Close
About
FAQ
Home
Collections
Login
USC Login
Register
0
Selected
Invert selection
Deselect all
Deselect all
Click here to refresh results
Click here to refresh results
USC
/
Digital Library
/
University of Southern California Dissertations and Theses
/
The kinetic study of engineered MBD domain interactions with methylated DNA: insight into binding of methylated DNA by MBD2b
(USC Thesis Other)
The kinetic study of engineered MBD domain interactions with methylated DNA: insight into binding of methylated DNA by MBD2b
PDF
Download
Share
Open document
Flip pages
Contact Us
Contact Us
Copy asset link
Request this asset
Transcript (if available)
Content
THE KINETIC STUDY OF ENGINEERED MBD DOMAIN INTERACTIONS WITH
METHYLATED DNA:
INSIGHT INTO BINDING OF METHYLATED DNA BY MBD2B
by
Po-Han Chen
A Thesis Presented to the
FACULTY OF THE USC GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
MASTER OF SCIENCE
(BIOCHEMISTRY AND MOLECULAR BIOLOGY)
December 2009
Copyright 2009 Po-Han Chen
ii
ACKNOWLEDGEMENTS
I would like to thank my advisor, Dr. Ite Laird-Offringa, for patiently advising and
encouraging me in my studies. She always gave me valuable responses and taught me to
be a good scientist. I also thank Dr. Binh Trinh for his experimental support and guidance.
All the members in my lab, Meleeneh Kazarian, Suhaida Selamat, Candace Johnson, and
Devon Pryor are appreciated. They gave me a lot of help during my Master’s year. I
would also like to thank Dr. Zoltan Tokes for his help in both research and academic
problems. Finally, I appreciate my parents for their support during my studies at USC.
iii
TABLE OF CONTENTS
Acknowledgements ii
List of Figures iv
List of Tables vi
Abbreviations vii
Abstract viii
Chapter 1: Introduction 1
Chapter 2: Materials and Methods 6
2.1 Plasmid 6
2.2 Protein expression and purification 7
2.3 DNA preparation 9
2.4 Generation of Biacore sensor chip surface 10
2.5 SPR analysis 10
2.6 Gel Shift Assay 12
Chapter 3: Results 13
3.1 The examination of MDB2b binding to a fully methylated CpG 13
dinucleotide
3.2 MBD domain is important in MBD2b binding to a fully methyl-CpG 18
dinucleotide
3.3 2xMBDs show stronger binding than 1xMBD to fully methyl-CpG 23
dinucleotides
Chapter 4: Conclusion and discussion 39
Bibliography 43
iv
LIST OF FIGURES
Figure 1 Oligonucleotides used in the study. 14
Figure 2 Interaction of MBD2b with non-methylated, hemimethylated and fully 14
methylated DNA.
Figure 3 The cartoon shows MBD2b and MBD3L1 domain structures. 16
Figure 4 The yeast two-hybrid system shows the interactions between MBD proteins. 17
Figure 5 Interaction of MBD3L1 with DNA, and effect of MDB3L1 on DNA 18
binding by MBD2b.
Figure 6 The cartoon shows two types of MBD2: MBD2a and MBD2b. 19
Figure 7 1xMBD binding to non-, hemi-, and fully methylated DNA surfaces. 21
Figure 8 SPR data from the Inomata manuscript. 21
Figure 9 Binding of 1xMBD to increased density surfaces. 22
Figure 10 1xMBD and 2xMBDs shown in the cartoon. 26
Figure 11 Binding of 2xMBDs to non-methylated, hemi-methylated or singly fully 27
methylated DNA.
Figure 12 The alternative binding model. 27
Figure 13 Schematic summary of protein-DNA contacts. 30
Figure 14 The rebinding model. 30
v
Figure 15 3D structural study of MBD domain binding to methylated DNA 31
Figure 16 The short version of 8mDNA may reduce the affinity of 2xMBDs bind to 31
a fully single methylated CpG of DNA.
Figure 17 Binding of 1xMBD and 2xMBD to a shorter singly methylated DNA 32
target.
Figure 18 Binding of 2xMBDs to singly methylated, trans-dimethylated or 34
cis-dimethylated DNA.
Figure 19 Binding of MBD2b to singly methylated or dimethylated DNA. 35
Figure 20 Gel shift of 2xMBD with a singly methylated DNA target (right DNA 37
helix).
