could not be demonstrated with desmin, actin, tubulin, keratin, or vimentin. There was also no clear colocalization of MTM1 with markers for the endocytic pathway including dynamin
(clathrin coated vesicles), caveolin (uncoated vesicles) or Rab5 (endosomes).
Mandel et al noted that in a subset of cells overexpressing myotubularin, there was extensive actin remodeling as evidenced by the appearance of numerous filopodia 39.
These plasma membrane extensions were clearly enriched with MTM1 protein. This phenomenon was also noted for MTM1 (C375S) and MTM1 (D278A) indicating that this localization is not dependent on enzyme activity. When actin remodeling was destroyed
26 by treatment of the cells with cytochalasin D, myotubularin and the D278A mutant
protein were still seen at the plasma membrane, which suggested that this localization
was not dependent on actin fibers. In contrast, when extensive plasma membrane
remodeling was induced in the form of membrane ruffles with constitutively active Rac1,
MTM1 and the C375S and D278A mutants were clearly localized to them. Membrane
ruffles are actin dense lateral protrusions of the plasma membrane associated with
migration and phagocytosis. The region of MTM1 responsible for this localization, amino
acids 179-248, was mapped and designated the Rac Induced localization Domain (RID).
1.15 Mouse Models for MTM1
Based on the clinical features and perinatal onset of X-linked myotubular myopathy, it
was anticipated that a targeted MTM1 null allele in the mouse would be lethal. In 2002, a
viable classical knockout mouse for MTM1 was reported 5. To generate the null, a mouse line with flox sites flanking exon 4 of MTM1 was generated and mated to a mouse line expressing Cre recombinase from the CMV promoter. The progeny were selected for germline excision. Exon 4 of MTM1 was targeted since excision leads to a frameshift and truncated protein, resulting in the absence of myotubularin. These MTM1 deficient male mice are asymptomatic at birth, but progress over time through four phases: During Phase
I animals grow more slowly as determined by weight, but are asymptomatic. Phase II animals display decreased hindlimb strength followed by decreased forelimb strength and kyphosis in Phase III. The final phase (IV) is marked by total hindlimb paralysis,
27 decreased muscle mass and death from respiratory distress and cachexia. As the disease
develops spatiotemporally, so too does the characteristic muscle pathology, including
hypotrophy and centrally located nuclei with a central clearing lacking myofibrils that is
positive for mitochondria. These findings suggested that in the mouse, myotubularin is not essential for myogenesis, since the animals appear asymptomatic at birth. However,
myotubularin is necessary for muscle maintenance since animals exhibit progressive
muscle degeneration. They also challenge the hypothesis that MTM1 results from an
arrest in muscle cell differentiation. In addition to the classic knockout, two conditional
MTM1 knockout lines were generated to knockout myotubularin expression in either a
muscle specific or neural specific manner. The muscle specific MTM1 knockout mouse
displayed a phenotype essentially identical to the original model, while the neural
specific knockout mouse exhibited no obvious phenotype or muscle pathology. These
mouse models of myotubularin deficiency demonstrate that the muscle symptoms of
MTM1 are strictly a lesion of muscle cells and provide a good, although not perfect,
animal model for the disease.
28
CHAPTER 2
SCREENING FOR PROTEIN-PROTEIN INTERACTIONS INVOLVING MYOTUBULARIN
2.1 Introduction
Examination of protein-protein interactions allows for the functional characterization of a
protein of interest in the context of the cell. They can place a protein in a known pathway,
suggest a mechanism of action, or otherwise provide clues to its function. Several
methods have been developed to screen for protein-protein interactions including the genetic method yeast 2-hybrid, which was first described by Fields and Song in 1989 74.
The major goal of this project was to identify proteins that interact with myotubularin. At the initiation of this work, much less was known about MTM1 and other members of this gene family than is now. It was hoped that uncovering proteins that interact with myotubularin would help place it in a particular pathway, identify a substrate, or give us clues as to why a deficiency in myotubularin leads to the human disease MTM1.
At the beginning of this project, the only protein-protein interaction described in the
MTM gene family was the interaction between the MTMR5 protein and the SET domains
29 contained in HRX, Enhancer of zeste, and Set1 37. Using coimmunoprecipitation in
mammalian cells, overexpressed MTMR5 interacted with overexpressed HRX, but not
HRX with a deleted SET domain. The domain of MTMR5 necessary for this protein- protein interaction was identified and referred to as the Set Interacting Domain (SID).
One of the major speculations that came from this work was that members of the MTM family with inactive phosphatase motifs may act to sequester substrate and act in an opposing manner to the active phosphatases in the family, such as MTM1. Subsequently, it was discovered that MTMR5, was in fact not the complete protein 38. The complete
MTMR5 protein no longer localized to the nucleus, and its interaction with HRX or SET
domains in general has not been verified; the notion that this interaction is
physiologically relevant has become suspect.
Subsequent work on MTMR5 has emerged from the Dixon Laboratory describing an
interaction with MTMR2 41. This interaction was revealed by immunoprecipitating
tagged MTMR2 protein from transfected HEK 293 cells. The resultant band coimmunoprecipitating with MTMR2 was analyzed by mass spectroscopy and found to be MTMR5. This interaction was confirmed with a GST based strategy as well as coimmunoprecipitation from doubly transfected cells. Mapping domains using a GST based strategy revealed that the coiled coil domains of both MTMR5 and MTMR2 were necessary for this interaction to occur. MTMR2 was found to exist as a homodimer by crosslinking experiments. Both the homodimerization of MTMR2 and its heterodimerization with MTMR5 were dependent upon the coiled coil domain. In direct
30 contrast to the speculation of Cui et al 37, the dead phosphatase MTMR5 was found to
cooperate with the active phosphatase MTMR2; when tested in vitro, MTMR2 showed a
4-5x increase in activity towards the substrate PI3P when MTMR5 was present. In
addition, the subcellular localization of MTMR2 is changed when its interaction with
MTMR5 is disrupted by deletion of the coiled coil domain of MTMR5; this indicated that
the dead phosphatase MTMR5 can actually target the active phosphatase MTMR2 to the
correct subcellular localization.
Since the initiation of this project, MTM1 has been found to interact with the inactive
phosphatase MTMR12 72. This discovery came with the cloning of a new gene, 3-PAP
(MTMR12). When first isolated from rat brain lysates, MTMR12 was predicted to be an
inactive phosphatase, yet in platelets it immunoprecipitated with a protein that demonstrated phosphatase activity towards PI3P 71. A large scale immunoprecipitation of
endogenous MTMR12 was performed from K562 platelet cells and the resultant band
that coprecipitated was subjected to mass spectroscopy; it was found to be myotubularin,
explaining why initial attempts to purify the inactive phosphatase MTMR12 yielded
phosphatase activity. This interaction was confirmed by coimmunoprecipitation of both
proteins using an antibody to myotubularin. Coimmunoprecipitation also revealed that
the SID of MTMR12 was necessary for the interaction with myotubularin. Additionally,
overexpressed tagged MTM1 protein in Cos-7 cells localized to the plasma membrane
and cells demonstrated filopodia; when MTMR12 was overexpressed with MTM1, the
filopodia and subcellular localization at the plasma membrane were absent. This effect
31 was not seen with the overexpression of MTMR12 with a deleted SID domain. This
suggested to the investigators that, much like MTMR2 and MTMR5, MTMR12 regulates
the subcellular localization of MTM1.
Our initial strategy to identify proteins with which myotubularin interacted was to screen
a pretransformed cDNA library using the yeast 2-hybrid method 74. There are two basic
systems available for performing yeast 2-hybrid: LexA and Gal4 based systems. The
system we chose is based on the Gal promoters in yeast (Figure 2.1A). Biologically, the
Gal promoters are activated by the Gal4 transcription factor which is composed of a DNA
activation domain (DNA-AD) and DNA binding domain (DNA-BD). The principle
behind the yeast 2-hybrid technique is that the DNA-AD and DNA-BD can actually be
part of separate proteins and still function to activate the Gal4 promoter if brought into the physical vicinity of one another, as with a protein-protein interaction. Traditionally in
yeast 2-hybrid, the gene of interest is fused to the DNA-BD by cloning its cDNA (“bait
vector”), and a cDNA library is generated and ligated to a vector providing the DNA-AD
(“fish vector”). These two plasmids are then cotransformed into a yeast strain that is
engineered with the Gal promoter upstream of lacZ and nutritional reporter genes. In
yeast colonies where a protein from the library fused to the DNA-AD is interacting with
the protein of interest fused to the DNA-BD, the DNA-AD and DNA-BD motifs come
together and activate the reporter genes. In our system, there is not only lacZ, but also
ADE2 and HIS3 nutritional reporter genes. The LexA based system uses the same
principle as the Gal based system except the BD is the prokaryotic protein LexA, the
32 activation domain is the acidic E coli peptide B42, and the reporters for the system are tethered to LexA operators 75.
Rather than the traditional method of building a cDNA library for cotransformation with our DNA-BD vector, we chose to use a human fetal brain library pretransformed into
Y187, a MATα yeast strain. This strategy necessitated the development of not only a
DNA-BD vector containing our gene of interest, but also a DNA-BD or “bait” strain to be mated with the library DNA-AD or “fish strain” (Figure 2.1B). Thus, unlike traditional systems, our system relies on the ability of the pretransformed DNA-AD strain or “fish” strain to mate with the DNA-BD strain. While a human skeletal muscle library may have been ideal for this screening, it was unavailable. Thus, we chose brain because this tissue is associated with a large number of transcripts.
33 A Gal4 protein
DNA- DNA- BD AD Gal 1 UAS lac Z
AD Bait Fish DNA
Gal 1 UAS lac Z B MTM1 341-544
DNA- “Bait” BD (trp1)
In MATa PJ692a Yeast Mating & Plate PRETRANSFORMED trp-/leu-/ade-/his- cDNA library ] X-gal “Fish” AD (leu2)
In MATα Y187
Figure 2.1: A. Traditional Yeast 2-Hybrid Scheme. The Gal4 protein consists of a DNA-BD and a DNA-AD and activates the Gal 1 promoter in yeast. However, the DNA-BD and DNA-AD domain need not be part of the same protein but only in close proximity to one another to activate the Gal 1 promoter. This principle was exploited in the yeast 2-hybrid technique by fusing MTM1 to a DNA- BD and screening for protein-protein interactions with a cDNA library fused to a DNA-AD. B. Scheme of Mating Based Yeast 2-Hybrid System from Clontech. The gene of interest is cloned into the DNA-BD vector and transformed into PJ692A (MATa). The cDNA library was cloned into the DNA-AD vector and transformed into Y187 (MATα) commercially by Clontech. The two strains are mated together and plated on QDO (SD-trp/-leu/-his/-ade) to detect the presence of the 2 vectors (trp-/leu-) and the activation of reporter genes indicating a positive protein-protein interaction (ade- /his-). Clones selected by QDO are then tested for lacZ activity, the third reporter gene in this system. Figure modified from Clontech Technical Materials.
34 2.2 Materials & Methods
2.2.1 General Reagents.
PCR and sequencing reagents were obtained from Perkin Elmer. All primers were
generated by Sigma-Genosys and all restriction enzymes were obtained from New
England Biolabs. All chemicals were obtained from either Sigma or Acros (Fisher).
Tissue culture reagents were obtained from Gibco. Reagents for yeast media were obtained from Fisher, with the exception of Synthetic Dropout Supplements which were from Clontech. All parent vectors used in this work are described in Appendix 1. All E. coli strains used are described in Appendix 2.
2.2.2 Yeast 2-Hybrid:
2.2.2.A. Yeast Strains and General Yeast Protocols:
i. Media: YPDA was used to grow all yeast strains when nutritional selection was not
needed. This media consisted of 20g/L peptone, 10g/L-yeast extract and 20g/L agar if
plates were prepared. The pH was adjusted to pH 5.8 and the solution was autoclaved
prior to the addition of 50mL of a sterile 40% glucose solution and 15mL of a sterile
0.2% adenine hemisulfate solution. To test strains for phenotype or transformation,
nutritional selection was desired and accomplished with the use of synthetic dropout (SD)
35 media supplemented with the appropriate amino acids. SD media consisted of 6.7g/L
yeast nitrogen base without amino acids, plus the prescribed amount of dropout supplement powder according to Clontech’s instructions (Table 2.1). The solution pH was adjusted to 5.8 and autoclaved prior to the addition of sterile glucose solution to a final concentration of 2%. Synthetic dropout media included single dropout (trp-, leu- or ura-) for phenotyping haploid transformants, double dropout (DDO= trp-/leu-) for selecting mated diploids, and quadruple dropout (QDO=trp-/leu-/ade-/his-) for selecting diploids with a positive protein-protein interaction. All yeast strains were grown at 30oC.
Amount of Dropout Synthetic Supplement Use Dropout Media Required (g/L)
Trp- 0.74 Select for presence of DNA-BD
Leu- 0.69 Select for presence of DNA-AD
Ura- 0.77 Differentiate Y187 from PJ692A
Select for mating of DNA-BD & DNA- DDO 0.64 AD strains Select for mating of DNA-BD & DNA- QDO 0.60 AD strains & positive yeast 2-hybrid reaction
Table 2.1 Amount of Dropout Supplement Required for Each Selective Media Used in Yeast 2-Hybrid. Dropout supplements (Clontech) are added at different amounts per liter of synthetic dropout media (SD). Double Dropout = DDO = SD/-trp/-leu. Quadruple Dropout = QDO = SD/-trp/-leu/-his/-ade.
36 ii. Strains: The yeast 2-hybrid system used in these experiments was Matchmaker 3 with a human fetal brain Matchmaker Pretransformed Library (Clontech). See Table 2.2 for parent strain genotype and Table 2.3 (p. 61) for explanation of strains generated from
parent strains, their use and their nutritional selection markers.
Parent Strain Genotype Use
PJ692A MATa, trp1-901, leu 2-3, 112, ura3-52, his3-200, gal4∆, gal80 ∆, DNA-BD LYS2:: GAL1uas-GAL1tata-HIS3, GAL2uas-GAL2tata-ADE2
Y187 MATα, ura3-52, his 3-200, ade2-101, trp1-901, leu2-3, 112, DNA-AD gal4∆, gal80 ∆, met-, URA3:: GAL1uas-GAL1tata-lacz
Table 2.2 Genotypes of Parent Strains used in Yeast 2-Hybrid. PJ692A is the parent strain for all DNA-BD or “bait” strains. Y187 is the parent strain for all DNA-AD strain or “fish” strains. All resultant strains made from these two parent strains along with their additional features are listed in Table 2.3.
2.2.2.B. Cloning:
i. Generation of Bait Vectors for Use in Yeast 2-Hybrid Screens: To generate a yeast
fusion protein coding for the Gal4 DNA binding domain (DNA-BD) fused to the
phosphatase and SID domains of myotubularin, pGBKT7 (Clontech) was used as the
parent vector. The source of human MTM1 cDNA used as a template for PCR was Ig38,
a human EST containing the entire MTM1 coding sequence (obtained from the laboratory
37 of J.L. Mandel). A fragment containing amino acids 341-544 of the human MTM1 cDNA
was generated by PCR using primers containing NcoI linkers for subsequent cloning.
The primers used to amplify this fragment of MTM1 were NcoDom3+4for
(5’CATGCATCCATGGAGTCTACT-3’) and NcoDo3+4rev (5’CGATATCCATGGA-
TTCGGCTG-3’), with NcoI restriction sites underlined. Reactions were carried out in
25ul total volume with 60ng of Ig38, 1.5 mM MgCl2 , 0.4 uM primers, 0.2 uM dNTPS and 1 unit Amplitaq Gold (Perkin Elmer). PCR was performed as follows: 95 oC for 12 min for one cycle; denaturation at 94 oC for 1 min, annealing at 65oC for 1 min, elongation for 1 min at 72 oC for 35 cycles; and a final extension at 72 oC for 7 min.
Once amplified, this PCR fragment was directly subcloned into pCR2.1 using a TA
Cloning Kit (Invitrogen). Briefly, 1uL of fresh PCR product was mixed with 1uL of
Ligation Buffer, 50ng of pCR2.1 (Invitrogen) and 1uL of T4 ligase in a 10uL reaction
volume. The reaction was incubated at 14oC overnight then transformed into DH5α One
Shot E.coli (Invitrogen). For transformation, 100uL of One Shot cells were incubated with 2uL of the ligation for 30 minutes on ice and then shocked at 42oC for 30 seconds.
Cells were allowed to recover in 250uL of SOC medium (2% tryptone, 0.5% yeast
o extract, 10mM NaCl, 2.5mM KCl, 10mM MgCl2, 20mM glucose) for 1 hr at 37 C with shaking at 250rpm (rotations per minute), before plating on LB (1% tryptone, 0.5% yeast extract, 1% NaCl, pH 7) plates supplemented with 100ug/mL ampicillin. Individual clones were propagated in liquid culture, and their DNA was isolated by Qiagen miniprep and analyzed by EcoRI and NcoI restriction digests. 38 The resulting vector, pCR2.1-hMTM1341-544 was used as a source of insert for the final ligation yielding pGBKT7-MTM1341-544. The hMTM1341-544 insert was prepared by
o cutting 20ug of pCR2.1-hMTM1341-544 in a 50uL reaction at 37 C for 4 hours with 5uL of
NEB (New England Biolabs) buffer #4 and 50 units of NcoI. The restriction fragment hMTM1341-544 was excised from a 1% TAE-agarose gel (40mM tris acetate, pH 8.0, 1mM
EDTA, 1% agarose) and purified through glass wool by centrifuging at 6000rpm for 30
minutes. The fragment was further purified by ethanol precipitation and used as insert in a ligation with prepared pGBKT7. To prepare vector, 10ug of pGBKT7 was cut in a
100uL reaction at 37oC for 4 hours with 10uL of NEB buffer #4 and 50U of NcoI. The
cut vector was ethanol precipitated and dephosphorylated in a 100uL reaction at 65oC for
1 hr with 3uL of Bacterial Alkaline Phosphatase (Gibco) and 10uL of 500mM tris, pH8.
A final purification was performed with phenol/chloroform treatment and ethanol precipitation before using as the vector for ligation. To ligate, 140ng of cut and dephosphorylated pGBKT7 was mixed with 50ng of hMTM1341-544 insert (approximate molar ratio insert:vector = 4:1), 1uL of 10x NEB ligation buffer and 1uL of T4 DNA
Ligase (NEB). Ligations were incubated overnight at 14oC before transforming 2uL into
Gibco Library Efficiency DH5α E. coli (Gibco) as described above. Transformations were plated on LB plates supplemented with 50ug/mL kanamycin and grown overnight at
37oC. Individual clones were screened for the correct ligation product by propagation in
liquid culture, followed by Qiagen miniprep and restriction digests with NcoI and EcoRI.
From this ligation, pGBKT7-hMTM1341-544 was identified and sequenced. Sequencing
was performed in 10ul total volume using DiRhodamine Cycle Sequencer Terminator
39 Reaction reagents as recommended by Perkin Elmer. The cyclo-sequencing reaction was
carried out as described 2.2.2 D ii c. The T7 primer (5’TAATACGACTCA-
CTATAGGG3’) was employed for forward sequencing, while 3’DNA-BD (5’TTTTCG-
TTTTAAAACCTAAGAGTC3’) was used in the backward reaction. All sequencing was
performed on either an ABI 377 automated sequencer or ABI 3100 Genetic Analyzer by
the Children’s Hospital Sequencing Core.
ii. Generation of pGBKT7-hMTM1341-544 (C375S) by Site-Directed Mutagenesis: The
Quikchange XL System (Stratagene) was used to perform site-directed mutagenesis on
cysteine 375 of hMTM1, converting it to a serine. Primers C375SMTM1F (5’-
CAGTGCTTGTGCATTCTAGTGACGGATGGGACAGG-3’) and C375SMTM1R (5’-
CCTGTCCCATCCGTCACTAGAATGCACAAGCACTG-3’) were generated and
PAGE purified by Sigma Genosys for use in the PCR reaction. These primers were
designed to generate a BtsI site in the final product for screening purposes. The reaction
was carried out in 50uL total volume with 10ng of pGBKT7-hMTM1341-544, 125ng of
each primer, 5uL of 10x buffer, 1uL of dNTP mix, 3uL of Quik Solution (Stratagene),
and 1uL of Pfu (Stratagene). PCR was carried out as follows: 95oC for 1 min for one
cycle; denaturation at 95oC for 50 sec, annealing at 60oC for 50 sec, elongation for 16
min at 68oC for 18 cycles; and a final extension at 68oC for 7 min. After amplification,
the entire PCR reaction was digested with 10U of DpnI at 37oC for 1 hr to remove the
parent vector. For transformation, 45uL of XL10 Gold Ultra competent E. coli
(Stratagene) were pretreated with 2uL of β-mercaptoethanol for 10 min and then mixed
40 with 2ul of the digested PCR reaction on ice for 30 min. A heat shock was performed for
30 sec at 42oC followed by the addition of 500uL of NZY+ broth (1% NZ amine, 0.5%
yeast extract, 0.5% NaCl, 4% glucose, 12.5 uM MgCl2, 12.5MgSO4 pH7.5) and a
recovery period of 1 hr at 37oC, with shaking at 225rpm. Transformations were incubated overnight on LB/ 50ug/mL kanamycin plates at 37oC. DNA was isolated from multiple
colonies and screened for mutagenesis by restriction digest with 10U of BtsI in a 20uL
reaction (NEB Buffer #4 + Bovine Serum Albumin (BSA)). A clone displaying the
predicted restriction pattern was sequenced as described for pGBKT7-hMTM1341-544 and once verified was named pGBKT7-hMTM1341-544 (C375S).
2.2.2.C. Strain Development:
i. Generation of Bait Strains PJ692A- pGBKT7-hMTM1341-544 and PJ692A-pGBKT7-
hMTM1341-544 (C375S): To make competent PJ692A yeast cells, 2-3 large colonies from
a stock plate were inoculated into 50mL of YPDA media and grown overnight at 30oC
with shaking at 250rpm until OD600 > 1.5. 30mL of this solution was diluted to an OD600
of 0.2-0.3 (300mL) with YPDA and then allowed to grow at 30oC with shaking at
250rpm until an OD600 of 0.4-0.6 was achieved. Cells were centrifuged at 1000g for 5
min and washed once with 25mL of distilled water. The pellet was then resuspended in
1.5mL of TE/LiAC (10mM tris-HCl, 1mM EDTA, 100mM lithium acetate, pH 7.5),
yielding competent cells.
41 To transform the PJ692A line, 100uL of competent cells were mixed with 100ng of either
pGBKT7-hMTM1341-544 or pGBKT7-hMTM1341-544 (C375S) and 100ug of herring testes
carrier DNA. Additionally, 0.6mL of PEG/LiAC (TE/LiAc with 40% PEG4000) were
added, and the cells were incubated at 30oC, with 200rpm shaking for 30 min. DMSO
(70uL) was added and the cells were gently inverted and then shocked at 42oC for 15
min. The cells were rested on ice for 2 min and then concentrated at 14000g for 5 sec.
They were resuspended in 500uL of TE (10mM tris-HCl, 1mM EDTA, pH7.5) and plated
on SD-trp plates to select for transformants. To confirm the presence of the bait vector in
the PJ692A strain, individual clones were amplified by whole cell yeast PCR using the
vector specific primers “T7” and “3’DNA-BD” as described in 2.2.2 Bi. Clones were propagated in selective liquid culture and 1uL was used as template. PCR was performed as described above in section 2.2.2 D.ii.b. The strain PJ692A-pGBKT7 was generated in the same manner.
ii. Verification of Yeast Bait Strains:
a. Growth: Growth curves were measured for the strains PJ692A-pGBKT7-hMTM1341-
544, PJ692A-pGBKT7-hMTM1341-544 (C375S), PJ692A-PGBKT7, and PJ692A-pVAC3-1.