Figure 21 Gel shift assay of 2xMBD with trans- or cis-dimethylated DNA targets 38
(the position of methylated CpGs of DNA helix shown below).
Figure 22 A cartoon showing the model we propose to explain the ability of the 42
MBD2b/MBD3L1 complex to bind to methylated DNA.
vi
LIST OF TABLES
Table 1 Primers used for the construction of 1xMBD and 2xMBDs. 7
Table 2 Comparison of kinetic data from Inomata et al. to our data using the MBD 21
from MDB2b.
Table 3 Binding constants of single 1x and 4xMBDs to DNA targets with 1, 2, or 3 23
methyl groups.
vii
ABBREVIATIONS
BSA bovine serum albumin
DTT dithiothreitol
EDTA ethylenediaminetetraacetic acid
IPTG isopropyl β-D-1-thiogalactopyranoside
MBD methyl-CpG binding domain
MBD1-4 methyl-CpG binding domain protein 1-4
MBD3L1 methyl-CpG binding domain protein 3-like-1
MIRA methylated-CpG island recovery assay
RU response unit
SA chip streptavidin chip
SPR surface plasmon resonance
viii
ABSTRACT
Mehtyl-CpG binding domain protein 2 and 3 (MBD2 and MBD3) contain
methyl-CpG binding domains (MBD) and belong to a family of MBD proteins.
Methyl-CpG binding domain protein 3-like-1 (MBD3L1) is a protein with highly
homology to MBD2 and MBD3, but it lacks the MBD domain in the N-terminus. A new
method, methylated-CpG island recovery assay (MIRA), was developed to purify
methylated DNA by using the mixture of MBD2b and MBD3L1. MBD3L1 was found to
be able to interact with MBD2b and increase the affinity of MBD2b binding to
methylated DNA. However, the mechanism is unknown. Our goal is to explore the
mechanism of MBD3L1 involved in MBD2b binding to methylated DNA. We created
the engineered 1xMBD and 2xMBDs from MBD2. 2xMBDs showed higher affinities
than 1xMBD binding to the single and double fully methylated DNA. The alternative
binding model and the rebinding model had been hypothesized to increase the affinities
of 2xMBDs, but we observed that the alternative binding model was more preferential for
2xMBDs binding to a fully single methyl-CpG by further studies. In addition, the two
linked MBDs were able to contact both methyl-CpGs of cis-methylated DNA, which
increased a much more stable binding than trans-methylated DNA. The orientation of
2xMBDs binding suggested the spacing between two MBDs may be an important factor
involved in the binding mechanism of MBD2b/MBd3L1 complex. Our findings reveal
essential properties of MBDs in binding to methylated DNA. By these results, we can
depict a clearer mechanism of how an MBD3b/MBD3L1 complex increases the affinity
and binds to methylated DNA.
1
CHAPTER 1: INTRODUCTION
Methylation at position 5 of cytosine in DNA is a major epigenetic modification
in mammalian cells. It occurs in the context of a 2-nucleotide palindrome, referred to as a
CpG dinucleotide, that is methylated on the two opposing Cs. The nature of the CpG
allows DNA methylation to be copied to the nascent strand following DNA replication.
DNA methylation has been found to play an important role in the regulation of gene
activity, genomic stability, and chromatin structure (Bird and Wollfe 1999 and Jones and
Takai 2001). Aberrant DNA methylation is involved in the early stages of cancer
formation (Baylin et al., 1998 and Costello et al., 2000). A family of methyl-CpG-binding
domain (MBD) proteins recognizes methylated DNA and recruits proteins that modify
the histones in the nucleosome. MBD proteins are characterized by an MBD domain.
Five MBD proteins have been identified: MeCP2, MBD1, MBD2, MBD3, and MBD4.
Additionally, MBD3L1 and MBD3L2 have been identified from the human gene.
MBD3L1 is 38% identical with MBD2 and 42% identical with MBD3 but lacking MBD
domain. Thus, MBD3L1 has a very weak interaction with CpG-methylated DNA.
Recently, studies have shown novel MBD proteins as MBD5 and MBD6 in human and
Arabidopsis. The Arabidopsis MBD5 and MBD6 (AtMBD5 and AtMBD6) contain a
MBD domain. They have been discovered to localize to the highly CpG-methylated
chromosomes and to bind methyl-CpG sites (Zemach and Grafi 2007). In 2008, Dr.
Grafi’s lab reported that the Arabidopsis MBD7, AtMBD7, contains three MBD domains
and has less mobility on chromocneters than AtMBD5 and AtMBD6, suggesting that
2
multiple MBD domains make a MBD protein have a more stable binding to
CpG-methylated DNA and constrain its mobility and localization (Zemach et al., 2008).
The first MBD protein discovered is MeCP2, which can associate with a
transcriptional repressor complex and recruit the complex on the promoter-dependent
methylated CpG sites (Nan et al., 1998). MBD3 forms a Mi-2/NuRD complex involved
in nucleosome remodeling and histone deacetylase activities. Due to changes in two
highly conserved amino acids, MBD3 has lost its ability to bind to methylated CpGs,
showing no affinity for methylated DNA except in the presence of MBD2 (Wade et al.,
1999 and Zhang et al., 1999). MBD4 is a thymine glycosylase that can remove the
product of deamination and function in DNA repair. Although MBD4 can bind at
methyl-CpG sites, it is not involved in transcriptional repression (Hendrich et al., 1999).
MBD1 has MBD, CXXC, and transcriptional repression domains and binds to methylated
CpG islands as a transcriptional regulator (Ng et al., 2000 and Fujita et al., 2000). The 3D
structure of the MBD domain from MBD1 has been solved, and the binding causes a loop
in the MBD domain to fold into DNA major groove at methylated sites and to form a
small contact area. The small contact surface allows the MBD domain to avoid steric
interfence from core histones (Ohki. et al., 2001). MBD2 has a coding sequence highly
similar to MBD3, and was found to associate with the MeCP1 complex that can bind to
methylated DNA (Ng et al., 1999). Two starting codes of MBD2 RNA lead to the
production of two variant types of MBD2 protein. MBD2b, the short version of MBD2,
lacks an N-terminal domain. Thus, MBD2b has its MBD domain in the N terminus
instead of MBD2a (MBD2) has the MBD domain located at amino acid 145-217. They
3
also carry a C-terminal coiled-coil domain. Recently, a new approach, methylated-CpG
island recovery assay (MIRA), was developed for DNA methylation analysis in
mammalian genomes (Rauch and Pfeifer 2005). The technique requires MBD3L1 to be
mixed with MBD2b to form an MBD2b/MBD3L1 complex, which can apparently
increase the ability of MBD2 to bind to methylated DNA. It is used to purify methylated
DNA for analysis on a microarray platform, which allows methylated regions of the
genome to be mapped. Although MIRA has become a commercial tool to allow
methylated regions of the genome to be mapped, how MBD3L1 increases MBD2b
binding to methylated DNA is unknown. The mechanism of MBD3L1 interacting with
MBD2b is also unclear. Thus, to study the kinetic properties of MBD2b and MBD3L1 on
the methylated DNA is a way to realize the story. For this purpose, I decided to use
BIACORE 2000 to challenge the maze. Indeed, in the absence of crystal structure of
MBD2b or MBD3L1, BIACORE 2000 can provide us with more complete and useful
insights of kinetics that was not available before.
Surface plasmon resonance (SPR) in BIACORE 2000 is a powerful tool to
analyze the kinetics of molecular interaction (Myszka, D.G. and Rich, R.L., 2001). The
principle of SPR is based on the molecular mass at the sensor surface would influence the
refractive index. Because every sensor surface is coated with a thin membrane of golden,
light projecting on the surface would leak an electrical field intensity called evanescent
wave into the golden layer, which could determine one of beams of light to be absorbed.
How much electrical field intensity lost is dependent on the refractive index. Therefore,
which one light beam absorbed by golden surface is determined by the refractive index
4
and the wavelength that both are influenced by the mass change at the sensor surface.