Colonies from each strain were inoculated from a stock plate to 3mL cultures of SD-trp
and grown overnight at 30oC. From these overnight cultures, 0.75mL was diluted to
o 10mL, incubated at 30 C with shaking at 250rpm, and measured for growth via OD600
over 0-26 hrs.
42 b. Autonomous activation: The PJ692A-pGBKT7-hMTM1341-544 (C375S) or PJ692A-
pGBKT7-hMTM1341-544 strains were plated alongside the PJ692A-pGBKT7 strain on
SD- trp, SD -ade/-trp, and SD -his-/-trp plates. Growth on trp- plates indicated the phenotype of the strains, whereas ade-/trp- and his-/trp- plates tested for the activation of the ADE2 and HIS3 reporters respectively.
c. Mating efficiency: To test mating efficiency, one large (>1mm) colony from the
appropriate stock plate was vortexed for 1 min in YPDA. From this initial liquid culture,
0.4mL was removed and diluted to a total volume of 3mL in YPDA. Haploid cultures contained either the bait strain for testing or Y187-pTD1-1, while mating cultures contained both bait and fish strains. Cultures were “mated” in a 125mL Erlenmeyer flask overnight at 30oC with shaking at 50rpm. Single strain haploid cultures were plated onto
SD-trp, SD-leu, and SD-ura plates at dilutions of 1:4000 and 1:8000 to verify the
phenotype. Mating cultures were plated at dilutions of 1:100 and 1:1000 onto SD-trp,
SD-leu, SD-ura, and SD-leu/-trp. Viability was determined for the bait and fish strains
with the following equation: viability (cfu/mL) = (# colonies on Single Dropout plate *
1000 uL/mL)/(volume plated (uL)* dilution factor). Whichever strain had the lowest viability was considered the limiting strain. Viability of the diploids was determined by using the # of colonies on the SD-trp/-leu plate in the viability equation described above.
Mating efficiency is defined as the percent diploid and was calculated with the following
equation: Mating efficiency (ME) = (viability of diploid * 100)/ (viability of the limiting strain). The mating efficiency of the bait strains PJ692A-pGBKT7-hMTM1341-544
43 (C375S) or PJ692A-pGBKT7-hMTM1341-544, was performed alongside control matings with PJ692A-pGBKT7 and PJ692A-pVA3-1.
2.2.2.D. Yeast 2-Hybrid Screening:
i. Library mating: To screen the human fetal brain cDNA library pretransformed into the
Y187 fish strain, the following procedure was performed: Either PJ692A-pGBKT7-
hMTM1341-544 (C375S) or PJ692A-pGBKT7-hMTM1341-544 was inoculated into a 50mL
o SD-trp culture and grown overnight with shaking at 250 rpm, 30 C until the OD600 was greater than 0.8. The culture was centrifuged at 1000g for 5 min and resuspended in 2-3 mL of media. The pretransformed cDNA library was thawed on ice and 10uL were removed for later titering. The remainder of the 1mL aliquot of library stain was added to the bait strain and the volume was brought up to 45mL with 2xYPDA plus 60ug/mL kanamycin. This culture was incubated in a 2L Erlenmeyer flask for 24 hrs at 30oC, with shaking at 40 rpm. The portion of the pretransformed library aliquoted for titering was diluted 1:100, 1:500 and 1:10,000 and plated on SD-leu plates for incubation at 30oC for
3-5 days. After 24 hours of mating, the mating culture was centrifuged for 10 min at
1000g and resuspended in 10mL of 2xYPDA/60ug/mL kanamycin. A small amount of this resuspension was diluted 1:10, 1:100, 1:1000 and 1:10,000 for plating onto control plates SD-trp, SD-leu, SD- trp/-leu. The remainder of the resuspension was plated onto
60 quadruple dropout (SD-leu/-trp/-his/-ade) large plates using 200ul of culture per plate.
These plates were incubated at 30oC for 3 to 4 weeks.
44
ii. Isolation of positive clones:
a. Generation of Stock Plates and X-gal Testing: As colonies appeared and grew to 2 mm
in diameter, they were removed from the mating plates and restreaked onto another
quadruple dropout plate to generate a stock plate, and confirm that the colonies were able to grow on quadruple dropout media. For X-gal testing, colonies were lifted off of the master plate with a nylon membrane that was subsequently submerged in liquid nitrogen for 10 sec. The membrane was thawed and placed colony side up onto 1mm Whatman paper soaked in Buffer Z (60mM Na2HPO4, 40mM NaH2PO4, 10mM KCl, 1mM MgSO4, pH 7.0) with β-mercaptoethanol (270uL/100mL buffer Z) and X-gal (Fisher) (1.67mL of
20mg/mL in dimethylformamide (DMF) / 100mL Buffer Z). The filter and Whatman paper were incubated at 30oC for 1-8 hr, checking for color change every hour. All X-gal
experiments on master plates were performed alongside a negative control plate to verify
that false positives were not appearing after prolonged incubation. Additionally, a control
mating plate (PJ692A-pVA3-1 x Y187-pTD1-1) was used as a positive control.
b. Whole Cell PCR to Amplify cDNA Inserts from Positive Clones: Colonies
testing positive for X-gal activation were grown from the master plate in 500uL of
quadruple dropout media overnight at 30oC, with shaking at 250rpm. For template in a
whole cell PCR reaction, 10ul of the overnight culture was centrifuged at 14,000g for 6
min and decanted, leaving a pellet of yeast cells. The primers used in these reactions were
45 ADLDINSERTSCREEN (5’-CTATTCGATGATGAAGATACCCCACCAAACCC-3’)
and ADBACHSCREEN (5’-GTGAACTTGCGGGGTTTTTCAGTATCTACGA-3’)
which amplify the cDNA inserts from the parent vector (pACT2). The pTD1-1 vector
(10ng) was used as a positive control alongside these PCR reactions. Reactions were
carried out in 25ul total volume with 1.5 mM MgCl2, 0.4 uM primers, 0.2 uM dNTPS and
1 unit Amplitaq Gold (Perkin Elmer). PCR was carried out as follows: 95 oC for 12 min
for one cycle; denaturation at 94 oC for 30 sec and annealing/elongation at 68oC for 3 min
for 35 cycles; and a final extension at 68 oC for 7 min. PCR products were visualized on
1% agarose gels and cloned into pCR2.1-TOPO (Invitrogen TOPO TA Cloning Kit). To
perform this cloning, 2ul of fresh PCR product was mixed with 1uL of pCR2.1-TOPO,
1uL of salt solution and 1uL of distilled water. The mixture was incubated at room
temperature for 5 min before transformation into One-Shot TOP10 cells. For
transformation, 2uL of the mixture were incubated on ice with 60uL of TOP10 cells for 5
min. The cells were shocked at 42oC for 30 sec and allowed to recover in 250uL of SOC media at 37oC, with shaking at 250rpm for 1 hr. The transformation mixture was plated
on LB/ amp 100ug/mL plates and incubated overnight at 37oC. Individual colonies were
grown in LB/amp liquid culture, miniprepped (Qiagen), and cut with EcoR1 to verify the
presence of insert.
c. Clone Sequencing: For sequencing of yeast 2-hybrid clones in pCR2.1 TOPO, either
25ng of the forward T3 primer (5’-ATTAACCCTCACTAAAGGGA-3’) or reverse T7 primer (5’TAATACGACTCACTATAGGG-3’) was added to 4uL of DiRhodamine, 3uL
46 of distilled water and 2uL of Qiagen miniprep (approximately 500ng). The cyclo- sequencing reaction was performed as follows: 96oC for 2 min for 1 cycle; denaturation
at 96oC for 10 sec, annealing at 50oC for 5 sec, and elongation at 60oC for 4 min for 25
cycles.
d. Yeast Plasmid Rescue: Inserts in some clones identified in yeast 2-hybrid screenings
were unable to be amplified by whole cell PCR. For these cDNAs, yeast plasmid
recovery was used. To recover the DNA-AD plasmid from individual clones, a 2mL
liquid culture was grown overnight in QDO media and centrifuged for 5 min at 16,000g.
The pellet was incubated at 37oC, with shaking at 230rpm for 1 hr with 10uL of lyticase
(25u/uL) (Fisher). The sample was vortexed for 1 min after the addition of 10uL of 20%
SDS, and put through one –20oC freeze thaw cycle. The sample was further purified
using a Chroma-Spin (Clontech) column by placing the sample on the column and
centrifuging at 700g for 5 min. Once column purified, 5uL of the sample were
transformed into 200uL of XL-1 blue E.coli by incubating for 30 min on ice, shocking for
45 sec at 42oC, and recovering for 1 hr in 900uL of LB at 37oC, with shaking at 250 rpm.
Transformations were grown on LB/100ug/mL amp plates at 37oC overnight. Resultant
colonies were propagated in liquid culture and plasmid DNA was isolated via Qiagen miniprep. To verify the presence of a library cDNA insert, plasmid DNA was restriction digested with BglII. DNA-AD vectors were sequenced as described above with the forward primer GalADseq (5’-TACCACTACAATGGAT-3’) on the parent vector pACT2.
47
e. Specific yeast2-hybrid: To retest the DNA-AD or fish strains obtained in both
wildtype and mutant library mating, Y187-PP2Ac and Y187-SCG10 strains were mated
to PJ692A-pGBKT7, PJ692A-pGBKT7-hMTM1341-544 (C375S) or PJ692A-pGBKT7- hMTM1341-544. Y187 based strains were generated as described in section 2.2.2Ci.
Colonies of all individual strains were resuspended in 1mL of YPDA and plated on SD-
trp and SD-leu plates to verify phenotype. Matings were performed in 50mL conical
tubes with a total volume of 1.2mL. Experimental matings were incubated overnight at
30oC, with shaking at 50rpm along side a control mating between PJ692A-pVA3-1 and
Y187-pTD1-1. X-gal assays were performed as described in section 2.2.2 D.ii.a.
2.2.3 Coimmunoprecipitation Strategy:
2.2.3.A Cloning:
i. Cloning of pcDNA4/HisMax TOPO-SCG10 and pcDNA4/HisMax TOPO-PP2Ac: To generate an SCG10 cDNA insert for ligation into the pcDNA4/HisMax-TOPO
(Invitrogen) expression vector, SCG10 was amplified from pACT2 using the primers
SCG10F (5’ATGGCTAAAACAGCAATGGCC-3’) and SCG10R (5’-
CTTGCTTCAGCCAGACAGTTC-3’). Reactions were carried out in 25ul total volume with 1.5 mM MgCl2, 0.4 uM primers, 0.2 uM dNTPS and 1 unit Amplitaq Gold. The
PCR reaction was performed as follows: 95oC for 12 min for 1 cycle; denaturation at
48 94oC for 30 sec, annealing at 57oC for 30 sec, and elongation at 72oC for 30 sec for 35 cycles; and a final extension at 72oC for 7 min. The SCG10 PCR product (4uL) was mixed with 1uL of Salt Solution and 1uL of pcDNA4/HisMax-TOPO and allowed to incubate at room temperature for 5 min before transforming into One Shot TOP10 cells as described in 2.2.2 D.ii.b. To verify properly ligated clones, the restriction digest pattern with PstI was observed. The clones were sequenced as described in 2.2.2 D.ii.c. using either the T7 forward primer (5’-TAATACGACTCACTATAGGG-3’) or the BGH reverse primer (5’-TAGAAGGCACAGTCGAGG-3’).
To generate a PP2A cDNA insert for ligation into the same expression vector, PP2A was amplified via RT-PCR. Human total RNA (5ug) generated by another member of the lab was mixed with 500ng of oligo-dT (Invitrogen) and 7uL of RNase-free water
(Aquanase). The mixture was incubated at 70oC for 10 minutes before the addition of
1uL RNAGuard (Invitrogen), 1uL random hexamers (500mM), 4uL of 5x First Strand
Buffer, 2uL of 0.1M dTT and 1uL of 10mM dNTP (Invitrogen RT Kit). The mixture was incubated at 42oC for 2 minutes followed by the addition of 1uL of Superscript II
(Invitrogen). This RT reaction proceeded for 50 minutes at 42oC and was then used for
PCR with primers PP2AF (5’-ATGGACGAGAAGGTGTTCACC-3’) and PP2AR (5’-
TCATTACAGGAAGTAGTCTGGG-3). PCR was carried out in 25ul total volume with
2uL of RT, 1.5 mM MgCl2, 0.4 uM primers, 0.2 uM dNTPS and 1 unit Amplitaq Gold.
The PCR reaction was performed as follows: 95oC for 12 min for 1 cycle; denaturation at
94oC for 30 sec, annealing at 59oC for 30 sec, and elongation at 72oC for 1 min for 35
49 cycles; and a final extension at 72oC for 7 min. The resultant PCR fragment was cloned
into pCDNA4/HisMax-TOPO and sequenced as described above.
ii. Generation of pCMV tag2b-hMTM1 and pTRE2pur-hMTM1: Full length human
MTM1 cDNA was amplified from Ig38 (200ng) with primers EcoRVFlagF (5’-
GGGATATCATGGCTTCTGCATC-3’) and EcoRVFlagR (5’-GGGATATCTC-
AGAAGTGAGTTTG -3’) in a 25ul reaction volume with 1.5 mM MgCl2, 0.4 uM primers, 0.2 uM dNTPS and 1 unit Pfu (Stratagene). The reaction was performed as follows: 95oC for 5 min for 1 cycle; denaturation at 95oC for 45 sec, annealing at 55oC for
45 sec, and elongation at 72oC for 4 minutes for 30 cycles; and a final extension at 72oC
for 10 min. The resultant PCR product was purified through a Qiagen PCR purification
column and then cut with EcoRV in 50uL reaction for 1 hr. The restricted PCR product
was purified again with a PCR purification column and used as insert in a ligation with
pCMVtag2B (Stratagene). The vector was prepared and used in a 10uL ligation with the
human MTM1 PCR product and described in 2.2.2.B.i.. Upon sequencing, this insert was
found to have a H315Y mutation in MTM1 which was corrected to wildtype by site-
directed mutagenesis as described in 2.2.2 B.ii. with primers hMTM1Y315Htop (5’-
CTTTTCTTCTTAGACATTCATAATATTCATGTTATGCGG-GAATC-3’) and
hMTM1Y315Hbot (5’GATTCCCGCATAACATGAATATTATGAATGTCTAAGAAG-
TTTTG-3’). The wildtype insert was confirmed to be mutation free and in the correct
frame and orientation with the N-terminal Flag tag of the parent vector by sequencing as
described in 2.2.2D.i. using primer CMVF (5’-CGCAAATGGGCGGTAGGCGTG-3’).
50 From pcDNAtag2B-hMTM1, the Flag-hMTM1 cassette was excised with NotI and SalI digestion, and the sticky ends were filled in with T4 polymerase (NEB) and 100mM dNTPs for 20 min at 12oC. This fragment was ligated into pTRE2pur (Clontech) restricted with EcoRV and further prepared for ligation as described in 2.2.2.B.i.
Ligations and transformations were carried out as described in 2.2.2.b.i. Correctly ligated vectors were determined from EcoRI digestion and sequence was confirmed as described in 2.2.2D.ii.a using primers pTRE2purCMVF (5’-CTCCATAGAAGACACCGGG-3’) and pTRE2purPolyAr (5’-CTTTGCCCCCT-CCATATAAC-3’).
2.2.3 B. Generation of Stable Clones that Express Flag-MTM1 Protein in a Tet-Inducible
Fashion: To generate stable cell lines producing Flag-MTM1 protein in a tet-inducible fashion, HeLa Tet-on cells were purchased from Clontech. These HeLa cells are stably transfected with the pTet-on vector and maintained in DMEM (Mediatech) with 10%
FBS (Fetal Bovine Serum; tet- free/Clontech), 100ug/uL G418, 50 I.U./mL penicillin,
50ug/mL streptomycin, and 2mM L-glutamine. HeLa-Tet-on cells (500,000 cells/6well) were transfected with 1ug of pTRE2pur-hMTM1 using Effectene (Qiagen) (10uL) as described in 2.2.3D. Forty-eight hours after transfection, the cells were split into puromycin (0.5ug/mL) containing media at dilutions of 1:10, 1:20, 1:50, 1:100. After 10-
14 days of puromycin selection, clones were picked off of the master plates and seeded into duplicate 24 well plates for expression testing. To test for tetracycline inducible expression of Flag-MTM1 protein, individual clones were treated with doxycycline (2000 ug/mL) for 48 hours and then lysed with Passive Lysis Buffer (Promega) and
51 immunoblotted. A second transfection was performed and clones were selected in 0.1 and 0.2 ug/mL puromycin to yield more clones.
2.2.3C. Immunoblotting and Antibodies:
For all immunoblotting, a discontinuous mini-gel protocol was used. Separating gels contained 10-12% acrylamide (Bio-Rad), 0.26-0.32% bis-acrylamide (Bio-Rad), 125mM
tris (pH 8.8), 0.1% SDS, and were polymerized with 10uL of TEMED (Bio-Rad) and
50uL of 10% ammonium persulfate per 15mL of gel. A 3.9% stacking gel was used at pH
8.8 (375mM tris, 0.1% SDS). Proteins for SDS-PAGE were boiled in SDS Loading
Buffer (125mM tris, pH 6.8, 4% SDS, 10% glycerol, 0.006% bromophenol blue, 2% β- mercaptoethanol) at 100oC for 5 min before loading onto the gel. Protein concentrations
were detected by Bradford analysis with 200uL of Protein Assay Reagent (BioRad) in
800uL of 0.25M tris, pH 7.5 and 4-16uL of total protein. Spectrophotometer readings for
Bradford analysis were taken on an Eppendorf Biophotometer and compared against a
standard curve using Bovine Serum Albumin. 10-50ug of proteins were routinely loaded
for each sample. Gels were electrophoresed with a BioRad Mini-Protean II in Running
Buffer (25mM tris, 1% SDS, 200mM glycine) at 200V and monitored for separation with
Precision Plus Protein Markers (BioRad) or Kaleidoscope Markers (Amersham). Gels
were transferred to Hybond ECL nitrocellulose membranes with the Mini-Protean II in
Transfer Buffer (25mM tris, 190mM glycine, 20% v/v% methanol) at 350mA (constant)
for 1 hr at room temperature. Membranes were blocked overnight with 5% milk in TBST
52 (10mM tris pH 7.5, 95mM NaCl, 0.1% Tween-20). Membranes were incubated with
primary antibodies diluted in blocking buffer for 1 hr at room temperature. Membranes
were washed 3 x 5 minutes with TBST and incubated with secondary antibodies diluted
in blocking buffer for 1 hr at room temperature. Membranes were washed again as
previously described before treatment with a 1:1 mixture of chemiluminescent reagents A
and B (solution A: 2.5mM 3-aminophthalhydrazide, 0.4mM coumaric acid, 100mM tris,
pH 8.8/ solution B: 0.02% hydrogen peroxide, 100mM tris pH 8.8). Membranes were
exposed to film for 10 sec to 10 min prior to developing.
For detection of Flag-tagged proteins, anti-Flag mouse M2 (Sigma #F3165) or anti-Flag
rabbit polyclonal (Sigma #F7425) antibodies were used at a 1:1000 dilution. Anti-Xpress
or Anti-HisG antibodies (Invitrogen) were used at a dilution of 1:5000 to detect SCG10
or PP2Ac proteins. Mouse and rabbit secondary HRP-conjugated (horseradish
peroxidase) antibodies were purchased from Cell Signaling and used at a 1:1000 dilution.
2.2.3 D. Coimmunoprecipitation Studies: For coimmunoprecipitation of either SCG10 or
PP2Ac with Flag-MTM1 protein, HeLa clone 83 was induced with doxycycline (2000
ug/mL) and transfected with SCG10 or PP2Ac expression vectors. Clone 83 cells were
plated the night before transfection at a density of 4 x 105 cells per well of a six well dish
in DMEM + 10%FBS, 50 I.U/mL penicillin, 50ug/mL streptomycin, 2mM-glutamine,
0.2ug/mL puromycin and 100ug/mL G418 (Fisher). The next day, the cells were washed
with PBS (137mM NaCl, 2.7mM KCl, 10mM Na2HPO4, 2mM KH2PO4, pH 7.4) and left
53 in 1.5mL of media. For transfection, either pcDNA4/HisMax TOPO-SCG10 or
pcDNA4/HisMax TOPO-PP2Ac (1ug) was added to 100uL of EC Buffer and 3.2uL of
Enhancer and incubated for 5 min at room temperature in a 15mL polystyrene conical
tube. After vortexing for 10 sec, 10uL of Effectene was added and the sample was
incubated at room temperature for an additional 10 min. Media (0.6mL) was mixed with
the sample and added dropwise to the well. Transfected cells were incubated at
o 37 C/5%CO2 for either 24 or 48 hrs for SCG10 or PP2Ac, respectively. Cells were then
lysed with RIPA, Triton or Empigan buffers (1mL) at 4oC for 10 min. RIPA buffer
consists of 50mM tris pH 7.5, 1% NP-40, 0.5% deoxycholate, 0.1% SDS, 5mM EDTA,
150mM NaCl, 1mM NaF, 1mM Na4P2O7, 1mM Na3VO4 in water. Triton buffers (0.1%)
were either high salt (50mM tris pH 7.5, 100mM NaCl) or low salt (50mM tris pH 7.5,
25mM NaCl). Empigan buffers (0.1%) were either high salt (50mM tris pH 7.5, 100mM
NaCl) or low salt (50mM tris pH 7.5, 25mM NaCl). Lysates were cleared by
centrifugation at 4oC, 16000g, for 10 minutes and the supernatant was measured by
Bradford assay for total protein concentration. For each coimmunoprecipitation, 1mg of
total protein was nutated at 4oC with 2uL (9 ug) of mouse monoclonal anti-Flag antibody
(Sigma) for 1 hour. The protein complexes were precipitated with 40uL of Protein G+
agarose (Santa Cruz) by nutating at 4oC for 1 hour, followed by a 10 minute 16000g
centrifugation. The pellet was washed 5x at 4oC with the same buffer used in lysis.
Proteins were liberated from the agarose beads by boiling for 5 minutes in SDS Loading
Buffer and analyzed by SDS-PAGE and immunoblotting. Westerns were performed with
54 mouse monoclonal anti-Flag antibody (1:1000) and mouse anti-HisG antibody (1:5000;
Invitrogen).
2.2.4. GST Pulldown Assays:
2.2.4.A. Purification of GST-MTM1 Fusion Proteins:
All GST vectors used in this study were a gracious gift from Gregory Taylor and Jack
Dixon (University of California, San Diego). The pGSTx-MTM1-His6 and pGSTx-
MTM1-(C375S)-His6 plasmids were transfected into BL21(DE3) CodonPlus RIL
(Stratagene Cat. # 230245) cells according to the manufacturer’s protocols. Briefly,
100uL of competent cells were thawed on ice in pre-chilled round-bottomed tubes. Two microliters of β-mercaptoethanol from the manufacturer was added after a 1:10 dilution in distilled water and cells were incubated for 10 min on ice. The appropriate plasmids
(50ng) were incubated with the cells for 30 min on ice before shocking at 42oC for 20 sec. LB was added (1mL) and the cells were allowed to recover for 1 hr at 37oC, with shaking at 250rpm and were plated on LB plates containing ampicillin (100ug/mL) and chloramphenicol (34ug/mL). From this stock plate, 4 colonies were inoculated into 1mL of 2YT (16g tryptone, 10g yeast extract, 5g NaCl per liter of solution) with the necessary antibiotics and grown overnight at 37oC, with shaking at 250rpm. For each pulldown,
75mL of 2YT was inoculated with the 1mL starter culture and grown at 37oC, with shaking at 250rpm until the OD600nm reached 0.6-0.7. Protein expression was induced by
55 adding IPTG to a final concentration of 0.5M and incubating at 25oC, with shaking at
150rpm, overnight. Following induction, the bacteria were centrifuged at 4000rpm for 10
min at 4oC and lysed by sonication in His Lysis Buffer (50mM tris pH 8, 300mM NaCl,
20mM imidazole pH8, 0.05% β-mercaptoethanol, 1mM benzamidine, 1mM PMSF, and
1ug/mL each of pepstatin, leupeptin, and apoprotinin). Typically, 15mL of His Lysis
Buffer was used per 75mL of induced culture. Sonication was performed in 15mL
aliquots with a Sonics and Materials Vibracell sonicator using 50x 6-second pulses at
80% output and a microtip. Triton X-100 was added to 0.5% (v/v%) and the lysis was
centrifuged at 16,000g for 30 min at 4oC for insoluble debris removal. Before use in
protein purification, Ni2+-NTA-agarose (Qiagen) was washed 3x in His Lysis Buffer to
remove the storage buffer. Typically, 400uL of Ni2+-NTA-agarose were used per 75mL
culture. The soluble fraction from sonication was incubated with washed Ni2+-NTA- agarose for 2 hours at 4oC with nutation. After incubation, the Ni2+-NTA-agarose was
centrifuged at 400g, 4oC for 2 minutes and the supernatant was discarded. The pellet was
washed 5x with His Lysis Buffer plus 0.5% Triton X-100, followed by 2 washes with His
Lysis Buffer without detergent. Fusion protein was eluted by incubating the pellet with 2 x 1mL of Elution Buffer (His Lysis Buffer + 300mM imidazole, pH 8) at 4oC for 10 min.
The two fractions were pooled and filtered through a 0.2um Millex-GV filter. This protein was then immobilized and further purified on glutathione agarose.
Glutathione affinity agarose resin (Sigma G4510) was prepared as 50% slurry by washing the lyophilized powder several times and resuspending the pellet in tris-buffered saline
56 pH 8 (TBS). Unused resin was stored at 4oC with 0.1% sodium azide as a preservative.
Before immobilizing fusion protein to these beads, they were washed 3x in GST Buffer
(TBS pH 8, plus 2mM dTT, 1mM benzamidine, 1mM PMSF, and 1ug/mL each of
leupeptin, pepstatin, and apoprotinin) to remove azide. The fusion protein eluted from the
Ni2+-NTA-agarose was diluted in 2mL of GST Buffer and nutated with washed
glutathione beads for 1 hour at 4oC. Typically 800uL of glutathione agarose was used per purification. After incubation, the glutathione agarose was washed 5x with GST Buffer, yielding GST fusion protein immobilized on the agarose support.
2.2.4.B. Generation of SCG10 and PP2Ac as Probes for GST Pulldowns:
i. In vitro Transcription and Translation: Human SCG10 and PP2Ac proteins were
generated using the TnT Coupled Reticulocyte Lysate System (Promega). All plasmids
used in this system were prepared by Qiagen midi-prep followed by phenol/chloroform
treatment and ethanol precipitation. Briefly, either pcDNA4/HisMax TOPO-SCG10 or
pcDNA4/HisMax TOPO-PP2Ac (1ug) were added to 25uL of lysate, 2uL of TnT buffer,
30uCi of 35S-Methionine (Amersham), 1uL of the provided amino acids without
methionine solution, 1uL of RNase inhibitor (Invitrogen), and 1uL of T7 RNA
polymerase (Promega). The total volume was brought up to 50uL with RNase free water.
The reaction was thoroughly mixed and incubated at 30oC for 1 hr. Protein generation was confirmed by autoradiography.
57 ii. Transient Transfection and Radiolabeling HeLa Cells: HeLa cells were transfected
using the liposome-based reagent Effectene (Qiagen). Cells were plated the night before
transfection at a density of 4 x 105 cells per well of a six well dish in DMEM with
10%FBS, 50 I.U./mL penicillin, 50ug/mL streptomycin, and 2 mM L-glutamine. The next day, the cells were washed with PBS and left in 1.5mL of media. For transfection,
either pcDNA4/HisMax TOPO-SCG10 or pcDNA4/HisMax TOPO-PP2Ac (1ug) was
added to 100uL of EC Buffer and 3.2uL of Enhancer and incubated for 5 min at room
temperature in a 15mL polystyrene conical tube. After vortexing for 10 sec, 10uL of
Effectene was added and the sample was incubated at room temperature for an additional
10 min. Media (0.6mL) was mixed with the sample and added dropwise to the well.
o Transfected cells were incubated at 37 C/5%CO2 for either 24 or 48 hrs for SCG10 or
PP2Ac respectively. Four hrs prior to harvesting the cells, they were washed 3x with PBS and given serum free OPTIMEM (Gibco/Methionine free) for 30 min. After this brief period of methionine starvation, OPTIMEM with 50uCi/mL 35S-methionine was added
o to the cells. After incubating for 3 hrs at 37 C/5%CO2, cells were harvested by lysis for
10 min at 4oC with 1mL of JBC Buffer (25mM tris pH 7.6, 150mM NaCl, 1% triton X-
100, 50mM NaF, 2mM NaVO4, 60ug/mL PMSF, 4ug/mL leupeptin, 1ug/mL apoprotinin). SDS-PAGE and immunoblotting confirmed the presence of transfected
SCG10 or PP2Ac. Acrylamide gels were dried with a Savant Slab gel dryer and
autoradiography was used to visualize the quality of cellular protein labeling.
58 2.2.4.C. Pulldown Assay:
Pulldown assays were performed using approximately 1ug of either GST-MTM1-His6 or
GST-MTM1-(C375S)-His6 fusion proteins as estimated by Coomassie (Bio-Rad)
staining. The appropriate amount of fusion protein bound to glutathione agarose was
probed either with 35S-radiolabelled cell lysis from one transiently transfected well (6-
well=approximately 1mg), or 48uL of 35S-radiolabelled protein made in vitro. In the case
of probing with an in vitro protein, 200uL of Binding Buffer (25mM tris pH7.5, 10%
glycerol, 0.1% NP-40, 5mM dTT, 0.1M NaCl, 5mM MgCl2, 1mM PMSF) was added to
the mixture. All pulldowns were nutated at 4oC for 1 hr to allow association to occur. The
fusion protein agarose pellet was washed 5x with either Binding Buffer supplemented
with 400mM NaCl if probed with in vitro produced probe, or with JBC buffer
supplemented with 500mM NaCl if probed with cell lysis. The agarose was then boiled in
SDS loading buffer for 5 min for analysis by 12% SDS-PAGE, followed by Coomassie
staining, and autoradiography or immunoblotting.
59
2.3 Results
2.3.1 Yeast 2-Hybrid Screening:
To generate our DNA-BD vector, MTM341-544 was cloned into the NcoI site of pGBKT7,
yielding the DNA-BD fused to the N-terminal of MTM1. This fragment contains the
phosphatase and SID domains of MTM1 and was chosen because it possessed functional
domains known to be sites of mutation in MTM1 patients. When we began these
experiments, no proteins were known to interact with myotubularin and the physiological
substrate of this enzyme was believed to be another protein. Accordingly, we were
concerned that substrates for myotubularin might be missed due to the transient nature of
some enzymatic reactions. Thus we also generated pGBKT7-MTM341-544 (C375S), a
mutant known to be catalytically inactive due to a mutation in the active site 51, by site- directed mutagenesis.
Once generated and sequenced, these DNA-BD vectors were transformed into PJ692A, the parent yeast strain for all DNA-BD vectors. PJ692A carries the HIS3 and ADE2 nutritional reporters downstream of the Gal1 and Gal2 promoters respectively (Table
2.1). Once transformed with a DNA-BD vector, the resultant strain can be maintained and verified by growth on Synthetic Dropout minus tryptophan (SD-trp). Synthetic Dropout media contains all of the nutrients needed for wildtype yeast to grow except for one or
60 ns are are ns
Use pos Y2H Neg. control trains were DNA-BD parent DNA-AD parent control DNA-AD control DNA-BD candidate DNA-AD candidate DNA-AD MT mating DNA-BD WT mating DNA-BD
-hybrid interaction interaction -hybrid
no no no no no no no no no yes LacZ
no no no no no yes yes yes yes yes ura- ura- (C375S) as “bait”. as “bait”. (C375S)
no no no no no no no no no 341-544 yes ade-
no no no no no no no no no yes his- his-
no no no no no no yes yes yes yes leu-
Growth on SD Media
no no no no no yes yes yes yes yes trp-
Reporter Reporter HIS3/ADE2 LacZ HIS3/ADE2 HIS3/ADE2 HIS3/ADE2 HIS3/ADE2 HIS3/ADE2/LacZ LacZ LacZ ting. The "controldiploid strain" results in a positiveyeast 2 as the “bait”. Mutant (MT) mating used MTM
Gal 1 Gal 1 Gal 1 Promoter Promoter Gal1/Gal2 Gal1/Gal2 Gal1/Gal2 Gal1/Gal2 Gal1/Gal2 341-544 gal1/gal2/gal1
)
Yeast Strains Employed in Yeast 2-Hybrid
none none Vector pVA3-1 pTD1-1 pGBKT7 pACT2-SCG10 pVA3-1 & pTD1-1
pACT2-PP2Axxx-xxx pGBKT7-MTM434-541 pGBKT7-MTM1 344-541(C375S
(C375S) (C375S) 341-544
Y187 341-544 Strain PJ692A hMTM Y187-PP2Ac Y187-SCG10 Y187-pTD1-1 Control Diploid PJ692A-pGBKT7- PJ692A-pGBKT7- PJ692A-pVA3-1 PJ692A-pGBKT7 hMTM generatedyeast transfection byof the appropriatema or vector andgenerated was mating Y187-pTD1-1 by and PJ692A-pVA3-1, whichexpress T cell antigen SV40 andrespectively. p53 These protei Table 2.3: Yeast Strains Employed in Yeast 2-Hybrid. The first 4 strains listed were purchased from Clontech. The rest of the s the of rest The Clontech. from purchased were listed 4 strains first The 2-Hybrid. Yeast in Employed Strains Yeast 2.3: Table knownMTM used to (WT) mating Wildtype interact.
61 more amino acids. To verify the suitability of PJ692A-pGBKT7-MTM341-544 and PJ692A-
pGBKT7-MTM341-544(C375S) in our yeast 2-hybrid system, four parameters were
evaluated for each strain: presence of vector/ phenotype, toxicity, mating efficiency, and autonomous activation. Initially after transformation, PJ692A-pGBKT7-MTM341-544
colonies were successfully selected on SD-trp plates and the phenotype was confirmed
from growth patterns on SD–leu and SD-ura. The presence of pGBKT7-MTM341-544 in
PJ692A was demonstrated using whole cell PCR with DNA-BD vector specific primers
(Fig 2.2). Similarly, whole cell PCR verified the presence of pGBKT7 and pGBKT7-
MTM341-544(C375S) (data not shown) in PJ692A.
1 2 3 4 5 6 7 8
800bp
200bp
Figure 2.2: Whole Cell PCR Demonstrating the Presence of DNA-BD Vectors in PJ692A Lane 1: 100bp ladder, Lane 2: pGBKT7-MTM1341-544, Lane 3: pGBKT7, Lane 4: PJ692A- pGBKT7-MTM 341-544, Lane 5: PJ692A-pGBKT7, Lane 6: pVA31, Lane 7: PJ692A-pLAM, Lane 8: No template control
62 The presence of pGBKT7-MTM341-544 in yeast could produce a fusion protein toxic to the yeast cells, which would be difficult to carry through a cDNA library screening. PJ692A- pGBKT7-MTM341-544 was grown alongside PJ692A-pGBKT7 and the control strains
pJ692A-pVA3-1 (p53) and PJ692ApLAM. All were found to have a similar growth rate
over 48 hours (Fig 2.3A), indicating no toxicity. Similar data were obtained for PJ692A- pGBKT7-MTM341-544(C375S) (data not shown).
The presence of DNA-BD fusion proteins in yeast could interfere with the cells’ ability to
mate. This parameter is especially important, since unlike traditional yeast 2-hybrid systems, our system relies on mating. Both PJ692A-pGBKT7-MTM341-544 and PJ692A-
pGBKT7-MTM341-544(C375S) strains were mated to the control DNA-AD strain Y187-
pTD1-1. The scheme for determining mating efficiency is represented in Figure 2.3C.
Mating efficiencies were calculated and compared to those obtained with pJ692A-
pGBKT7 (13%) and PJ692A-pVA3-1 (20%). PJ692A-pGBKT7-MTM344-541 and
PJ692A-pGBKT7-MTM341-544(C375S) mated with 7.8% and 17%, respectively.
Importantly, all strains mated with efficiencies greater than 5%, which is the minimum considered acceptable by Clontech for screening their pretransformed library.
The MTM341-544 or MTM341-544(C375S) protein fragments could contain domains that
mimic an activation domain (DNA-AD) and therefore would be capable of autonomously
63 A B
2
1.8 Verify PJ692A-pGBKT7-hMTM1344-541 1.6 as experimental & PJ692A-pGBKT7 as neg contol 1.4
1.2 Plate on trp- Plate on Plate on 1 pgbkt7nco to verify presence trp-/ade- trp-/his- OD 600 OD of vector to test for activation to test for activation pgbkt7 0.8 at ade promoter at his promotor p53 0.6 plam
0.4 Both experimental No growth No growth and negative 0.2 control grow well
0 0 105 255 390 495 620 1560 1890 2880 Time (min)
Determining Mating Efficiency
C PJ692A-pGBKT7-hMTM1 344-541 Take half of each Pick 1 Y187-TD1 and resuspend and put in same flask as colony and resuspend in in YPDA the mating culture YPDA
Grow in10ml Grow in10ml Grow in10ml culture at 30degrees culture at 30degrees culture at 30degrees overnight overnight overnight
Plate as 1:1000 & 1:2000 dilution Plate as 1:10 and 1:100 Plate as 1:1000 & 1:2000 dilution trp- (select for plasmid) dilutions on leu- (select for plasmid) leu-, his-. ura- DDO (select for diploid) trp-, his-. ura- to verify phenotype QDO(select for pos interaction) to verify phenotype
Count trp- plate and calculate Count DDO plate and calculate Count Leu- plate and calculate viable cfu/ml viable cfu/ml viable cfu/ml =(cfu*1000)/(vol plated*dilut factor) =(cfu*1000)/(vol plated*dilut factor) =(cfu*1000)/(vol plated*dilut factor)
Mating efficiency cfu/mL= 1.07x107 =% diploid cfu/mL=2.05x107 =diploid cfu/limiting partner cfu
cfu/mL= 8.4x105
Figure 2.3: Verification of DNA-BD Strains Used for Library Matings A) Growth Curve for DNA-BD Yeast Strains PJ692A-pGBKT7-hMTM341-544 (marked as pGBKT7-nco) demonstrates similar growth rate as control DNA-BD strains indicating that there is no toxicity associated with the production of fusion protein. B) Scheme and Results of Autonomous Activation Assay: PJ692A-pGBKT7hMTM341-544 does not self activate either the HIS3 or ADE2 reporter genes of PJ692A. The same growth pattern was also found for the C375S mutant strain C) Scheme for Computing Mating Efficiency. PJ692A-pGBKT7-MTM341-544 was mated to the control strain Y187-pTD1-1. The DNA-BD strain was determined to be the limiting partner due to its lower viability and was used to calculate mating efficiency as 7.8%.
64 activating the Gal promoters. If this were the case, every library cDNA screened would
appear positive because the DNA-BD vector alone would activate the system. To test for
autonomous activation, both PJ692A-pGBKT7-MTM341-544 and PJ692A-pGBKT7-
MTM341-544(C375S) were grown on SD-trp/-ade and SD-trp/-his since PJ692A carries the
HIS3 and ADE2 reporters downstream of the Gal1 and Gal2 promoters, respectively.
Both strains phenotyped correctly by growing on SD-trp, but failed to grow on SD-trp/- ade and SD-trp/-his after 48 hours, indicating that neither MTM341-544 nor MTM341-
544(C375S) contain an occult activation domain.
Once both MTM DNA-BD strains had been verified for use in our yeast 2-hybrid system,
they were mated to our human fetal brain cDNA library cloned into the pACT2 vector
and pretransformed into Y187. These large scale matings were plated on SD-trp/-leu/-
his/-ade (QDO) to identify clones containing a potential protein-protein interaction. A
small amount of the mating was plated to determine mating efficiency and the library was
titered. The wildtype and mutant screenings had mating efficiencies of 18% and 9%,
respectively, with 3.2x107 and 3.6x106 clones screened in the wildtype and mutant matings, respectively.
All positive diploid colonies isolated by nutritional selection on QDO, were phenotyped again and tested for lacZ expression via X-gal assays. LacZ expression is another confirmation that a positive protein-protein interaction is occurring in the diploid because one of the parents, Y187, contains lacZ downstream of the Gal1 promoter. Once verified
65 in this manner, the DNA-AD cDNA inserts from positive clones were either directly amplified from yeast and subcloned, or the DNA-AD vector itself was isolated by yeast miniprep and plasmid rescue in E.coli. All positive cDNAs were sequenced from the forward direction and resultant sequences were analyzed against the Genbank Databases for identification. A summary of cDNA clones obtained from both matings by nutritional selection and X-gal assays is presented in Table 2.4.
cDNA Wildtype Mutant Screening Screening
Superior Cervical Ganglion 10 1 4 PP2Α Catalytic Subunit 2 0 Apparently novel clones 3 1 Parathymosin 3’ UTR 0 3 Various vector sequences 4 0 Neuronatin (out of frame) 2 3 Golgi autoantigen 0 1 LINE 1 RT 0 1 Alu repeat 0 1 Totals 12 14
Table 2.4: Clones Obtained in Mutant and Wildtype Yeast 2-Hybrid Matings: Clones were isolated either by yeast whole cell PCR followed by TA cloning or by direct isolation from yeast followed by rescue in E.coli. All clones were sequenced from the forward direction from the holding vector.
66 Many of the clones obtained represented various vector sequences and were discarded.
Five clones obtained in both matings represented the protein neuronatin, yet the cDNAs
were all out of frame with the activation domain. These clones were repeatedly evaluated
against the Genbank Databases, but no other gene product was identified. The clones representing “neuronatin” were discarded. Four clones represented unknown or novel cDNAs. None of these cDNAs aligned with one another, indicating that each was an independent clone. These novel clones were repeatedly evaluated against the Genbank databases, but no good candidate for further study emerged. Two proteins, superior cervical ganglion 10 (SCG10) and the catalytic subunit of protein phosphatase 2A
(PP2Ac), were represented 5 and 2 times between the two matings, respectively, and were considered for further study.
The first candidate, SCG10, was represented once in the wildtype MTM341-544 mating and
4 times in the MTM341-544(C375S) mating. This indicated that SCG10 was able to interact
with both MTM1 fragments, at least in the context of the yeast cell. In the mutant mating,
the 4 SCG10 clones represented at least 3 independent clones: The first clone represented
61bp of 5’UTR followed by the entire coding sequence of SCG10 (179 aa). This was the
same clone seen in the wildtype screening. The second clone seen in the mutant mating
had the complete SCG10 coding sequence, but no 5’UTR. The third clone seen began at amino acid 4 and included at least through amino acid 128. The fourth clone obtained in
the mutant mating seemed to be similar to the first clone as it possessed 5’UTR followed
67 by SCG10 coding sequence; however, only the first 138 amino acids of SCG10 were confirmed in the forward sequencing reaction.
The second candidate, PP2Ac, was represented 2 times in the wildtype MTM344-541 mating, but not detected in the MTM341-544(C375S) mating. This suggested that PP2Ac was able to interact with the wildtype MTM1 fragment at least in the context of the yeast cell. The two clones found in the wildtype mating represented the C-terminal 66 amino acids of PP2Ac. From the forward sequencing reaction, these two positive clones appear to be from the same cDNA clone in the construction of the library, but the complete cDNA inserts were not sequenced, thus their identity has not been confirmed.
MATINGS PERFORMED DNA-BD* DNA-AD* DDO** QDO** Xgal pJ692A-WT Y187-pp2a + + + pJ692A-MT Y187-pp2a + + + pJ692A-empty Y187-pp2a + - -
pJ692A-WT Y187-scg10 + + + pJ692A-MT Y187-scg10 + + + pJ692A-empty Y187-scg10 + - -
Table 2.5. Specific Yeast 2-Hybrid. To confirm candidate interactions, the yeast 2-ybrid interaction was recapitulated either with wildtype or mutant DNA-BD and the indicated DNA-AD vector. The yeast 2-hybrid interaction was unable to be recapitulated with empty DNA-BD vector, indicating that the reaction requires either MTM1 fragment. *All DNA-BD and DNA-AD strains were verified by nutritional selection for phenotype. All . strains phenotyped correctly. PJ692A-WT= PJ692A-pGBKT7-MTM341-544; PJ692A-MT = PJ692A-pGBKT7-MTM341-544 (C375S). PJ692A-empty = PJ692A-pGBKT7, the parent DNA-BD vector. **QDO=-trp/-leu/-ade/-his and selects for the presence of the yeast two-hybrid interaction; DDO=-leu/-trp and selects for the presence of both constructs.
68 Both candidate proteins were retested in the yeast 2-hybrid system by conducting a
specific yeast 2-hybrid experiment (Table 2.5). For both the SCG10 and PP2A clones, the
yeast 2-hybrid interaction was recapitulated by mating the PJ692A-pGBKT7-MTM341-544 and PJ692A-pGBKT7-MTM341-544(C375S) strains directly with Y187-SCG10 (whole
coding sequence without 5’UTR) and Y187-PP2Ac, the clone representing the last 66
amino acids of PP2Ac. To verify that neither the SCG10 nor PP2Ac clones had an occult
DNA-BD, both Y187-SCG10 and Y187-PP2Ac strains were mated to PJ692A-pGBKT7.
As anticipated, when the MTM1 fusion proteins were removed from the system by
mating with the empty vector strain, the mating was successful as indicated by growth on
SD-trp/-leu (DDO), but the positive yeast 2-hybrid interaction was lost, as indicated by
failure to grow on QDO and loss of X-gal staining (Table 2.5). Notably, although PP2Ac
was not seen in the mutant screening, MTM1341-544 (C375S) was able to interact with
PP2Ac in the specific yeast 2-hybrid experiment.
2.3.2 Coimmunoprecipitation Strategy to Confirm Candidate Protein-Protein Interactions
Based on the results of the yeast 2-hybrid screenings, both SCG10 and PP2Ac were evaluated for protein-protein interactions with myotubularin using
coimmunoprecipitation. Protein-protein interactions from yeast 2-hybrid can be false
positives and are considered candidate interactions until confirmed by an independent
method. Coimmunoprecipitation was chosen for confirmation because the interaction is
detected in a mammalian cell line which lends credibility to the physiological
69 significance of it. To facilitate the immunoprecipitation of myotubularin with the
candidate proteins for interaction, we attempted to generate a cell line stably expressing
Flag-tagged myotubularin. Initially we transfected Cos-7 cells with pCMV-Tag2b-
MTM1 and selected with G418. The cells expressed Flag-myotubularin transiently as
expected, but expression was lost within a few passages in G418. After several attempts,
we hypothesized that Flag-myotubularin (Flag-MTM1 protein) may be toxic to the cells.