When the mass change caused by ligand/analyte association and dissociation, different
light beams with specific angles would be continuously absorbed by the surface, which
recorded by a detector. A detector which records the reflected light will detect the
missing angle of beams of light. Changes in mass affect the refractive index, which in
turn affects the angle of observed light. Therefore, the detector which monitors the angles
of observed light can be used to calculate the mass change on the chip surface. Because it
can reflect the mass change during a whole experiment, SPR can provide real-time data
for interactions and can provide information on the association and dissociation of
complexes. To carry out an SPR experiment, we need to coat appropriate amount of
ligand such as DNA, RNA, or protein on the chip surface. We usually use biotin-labeled
ligand that can strongly interact with commercially available streptavidin-coated sensor
chips. Each chip has four blank surfaces, and one blank (uncoated) surface left to allow
us to subtract background noise. Once ligands have been coated on the three other flow
cells of the chip, analytes can be injected over the four surfaces consecutively. By
injecting different concentrations of analytes, the association and dissociation properties
of complexes on the surfaces could be observed and analyzed. The computer records the
changes of mass on the chip surface in a “sensorgram”. The sensorgram can be analyzed
to obtain the kinetics describing each interaction. Surface plasmon resonance is widely
applied in many fields of research. Our lab has gained many interesting insights into
intermolecular interactions through the SPR experiments. Therefore, we tried to explore
the mechanism of MBD2b binding to methylated DNA and the effect of MBD3L1 on this
interaction.
5
Since we found that MBD3L1 did not increase binding of MBD2b to
methylated DNA, we changed my focus to study the binding kinetics of a single MBD
(1xMBD) and several engineered dimers (2xMBDs) on methylated DNA. Our studies of
the binding kinetics of the MBD help to provide insight into how this domain interacts
with unmethylated, hemi methylated and fully methylated DNA, and has allowed us to
propose a model for how MBD3L1 might stabilize MBD2b on methylated DNA.
6
CHAPTER 2: MATERIALS AND METHODS
2.1 Plasmids
The human MBD2b-pGEX-5X-1 plasmid, which was used to extract the MBD
domain, was a gift from Dr. Gerd P. Pfeifer (City of Hope, Duarte, CA) (Jiang et al.,
2004). For expression of the recombinant proteins in E. coli, all PCR products of the
MBD domain (MBD2b
145-217
) were cloned into a PET3d-based vector, which has a
(His)
6
-tag at the C-terminal side of the inserted protein-encoding fragments. The 1xMBD,
2xMBD
n
, 2xMBD
5
, and 2xMBD
10
plasmids were engineered as follows: The first MBD
domain in 2xMBDs was constructed and inserted in the vector before the insertion of
second MBD domain. The 5’ primer used in all first domain constructs is primer 1 (All
primers shown in Table 1). This primer creates a BspEI site at the 5’ end of the fragment
and keeps the original ATG in the NcoI site. In order to add different linkers between two
MBD domains, the 3’ primers were designed in different ways. Primer 2 was used as the
3’ primer to amplify the 2xMBD
n
PCR fragment (primer 1 + primer 2), and resulted in a
NotI site at the 3’ end. Primer 3 was used as the 3’ primer to amplify the 2xMBD
5
PCR
fragment (primer 1 + primer 3), and primer 4 was used as the 3’ primer to amplify the
2xMBD
10
PCR fragment (primer 1 + primer 4). A PCR fragment encoding the second
MBD domain was amplified by using primers 5 as the 5’ primer, which deletes NcoI site
but creates an EagI site (CGGCCG) that can anneal with NotI site (GCGGCCGC).
Primer 2, 3, and 4 are also used as the 3’primer to amplify the second MBD domain
fragments (primer 5 + primer 2, primer 5 + primer 3, primer 5 + primer 4). Therefore, the
second MBD domains can be inserted into the NotI cutting site. The 1xMBD plasmid
7
lacked an insertion of a second MBD fragment. All plasmids were checked by
sequencing.
Primer Sequence
1
5’-AATTTCCATGGAATCCGGAAAACGAATGGATTGCCCGGCCCT-3’
BspEI
2
5’-GATGTGCGGCCGCCTTCGAAGGCATCATCTTTCCAGTTCT-3’
NotI
3
5’-GATGTGCGGCCGCACCACCAGAGCCGCTAGCCTTCGAAGGCATCATCTTTCCAGTTCT-3’
NotI
4
5’-GATGTGCGGCCGCAGAACCACCACCACCAGAACCACCACCGCTAGCCTTCGAAGGCATCATCTT
NotI
TCCAGTTCT-3’
5
5’-AATTTCGGCCGAATCCGGAAAACGAATGGATTGCCCGGCCCT-3’
EagI BspE1
Table 1. Primers used in the study. All primers were designed by Dr. Binh Trinh in our
lab and purchased from Integrated DNA Technologies, Inc. Each primer is stored at -80ºC.