We adopted a second approach: the tet-on system. The tet-on system, first described by
Bujard et al in mammalian cell culture, is based on the regulatory elements of the Tn10-
specified tetracycline resistance operon in E. coli. 76,77. In E. coli, in the absence of
tetracycline, the tetR repressor protein will bind to the operon’s tetO operators and
prevent transcription. In tet-on inducible systems, tetR was modified at four amino acids
and fused to the C-terminal domain of a HSV protein, VP16, creating the reverse
tetracycline-controlled transactivator (rtTA). This protein will only activate transcription
of a gene of interest from a tetO promoter in the presence of doxycycline, a tetracycline.
To use this system, HeLa Tet-on cells, stably and constitutively expressing rtTA (p-tet-
on), were purchased from Clontech. When a cDNA of interest is cloned into a vector with
a minimal promoter containing tetO sequences (pTRE2pur) and transfected into the HeLa
Tet-on cell line, the rtTA protein will bind and turn on its transcription only if induced by
doxycycline. In the absence of doxycycline, the rtTA protein is unable to bind to the tetO
promoter and activate transcription. Flag-MTM1 was cloned into the pTRE2pur vector
and the resultant product was transfected into HeLa Tet-on cells. After selection with
70 puromycin, multiple clones were identified that stably expressed Flag-myotubularin in a doxycycline inducible fashion (See Chapter 5). For coimmunoprecipitation experiments, clone 83, the highest expressing clone obtained was used.
To generate PP2Ac and SCG10 proteins in cell culture for coimmunoprecipitation, mammalian expression constructs containing the full length protein sequence of either human PP2Ac or SCG10 fused to N-terminal Xpress and His tags were generated by cloning into the pcDNA4-HisMax-TOPO vector.
The coimmunoprecipitation of either Xpress/HisG tagged PP2A or SCG10 with Flag tagged MTM1 was attempted according to Figure 2.4. HeLa clone 83 cells were transiently transfected with either Xpress/HisG-SCG10 or Xpress/HisG-PP2Ac and stimulated with doxycycline to induce Flag-MTM1 expression. Cells were lysed in RIPA buffer, a harsh buffer containing multiple detergents and high salt. MTM1 was immunoprecipitated with the Anti-Flag antibody and Protein G agarose; neither SCG10 nor PP2Ac were bound. There was concern that the cell lysis and coimmunoprecipitation conditions may be too harsh to preserve protein-protein interactions; therefore, we tried coimmunoprecipitation with additional buffers containing high or low salt with either anionic (Triton X 100) or zwitterionic (Empigan) detergents (Table 2.4). In all cases,
MTM1 was immunoprecipitated and either SCG10 or PP2Ac were transiently expressed in the cell lysate, but no interactions could be demonstrated (Figures 2.5 and 2.6).
71
Cell lysate protein G Add Flag ATB agarose (mouse)
MTM1 -Flag Pull down Flag-MTM1
Detect binding partner with Anti- Xpress (mouse) with SDS-PAGE Boil to liberate MTM1, binding partner, & Flag antibody
Figure 2.4: Scheme for Coimmunoprecipitation of Xpress/HisG-PP2Ac or Xpress/HisG-SCG10 with Flag MTM1. HeLa cells that express Flag-MTM1 upon induction with doxycycline are transiently transfected with Xpress/HisG-PP2Ac or Xpress/HisG-SCG10 plasmids. After 48 hr, the cells are lysed and incubated for 1 hour with anti-Flag antibody. Protein G agarose is added to the mixture to capture the Flag antibody and its antigen. The lysates are incubated for 1 hour before pelleting and washing the agarose. The pellet is boiled in SDS Loading buffer and immunoblotted to determine if PP2Ac or SCG10 coimmunoprecipitated with MTM1.
72 Buffer pH Salt Detergent 1 50mM tris pH7.5 25mM NaCl 0.1% Triton 2 “ 100mM NaCl “ 3 “ 25mM NaCl 0.1% Empigen 4 “ 100mM NaCl “ 5 Ripa*
Table 2.6: Buffers Used in Coimmunoprecipitation Studies. The five buffers described were used for cell lysis and coimmunoprecipitation experiments. No protein-protein interactions were established for either SCG10 or PP2Ac with MTM1 protein using any of these buffers.* for RIPA recipe see Materials and Methods
73 Anti-Flag Anti-Xpress Lysate CoIP Lysate CoIP 75 75 Flag-MTM1 50
Xpress- 50 PP2Ac 37
25 25
Dox + - + - Dox + + PP2Ac + + + - PP2Ac + +
Figure 2.5 PP2Ac Does Not Coimmunoprecipitate with Myotubularin Under Low Salt/Triton Lysis Conditions: Tet-inducible HeLa cells were treated with doxycycline and transfected with an Xpress/HisG tagged PP2Ac mammalian expression vector. Control cells were either not treated with doxycycline or not transfected. Cells were lysed in Low Salt/Triton Buffer, immunoprecipitated with Anti-Flag antibody, and immunoblotted with Anti Flag and Anti- Xpress antibodies to detect immunoprecipitation and coimmunoprecipitation, respectively. In the left panel, Flag MTM1 protein was present in cells stimulated with doxycycline and absent in untreated cells. Flag-MTM1 protein was successfully immunoprecipitated with the Flag antibody as indicated in left panel, CoIP lanes. PP2Ac was also present in the whole cell lysates after delivery by transient transfection as indicated in the right panel, Lysate lane, but SCG10 did not coimmunoprecipitate with Flag-MTM1 protein as is indicated by its absence the right panel, CoIP lane. The Anti-Flag CoIP lanes were exposed for 10 sec, whereas the rest of the lanes were exposed for 30 sec. Additional bands are presumably non specific and/or IgG bands since antibodies to immunoprecipitate and detect are from the same species.
74 Whole Cell CoIP
Dox + - + + - + SCG10 + + - - + +
MTM1 Anti-Flag
SCG10 Anti-Xpress
Figure 2.6: SCG10 does not Coimmunoprecipitate with Myotubularin Under RIPA Buffer Lysis Conditions: Tet-inducible HeLa cells were treated with doxycycline and transfected with an Xpress/HisG tagged SCG10 mammalian expression vector. Control cells were either not treated with doxycycline or not transfected. Cells were lysed in RIPA buffer, immunoprecipitated with Anti-Flag antibody, and immunoblotted with Anti Flag and Anti- Xpress antibodies to detect immunoprecipitation and coimmunoprecipitation respectively. As is seen in the top panel on the left, Flag MTM1 protein was present in cells stimulated with doxycycline and absent in untreated cells. Flag-MTM1 protein was successfully immunoprecipitated with the Flag antibody as indicated on the top right panel. SCG10 was also present in the whole cell lysates after delivery by transient transfection as indicated in the bottom left panel, but SCG10 did not coimmunoprecipitate with Flag-MTM1 as is indicated by its absence the lower right panel.
75 2.2.3 GST Strategy to Confirm Candidate Protein-Protein Interactions
Since we were not successful in demonstrating interactions by coimmunoprecipitation,
we attempted a GST-based strategy to try to confirm the protein-protein interactions. The
successful expression and purification of GST-MTM1-His6 and GST-MTM1-(C375S)-
51 His6 had been reported in the literature . GST-MTM1-His6 fusion proteins were
expressed in CodonPlus RIL E coli by stimulating the cells with IPTG. CodonPlus RIL
cells are engineered to express additional tRNAs for arginine (R), isoleucine (I) and
leucine (L) with mammalian codon bias. Fusion proteins were purified with Ni-NTA agarose then immobilized and washed on glutathione agarose.
Immobilized, purified fusion proteins were probed with either Xpress/HisG-SCG10 or
Xpress/HisG-PP2Ac proteins. Initially, we generated Xpress/HisG-SCG10 and
Xpress/HisG-PP2Ac using in vitro transcription and translation with 35S-methionine
labeling. No protein-protein interactions were demonstrated when immobilized GST-
MTM1-His6 was probed with these in vitro made proteins (Figure 2.7). NSDHL, a non-
related protein used as a negative control showed a small amount of nonspecific
interaction (Figure 2.7, black arrow). No binding between GST fusion proteins and
SCG10 or PP2Ac was seen when the immobilized GST protein was the C375S mutant
(data not shown).
76 35S autoradiograph
input elution wash
GST-MTM-H6
P S N
Figure 2.7: GST-MTM1 Protein Probed with PP2A & SCG10 Made via in vitro Transcription /Translation. Wildtype GST-MTM1-His6 fusion protein was immobilized on glutathione agarose and probed with Xpress/HisG-SCG10 (S), Xpress/HisG-PP2Ac (P), or an unrelated control protein NSDHL (N) labeled with 35S methionine. Input lanes represent 1/24th of probe used in the pulldown. Elutions represent ½ of the pulldown mixture. Proteins were eluted by boiling for 5 min. Wash lanes represent the whole last wash performed prior to elution. The amount of GST fusion protein (red arrow) in each pulldown was estimated to be 1 mg by Coomassie staining shown below elution lanes. Black arrow represents very small amount of NSDHL eluted from fusion protein, representing a nonspecific reaction.
77
At this point there was concern that in vitro protein probes might lack additional proteins in a complex or post-translational modifications required for interaction. To address this concern, PP2Ac and SCG10 were generated in vivo and cellular extracts were used in pulldown assays. These lysates came from HeLa cells transfected with either
Xpress/HisG-PP2Ac or Xpress/HisG-SCG10 and grown in 35S-methionine. Again, no
interaction between either SCG10 or PP2Ac was demonstrated with either wildtype or
mutant GST-MTM1 fusion proteins. (Figure 2.8- mutant MTM1 probed with SCG10,
Figure 2.9- wildtype MTM1 probed with SCG10 and PP2Ac and mutant probed with
PP2Ac.) It was also noted that no other proteins from the HeLa cell lysate specifically
bound to the GST-MTM1-His6 fusion proteins as detected by autoradiography.
78
35S autoradiograph Western Input Wash Elution Input
SCG10
SCG10 - + - + + -
Figure 2.8: GST-MTM1 (C375S) Protein Probed with Xpress/HisG-SCG10 Transfected into HeLa. HeLa cells were transfected with Xpress/HisG-SCG10, labeled with 35S- methionine and lysed. The lysate was used to probe immobilized GST-MTM1(C375S)-His6 fusion protein. 10uL of lysate was loaded directly for autoradiography and immunoblotting, while 990ul was used as probe for the pulldown assay. Lanes marked “wash” represent the last wash performed before elution. Lanes marked “elution” represent 1/3 of the total elution. Proteins were eluted by boiling for 5 min. The amount of GST fusion protein in each pulldown sample was estimated to be 1mg by Coomassie staining (lower panel).
79
35S autoradiograph Western Elution Wash Input Input
PP2Ac
SCG10
S P H P H S P H S P H
WT MT
Figure 2.9: GST-MTM1 and GST-MTM1(C375S) Proteins Probed with Xpress/HisG- SCG10 and Xpress/HisG-PP2Ac Transfected Into HeLa Cells. HeLa cells were transfected with Xpress/HisG-SCG10 (S), Xpress/HisG-PP2Ac (P), or left untransfected (H) and labeled with 35S-methionine before lysis. The lysate was used to probe immobilized GST- MTM1-His6 (Green) or GST-MTM1(C375S)-His6 (Blue) fusion proteins. 10uL of lysate was loaded directly for autoradiography and immunoblotting, while 990ul was used as probe for the pulldown assay. Lanes marked “wash” represent the last wash performed before elution. Lanes marked “elution” represent 1/3 of the total elution. Proteins were eluted by boiling for 5 min. The amount of GST fusion protein in each pulldown sample was estimated to be 1mg by Coomassie staining (Lower panel).
80
2.4 CONCLUSIONS AND DISCUSSION
To discover proteins that interact with myotubularin, yeast 2-hybrid was employed and a
human fetal brain library was screened. Two screenings were performed with two
different DNA-BD proteins, MTM 341-544 and the catalytically inactive mutant MTM 341-
544(C375S). Although both wildtype and mutant MTM1341-544 proteins appeared to
interact with PP2Ac and SCG10 in yeast, confirmation of whole myotubularin interacting
with either of these candidates could not be demonstrated. Both coimmunoprecipitation
and GST pulldown strategies have been attempted with whole MTM1 and the
catalytically inactive C375S mutant, but no interactions with either PP2Ac or SCG10
have been demonstrated. With these data, one can only conclude that these candidate
protein-protein interactions represent false positives derived from our yeast 2-hybrid
screenings.
There are several factors that may contribute to the generation of false positives in this
system. One possibility is the choice of baits. Although we used specific yeast 2-hybrid
to recapitulate the candidate protein-protein interactions in yeast, we used the original
DNA-BD vectors with MTM1341-544 and MTM1341-544 (C375S). In retrospect, we did not
look at these candidate interactions using the complete coding sequence for either
wildtype or mutant myotubularin; thus, we cannot rule out that the use of a fragment of
myotubularin allowed domains that would usually be buried in the whole protein to
interact. Such an interaction is unlikely to be physiologically significant. Initially the
81 MTM1341-544 fragment was chosen because it contained the putative Set Interacting
Domain (SID) and protein phosphatase domain. At the time, the SID domain of MTMR5 had been found to interact with HRX, a SET domain containing protein 37 and the
phosphatase domain was anticipated to interact with a protein substrate. This fragment of
MTM1 was thought to be an important portion of the protein and we were concerned
about working with the whole MTM1 sequence due to its size (68kD). We also had
concerns that the PDZ-binding domain of MTM1, if included in the bait, would lead to
several false positives (personal communication, Jill Rafael, Molecular and Cellular
Biochemistry, The Ohio State University). The MTM1341-544(C375S) fragment was
chosen as our second bait out of concern that a protein substrate for myotubularin may
interact only transiently and thus not be detected with the wildtype enzyme.
Use of a partial protein to study protein-protein interactions has proved confusing in the
MTM field: Initially, MTMR5 was identified in a yeast 2-hybrid screen using a SET
domain as the bait 37. Even in coimmunoprecipitation studies, MTMR5 was found to
interact with the complete HRX protein and have a nuclear localization. Further study of
MTMR5 revealed that the protein as first described only represented the C-terminal half
of the full protein 38. This full length protein no longer localized to the nucleus, its interaction with HRX has not been revisited, and the authors speculate that this protein-
protein interaction is not relevant 38.
82 The question of why PP2Ac and SCG10 in particular were identified in this system must
also be addressed. The whole coding sequence of SCG10 was revealed in these yeast 2-
hybrid screenings, but only the last 66 C-terminal amino acids of PP2Ac were found. In
retrospect, the entire cDNA of PP2Ac was not revealed and probably should have been
used to recapitulate the interaction in the specific yeast 2-hybrid experiments. It is
possible that a domain of PP2Ac that is usually not exposed was accessible by using only
a fragment of PP2Ac. Interestingly, two proteins known to interact with PP2Ac, α4 and
ERα, have been identified 78-80. The regions of these two proteins responsible for
interaction with PP2Ac have been described and investigators have noted an eight amino
acid stretch of similarity between them 78-80 (Figure 2.9). Upon comparison of this amino acid stretch to MTM1341-544, one region of similarity was noted. Unfortunately, the region
of PP2Ac required for interaction with ERα or α4 has not been mapped. Whether or not
this similarity can explain why PP2Ac was revealed in our yeast 2-hybrid screenings is completely unclear, yet the observation of this similarity in our bait was made.
It is not clear why true binding partners for myotubularin such as MTMR12 were not identified. One possibility is the choice of cDNA library. We chose to use a human fetal brain library because a skeletal muscle library was not available and brain tissue typically expressed an abundance of different transcripts. A skeletal muscle library would have been a more ideal choice since MTM1 deficiency in skeletal muscle clearly causes a phenotype. In addition, protein levels of myotubularin in the brain are very low compared to muscle and liver at least in mouse 39. Two protein isoforms of MTM1 are seen in
83
NCESARER MTM1482-489 AMESAKET ERα AMKSAVES α4
Figure 2.10: Comparison of a Similar Domain in ERα and α4 to an MTM1 Amino Acid Sequence Contained in the Baits Used for Yeast 2-Hybrid Screenings: Investigators recognized similarity across amino acids between regions of ERα and α4, two proteins known to interact with PP2Ac. Five amino acids were identical between the two proteins, three of which are found in MTM1 (red). A fourth amino acid in MTM1 was found in ERα only.
skeletal muscle. The difference between the two forms is unclear, but the slower
migrating form is predominant after myotube formation 39. This expression pattern may
suggest that myotubularin has a specialized function in skeletal muscle rather than in
brain where the protein level is lower.
The reasons for choosing MTM1341-544 and MTM1341-544(C375S) have been discussed
above in the context of generation of false positives; yet these bait choices may have
prohibited finding true positives as well. These baits both contain the phosphatase domain and SID of myotubularin, but lack the coiled coil domain. Since the start of this work, myotubularin has been demonstrated to bind with MTMR12 (3-PAP) 72. Notably,
this interaction was first seen in rat brain lysates and then fully characterized in platelets.
The SID of MTMR12 was found to be necessary for this interaction in vivo, but the
necessary region of myotubularin for binding was not addressed. Of note though is that
84 the interaction between MTMR2, one of myotubularin’s closest relatives, and MTMR5 required coiled coil domains of both proteins 41. Additionally, the interaction described
between the MTMR9 and MTMR7 proteins required the coiled coil domain of MTMR9
in GST analysis 70. These data suggest that the coiled coil domain is an important domain
for protein-protein interactions and its exclusion in our yeast 2-hybid screens may explain
why true positives were not revealed.
Finally, whether yeast 2-hybrid was an ideal method for screening protein-protein
interactions must be addressed. Membrane bound proteins are considered poor candidates
for yeast 2-hybrid analysis because they are unable to localize to the nucleus to activate
reporter genes. There is no evidence that myotubularin is membrane bound, yet it
interacts with a molecule that is associated with vesicular trafficking, PI3P 66. MTM1 has
been localized to the plasma membrane 39 and also endosomes 55. These characteristics of
myotubularin may make it a less than ideal candidate for analysis by yeast 2-hybrid. It is
noteworthy that with the exception of MTMR5 and its initial description as binding with
SET domains, all other protein-protein interactions described for this gene family have
come from immunoprecipitation of the gene of interest from cells followed by mass
spectroscopy of the resultant binding partners for identification 41,70,72. Yeast 2-hybrid with full length myotubularin has been tried by the Mandel Laboratory and no publishable interactions were described (personal communication between GE Herman and J Laporte).
85 The choice of bait for the yeast 2-hybrid screenings represented the phosphatase and SID domain of myotubularin. Given the high level of homology of MTM1 with other genes in the family at these domains and the use of a brain cDNA library where myotubularin is not highly expressed, we cannot rule out the possibility that we inadvertently identified proteins that interact with other members of the MTM1 gene family.
86
CHAPTER 3
DEVELOPMENT AND CHARACTERIZATION OF AN ANTIBODY FOR MYOTUBULARIN
3.1 Introduction
The development of an antibody to the protein product of a gene of interest can prove very useful in understanding the physiological function of the protein. An antibody to myotubularin could prove useful in several ways. For instance, an antibody may allow protein-protein interactions to be confirmed or identified by coimmunoprecipitation or allow investigators to track the subcellular localization of the endogenous protein. From a clinical standpoint, a good antibody to myotubularin could be developed into a tool for molecular diagnosis, by allowing some patients to be diagnosed based on their level of myotubularin rather than by sequencing of MTM1.
At the initiation of our work with myotubularin, there were no descriptions of antibodies in the literature. This necessitated overexpression of tagged myotubularin in experiments investigating subcellular localization and protein-protein interactions. The goal of the work presented in this chapter was to make an antibody that would specifically recognize
87 human myotubularin without cross-reacting with other known members of the MTM
family of proteins. Before joining this project, maltose binding fusion proteins
representing full length, N-terminal, and C-terminal MTM1 proteins had been purified
and injected into chickens and rabbits. This strategy did not yield a specific antibody to
myotubularin, thus a peptide antibody strategy was chosen 81,82.
Traditionally, in this strategy, one predicts an 11-15 amino acid region of high antigenicity and injects that peptide conjugated to a carrier protein into animals. Sera are then tested for an immune response to the peptide that also will recognize the whole protein. The prediction of an appropriate peptide from a whole protein of interest is an
empirical process. However, over time some general rules have been suggested for
increasing the chance that the peptide will lead to a successful antibody 83,84. The first
guideline is to choose a peptide that has several hydrophilic residues to increase the
chance that it exists on the surface of the whole protein and is soluble in aqueous
solution. Secondly, C-terminal and N-terminal regions of a protein are more likely to be
exposed and flexible and are better candidates for antigenic peptides. Protein regions with
consensus sequence sites for post-translational modifications should be avoided since the
modification will not be part of the peptide. Generally, peptide sequences are 11-15
amino acids long. Specifically, we enlisted the assistance of Dr. Pravin Kaumaya at The
Ohio State University to help predict an appropriate peptide from the human
myotubularin sequence. Although this is an empirical exercise, Dr. Kaumaya possesses
several algorithms 85-92 for antigenic prediction and has experience analyzing the data
88 from them. Many of these programs evaluate hydrophilicity of the protein of interest, but
each weighs regions of the protein slightly differently. Thus, using multiple algorithms
and choosing a region predicted by many of them increases the chance that the region
will truly be antigenic.
The peptide antigen strategy has advantages in that it is relatively easy to generate a small peptide commercially. Purification of a fusion protein from E. coli may be complicated by protein toxicity, low expression, codon bias, or inclusion bodies. Additionally, by using a very small epitope, one can attempt to generate an antibody that will be very specific for the protein of interest and not react with other members of a gene family or with the same protein from another species. There are also potential disadvantages to the peptide antibody approach; the peptide chosen may not be very antigenic and may not elicit an immune response in the animal. Also, even if an immune response is elicited, the peptide antigen represents a very small portion of the whole protein and thus the resultant antibody may not recognize the whole protein. The peptide antigen may represent a portion of the native protein that is not usually exposed physiologically. While this may not be a problem for recognizing the whole protein when it is denatured, such as with immunoblotting, detection of the whole protein under native conditions (eg immunoprecipitation, immunohistochemistry) may not be possible.
We chose to use the antigenic peptide strategy to generate polyclonal antibodies in rabbits. Rabbits were chosen since larger volumes of sera can be obtained compared to
89 rodents. Polyclonal antibodies were chosen because they are less expensive and labor intensive to generate than monoclonal antibodies. Polyclonal antibodies result when an antigen is injected into an animal and sera are collected. Presumably, the animal will generate multiple different antibodies to the single antigen. This is in contrast to monoclonal antibodies which are collected from hybridoma cell lines. Hybridomas are generated by fusing a myeloma cell with an antibody secreting B-cell from the lymphoid tissue of an injected animal. A polyclonal antibody, which represents a set of antibodies to a single antigen, can often be successfully generated for use in immunoprecipitation, immunoblotting and immunohistochemistry. A monoclonal antibody is only one specific antibody which can be continuously regenerated; this may be advantageous for certain applications 83.
3.2 Materials & Methods
3.2.1 Generation of Peptide Antibodies to Human Myotubularin:
3.2.1 A. Peptide Prediction:
The amino acid sequence of human MTM1 was analyzed using several antigenicity prediction programs from the laboratory of Pravin Kaumaya (Ohio State University).
These programs included the Bulk Hydrophobic Scale, Hopp and Woods Hydrophilicity
Scale, Janin Accessibility Scale, Parker Hydrophilicity Scale, Fraga Global Scale,
Welling Antigenicity Scale, Kyte and Doolittle Hydropathy Scale, Novotny Large Sphere
90 Accessibility Scale, and the Fauchere and Pliska Scale 85-92. Multiple prediction programs were employed to uncover regions of human myotubularin that appeared to be ideal candidates for an antigenic peptide. The more programs that predicted a region to have strong antigenicity, the stronger that candidate was for consideration. Once strong candidate regions were identified, their amino acid sequences were compared to other members of the MTM family for homology using DNASTAR-Megalign
(https://www.dnastar.com/web/r10.php) and CLUSTALW (http://www.ebi.ac.uk/ clustalw/#) softwares. Candidate regions of myotubularin were considered stronger if they shared little amino acid identity with human MTMR1-7. Finally, candidate regions were evaluated for post-translational modification sites. The program to predict post- translational modification, ScanProsite, is available at http://expasy. cbr.nrc.ca/tools/scnpsit1.html. From this analysis of human myotubularin, two regions were identified: MTM1261-273 (LDVIRETNKQISK) and MTM1574-591
(LQLANSAKLSDPPTSPS). These peptides were generated by Sigma Genosys with an
N-terminal promiscuous T cell antigen and GPSL turn motif. The two peptides were
referred to as MTM1aa574-591h (LSEIKGVIVHRLEGVGPSLLQLANSAKLSD-
PPTSPS) and MTM1QQ261-273h (LSEIKGVIVHRLEGVGPSLLDVIRETNKQISK).
3.2.1 B. Antibody Production:
Once generated, the MTM1 peptides were supplied to Cocalico Biologicals, Inc. for
injection into rabbits. Each peptide was injected into two New Zealand white rabbits,
noted as CRI-1, CRI-2, CRI-3, and CRI-4. To make the peptide solution for injection, 2 91 mg of peptide were dissolved in 1mL of PBS and 1mL of CFA (Complete Freund’s
Adjuvant), yielding a 1mg/mL solution. One milliliter of this solution was injected into each animal on Day 1 after a prebleed. On Days 21 and 50, additional 1mL injections were given. To make these peptide solutions, 0.5mg of peptide was dissolved in 0.5mL of
PBS and 0.5mL of IFA (Incomplete Freund’s Adjuvant). Bleeds were taken from the
rabbits on Days 35, 49, 64, 71, and 78. Exsanguination bleeds were collected from
animals at the end of the study.
3.2.1 C. Antibody Purification:
To purify MTM1 polyclonal antibodies from rabbit sera, AminoLink (Pierce) affinity
purification columns and protocols were employed. The first step in purification was
immobilizing the peptide antigen to the column. Storage Buffer was replaced with
Coupling Buffer (Pierce). The MTM1 peptide (2mg) was dissolved in Coupling Buffer
and added to the column with 200uL of Reductant (32mg of sodium cyanoborohydride
reductant in 0.5mL of 10mM NaOH). The column was nutated at room temperature for 2
hr and sat overnight. The column was drained and washed with 5 mL of Coupling Buffer.
To block the remaining sites on the column, 4mL of Quenching Buffer (Pierce) was
passed over the column followed by 2mL of Quenching Buffer with 200uL of Reductant
which was left to nutate on the column for 30 min at room temperature. The column was
washed according to manufacturer’s instructions (Pierce). PBS (8mL) was washed through the column before applying 2mL of the rabbit sera. The column was covered
92 with 2mL of PBS to prevent drying. The serum was incubated at room temperature for 1 hr and then the column was washed with 14mL of PBS. The bound antibodies were eluted with 10mL of Elution Buffer (Pierce; ImmunoPure IgG Elution Buffer). The elution was collected as 1mL fractions, which were immediately neutralized by the addition of 50uL of 1M tris pH 9.5. The column was stored at 4oC with degassed PBS with 0.5% sodium azide as a preservative.
3.2.1 D. Peptide Competition:
The MTM1 peptides used as antigen in antibody production were diluted in water
(1mg/mL). 1uL of the peptide solution was incubated with 10uL of purified sera on ice for 1 hour. Mixing an excess of antigenic peptide with sera resulting from its injection should sequester the antibody and inactivate the serum. The competed sera sample was then used as primary antibody as described in Section 2.2.3 C.
3.2.2 Generation of a Mouse MTM1 Mammalian Expression Construct
To clone mouse MTM1, a Research Genetics EST clone (BQ918575; IMAGE ID
6398326) was identified by performing a BLAST search using mouse MTM1 cDNA
(AF073996). The primers used were mouseMTM1F (5’-ATGGCT-
TCTGCATCAGCATC-3’) and mouseMTM1R (5’-TCAGAAGTGCGTCTGCACATG-
3’) which amplifies the full-length mouse MTM1 cDNA. Reactions were carried out in
93 25ul total volume with 1.5 mM MgCl2, 0.4 uM primers, 0.2 uM dNTPS and 1 unit
Amplitaq Gold (Perkin Elmer). The EST clone (1 ng) was used as a source of template and the PCR reaction was carried out as follows: 95 oC for 10 min for one cycle; denaturation at 94 oC for 30 sec, annealing at 52oC for 30 sec, elongation for 2 min at 72 oC for 35 cycles; and a final extension at 72 oC for 7 min. This PCR fragment was cloned into pcDNA4-HisMax-TOPO as described in section 2.2.2.A.i. Sequencing was performed as described in 2.2.1.D.ii.c with the T7 and BGH reverse primers described in section 2.2.2.A.i. Additional sequencing was performed with the mouseMTM1seq primer (5’-CATTATGCGCTGCAGTCAGC-3’). The resultant clone, pcDNA4-HisMax-
TOPO-mMTM1 was transfected into Cos-7 cells for 48 hours as described in section
2.2.2.C. The cells were lysed using passive lysis buffer and the resulting protein was immunoblotted as described in section 2.2.2.C.
3.3 Results
The first step in generating a peptide antibody to a gene of interest is to predict a peptide sequence within the primary structure of the gene that when injected into an animal will illicit an immune response. The antibodies elicited in the animal must not only react with the peptide, but also recognize the whole protein. We consulted Pravin Kaumaya, PhD from The Ohio State University to help predict suitable antigenic peptides. We accessed
94 his set of 10 algorithms to analyze the sequence of human myotubularin for identification
of potentially antigenic regions. Each algorithm identified and scored several candidate
peptides for myotubularin based on criteria such has hydrophilicity and accessibility. We
analyzed which candidate peptides appeared in multiple algorithms and considered a
region stronger if it was predicted by more algorithms and had higher scores. In addition
to the original 10 algorithms, Dr. Kaumaya has an algorithm that predicts secondary
structure for a primary amino acid sequence. Candidate regions within myotubularin were
evaluated for the presence of β-sheets, α-helices and turns. Preference was given to
regions containing either α-helices or turns because these regions are considered more
flexible and potentially more antigenic 83,84. After determining the secondary structure of candidate peptides, they were evaluated for possible post-translational modifications, including glycosylation, phosphorylation, etc, with ScanProsite. Regions with sites for possible post-translational modifications were considered weaker candidates since the peptide produced to mimic the region would lack modifications. From all of these analyses of the primary structure of human myotubularin came several possible peptides; however, our goal was to generate an antibody to myotubularin that did not cross react with any other members of the MTM gene family. Each region identified by Dr.
Kaumaya’s method of peptide prediction was compared to protein sequence from other members of the MTM gene family for similarity. Regions that displayed high homology across the gene family were excluded to prevent the generation a cross-reacting antibody.
95 Two optimized candidate regions emerged for generation of an antigenic peptide:
MTM1261-278 and MTM1574-591. The first region of MTM1 identified, MTM1574-591, was predicted to be antigenic by five of the ten algorithms and lacked putative sites for post- translational modification. Additionally, this region was predicted to have a turn when analyzed for secondary structure, making it a stronger candidate. Upon comparison with other members of the MTM gene family, MTM1574-591 was found to have little homology
(Figure 3.1 A). The second peptide, MTM1261-278, also showed significant antigenicity in
five of the ten algorithms and possessed no putative sites for post-translational modification. Upon alignment with other members of the MTM family, MTM1261-278
possesses little homology with other proteins in that region (Figure 3.1 B).
Routinely, peptides prepared for injection into animals to produce antibodies are
covalently conjugated to a carrier protein such as KLH. At Dr. Kaumaya’s suggestion,
we designed the final peptides to contain an N-terminal promiscuous T cell antigen
(Figure 3.1C). Promiscuous T-cell antigens are epitopes that are universally active in
multiple MHC haplotypes. The rationale for such a design is that the T cell antigen of the
peptide will stimulate the animal’s T cells, which is a prerequisite for B cell stimulation
and antibody production. Thus including a promiscuous T cell antigen on the peptide
should encourage a stronger immune response to the peptide 93. The promiscuous T cell
antigen chosen was a region of measles virus (MVF) thought to be antigenic across a
96 A
MTM1 LQLANSAKLSDPPTS-PS MTMR1a LQREISNRS-TSSSERGS MTMR1b ------MTMR2 LQREISNRS-TSSSERAS MTMR7 LEEVRHT------CFV MTMR8 LEKKLKVRDEPPEEICTC MTMR6 LESKIKQ------RKN MTMR9 LRRQLAE------L MTMR3a VPLASRRCSDPSLNEKWQ MTMR3b VPLASRRCSDPSLNEKWQ MTMR3c VPLASRRCSDPSLNEKWQ MTMR4 SPLT-RTSSDPNLNNHCQ MTMR5 APQSRRRVVWPCYDSCPR MTMR13 GPRSQRRTVWPCYDDVSC MTMR10 FPSRV------MTMR12 FREETDHLIKNLLGKRIS
B
MTM1 LDVIRETNKQIS------KLTIYD MTMR1a LQTIMDANAQSH------KLIIFD MTMR1b LQTIMDANAQSH------KLIIFD MTMR2 LQAIMDSNAQSH------KIFIFD MTMR7 LQAIRKANPGSD------FVYVVD MTMR8 LEAISQT NPGSQ------FMYVVD MTMR6 LQAISKANPVNR------YMYVMD MTMR9 INATLRAGKRG------YLIDT MTMR3a VQSVAKACASD--SRSSGSKLSTRNTSRDFPNGGDLSDVEFDSSLSNASGAESLAIQPQKLLILD MTMR3b VQSVAKACASD--SRSSGSKLSTRNTSRDFPNGGDLSDVEFDSSLSNASGAESLAIQPQKLLILD MTMR3c VQSVAKACASD--SRSSGSKLSTRNTSRDFPNGGDLSDVEFDSSLSNASGAESLAIQPQKLLILD MTMR4 VTSIAKACALDPGTRATGGSLSTGNN-----DTGEACDADFDSSLTACSGVESTAA-PQKLLILD MTMR5 LQAVVSSMPRYADASGRNTLSG------FSSAHMG MTMR13 LQALLNAVSVHQKLRGNSTLT------VRPAFAL MTMR10 CNAITKS------MTMR12 LDGIYKTI------
C NH3-LSEIKGVIVHRLEGVGPSLLQLANSAKLSDPPTSPS-COOH MTM1574-591
NH3-LSEIKGVIVHRLEGVGPSLLDVIRETNKQISKLTIYD-COOH MTM1261-278
Figure 3.1: Alignment of MTM1 with Other Members of the Gene Family in the Peptides Used for Antibody Production. Human MTM1 sequences are shown in red. The sequence of MTMR11 was unavailable for this analysis. Alignments were performed with the entire sequences of all proteins using CLUSTALW. A. MTM1574-591. B. MTM1261-278. C. The final peptides generated contained a promiscuous T-cell antigen (green), a proline containing turn motif (blue), and the indicated region of human MTM1 (red).
97 broad range of MHC haplotypes 93. Peptides also contained a spacer region engineered
between the promiscuous T cell antigen and the epitope of MTM1. This spacer contains a
proline to generate a turn in the structure (Figure 3.1C).
After the design of the MTM1261-278 and MTM1574-591 peptides was finalized, they were
produced commercially by Sigma Genosys. Injection of the peptides into New Zealand
white rabbits was carried out by Cocalico Biologicals, Inc and at various time points sera
were delivered to us for further testing. MTM1261-278 was injected into two rabbits, CRI-1
and CRI-2, while MTM1574-591 was injected into rabbits CRI-3 and CRI-4. The injection schedule included an initial dose of peptide on Day 1 followed by boosts on Day 21 and
Day 50. Bleeds were taken at various times throughout the study and at exsanguinations,
but we routinely screened the “third bleeds” at Day 64 in all animals for antibody
production.
To screen the sera for production of antibodies specific for human myotubularin, Cos-7
cells were transfected with pCMV-Tag2b-MTM1, a vector designed to express full length
human myotubularin with an N-terminal Flag tag (Section 2.2.3 A). Untransfected Cos-7
cells served as controls. The cells were lysed and immunoblotted using test sera at a
1:1000 dilution or the Flag antibody (1:1000). Representative blots are shown for
unpurified sera in Figure 3.2. The presence of Flag-MTM1 protein at the expected size of
72kD on the anti-Flag immunoblot indicated that the transfection was successful. The other sera tested yielded blots with several bands in both transfected and untransfected
98 + - kD kD + - 216 160 132 105 75 75
50 45 + - kD 35 216 32 132 75 72
CRI-1 CRI-2 45 + - kD kD - + 32 160 160 105 105 75 75 50 50 Flag 35
35
CRI-3 CRI-4
Figure 3.2: Immunoblots Using Unpurified Sera. Cos-7 cells were transfected with pcDNA- tag2B-hMTM1 (+) which generates Flag tagged human myotubularin in mammalian cells. Control lanes (-) represent lysates from untransfected Cos-7 cells. After lysis, the resultant protein samples were used to test the purified sera from four rabbits (1:1000). The transfection and expression of Flag tagged MTM1 protein was verified using the Anti-Flag antibody (Chapter 2). Red arrow indicates Flag-MTM1 protein at 72 kD. Size markers used were Kaleidoscope Markers (Amersham) for CRI-2 and Flag blots or Dual Color Protein Markers (Bio-Rad) for CRI-1, CRI-3 and CRI-4 blots. 99 cells. None of the sera demonstrated a clear band at the predicted size only in the
transfected cells. Based on studies of previous antibodies to myotubularin seen in the
literature, we expected that endogenous myotubularin (68 kD) would be below the level
of detection in crude lysates 39,73.
We next affinity purified portions of the third bleeds from all of the animals using Amino
Link columns to which the appropriate peptide antigen had been covalently bound. To test the purified sera samples, the same immunoblot strategy with exogenous expression of Flag-MTM1 protein was used. Results are shown in Figure 3.3. CRI-1 and CRI-2 sera demonstrated multiple bands in both transfected and untransfected lanes with the fastest moving band potentially migrating at the correct size. For CRI-2, the lowest bands are of similar intensity and for CRI-1, the lowest band is more intense in the transfected cells.
Since for both of these sera samples the fastest migrating band is present in untransfected
cells, it is unlikely to be Flag-MTM1 protein. For CRI-1 we cannot completely rule out
that the fastest migrating bands represent Flag-MTM1 protein in the transfected sample
and endogenous myotubularin cross-reacting in the untransfected sample due to their
similar molecular weights (68 kD versus 72 kD). It is also possible that CRI-1 contains
an antibody to myotubularin, but also cross reacts nonspecifically with another protein of similar size. CRI-3 appeared specific since a band migrating at the same size as Flag
MTM1 protein was seen only in transfected cells and not in untransfected cells,
indicating specificity for Flag-MTM1 protein. The serum sample seemed very clean since
100 kD + - kD - + 255 160 160 105 105 75 75
+ - kD 50 255 50 160 105 CRI-1 CRI-2 75
- + kD - + kD 160 255 105 50 160 105 75 75 Anti-flag
50 50
CRI-3 CRI-4
Figure 3.3: Immunoblots Using Purified Sera. Cos-7 cells were transfected as in Figure 3.2 with pcDNA-tag2B-hMTM1 (+). After lysis, the resultant protein samples were used to test the purified sera from four rabbits (1:1000). Control lanes (-) represent lysates from untransfected Cos-7 cells. The transfection and expression of Flag MTM1 protein was verified using the Anti-Flag antibody. Red arrows at 72 kD represent the Flag-MTM1 protein.
101 minimal additional bands were observed (not well seen). The purified prebleed for CRI-3
showed no bands at all (data not shown). Additionally, purified sera CRI-3, was able to
detect doxycycline induced Flag MTM1 protein in HeLa cells (Figure 3.5). To further
confirm the specificity of the sera, the MTM1574-591 peptide itself was incubated with
CRI-3 sera prior to its use as a primary antibody (Figure 3.4). As expected, the MTM1574-
591 peptide was able to successfully compete for the antibody, resulting in lack of detection of the protein by immunoblotting.
Next we investigated whether the CRI-3 serum could be used for immunoprecipitation of the Flag-MTM1 protein (Figure 3.5). HeLa cells which express Flag-MTM1 protein in a tet inducible fashion were stimulated with doxycycline and immunoprecipitated with either the Flag antibody or CRI-3. All of the samples were then immunoblotted with the
Flag antibody for detection of Flag MTM1 protein. As expected, upon treatment with doxycycline, Flag-MTM1 was induced and detected by the Flag antibody in crude cell lysates. Tagged protein was detected by the Flag antibody following immunoprecipitation with either the Flag antibody itself or CRI-3. The immunoprecipitation of Flag-MTM1 protein by CRI-3 followed by detection with Flag antibody is cleaner than the immunoprecipitation and detection by the Flag antibody because CRI-3 and Flag are not antibodies from the same species.
102 Anti-FlagCRI-3 CRI-3
MTM1 + - - + - + Peptide - - - - + + 75 kD
50 kD
Figure 3.4: Competition of CRI-3 with the MTM1 574-591 Peptide. Cos 7 cells were transfected with pCMV-Tag2b-hMTM1. Untransfected cells served as controls. Lysates from these cells were probed with anti-Flag, CRI-3, or CRI-3 after competition with the peptide antigen hMTM1 574-591. The Flag MTM1 band detected by CRI-3 was lost when the sera was pre-incubated with the peptide used to raise the sera in animals.
IP IP lysate Flag CRCRI-3
Dox + - + - + - 75
50 o 1 antibody: Anti-Flag
Figure 3.5: Immunoprecipitation of Flag-MTM1 Protein with Purified CRI-3 Serum. HeLa cells that express Flag-MTM1 protein in a tet inducible fashion were stimulated with doxycycline (dox), lysed, and immunoprecipitated with either CRI-3 or the Flag antibody (see Chapter 2 for protocol). Control cells (-) were not stimulated. All samples were immunoblotted with the Flag antibody. Flag MTM1 protein was successfully immunoprecipitated with both the Flag antibody and CRI-3.
103 The CRI-3 serum resulted from the injection of MTM1574-591, a peptide from the human
myotubularin sequence. To further characterize this serum we asked whether or not it
would recognize mouse myotubularin. Based on the alignment between human MTM1574-
591 and the corresponding region of mouse MTM1 (Figure 3.6B), we predicted that CRI-3 would not cross react with mouse MTM1 protein because of differences in the C-terminal region of the epitope. Full length mouse MTM1 was amplified by PCR from an
appropriate EST clone (IMAGE ID 6398326) and cloned into pcDNA4-HisMax-TOPO
(Invitrogen), a mammalian expression vector designed to express mouse MTM1 with N-
terminal Xpress and HisG Tags. This expression vector was transfected into Cos-7 cells
for 48 hours before lysing the cells and immunoblotting (Figure 3.6A). Control protein
lysates were made from untransfected Cos-7 cells and tet-inducible HeLa cells producing
human Flag-MTM1 protein. While both proteins were present as shown by
immunoblotting with the Flag and HisG antibodies, CRI-3 detected only human
myotubularin, not mouse. Since the CRI-3 serum did not recognize the mouse MTM1
protein, it suggests that the epitopes present in human and mouse MTM1 are sufficiently
different.
104 A Anti-HisG Anti-Flag kD kD H+ C+ C- 75 H+ H- 75
50 50 37 37 CRI-3 C- C+ H+ kD 75
50
37
B Human MTM1 LQLANSAKLSDPPTSPS Mouse MTM1 LQLANSAKLADAPASTS
Figure 3.6: Purified CRI-3 Serum Detects Human Myotubularin, but not Mouse Myotubularin by Immunoblotting: A. Cos7 cells were transfected with pCDNA4-HisMax-Topo-mMTM1, which expresses HisG tagged full length mouse myotubularin and lysed for immunoblotting (C+). Control Cos7 cells were untransfected (C-). Lysates were also made from tet inducible HeLa cells expressing Flag-tagged human full length myotubularin that were either stimulated with doxycycline (H+) or not stimulated (H-). Immunoblotting with the HisG antibody detected HisG-mMTM1 protein at the expected size of 73kD (green arrow for C+ lane) and also a strong product slightly larger than 50kD, which may be a degradation product. Immunoblotting with the Flag antibody revealed the presence of Flag-hMTM1 protein in HeLa cells stimulated with doxycycline (red arrow). Flag-MTM1 protein was seen again when the lysates were probed with the purified CRI-3 serum (red arrow), but mouse myotubularin was not detected with this sera. B. Alignment of the MTM1 574-591 peptide used to generate CRI-3 with the analogous mouse MTM1 sequence.
105 3.4 Conclusions and Discussion
To develop an antibody recognizing human myotubularin, the primary structure of myotubularin was analyzed with a set of algorithms to reveal candidate regions optimal for peptide design. This analysis yielded two regions, scoring high in the prediction algorithms, lacking putative post translational modification sites, and lacking significant homology with other members of the MTM gene family. These two epitopes were synthesized commercially with promiscuous T cell antigens and injected into two rabbits each. Following affinity purification of the sera samples, CRI-3, a serum from one of the rabbits injected with the MTM1574-591 peptide, was found to specifically recognize human
myotubularin: It detected overexpressed Flag-MTM1 cleanly, and competition with the
antigenic peptide MTM1574-591 resulted in a loss of signal. Upon further investigation
CRI-3 was also found to be capable of immunoprecipitating exogenously expressed
human MTM1 protein. Due to the difference in sequence of the epitope in mouse, CRI-3
does not recognize myotubularin from this species
Concomitant with our efforts to generate an antibody to MTM1 protein, the laboratory of
J.L. Mandel generated a series of antibodies to myotubularin using injection of rabbits
with full length myotubularin as well as peptides antigens (amino acids 13-32, 365-385,
and 574-596)73. These antigens were also used to generate monoclonal antibodies using
Sf9 insect cells. Notably, one of their peptide antigens (aa 574-596) includes our C-
106 terminal MTM1 peptide (aa 574-591) 73. Their polyclonal antibody from this peptide was able to immunoprecipitate endogenous myotubularin from normal myotubes. One of their major findings was an inability to detect endogenous myotubularin unless the protein sample was first concentrated by an immunoprecipitation step. They immunoprecipitated endogenous myotubularin from 3mg of human myotube crude lysate. This is consistent with our inability to detect endogenous myotubularin in any of our cell samples containing approximately 1mg of crude lysate from either Cos-7 or HeLa cells.