2.2 Protein expression and purification
MBD-encoding plasmids were transformed into E. coli BL21 (J57), a variant of
the BL21 (DE3) strain containing a pACYC-derived plasmid encoding two rare tRNA
and one biotin protein ligase (BirA) and plated in LB plates with ampicillin (0.1 μg/ml)
and chlormaphenicol (20 μg/mL). Cells were inoculated in 200 ml LB with ampicillin
(0.1 μg/ml) and chloramphenicol (20 μg/ml) and grown overnight at 37
o
C with vigorous
shaking. The next day the culture was diluted five times to 1 L LB, and cells were grown
8
to A600= 0.6-0.8. IPTG was added to a final concentration of 1 mM and cells were
incubated at 37°C for 3 hours with vigorous shaking. Cells were collected by the
centrifuge using rotor SLA3000 (SORVALL) at 8000 rpm for 15 minutes. Bacterial
pellets were re-suspended in sonication buffer (10 mM Tris, pH 8.0, 150 mM NaCl, 1
mM DTT, 0.5 % Triton X-100, and 1 tablet of protease in 10 mL buffer [Roche, Cat. #:
11-836-153-001]). The cell suspension was sonicated for 1 minute followed by a
30-second rest to four times and then spun down by the centrifuge using rotor SS34
(SORVALL) at 13000 rpm for ten minutes. 500 μL Qiagen Ni-NTA agarose beads in
storage buffer were spun down at 1000 RPM for 1 minute, and the storage buffer was
discarded carefully. The beads were added 500 μL sonication buffer, mixed to wash, and
spun down at 1000 rpm for 1 minute, and the buffer was discarded. Washing of the beads
was repeated two times. Prepared beads were then added to the supernatant, and the
mixture was gently shaken by a rotator in the cold room at 4ºC for 1 hour. The mixture
was spun down by the centrifuge CL2 at 5000 rpm for two minutes at 4ºC, and the
supernatant was discarded. The beads were washed with wash buffer (sonication buffer
added by 50 mM imidazole and 10 % glycerol) three times for 3 minutes each. Finally, 1
M imidazole elution buffer (wash buffer added by 1 M imidazole) was used to elute
proteins from the beads, and the proteins were stored at -80°C. All purification steps were
carried out on ice or in the cold room.
9
2.3 DNA preparation
DNA targets were designed to allow the comparison of the affinity and kinetics
of the interaction between DNA with and without methylated CpG sites and MBD
domains. All DNAs were chemically synthesized with a 3’-biotin tag (Integrated DNA
Technologies, IDT), which can bind to commercially made streptavidin-coated sensor
chips (SA chips, GE Healthcare Life Sciences). Different pairs of oligo were annealed to
mimic methylated DNA. The first set of oligos 5’8U
5’-CAGTGCACGTTACCGCGCGCGACCGACATC-3' and 3’8U
5’-GATGTCGGTCGCGCGCGGTAACGTGCACTGA-B-3’ (B=biotin) were annealed
together to generate a 30 bps un-methylated DNA (8uDNA). The oligos 5’8M
5’-CAGTGCAC
m
GTTACCGCGCGCGACCGACATC-3’ and 3’8U were annealed
together to generate a 30 bps hemi-methylated DNA (8hDNA). The oligos 5’8M and
3’8M 5’-GATGTCGGTCGCGCGCGGTAAC
m
GTGCACTGA-B-3’ were annealed
together to generate a fully methylated DNA (8mDNA). Furthermore, in order to study
the topological effect of two methylated CpG sites at different positions of the DNA
helix, cis-methylated DNA (8m18mDNA: methyl groups on same side of the helix) and
trans-methylated DNA (8m14mDNA: methyl groups on opposite side of the helix) were
made. 8m18mDNA was made by annealing of oligos 5’8M18M
5’-CAGTGCAC
m
GTTACCGCGC
m
GCGACCGACATC-3’ and 3’8M18M
5’-GATGTCGGTCGC
m
GCGCGGTAAC
m
GTGCACTGA-B-3’. 8m14mDNA was made
by annealing of the oligos 5’8M14M
5’-CAGTGCAC
m
GTTACC
m
GCGCGCGACCGACATC and 3’8M14M
10
5’-GATGTCGGTCGCGCGC
m
GGTAAC
m
GTGCACTGA-B. To examine whether the
non-methylated part of the DNA sequence close to a methylated CpG can help or
stabilize binding of MBD protein to the methylated CpG, a 14-bps fully methylated DNA
(short 8mDNA) was designed to compare with the 8mDNA. It was composed of the
oligos s5’8M 5’-CAGTGCAC
m
GTTACC-3’ and the s3’8M
5’-GGTAAC
m
GTGCACTGA-B-3’.