The Mandel laboratory used this series of MTM1 antibodies to immunoprecipitate myotubularin from lymphoblasts, fibroblasts and muscle cell lines from both normal and
MTM1 patients 73. They found that 87% of the 24 patients with characterized mutations in the MTM1 gene displayed either an absence or lower levels of myotubularin on immunoblotting compared to normal control cell lines. Additionally, one patient of five lacking an MTM1 coding mutation, but presenting with the MTM1 phenotype, had no detectable myotubularin via this method of detection. Diagnosis of myotubular myopathy by immunoblotting could be performed for patients without a coding mutation in MTM1.
While there may be a place for diagnosis of myotubular myopathy on the protein level, the test is not available commercially nor is it practical for every patient.
107
CHAPTER 4
EXCLUSION OF MUTATIONS IN THE MTMR1 GENE AS A FREQUENT CAUSE
OF X-LINKED MYOTUBULAR MYOPATHY
4.1 Introduction
The initial description of the MTMR1 gene was published in 1996 as part of the positional cloning of MTM1 2. MTMR1 has 75% amino acid identity with MTM1 and maps to the Xq28, 20kB telomeric to MTM1. It spans 71 kB; both genes are transcribed in the same direction 4. MTMR1 contains 16 exons and the transcript is present by RT-
PCR in all tissues tested, including skeletal muscle, liver, brain and heart. An additional lower abundance transcript is detected by RT-PCR in kidney and lung and denoted
MTMR1a. This form includes an alternately spliced exon 13a (211bp) between exons 13 and 14. Exon 13a possesses a stop codon 16 amino acids downstream and its inclusion in the transcript is predicted to cause the premature truncation of MTMR1 and loss of 113 amino acids. The biological significance of this alternate transcript remains unknown.
The close proximity of MTMR1 and MTM1, along with their high degree of homology suggested that they result from an ancestral duplication 4. An analysis between the coding
108 regions of human MTM1 and MTMR1 genes were compared with mouse Mtm1. From
this analysis using DIVERGE software, it was predicted that a duplication event took
place long before the divergence of primate and rodents 4.
No disease association has been found for MTMR1. Approximately 20% of patients
diagnosed with X linked myotubular myopathy lack mutations in the coding sequence of
the MTM1 gene 12, some with clearly X-linked pedigrees. Given the location of MTMR1
on the X chromosome in Xq28 and its high identity to MTM1, we felt that MTMR1 may be a candidate gene for mutation in patients diagnosed with myotubular myopathy who lack a mutation in MTM1.
4.2 Materials & Methods
4.2.1 Patient Genomic DNA and Exon Sequencing:
Patient samples and clinical and family histories for this study were collected by Gail E.
Herman. Patients were classified as severe, moderate, or mild based on the guidelines in
Herman et al 3 (Table 4.2). Sequencing for MTM1 coding mutations was performed
previously in all patients and was negative. Promoter mutations in MTM1 were not
evaluated; however, most large deletions and rearrangements were previously excluded.
All sixteen exons of the human MTMR1 gene were sequenced from genomic DNA
including an alternatively spliced exon 13 (Genbank Accession # AF002223). Primers
109 were designed to include at least 30 bp adjacent to the exon-intron boundaries to detect
any splicing mutations. Due to the high GC content of exon 1, PCR primers for this exon
included 5’ “GC clamps” and M13 linkers for sequencing. Genomic DNA from MTM
patients was isolated from either whole blood or lymphoblast cell lines using Puregene
DNA Isolation Kit (Gentra) by David Zhao and Kevin Kopacz. Additional patients’
samples were graciously supplied to us as genomic DNA. Reactions were carried out in
25 ul total volume with 100 ng patient DNA, 1.5 mM MgCl2 , 0.4 uM primers, 0.2 uM
dNTPS and 1 unit Amplitaq Gold (Perkin Elmer). PCR for exons 2-16 was performed as
follows: 95 oC for 12 min for one cycle; denaturation at 95 oC for 30 sec, annealing at the
temperatures indicated in table 4.1 for 30 sec, elongation for 30 sec at 72 oC for 35 cycles;
and a final extension at 72 oC for 7 min. For exon 1, 10% DMSO and Promega
Mastermix were used for touchdown PCR as follows: 95 oC for 3 min for one cycle; denaturation at 94 oC for 30 sec, annealing for 30 sec starting at 64 oC and extension for 1
min at 72 oC for 40 cycles, decreasing annealing temperature by 0.2 oC every cycle; and
final extension at 72 oC for 8 min. Sequencing was performed in 10 ul total volume using dRhodamine Cycle Sequencer Terminator Reaction reagents for exons 2-16 and Big Dye
Terminator for exon 1, as recommended by Perkin Elmer. For exons 2-16, the same primers used for PCR amplification were used for DNA sequencing of both strands (see
Table 4.1). For exon 1, the M13 forward primer 5’CAGGAAACAGCTATGAC3’was
employed for sequencing with 5% DMSO added to the cycle sequencing reaction. All
sequencing was performed on an ABI 377 automated sequencer by the Children’s
Research Institute Sequencing Core.
110 Exon Forward Primer Reverse Primer Annealing 5' --> 3' 5' --> 3' Temp oC 1 GCGCCAGGAAACAGCTATG GCGCGTAAAACGACGGCCAG 64-56 ACCGGGCGGGCGGTATAGA AGGTCCGAGGGCAGCAGT 2 CACTGAGATCCTATGATATAACG CTTCTTGGCTACTTAGCATACTG 56 3 AGATTTACTCCCGTTTGGAG CATACATCACCTCCACTGGC 52 4 ATGGAAGCACAAGTCTATGG GTAATGGGATTTCAGGGGAC 52 5 TGAAAACTGTTGACTGACAC CTACTGTCCAAAGACCATG 52 6 GCACGGTATCTGGCTTATAG AACATAGGATATTTGGACCC 52 7 TTGTCAGTTCTGGGTAATGC TTGGCACAGGGAGAAGAGAG 52 8 GTAAAGTCATGCATTTCACC GAATCAAAGGTATGTTCAGG 55 9 AAGAGACCGTGCTTGAGAAG ACCAGGGAGGAATAGCTATG 55 10 AGCCTCCAGAACTAGGAGAG TTCATGCACACAGCCAAGAG 52 11 ACTTCCAAAGCTCTCCCCAG GTGTATATATGCGCTTACGTG 52-53 12 ACCTTCTGTTTCTGTTCAGC TGGACCTAGTGGACAAGATG 53-55 13 TTACATGGGCTCAGTCTTCC CCCTCCTACAGTAGCACTTC 55 13a ATTGCTGACTGTAAATGCCC GCCTTCTCACTTTTGACCAC 55 14 CTGGGACCTAGTAGACATTC TCAGAGTTGGCGTATGGGTG 52 15 AAAGCTCCTGAAGCCAACAG CAGCCATAACAGTGATCGCC 52 16 TGGGGACATTTGCCATTTTC GTGATTGACGCACAGTAAAC 55
Table 4.1: Primer Sequences for Amplification of MTMR1 Exons. Genomic DNA derived from blood samples of X-linked myotubular myopathy patients were sequenced for coding or splicing mutations in MTMR1 using these primers which are each located at least 30 base pairs outside of the intron exon boundaries. Table adapted from Copley et al 1.
4.3 Results
We have screened the MTMR1 gene for coding mutations in 14 patients for whom no mutation was found in MTM1 itself (Table 4.2). Pedigrees for two of the patients were consistent with an X-linked pattern of inheritance (Fig. 4.1). Two probands had affected male siblings and ten had a family history with no similarly affected male relatives. For all patients, a diagnosis of MTM was confirmed by muscle biopsy, and mutations in the coding region of the MTM1 gene itself had already been excluded by genomic
111 sequencing 12,46. MTM1 transcript or protein levels were not evaluated in this set of patients.
All sixteen exons of MTMR1 were sequenced from genomic DNA including an alternatively spliced exon 13a (see Table 4.1). No mutations were detected among the fourteen patients studied. These results suggest that mutations in MTMR1 are not a frequent cause of MTM1 in males for whom no mutation is found in the MTM1 gene itself. It is possible, however, that some of these patients possess mutations in regulatory elements or non-coding regions of MTM1 or MTMR1. This study was published in
Copley et al 1.
Family 2: Family 13:
I
II I
III II
IV
Figure 4.1: Pedigrees of Two MTM Families Studied That Are Consistent with an X-Linked Pattern of Inheritance. In family 13, individual III-6 had a muscle biopsy consistent with MTM and died shortly after birth. III-3 reportedly died with severe hyaline membrane disease following delivery at 30 weeks gestation. III-7 died with a neural tube defect. Figure adapted from Copley et al 1.
112 Patient Family History Clinical Phenotype 1 Negative Classic 2 X-linked (Fig. 4.1 ) Classic 3 Affected male sibling Classic 4 Negative Classic 5 Negative Very mild 6 Negative Moderate 7 Negative Classic 8 Negative Mild 9 Negative Moderate 10 Negative Classic 11 Negative Mild 12 Affected male sibling Mild 13 X-linked (Fig. 4.1) Classic 14 Negative Classic Table 4.2 Family history and Disease Severity in Patients Sequenced for MTMR1 Mutations. Table adapted from Copley et al 1. Patients originally described in Herman et al 3.
4.4 Conclusions and Discussion
No mutations in the coding region of MTMR1 were detected in patients lacking a mutation in MTM1. This finding excludes MTMR1 as a cause of X-linked myotubular myopathy in our set of fourteen patients. These fourteen patients, including two with clear X-linked inheritance, had the diagnosis of myotubular myopathy confirmed by muscle biopsy and more than half presented with the classic severe phenotype. It is
113 unclear what the genetic lesion is in these patients. We cannot rule out the presence of a
physiologically significant mutation in a promoter or enhancer element in MTM1. The
protein levels of myotubularin have not been tested in these patients. Doing so may shed
light on the diagnosis of these particular patients. As discussed in Chapter 3, there is at
least one documented case of a patient with a severe presentation of myotubular
myopathy who lacks a mutation in the MTM1 gene and also lacks protein as detected by
immunoprecipitation of myotubularin from lymphoblasts 73.
Since the publication of this work 1, Buj-Bello et al. demonstrated that there are three
alternatively spliced forms of exon 2 in MTMR1 called 2.1, 2.2, and 2.3 located between exons 2 and 3 5 (Figure 4.2). Exons 2.1, 2.2 and 2.3 are very small at 8, 9, and 17 amino
acids, respectively. Profiling by RT-PCR demonstrated expression of exon 2 (isoform A)
and 2 + 2.1 (isoform B) early during myoblast differentiation with expression of exons 2
+ 2.1 + 2.2 (isoform C) with myoblast fusion. This pattern was similar in mouse embryo
and adult skeletal muscle samples as well as human muscle biopsies from a 7 week old
infant and adult. Both isoforms A and C have phosphatase activity towards PI3P in vitro
and share a similar subcellular localization as myotubularin.
The additional exons described in the literature have not been investigated for mutations
in our series of patients. Thus, it remains possible that mutations in these exons could be
present. We believe this is unlikely since the isotype switching to isoform C, which
includes these additional exons, was not apparent in muscle obtained from a 7 week
114 infant. In addition, mutations in these alternatively spliced forms of exon 2 were excluded in another set of patients lacking MTM1 mutations 5.
1 2 Exons 3-12 13 Ex 14-16 MTMR1, isoform A 662 aa
1 2 Exons 3-12 13 13a MTMR1a 565 aa
1 2 2.1 Exons 3-12 13 Ex 14-16 MTMR1, isoform B 670 aa
1 2 2.1 2.2 Exons 3-12 13 Ex 14-16 MTMR1, isoform C 679 aa
1 2 2.1 2.2 2.3 Exons 3-12 13 Ex 14-16 MTMR1, isoform D 696 aa
Figure 4.2: Schematic Representation of Splice Forms of MTMR1 Described in the Literature. MTMR1 (isoform A) was first described in 2. A truncated splice form MTMR1a, resulting from the addition of exon 13a, was described in kidney and lung by RT-PCR4. Isoforms B, C, and D were first described in 5 and result from the splicing of exon 2.1, 2.2, and 2.3 which add 8, 9, and 17 amino acids, respectively. Exons are not drawn to scale.
115
CHAPTER 5
EFFECTS OF MYOTUBULARIN OVEREXPRESSION ON CELL SHAPE
5.1 Introduction
Cell shape changes during a variety of processes including adhesion, spreading and migration. These processes require the reorganization of actin filaments as well as other structural networks within cells. Cells reorganize actin into specialized structures to accomplish spreading and migration, including the development of filopodia, lamellipodia, and membrane ruffles. Filopodia are long, tentacle-like protrusions of the plasma membrane. Lamellipodia are wide lateral extensions of the plasma membrane and ruffles are dynamic, sinewy, lateral extensions rich in F-actin 94.
The Rho GTPases, including Rac1, cdc42, and RhoA, have all been implicated in processes affecting cell shape 95-97. These small proteins exist in two forms: an active
GTP-bound form and inactive GDP-bound form. Adhesion and growth factors such as
EGF, PDGF and CSF-1 have been found to stimulate Rac195,97,98. Bradykinin and LPA have been suggested to simulate cdc42 and Rho, respectively 96,99.
116
The effects of Rac1 were initially described by Ridley et al in 1992 97. A G12VRac1
mutant codes for a constitutively active form of the protein that induces membrane ruffles
when overexpressed in fibroblasts. Ruffles were also induced within 5 minutes when
either wildtype Rac1 or G12VRac1 recombinant proteins were injected into 3T3 cells.
Actin colocalization to the ruffles was also noted. A dominant negative mutant of Rac1
that binds GTP with reduced affinity, S17NRac1, failed to induce ruffles upon injection.
This S17NRac1 mutant also prevented PDGF stimulation of membrane ruffling. This
pattern of actin reorganization was also demonstrated in macrophages95. The constitutively activated Rac1 mutant induced ruffles and lamellipodia, whereas the dominant negative form prohibited ruffles and lamellipodia after stimulation with CSF-1.
Interestingly, it appears that cdc42 may be able to activate Rac1 95,99. Upon injection of
activated cdc42 into 3T3 or macrophages, microspikes and filopodia were quickly seen
within a few minutes and attributed to the increase in activated cdc42. Within 30 min,
lamellipodia and ruffles were seen, suggesting that Rac1 had been activated. When cdc42
was injected with RacGAP, a protein which sequesters Rac1 and inhibits activation, the
filopodia were still observed, but not the ruffles.
Cell spreading is the process by which a cell flattens out and begins to extend its plasma
membrane. The formation of lamellipodia and filopodia are involved in this process.
When 3T3 cells are plated in tissue culture dishes, they start to develop lamellipodia and
117 spread within 3 hours 100. They also demonstrate an increase in activated Rac1 and cdc42
that is sustained for at least 24 hours. One might expect this association of Rac1 and
cdc42 activation during cell spreading since these proteins induce lamellipodia and
filopodia, respectively. In addition to these observations, injection of dominant negative
Rac1 into 3T3 cells partially inhibited cell spreading and injection of dominant negative
cdc42 profoundly inhibited cell spreading 101. When dominant negative cdc42 was co-
injected with dominant active Rac1, the ability of cells to spread was partially restored.
These data suggest that both cdc42 and Rac1 participate in cell spreading.
Much work has been done to elucidate the mechanisms by which activated Rac1 and cdc42 contribute to actin reorganization. Significant progress has been made on one particular model suggesting that Rac1 binds to the protein IRSp53 which in turn binds to
WAVE2(Scar) 102,103. All three of these proteins localize to membrane ruffles 102-104.
When cells were transfected with activated Rac1 and IRSp53 lacking the critical region for WAVE2 binding, ruffles were lost. Ruffles were also suppressed when activated Rac1 was overexpressed with the PR region of WAVE2, which acts as a dominant negative and sequesters IRSp53. WAVE2 has been shown to stimulate the Arp2/3 complex in vitro and this activity was increased with the addition of IRSp53 103. The Arp2/3 complex
functions in actin nucleation and is concentrated at the leading edge of cells 105. A similar model has been postulated for cdc42, except cdc42 binds N-WASP which in turn activates the Arp2/3 complex, promoting actin reorganization 106.
118 Our interest in cell shape came with the observation by Laporte et al that myotubularin
localizes to the plasma membrane 107. This localization was strongly seen in ruffles upon
stimulation of cells with dominant active Rac1. Localization to Rac1-induced ruffles was
also seen with the D278A and C375S mutant proteins, suggesting that it does not require enzyme activity. Localization of myotubularin to endogenous ruffles and filopodia was observed, and in a small subset of cells, overexpression of myotubularin seemed to produce numerous filopodia. Based on these data, we decided to evaluate the effects of stable myotubularin overexpression on cell shape. We chose to observe cell spreading and membrane ruffling in the context of over expression of the Flag-MTM1 protein.
5.2 Materials and Methods
5.2.1 Cell Spreading Assay:
To characterize the amount of spreading at multiple times with various levels of MTM1 protein expression, tet-inducible HeLa clones were tested either with or without doxycycline treatment. Individual clones were fed fresh media with doxycycline every 24 hrs for 48 hrs prior to splitting for spreading assays. To assess spreading, cells were split
5 at a density of 1x10 cells per T25 flask at timepoint 0 into DMEM supplemented with
10% FBS (tetracycline free), 50 I.U./mL penicillin, 50ug/mL streptomycin,
100ug/mLG118, 0.2ug/mL puromycin, and 2mM L-glutamine either with or without
doxycycline (2000ug/mL). At 2 hr post splitting, the cells were rinsed with PBS (warmed
119 to 37oC) to remove any unattached cells. The appropriate media was then replaced for the
remainder of the experiment. At 2, 4, and 6 hours, the cells were viewed at 40x
brightfield magnification and images were captured with a digital camera and SPOT software. Approximately 18 images representing 150-200 cells were observed for each
sample at each timepoint. Cells were counted as spread if they appeared angular,
flattened, and non-light refractive or dim. Cells were counted as unspread if they were
predominantly round and brightly refractive. Highly clumped cells were disqualified in
the quantitation. Spreading was expressed as the percent of spread cells from total cell
number. Standard deviations were calculated over separate experiments with Microsoft
Excel.
5.2.2 Cell Ruffling Assay:
To investigate ruffling in tet inducible HeLa cell lines, individual clones were treated
with doxycycline (2000ug/mL) for 48 hr along side an untreated sample. Treated cells
were fed doxycycline-containing media every 24 hrs to assure a steady level of
doxycycline stimulation. Twenty-four hrs before fixation, cells were split onto glass
coverslips at a confluency of approximately 50%. For fixation, cells were washed 5x
with PBS followed by a 30 min room temperature incubation with 4% paraformaldehyde
in PBS. After washing 5x in PBS, cells were permeablized in 0.2% Triton-X100 in PBS
at room temperature for 20 min. Cells were washed 5x with PBS and labeled with
BODIPY-fl phallocidin (Molecular Probes B607) that binds to F-actin at a v/v ratio of
120 1:250 in PBS for 20 min in the dark. Coverslips were mounted with Vector mounting
medium with DAPI and visualized at 40x magnification using a Zeiss LSM 410 confocal
microscope with an argon-krypton laser. Digital images were captured with Zeiss
software. Ruffles were defined and counted as laterally protruding leading edges rich in
F-actin. Approximately 100 cells were counted for each sample, with the extent of
ruffling expressed as the total number of ruffles divided by the total number of cells. Cell
samples were evaluated for ruffling from digital images by 2 independent observers.
Standard deviations were calculated over separate experiments with Microsoft Excel.
5.2.3 Rac1 Activation Assay.
The level of activated, GTP-bound Rac1 protein was evaluated using a Rac1 Activation
Kit from Upstate Biotechnology. This assay exploits the observation that activated, GTP bound Rac1 binds the PDB(Pak-1 Binding Domain) region of PAK1, while the inactive,
GDP-bound form does not. Briefly, Clone 83 was cultured with doxycycline stimulation
(2000 ug/mL) in 100mm dishes for 48 hrs before lysis at 4oC in 800uL of MLB (5X
MLB: 125mM HEPES, pH 7.5, 750mM NaCl, 5% Igepal CA-630, 50mM MgCl2, 5mM
EDTA and 10% glycerol). Untreated cells served as controls. Dishes were approximately
90% confluent at the time of lysis. The resultant protein samples were pre-cleared with
100uL of GST-agarose for 10 min at 4oC with nutation. The samples were centrifuged at
16,000g for 2 min at 4oC and the supernatant was removed. To visualize activated Rac1
on immunoblot and verify that the immunoprecipitation was working, 500uL
121 (approximately 1.2mg total protein) of lysis from untreated cells were loaded in vitro
with GTPγS by adding 1uL of 0.5M EDTA and 5uL of GTPγS. This mixture was nutated
at room temperature for 30 min and the reaction was stopped by the addition of 32.5uL of
1M MgCl2. A similar in vitro reaction was performed on another aliquot of lysis from untreated cells using GDP as a negative control. Endogenous Rac1 was
immunoprecipitated from the experimental samples by adding 5uL (5ug) of PAK1-PBD
agarose and nutating at 4oC for 1 hr. The control samples preloaded with either GTPγS or
GDP were immunoprecipitated with PBD agarose in the same way except they were only
incubated for 30 min. All samples were centrifuged for 2 min at 16,000g at 4oC and washed with 500uL of MLB 3x before the addition of 40uL of SDS Loading Buffer
(Section 2.2.3 C). The samples were boiled for 5 min at 100oC and immunoblotted as
described in section 2.2.3 C, with the following exceptions. Membranes were blocked in
PBS- 3% milk for 30 min, and incubated with the anti-Rac1 antibody (1:500) overnight at
4oC in PBS-3% milk. Membranes were washed 3x with water and incubated for 1 hr at
room temperature with anti-mouse-HRP (1:1000). Membranes were washed 3x with
water and once with PBS-0.05% Tween 20 before performing chemiluminescence.
Protein concentrations were detected by Bradford Assay and 30ug of total protein were loaded for each sample.
122 5.3 Results
To evaluate cell spreading in the context of myotubularin overexpression, we developed several HeLa cell clones that express Flag-MTM1 protein when treated with doxycycline
(See Chapter 2). All of these clones failed to exhibit Flag-MTM1 protein expression when cultured without doxycycline. Clone 12 continuously lacked Flag-MTM1 protein expression in the presence of 2000ug/mL doxycycline, the highest dosage used, and thus served as a negative control in the spreading assays. HeLa clone 83 demonstrated a high level of Flag-MTM1 protein expression. HeLa clone 6 initially demonstrated an intermediate level of Flag-MTM1 protein expression with doxycycline stimulation, which significantly decreased after passaging. The initial question was whether any of the clones showed differences in cell spreading when treated with doxycycline to induce
Flag-MTM1 protein. To address this question, we performed cell spreading assays on the three independent HeLa clones. Clones were treated with doxycycline for 48 hrs, split at equal densities, and pictures were taken every 2 hr for 6 hr using 40x brightfield microscopy. Eighteen digital images were collected for each sample at each timepoint and were evaluated for spread cells. We initially performed cell spreading assays on
HeLa clones 12, 6 and 83 in triplicate (Figure 5.1a&b). Clone 6 initially expressed an intermediate level of Flag-MTM1 protein, but the levels decreased sharply in the second and third replicates. Cell spreading assays for these three clones showed that there was no significant difference in cell spreading regardless of doxycycline treatment for clones 12 and 6. Interestingly, HeLa cell clone 83, which exhibited high levels of Flag-MTM1
123 Figure 5.1: Cell Spreading in Three Independent Clones Over Three Independent Experiments. A. HeLa cell clones 12, 6 and 83 were isolated as tet-inducible stable transfectants that express Flag-MTM1 protein in response to doxycycline treatment. Clone 12, which did not express Flag-MTM1 protein, served as a negative control. Clone 6 expressed Flag-MTM1 protein at an intermediate level during the initial experiment, but lacked Flag-MTM1 protein expression in the second and third experiments in the series. Clone 83 demonstrated a high level of Flag-MTM1 protein in all three determinations of spreading. B. Clones 12, 6, and 83 were stimulated with doxycycline for 48 hours and then split at equal density into T25 flasks at time 0. Untreated cells served as controls. At 2,4, and 6 hours, digital images of the cells were captured during bright field microscopy. From these images, cells were counted as spread or unspread and the percent spreading over time was calculated by dividing the number of spread cells by the total cell count. Error bars represent standard deviations as calculated over three separate experiments.
124 A Trial 1 83 6 12 dox + - + - + - 75 kD Flag-MTM1 50 kD
Trial 2 83 6 12 dox - + - + - +
75 kD Flag-MTM1
Trial 3 83 6 12 dox + - + - + -
75 kD Flag-MTM1
125 Figure 5.1 Continued
B clone 12
100 90 80 70 60 50 40 % spread 30 clone 12 + dox 20 10 clone 12 - dox 0 0 hr 2 hr 4 hr 6 hr time
clone 6
100 90 80 70 60 50 40 % spread clone 6+ dox 30 clone 6 - dox 20 10 0 0 hr 2 hr 4 hr 6 hr time
clone 83
90 80 70 60 50 40 % spread 30 clone 83 + dox 20 clone 83 - dox 10 0 0 hr 2 hr 4 hr 6 hr time
126 Figure 5.2. Cell Spreading in Clone 83 Over a Range of Dosages: Compilation of Three Independent Experiments. A. Clone 83 cells were treated for 48 hours with the indicated dosages of doxycycline before splitting at equal densities at time 0 for evaluation of cell spreading. Error bars represent standard deviations calculated over three separate experiments. Table shows mean values obtained for each time point +/- standard deviation. B. Representative immunoblot showing that Flag-MTM1 is expressed in Clone 83 in a doxycycline dosage response fashion. A two-way Analysis of Variance (ANOVA) was performed by Drs. Soledad Fernandez and Nicole Kelbick (Children’s Research Institute). An overall model involving time and dox was statistically significant (p < 0.0001) but no significant interaction between dose and time was detected (p = 0.6323), as expected. Multiple comparisons involving time frames showed that percent spread was significantly different between 2 hours and 4 hours (p < 0.0001) and 2 hours and 6 hours (p < 0.0001) but not 4 hours and 6 hours. Percent spread was significantly different between all non-zero dose levels compared to 0 dox (p < 0.0001 for each pairwise comparison) except 100 dox versus 0 dox. The difference between percent spread between 100 dox and 2000 dox was also significantly different (p < 0.0001).
127 Figure 5.2
A 100 90 0 dox 80 100 dox 300 dox 70 500 dox 60 2000 dox 50 40 30 20 10 0 0 hr 2 hr 4 hr 6 hr
Doxycycline Dosage
Time (hr) 0 100 300 500 2000 0 0 0 0 0 0 2 62 ± 3.6 52 ± 4.4 46.5 ± 10.3 50.9 ± 5.1 45.5 ± 7 4 80.1 ± 2.9 74.9 ± 3.7 64.9 ± 3 63.3 ± 5.1 58.1 ± 11 6 82.3 ± 5.7 77.3 ± 4.8 71.4 ± 3.7 65.2 ±.3 58.3 ± 8.1
B Dox 0 100 300 500 2000
Flag MTM1
- -
128 protein, showed a decrease in cell spreading when treated with doxycycline. Since no
differences in cell spreading were noted with clones 12 and 6 upon doxycycline
stimulation, we hypothesized that overexpression of Flag-MTM1 protein and not the
mere presence of doxycycline may be responsible.
Since we could not replicate these findings with another high expressing clone, we
investigated the changes in cell spreading by performing doxycycline dosage studies in
HeLa clone 83. The tet-on system is known to behave in a dose responsive fashion76,77, thus by treating Clone 83 with various dosages of doxycycline, we could investigate whether spreading correlated with levels of Flag-MTM1 protein. Cell spreading was evaluated over several Flag-MTM1 protein levels using our standard cell spreading assay.
This experiment was performed on three separate occasions and is graphically represented in Figure 5.2a. The deficiency in cell spreading became more pronounced at higher dosages of doxycycline and, thus, higher levels of Flag-MTM1 protein. While the standard deviations across these data overlap, the pattern is clear. Additionally, at dosages of 300ug/mL of doxycycline and higher, the standard deviations do not overlap with the untreated cells as can be seen from the actual numbers in Figure 5.2a. A two- way Analysis of Variance (ANOVA) demonstrated statistical significance between all doxycycline dosages compared to untreated cells with the exception of the 100 ug/mL dosage. We attempted to establish this spreading deficiency in another tet-on HeLa cell line, clone 16. We performed a dosage study with one trial and found that while clone 16 had increased expression with doxycycline administration, there was no concordant
129 decrease in spreading (Figure 5.3a). These results seemed to directly contradict those
seen with clone 83; however, upon performing side by side immunoblots with equivalent
levels of total protein, we were able to establish that clone 16 has significantly lower
levels of Flag-MTM1 protein expression than clone 83 (Figure 5.3a). Upon comparison
of the two immunoblots, it can be seen that at the highest dosage of doxycycline (2000
ug/mL), clone 16 is expressing Flag-MTM1 protein at a level similar to levels obtained with the lowest dose of doxycycline (100ug/mL) in clone 83. Since there was no significant difference in the percent spreading between treatment of clone 83 with
100ug/mL doxycycline and absence of doxycycline, we would expect no difference in spreading in clone 16 due to its low level of Flag-MTM1 protein expression.
We also investigated another cellular process affecting cellular shape, membrane ruffling.
Myotubularin has been shown to localize to Rac1 induced membrane ruffles 39, but no studies have quantitatively evaluated ruffling in the context of myotubularin overexpression. We began by assessing the level of ruffling in three independent HeLa tet-on clones with or without induction of Flag-MTM1 protein expression with doxycycline treatment (Figure 5.4). Briefly, cells were induced with 2000ug/mL doxycycline and split onto glass coverslips 24 hours prior to fixation and permeablization. The coverslips were stained with BODIPY labeled phallocidin which stains F-actin. Ruffles were defined as laterally protruding leading edges rich in F-actin, a component commonly visualized in membrane ruffles 97 (See Figure 5.5). Three
independent clones were used for this experiment: the immunoblot demonstrating the
130 A
Cell Spreading VS. Time at Various Dox Doses (Clone16)
100
80
60 0 dox 100 dox 40 300 dox % spreading 20 500 dox 2000 dox 0 0246 Time (hr)
B Dox 2000 500 300 100 0
Clone 16
Clone 83
Figure 5.3: Cell Spreading in Clone 16 over a Range of Doxycycline Dosage. A. HeLa Clone 16 was treated with the indicated levels of doxycycline (ug/mL) for 48 hours prior to splitting at equal densities into T25 flasks for spreading assays. Percent spreading is represented for one experiment and no clear dose response is seen. B. Immunoblot of cells at time 0. Spreading assay shows that clone 16 exhibits dose dependent expression of Flag-MTM1 protein, but even at the highest dose (2000), there is significantly less than is seen with the lowest dose of doxycycline on Clone 83 (100). Equal quantities of protein were loaded and .immunoblots were exposed for equal amounts of time.
131 levels of expression of these three clones in this experiment is presented in Figure 5.1a, replicate 1. This experiment was performed while clone 6 was still producing Flag-
MTM1 protein at an intermediate level. Clones 83 and 6 demonstrated a decrease in the level of membrane ruffling that was detected by two independent observers. Interestingly, the higher expressing clone 83 appeared to have a lower level of membrane ruffling than clone 6, the intermediate expressing clone. Observer 2 noted a similar level of ruffling in non-expressing clone 12 regardless of doxycycline treatment, while Observer 1 noted increased ruffling with doxycycline treatment. Taken together, these observations suggest that mere treatment with doxycycline alone does not produce a decrease in ruffling since clone 12 fails to express Flag-MTM1 protein and fails to exhibit loss of ruffling with doxycycline treatment. From this initial study, we chose to further investigate the apparent decrease in ruffling with Flag-MTM1 protein overexpression in clone 83. The ruffling assay was performed on clone 83 on three separate occasions (Figure 5.6). The data from Observer 1 suggests that there is a statistically significant decrease in ruffling with overexpression of Flag-MTM1 after administration of doxycycline. Observer 2 also recorded a pattern of decreased ruffling with overexpression of doxycycline, but the calculated standard deviations were large, prohibiting a definitive conclusion.
From these ruffling and spreading data and considering the localization of myotubularin to plasma membrane ruffles 39, we hypothesized that myotubularin may affect levels of
activated Rac1 in cells. Increases in activated Rac1 levels are associated with increases in
lamellipodia and ruffling95,96. The small GTPase Rac1, when bound to GTP and thus
132 activated, binds to the PBD domain of Pak1, while the inactive, GDP-bound form does not. By immunoprecipitating cell lysates with PBD-agarose, the level of activated Rac1 can be gauged by immunoblotting. To test this hypothesis, we treated Clone 83 with doxycycline for 48 hr and then performed the Rac activation assay (Figure 5.7). Lysates were treated with either GTPγS or GDP in vitro prior to immunoprecipitation to serve as positive and negative controls, respectively. Treatment of protein lysates with the non- hydrolyzable GTPγS is expected to bind all of the Rac1 protein, whereas endogenously activated Rac1 in the context of the tissue culture cell is expected to be only a fraction of the available Rac. As expected, when the positive control lysate was pretreated with
GTPγS prior to immunoprecipitation, Rac1 was detected by immunoblotting.
Accordingly, when the negative control lysate was pretreated with GDP, no activated
Rac1 was seen after immunoprecipitation. Unfortunately, no activated Rac1 was detected in experimental Clone 83 lysates with or without doxycycline treatment. Since a small fraction of Rac1 is thought to be endogenously activated in cells, it may be that the levels in our cells are below the sensitivity of the system. Accordingly, no conclusion regarding the effects of MTM1 overexpression on the activation of Rac1 can be drawn at this time.
133 Observer #1
60
dox - 50 dox +
40
30
(ruffles/cell)*100 20
10
0 clone 83 clone 6 clone 12
Observer # 2
100
90 dox - 80 dox + 70
60
50
40
(ruffles/cell) * 100 * (ruffles/cell) 30
20
10
0 clone 83 clone 6 clone 12
Figure 5.4: Comparison of Percent Ruffling in Three Independent HeLa tet-on Clones Expressing Flag-MTM1 Protein. All three clones were cultured and prepared as described in Materials & Methods. Digital images were captured with confocal microscopy and evaluated by two independent observers. Observers counted the number of ruffles per field and divided it by the number of cells per field to yield percent ruffling. The level of Flag-MTM1 protein present in each clone after stimulation with doxycycline is shown in Figure 5.1 Trial 1.
134
Clone 83 - dox Clone 83 + dox
Figure 5.5: Clone 83 Cells Demonstrate Fewer Ruffles Qualitatively When Treated with Doxycycline to Induce Flag-MTM1 Expression. Clone 83 was cultured on coverslips for 24 hours before fixing with 4% paraformaldehyde in PBS and permeablizing with 0.2% triton in PBS. Permeablized cells were stained with BODIPY-phallocidin and visualized by confocal microscopy. Yellow arrows denote ruffles.
135 Observer #1 Ruffling Synopsis
70 60 50 40 30 20 10 0 dox + dox-
Observer #2 Ruffling Synopsis
70 60 50 40 30 20 10 0 dox + dox-
Figure 5.6: Percent Ruffling Noted by Two Observers in Three Experiments. Clone 83 cells were prepared as described in Figure 5.5 and Materials and Methods. Confocal microscopy images from three separate experiments were captured and evaluated by two independent observers. The Y axis represents percent ruffling calculated by dividing the number of ruffles per sample by the number of cells per sample. Error bars represent standard deviation.
136 Anti-Rac1 IP: PBD-agarose 25 kD Rac1
20 kD Dox + - - - GDP - - + - GTPγS - - - +
Figure 5.7. Rac1 Activation Assay Performed on Clone 83 Cellular Lysates after Treatment with Doxycycline. Clone 83 was treated with doxycycline for 48 hr prior to lysis with MLB, and immunoprecipitation of activated Rac1 with PBD-agarose. Untreated Clone 83 lysates were pretreated in vitro with either GTPγS or GDP for 30 minutes at room temperature, before immunoprecipitating activated Rac1 with PBD-agarose. Activated Rac1 (22kD) was detected, as expected, after treatment with GTPγS and immunoprecipitation. No activated Rac1 was detected in the sample pretreated with GDP, as expected. When Clone 83 lysates were immunoprecipitated with PBD-agarose, no activated Rac1 was seen. Upon in vitro stimulation of Rac1 with GTPγS, nearly all of the protein will become activated. In vivo, only a small fraction of Rac1 may be activated. This may indicate the level of activated Rac1 is below the detection limit of this system.
5. 4 Conclusions and Discussion
The goal of these experiments was to investigate the effects of myotubularin overexpression on cell shape. We made use of several HeLa clones that stably express
Flag-MTM1 protein when induced with doxycycline. To our knowledge, this is the only description of stable overexpression of myotubularin in tissue culture cells. With the study by Mandel et al showing that myotubularin localized to the cytoplasm and plasma membrane39, we hypothesized that myotubularin may affect cell spreading. We evaluated
cell spreading in the context of Flag-MTM1 protein overexpression in our multiple HeLa
137 Tet-on clones. Initial experiments revealed that clone 83, our highest overexpression clone, demonstrated a deficiency in cell spreading that was seen at 2, 4, and 6 hr after cell splitting. The difference in spreading between treated and untreated samples of clone 83 was statistically significant in three separate experiments. In contrast, clone 12, a non- expresser, demonstrated no change in cell spreading over the 6 hrs of observation. Taken together, we conclude that overexpression of Flag-MTM1 proteins results in a reduction of cell spreading at least in clone 83. We sought to verify this result in another HeLa tet- on clone, but we did not possess another independent, high expressing clone. Instead, we examined the dose responsiveness of spreading in clone 83 treated with increasing levels of doxycycline to induce increased levels of Flag-MTM1 protein. As the level of Flag-
MTM1 protein increased, the extent of spreading decreased in a dose dependent fashion.
The difference in spreading between untreated and treated cells could be seen statistically at doxycycline doses as low as 300ug/mL. From these data, we conclude that overexpression of Flag-MTM1 protein results in decreased cell spreading.
With the observation that myotubularin localized to plasma membrane ruffles 39, we hypothesized that overexpression of myotubularin may affect plasma membrane remodeling and affect membrane ruffling quantitatively. We conducted ruffling assays in clones 83, 6, and 12. Two independent observers recorded fewer ruffles in clones 83 and
6 when treated with doxycycline compared to untreated cells, suggesting that overexpression of Flag-MTM1 protein may result in reduced ruffling. On three separate occasions, the ruffling assays were performed on clone 83. Both observers saw a pattern
138 of decreased ruffling with doxycycline treatment, but only the standard deviations
derived from the recordings of observer number 1 were non-overlapping. From these
data, we conclude that overexpression of myotubularin may result in decrease in ruffling in clone 83, but we caution that the observations from only one observer were statistically significant. We hypothesized that a decrease in ruffling may be due to a decrease in levels
of activated Rac1 secondary to myotubularin overexpression. Because of the suggestive
ruffling data, we attempted to quantitate the levels of activated Rac-1 in treated and
untreated clone 83 cell lysates. Unfortunately, there was too little activated Rac1 to detect
by immunoprecipitation with PBD-agarose, preventing any conclusion from being drawn.
Traditionally, quantitation of activated Rac1 in vivo is performed after stimulation with known activators, such as EGF, PDGF or CSF-1. While stimulating our cells with one of
these growth factors may have allowed us to quantitate activated Rac1, we recognize that
this would not truly represent the conditions under which changes in spreading and
ruffling were seen. Alternatively, it may be interesting to evaluate activated cdc42 levels
in the context of myotubularin overexpression, since cdc42 may activate Rac1 and loss of
cdc42 results in a deficiency in cell spreading.
139
CHAPTER 6
FUTURE DIRECTIONS
In conclusion, we have attempted to identify proteins that interact with myotubularin as a way to understand its role in the cell. Although proteins initially found in our yeast 2- hybrid screening could not be confirmed by either coimmunoprecipitation or GST pulldown strategies, it remains important to recognize the progress made on protein- protein interactions in the MTM field. With the discovery of the interactions between
MTMR7 and MTMR9 70 and between MTMR2 and MTMR5 41, it was shown that the
coiled coil domains of these proteins were important for interactions. The interaction
between myotubularin and MTMR12 required the SID of MTMR12, but the necessary
region of myotubularin was not mapped 72. It remains to be seen if other proteins outside
the MTM family interact with myotubularin. If additional partners for myotubularin are
sought, it may be important to use the whole protein for analysis and carefully select the cDNA library if a screening method is employed.
It was unexpected that an active phosphatase such as MTMR2 interacts with a dead phosphatase such as MTMR5 and that this interaction regulates subcellular localization.
140 In fact, the interaction of MTMR2 with MTMR5 actually seems to increase the enzyme’s activity in vitro. A similar cooperation was also seen in vitro between MTMR7 and
MTMR9. These data directly challenge the hypothesis that the dead phosphatases
antagonize the effects of the active phosphatases. With these results observed amongst
MTM family members, it will be interesting to see what physiological role the interaction
between MTMR12 and myotubularin plays. Overexpression of MTMR12 is known to
affect the subcellular localization of overexpressed myotubularin 72. It will be interesting to see where MTMR12 localizes in cells lacking myotubularin from the knock out mice or if MTM1 proteins from patients have affected MTMR12 binding. It also may be interesting to generate both MTMR12 knockout and overexpression mouse models to further investigate the physiological consequences of this interaction. Additionally, enzyme kinetic studies evaluating a possible cooperation between MTMR12 and myotubularin and mapping the critical regions of both proteins for the interaction will yield insight into the nature of MTM1.
What myotubularin does in cells and how deficiency of the protein leads to X-linked myotubular myopathy is still elusive. Since the first work of Taylor et al suggesting that myotubularin acts on endosomal pools of PI3P in mammalian cells 61, additional studies suggesting that myotubularin affects vesicular trafficking have been reported. Blondeau et al initially reported that overexpression of human myotubularin in yeast generated several large vacuoles that were absent upon overexpression of the C375S mutant 53. This
phenotype is reminiscent of yeast lacking PI3K (VPS34), the enzyme that generates PI3P
141 from phosphotidylinositol, suggesting that down regulation of PI3P is responsible for the
vacuoles 53. A similar yeast vacuolar phenotype was seen with the overexpression of
human MTMR2 and MTMR3 proteins 60.
Further studies to demonstrate a role for MTM proteins in vesicular trafficking have been
undertaken in C. elegans. Dang et al made use of a C. elegans strain engineered to
secrete GFP into the body cavity and observed coelomocytes, scavenger cells with high
endocytic activity. Mutations in MTMR6 and MTMR9 in C. elegans resulted in
decreased endocytosis and disrupted PI3P staining in vivo 108. Using the same system,
myotubularin, possibly in conjunction with MTMR6, when knocked down with RNAi
could restore an endocytosis defect generated by mutating the PI3K homologue Let-
512109. This suggested that PI3P levels are key to proper endocytosis and a balance between MTM proteins and PI3K must be struck. While these two studies in C. elegans do not explain the completely the role of MTM proteins in cells, they demonstrate that these proteins do play a role in vesicular trafficking.
Studies on yeast and worm have been followed by an interesting study in differentiated myoblasts 110. Chaussade et al delivered the MTM1 cDNA at very high efficiency to
differentiated myoblast using an adenoviral approach. To study vesicular trafficking in
this system, they observed translocation of GLUT4 and glucose uptake. GLUT4 is a
glucose uptake protein that localizes from the intracellular compartment to the plasma
membrane via GLUT4 coated vesicles and increases glucose uptake upon stimulation
142 with insulin. With overexpression of myotubularin, a decrease in endosomal PI3P was
seen by a marked disruption of staining for EEA1 (early endosomal antigen 1), a protein known to interact with PI3P. Also, a decrease in GLUT4 translocation and glucose uptake was also observed with insulin stimulation. Not only do these data suggest that myotubularin plays a role in vesicular trafficking, but they also suggest that studying
GLUT4 translocation and glucose uptake in MTM1 patient muscle cell lines or MTM1 knockout mice may reveal further information regarding X-linked myotubular myopathy.
While the studies performed on yeast, C. elegans, and differentiated muscle cell lines do not coalesce into a clear model for the function of MTM proteins in cells, it is clear that the MTM family of proteins do play a role in the complex process of vesicular
trafficking. It will be interesting to see how observations made by us and others
regarding myotubularin and ruffles may fit into a vesicular trafficking model since ruffles
have been associated with phagocytosis 111 and activated Rac1 has been associated with
pinocytosis 97.
.
143
LIST OF REFERENCES
1. Copley, L.M. et al. Exclusion of mutations in the MTMR1 gene as a frequent cause of X-linked myotubular myopathy. Am J Med Genet 107, 256-8 (2002).
2. Laporte, J. et al. A gene mutated in X-linked myotubular myopathy defines a new putative tyrosine phosphatase family conserved in yeast. Nat Genet 13, 175-82 (1996).
3. Herman, G.E., Finegold, M., Zhao, W., de Gouyon, B. & Metzenberg, A. Medical complications in long-term survivors with X-linked myotubular myopathy. J Pediatr 134, 206-14 (1999).
4. Kioschis, P. et al. Genomic organization of a 225-kb region in Xq28 containing the gene for X-linked myotubular myopathy (MTM1) and a related gene (MTMR1). Genomics 54, 256-66 (1998).
5. Buj-Bello, A. et al. Muscle-specific alternative splicing of myotubularin-related 1 gene is impaired in DM1 muscle cells. Hum Mol Genet 11, 2297-307 (2002).
6. Wallgren-Pettersson, C. et al. The myotubular myopathies: differential diagnosis of the X linked recessive, autosomal dominant, and autosomal recessive forms and present state of DNA studies. J Med Genet 32, 673-9 (1995).
7. Joseph, M., Pai, G.S., Holden, K.R. & Herman, G. X-linked myotubular myopathy: clinical observations in ten additional cases. Am J Med Genet 59, 168- 73 (1995).
8. McEntagart, M. et al. Genotype-phenotype correlations in X-linked myotubular myopathy. Neuromuscul Disord 12, 939-46 (2002).
9. Yu, S. et al. X-linked myotubular myopathy in a family with three adult survivors. Clin Genet 64, 148-52 (2003).
144 10. Biancalana, V. et al. Characterisation of mutations in 77 patients with X-linked myotubular myopathy, including a family with a very mild phenotype. Hum Genet 112, 135-42 (2003).
11. Fan, H.C., Lee, C.M., Harn, H.J., Cheng, S.N. & Yuh, Y.S. X-linked centronuclear myopathy. Am J Perinatol 20, 173-9 (2003).
12. Herman, G.E., Kopacz, K., Zhao, W., Mills, P., Metzenberg, A., Das, S. Characterization of Mutations in Fifty North American Patients with X-Linked Myotubular Myopathy. Human Mutation (2001).
13. Helliwell, T.R., Ellis, I.H. & Appleton, R.E. Myotubular myopathy: morphological, immunohistochemical and clinical variation. Neuromuscul Disord 8, 152-61 (1998).
14. Sarnat, H.B. Myotubular myopathy: arrest of morphogenesis of myofibres associated with persistence of fetal vimentin and desmin. Four cases compared with fetal and neonatal muscle. Can J Neurol Sci 17, 109-23 (1990).
15. van Wijngaarden, G.K., Fleury, P., Bethlem, J. & Meijer, A.E. Familial "myotubular" myopathy. Neurology 19, 901-8 (1969).
16. Larkin, K. & Fardaei, M. Myotonic dystrophy--a multigene disorder. Brain Res Bull 56, 389-95 (2001).
17. Reardon, W., Newcombe, R., Fenton, I., Sibert, J. & Harper, P.S. The natural history of congenital myotonic dystrophy: mortality and long term clinical aspects. Arch Dis Child 68, 177-81 (1993).
18. Spiro, A.J., Shy, G.M. & Gonatas, N.K. Myotubular myopathy. Persistence of fetal muscle in an adolescent boy. Arch Neurol 14, 1-14 (1966).
19. Francis-West, P.H., Antoni, L. & Anakwe, K. Regulation of myogenic differentiation in the developing limb bud. J Anat 202, 69-81 (2003).
20. Parker, M.H., Seale, P. & Rudnicki, M.A. Looking back to the embryo: defining transcriptional networks in adult myogenesis. Nat Rev Genet 4, 497-507 (2003).
21. Yun, K. & Wold, B. Skeletal muscle determination and differentiation: story of a core regulatory network and its context. Curr Opin Cell Biol 8, 877-89 (1996).
22. Muntoni, F., Brown, S., Sewry, C. & Patel, K. Muscle development genes: their relevance in neuromuscular disorders. Neuromuscul Disord 12, 438-46 (2002).
145 23. Jungbluth, H. et al. Early and severe presentation of X-linked myotubular myopathy in a girl with skewed X-inactivation. Neuromuscul Disord 13, 55-9 (2003).
24. Kristiansen, M. et al. X-inactivation patterns in carriers of X-linked myotubular myopathy. Neuromuscul Disord 13, 468-71 (2003).
25. Sutton, I.J., Winer, J.B., Norman, A.N., Liechti-Gallati, S. & MacDonald, F. Limb girdle and facial weakness in female carriers of X-linked myotubular myopathy mutations. Neurology 57, 900-902 (2001).
26. Hammans, S.R. et al. A clinical and genetic study of a manifesting heterozygote with X-linked myotubular myopathy. Neuromuscul Disord 10, 133-7 (2000).
27. De Angelis, M.S., Palmucci, L., Leone, M. & Doriguzzi, C. Centronuclear myopathy: clinical, morphological and genetic characters. A review of 288 cases. J Neurol Sci 103, 2-9 (1991).
28. Tanner, S.M. et al. Skewed X-inactivation in a manifesting carrier of X-linked myotubular myopathy and in her non-manifesting carrier mother. Hum Genet 104, 249-53 (1999).
29. Darnfors, C. et al. X-linked myotubular myopathy: a linkage study. Clin Genet 37, 335-40 (1990).
30. Lehesjoki, A.E. et al. X linked neonatal myotubular myopathy: one recombination detected with four polymorphic DNA markers from Xq28. J Med Genet 27, 288- 91 (1990).
31. Thomas, N.S. et al. X linked neonatal centronuclear/myotubular myopathy: evidence for linkage to Xq28 DNA marker loci. J Med Genet 27, 284-7 (1990).
32. Dahl, N. et al. X linked myotubular myopathy (MTM1) maps between DXS304 and DXS305, closely linked to the DXS455 VNTR and a new, highly informative microsatellite marker (DXS1684). J Med Genet 31, 922-4 (1994).
33. Dahl, N. et al. Myotubular myopathy in a girl with a deletion at Xq27-q28 and unbalanced X inactivation assigns the MTM1 gene to a 600-kb region. Am J Hum Genet 56, 1108-15 (1995).
34. Hu, L.J. et al. Deletions in Xq28 in two boys with myotubular myopathy and abnormal genital development define a new contiguous gene syndrome in a 430 kb region. Hum Mol Genet 5, 139-43 (1996).
146 35. Laporte, J. et al. Cloning and characterization of an alternatively spliced gene in proximal Xq28 deleted in two patients with intersexual genitalia and myotubular myopathy. Genomics 41, 458-62 (1997).
36. Hu, L.J. et al. X-linked myotubular myopathy: refinement of the gene to a 280-kb region with new and highly informative microsatellite markers. Hum Genet 98, 178-81 (1996).
37. Cui, X. et al. Association of SET domain and myotubularin-related proteins modulates growth control. Nat Genet 18, 331-7 (1998).
38. Firestein, R. & Cleary, M.L. Pseudo-phosphatase Sbf1 contains an N-terminal GEF homology domain that modulates its growth regulatory properties. J Cell Sci 114, 2921-7 (2001).
39. Laporte, J. et al. The PtdIns3P phosphatase myotubularin is a cytoplasmic protein that also localizes to Rac1-inducible plasma membrane ruffles. J Cell Sci 115, 3105-17 (2002).
40. Fabre, S., Reynaud, C. & Jalinot, P. Identification of functional PDZ domain binding sites in several human proteins. Mol Biol Rep 27, 217-24 (2000).
41. Kim, S.A., Vacratsis, P.O., Firestein, R., Cleary, M.L. & Dixon, J.E. Regulation of myotubularin-related (MTMR)2 phosphatidylinositol phosphatase by MTMR5, a catalytically inactive phosphatase. Proc Natl Acad Sci U S A 100, 4492-7 (2003).
42. Doerks, T., Strauss, M., Brendel, M. & Bork, P. GRAM, a novel domain in glucosyltransferases, myotubularins and other putative membrane-associated proteins. Trends Biochem Sci 25, 483-5 (2000).
43. Herman, G.E. et al. Characterization of mutations in fifty North American patients with X-linked myotubular myopathy. Hum Mutat 19, 114-21 (2002).
44. Laporte, J. et al. MTM1 mutations in X-linked myotubular myopathy. Hum Mutat 15, 393-409 (2000).
45. Buj-Bello, A., Biancalana, V., Moutou, C., Laporte, J. & Mandel, J.L. Identification of novel mutations in the MTM1 gene causing severe and mild forms of X-linked myotubular myopathy. Hum Mutat 14, 320-5 (1999).
46. de Gouyon, B.M. et al. Characterization of mutations in the myotubularin gene in twenty six patients with X-linked myotubular myopathy. Hum Mol Genet 6, 1499- 504 (1997). 147 47. Laporte, J. et al. Mutations in the MTM1 gene implicated in X-linked myotubular myopathy. ENMC International Consortium on Myotubular Myopathy. European Neuro-Muscular Center. Hum Mol Genet 6, 1505-11 (1997).
48. Tanner, S.M. et al. Characterization of 34 novel and six known MTM1 gene mutations in 47 unrelated X-linked myotubular myopathy patients. Neuromuscul Disord 9, 41-9 (1999).
49. Vincent, M.C., Guiraud-Chaumeil, C., Laporte, J., Manouvrier-Hanu, S. & Mandel, J.L. Extensive germinal mosaicism in a family with X linked myotubular myopathy simulates genetic heterogeneity. J Med Genet 35, 241-3 (1998).
50. Tanner, S.M., Laporte, J., Guiraud-Chaumeil, C. & Liechti-Gallati, S. Confirmation of prenatal diagnosis results of X-linked recessive myotubular myopathy by mutational screening, and description of three new mutations in the MTM1 gene. Hum Mutat 11, 62-8 (1998).
51. Taylor, G.S., Maehama, T. & Dixon, J.E. Inaugural article: myotubularin, a protein tyrosine phosphatase mutated in myotubular myopathy, dephosphorylates the lipid second messenger, phosphatidylinositol 3-phosphate. Proc Natl Acad Sci U S A 97, 8910-5 (2000).
52. Gillooly, D.J. et al. Localization of phosphatidylinositol 3-phosphate in yeast and mammalian cells. Embo J 19, 4577-88 (2000).
53. Blondeau, F. et al. Myotubularin, a phosphatase deficient in myotubular myopathy, acts on phosphatidylinositol 3-kinase and phosphatidylinositol 3- phosphate pathway. Hum. Mol. Genet. 9, 2223-2229 (2000).
54. Schaletzky, J. et al. Phosphatidylinositol-5-phosphate activation and conserved substrate specificity of the myotubularin phosphatidylinositol 3-phosphatases. Curr Biol 13, 504-9 (2003).
55. Kim, S.A., Taylor, G.S., Torgersen, K.M. & Dixon, J.E. Myotubularin and MTMR2, phosphatidylinositol 3-phosphatases mutated in myotubular myopathy and type 4B Charcot-Marie-Tooth disease. J Biol Chem 277, 4526-31 (2002).
56. Wishart, M.J., Taylor, G.S., Slama, J.T. & Dixon, J.E. PTEN and myotubularin phosphoinositide phosphatases: bringing bioinformatics to the lab bench. Curr Opin Cell Biol 13, 172-81 (2001).
57. Laporte, J., Bedez, F., Bolino, A. & Mandel, J.-L. Myotubularins, a large disease- associated family of cooperating catalytically active and inactive phosphoinositides phosphatases. Hum. Mol. Genet. 12, 285R-292 (2003). 148 58. Laporte, J. et al. Characterization of the myotubularin dual specificity phosphatase gene family from yeast to human. Hum Mol Genet 7, 1703-12 (1998).
59. Azzedine, H. et al. Mutations in MTMR13, a new pseudophosphatase homologue of MTMR2 and Sbf1, in two families with an autosomal recessive demyelinating form of Charcot-Marie-Tooth disease associated with early-onset glaucoma. Am J Hum Genet 72, 1141-53 (2003).
60. Laporte, J. et al. Functional redundancy in the myotubularin family. Biochem Biophys Res Commun 291, 305-12 (2002).
61. Kim, S.-A., Taylor, G.S., Torgersen, K.M. & Dixon, J.E. Myotubularin and MTMR2, Phosphatidylinositol 3-Phosphatases Mutated in Myotubular Myopathy and Type 4B Charcot-Marie-Tooth Disease. J. Biol. Chem. 277, 4526-4531 (2002).
62. Bolino, A. et al. Charcot-Marie-Tooth type 4B is caused by mutations in the gene encoding myotubularin-related protein-2. Nat Genet 25, 17-9 (2000).
63. Nelis, E. et al. A novel homozygous missense mutation in the myotubularin- related protein 2 gene associated with recessive Charcot-Marie-Tooth disease with irregularly folded myelin sheaths. Neuromuscul Disord 12, 869-73 (2002).
64. Berger, P., Bonneick, S., Willi, S., Wymann, M. & Suter, U. Loss of phosphatase activity in myotubularin-related protein 2 is associated with Charcot-Marie-Tooth disease type 4B1. Hum. Mol. Genet. 11, 1569-1579 (2002).
65. Kim, S.-A., Vacratsis, P.O., Firestein, R., Cleary, M.L. & Dixon, J.E. Regulation of myotubularin-related (MTMR)2 phosphatidylinositol phosphatase by MTMR5, a catalytically inactive phosphatase. PNAS 100, 4492-4497 (2003).
66. Bart Vanhaesebroeck, S.J.L., Khatereh Ahmadi, John Timms, Roy Katso, Paul C. Driscoll, Rudiger Woscholski, Peter J. Parker, Michael D. Waterfield. Synthesis and Function of 3-Phosphorylated Inositol Lipids. Annual Review of Biochemistry 70, 535-602 (2001).
67. Leevers, S.J., Vanhaesebroeck, B. & Waterfield, M.D. Signalling through phosphoinositide 3-kinases: the lipids take centre stage. Curr Opin Cell Biol 11, 219-25 (1999).
68. Walker, D.M. et al. Characterization of MTMR3. an inositol lipid 3-phosphatase with novel substrate specificity. Curr Biol 11, 1600-5 (2001).
149 69. Zhao, R., Qi, Y. & Zhao, Z.J. FYVE-DSP1, a dual-specificity protein phosphatase containing an FYVE domain. Biochem Biophys Res Commun 270, 222-9 (2000).
70. Mochizuki, Y. & Majerus, P.W. Characterization of myotubularin-related protein 7 and its binding partner, myotubularin-related protein 9. Proc Natl Acad Sci U S A 100, 9768-73 (2003).
71. Nandurkar, H.H. et al. Characterization of an adapter subunit to a phosphatidylinositol (3)P 3-phosphatase: identification of a myotubularin-related protein lacking catalytic activity. Proc Natl Acad Sci U S A 98, 9499-504 (2001).
72. Nandurkar, H.H. et al. Identification of myotubularin as the lipid phosphatase catalytic subunit associated with the 3-phosphatase adapter protein, 3-PAP. PNAS 100, 8660-8665 (2003).
73. Laporte, J., Kress, W. & Mandel, J.L. Diagnosis of X-linked myotubular myopathy by detection of myotubularin. Ann Neurol 50, 42-6 (2001).
74. Fields, S. & Song, O. A novel genetic system to detect protein-protein interactions. Nature 340, 245-6 (1989).
75. Fashena, S.J., Serebriiskii, I.G. & Golemis, E.A. LexA-based two-hybrid systems. Methods Enzymol 328, 14-26 (2000).
76. Gossen, M. et al. Transcriptional activation by tetracyclines in mammalian cells. Science 268, 1766-9 (1995).
77. Kistner, A. et al. Doxycycline-mediated quantitative and tissue-specific control of gene expression in transgenic mice. Proc Natl Acad Sci U S A 93, 10933-8 (1996).
78. Inui, S. et al. Ig receptor binding protein 1 (alpha4) is associated with a rapamycin-sensitive signal transduction in lymphocytes through direct binding to the catalytic subunit of protein phosphatase 2A. Blood 92, 539-46 (1998).
79. Lu, Q. et al. Regulation of estrogen receptor alpha-mediated transcription by a direct interaction with protein phosphatase 2A. J Biol Chem 278, 4639-45 (2003).
80. Murata, K., Wu, J. & Brautigan, D.L. B cell receptor-associated protein alpha4 displays rapamycin-sensitive binding directly to the catalytic subunit of protein phosphatase 2A. Proc Natl Acad Sci U S A 94, 10624-9 (1997).
81. Getzoff, E.D., Tainer, J.A., Lerner, R.A. & Geysen, H.M. The chemistry and mechanism of antibody binding to protein antigens. Adv Immunol 43, 1-98 (1988). 150 82. Lerner, R.A. Antibodies of predetermined specificity in biology and medicine. Adv Immunol 36, 1-44 (1984).
83. Harlow, E. & Lane, D. Chapter 5: Immunizations. in Antibodies: A Laboratory Manual 53-139 (Cold Spring Harbor Laboratory, Cold Spring Harbor, 1988).
84. Detection Systems. in Molecular Cloning: A Laboratory Manual, Vol. 3 (eds. Sambrook, J. & Russel, D.) A9.1-A9.49 (Cold Spring Harbor Laboratory Press, Cold Spring Harbor, 2001).
85. Kyte, J. & Doolittle, R.F. A simple method for displaying the hydropathic character of a protein. J Mol Biol 157, 105-32 (1982).
86. Janin, J. Surface and inside volumes in globular proteins. Nature 277, 491-2 (1979).
87. Hopp, T.P. & Woods, K.R. A computer program for predicting protein antigenic determinants. Mol Immunol 20, 483-9 (1983).
88. Novotny, J. A static accessibility model of protein antigenicity. Int Rev Immunol 2, 379-89 (1987).
89. Fauchere, J.L., Charton, M., Kier, L.B., Verloop, A. & Pliska, V. Amino acid side chain parameters for correlation studies in biology and pharmacology. Int J Pept Protein Res 32, 269-78 (1988).
90. Parker, J.M., Guo, D. & Hodges, R.S. New hydrophilicity scale derived from high-performance liquid chromatography peptide retention data: correlation of predicted surface residues with antigenicity and X-ray-derived accessible sites. Biochemistry 25, 5425-32 (1986).
91. Singh, B., Lee, K.C. & Fraga, E. Hydrophobic amino acid residues in peptide antigens determine the genetic control of immune responses. Pept Res 2, 120-7 (1989).
92. Welling, G.W., Weijer, W.J., van der Zee, R. & Welling-Wester, S. Prediction of sequential antigenic regions in proteins. FEBS Lett 188, 215-8 (1985).
93. Kaumaya, P.T.P., Kobs-Conrad, Susan, DiGeorge, Ann M., Stevens, Vernon C. "De Novo" Engineering of Peptide Immunogenic and Antigenic Determinants as Potential Vaccines. Peptides: Design, Synthesis, and Biological Activity, 133-164 (1994).
151 94. Small, J.V., Stradal, T., Vignal, E. & Rottner, K. The lamellipodium: where motility begins. Trends Cell Biol 12, 112-20 (2002).
95. Allen, W.E., Jones, G.E., Pollard, J.W. & Ridley, A.J. Rho, Rac and Cdc42 regulate actin organization and cell adhesion in macrophages. J Cell Sci 110 ( Pt 6), 707-20 (1997).
96. Ridley, A.J. & Hall, A. The small GTP-binding protein rho regulates the assembly of focal adhesions and actin stress fibers in response to growth factors. Cell 70, 389-99 (1992).
97. Ridley, A.J., Paterson, H.F., Johnston, C.L., Diekmann, D. & Hall, A. The small GTP-binding protein rac regulates growth factor-induced membrane ruffling. Cell 70, 401-10 (1992).
98. del Pozo, M.A., Price, L.S., Alderson, N.B., Ren, X.D. & Schwartz, M.A. Adhesion to the extracellular matrix regulates the coupling of the small GTPase Rac to its effector PAK. Embo J 19, 2008-14 (2000).
99. Kozma, R., Ahmed, S., Best, A. & Lim, L. The Ras-related protein Cdc42Hs and bradykinin promote formation of peripheral actin microspikes and filopodia in Swiss 3T3 fibroblasts. Mol Cell Biol 15, 1942-52 (1995).
100. Kato, S., Kidoaki, S. & Matsuda, T. Substrate-dependent cellular behavior of Swiss 3T3 fibroblasts and activation of Rho family during adhesion and spreading processes. J Biomed Mater Res 68A, 314-24 (2004).
101. Price, L.S., Leng, J., Schwartz, M.A. & Bokoch, G.M. Activation of Rac and Cdc42 by integrins mediates cell spreading. Mol Biol Cell 9, 1863-71 (1998).
102. Miki, H., Suetsugu, S. & Takenawa, T. WAVE, a novel WASP-family protein involved in actin reorganization induced by Rac. Embo J 17, 6932-41 (1998).
103. Miki, H., Yamaguchi, H., Suetsugu, S. & Takenawa, T. IRSp53 is an essential intermediate between Rac and WAVE in the regulation of membrane ruffling. Nature 408, 732-5 (2000).
104. van Hennik, P.B. et al. The C-terminal domain of Rac1 contains two motifs that control targeting and signaling specificity. J Biol Chem 278, 39166-75 (2003).
105. Mullins, R.D., Heuser, J.A. & Pollard, T.D. The interaction of Arp2/3 complex with actin: nucleation, high affinity pointed end capping, and formation of branching networks of filaments. Proc Natl Acad Sci U S A 95, 6181-6 (1998).
152 106. Rohatgi, R. et al. The interaction between N-WASP and the Arp2/3 complex links Cdc42-dependent signals to actin assembly. Cell 97, 221-31 (1999).
107. Laporte, J. et al. The PtdIns3P phosphatase myotubularin is a cytoplasmic protein that also localizes to Rac1-inducible plasma membrane ruffles. J Cell Sci 115, 3105-3117 (2002).
108. Dang, H., Li, Z., Skolnik, E.Y. & Fares, H. Disease-related myotubularins function in endocytic traffic in Caenorhabditis elegans. Mol. Biol. Cell, E03-08- 0605 (2003).
109. Xue, Y. et al. Genetic Analysis of the Myotubularin Family of Phosphatases in Caenorhabditis elegans. J. Biol. Chem. 278, 34380-34386 (2003).
110. Chaussade, C. et al. Expression of myotubularin by an adenoviral vector demonstrates its function as a PtdIns(3)P phosphatase in muscle cell lines. Involvement of PtdIns(3)P in insulin-stimulated glucose transport. Mol Endocrinol, me.2003-0261 (2003).
111. Castellano, F., Montcourrier, P. & Chavrier, P. Membrane recruitment of Rac1 triggers phagocytosis. J Cell Sci 113 ( Pt 17), 2955-61 (2000).
153
APPENDIX A
154
Name Source Size Promoter Resistance website Use (kB) Expresses rtTA; pTet-on Clontech 7.4 CMV Amp/Neo www.bdbiosciences.com stably transfected into HeLa Tet-on Control expression of pTRE2- Clontech 5.1 Tet- Amp/Puro www.bdbiosciences.com GOI in tet pur hCMV inducible systems DNA-AD vector pACT2 Clontech 8.1 ADH Amp/leu www.bdbiosciences.com used in cDNA library DNA-BD vector pGBKT7 Clontech 7.3 ADH kan/trp www.bdbiosciences.com used to clone bait Control DNA- BD vector pVA3-1 Clontech 9.4 ADH Amp/trp www.bdbiosciences.com expressing murine p53 Control DNA- AD vector p-TD1-1 Clontech 10.1 ADH Amp/leu www.bdbiosciences.com expressing SV40 large T antigen Holding vector pCR2.1- Invitrogen 3.9 T7 Amp/kan www.invitrogen.com with TOPO TOPO technology Mammalian expression with N terminal 6His pcDNA4- tag, Xpress tag HisMax Invitrogen 5.27 CMV Amp/zeo www.invitrogen.com and enterokinase TOPO cleavage site/TOPO technology Mammalian pCMV- expression with Stratagene 4.3 CMV Kan/neo www.stratagene.com Tag2B N-terminal Flag tag For making GST fusion proteins with C-terminal pGSTx Dixon CMV NA His tag GST was cloned from pGEX-KT into pET21a
Table A.1: General Information for Parent Vectors. GOI = Gene of interest.
155
Strain Source Genotype Tien-Hsien supE44 hsdR17 recA1 endAI gyrA46 thi relA1 lac- XL1-Blue Chang, OSU F’[proAB+ lacIq lacZ ∆M15 Tn10 (tetr)]
BL21- E.coli B F– ompT hsdS(r – m –) dcm+ Tetr gal λ Stratagene B B CodonPlus(DE3)-RIL (DE3) endA Hte [argU ileY leuW Camr] TetR .(mcrA)183 .(mcrCB-hsdSMR-mrr)173 endA1 Xl10-Gold Stratagene supE44 thi-1 recA1 gyrA96 relA1 lac Hte [F´ proAB lacIqZ.M15 Tn10 (TetR) Amy CamR] F- 80dlacZ M15 (lacZYA-argF) U169 recA1 - + - DH5α Invitrogen endA1 hsdR17(rk , mk ) phoA supE44 thi-1 gyrA96 relA1
F- mcrA (mrr-hsdRMS-mcrBC) 80lacZ M15 OneShot Top10 Invitrogen lacX74 recA1 ara 139 (ara-leu)7697 galU galK rpsL (StrR) endA1 nupG
Table A.2: Genotypes and Sources of E. coli Strains. Genotype information was provided by distributor of the cell line. Genotype information for XL1-Blue cells was obtained from Molecular Cloning, Sambrook and Russel, 3rd Ed 2001.
156