2.4 Generation of Biacore sensor chip surface
Streptavidin-coated sensor chips (SA chip, GE Healthcare) were used to
generate the DNA surface. For the analysis of singly-methylated DNA, flow cells were
coated with 8uDNA, 8hDNA, and 8mDNA, and one flow cell was unmodified. The next
chip was generated for cis- and trans-methylated sites analysis and consisted of 8mDNA,
8m14mDNA, 8m18mDNA, and unmodified surface. The final chip was generated to
examine the role of the non-methylated portion of DNA in the 8mDNA, and contained
short 8mDNA, 8mDNA, 8m18mDNA, and an unmodified surface.
2.5 SPR Analysis
Kinetic studies were performed using the Biacore 2000 instrument (GE
Healthcare). Pairs of DNA oligo were mixed with each other at equal concentration and
volume, heated at 85ºC for 10 min, and allowed to anneal at room temperature for 1 hour.
The annealed DNA oligo nucleotides were diluted to a final concentration of 0.1 μM in
HBS buffer (10 mM HEPES, pH 7.4, 150 mM NaCl, 3 mM EDTA, 0.005% surfactant
P20). The resultant DNA was then diluted to a concentration of 1 nM in Biacore running
11
buffer (10 mM Tris-HCl, pH 8.0, 150 mM NaCl, 5% glycerol, 62.6 μg ml
-1
acetylated
BSA, 1 mM DTT, 20 μg ml
-1
DNA, 0.05% surfactant P20). E. coli DNA was added into
the running buffer to minimize background binding. It was extracted from the
methylation deficient Dam/Dcm minus E. coli GM2163 and sonicated to small
fragments. In this step, 1 mL extracted DNA in the elution buffer was put in a 1.5 mL
tube. We usually use the sonic dismembrator (Fisher Scientific, Pittsburgh, PA) at level
three to sonicate extracted DNA for one minute with 30 seconds stop, and the total
sonication time is 2 minutes. To avoid overly strong binding caused by sample proteins
re-binding to DNA, the DNA was coated on commercial streptavidin sensor chips (SA
chip, GE Healthcare) at very low density, typically about 50 response units (RU) on each
cell flow surface. There are four flow cells on a chip that can be coated by different kinds
of DNA nucleotides, but one flow cell was left blank as a reference surface for
subtraction of signal noise from nonspecific binding and baseline drift. In addition,
random injections of Biacore running buffer were included at regular intervals to allow
double referencing of samples (Myszka, D.G., 1999). Proteins were serially diluted in
Biacore running buffer in five 1:3 dilutions. Three injecting of each concentration were
included, and injection order was randomized to avoid data bias. The whole experiment is
kept at 20ºC, and the flow rate is at 50 μL min
-1
for each one- minute injection followed
by a five-minute dissociation with running buffer. 2 M NaCl was injected for 1 min at 20
μL min
-1
to regenerate the surface. All experiments were repeated two or three times.
Data were processed by scrubber and analyzed by CLAMP XP (Myszka, D.G. and
12
Morton, T.A. 1998) (developed by the Center for Biomolecular Interaction Analysis at
the University of Utah, www.cores.utah.edu/interaction).
2.6 Gel Shift Assay
Appropriate concentrations of DNA oligos were annealed as described in DNA
preparation and diluted to 10 μM. 2 μL 10 μM DNA oligo and MBD proteins were
mixed in binding buffer (10 mM Tris, pH 8.0, 50 mM NaCl, 5 mM MgCl
2
, 0.1 mM
EDTA, 0.1 % NP40, 1 mM DTT, and 5 % glycerol) to total 20 μL volume at room
temperature for 30 minutes. Reactions were electrophoresed through 6 % polyacrylamide
gels in 0.5x TBE buffer at 70 V for 80 minutes, and electroblots were transferred to a Pall
Biodyne B 0.45 μm memberane at 40 V for 1 hour. The membrane was blocked with 5 %
milk dissolved in TBST buffer and probed with extravidin- horseradish peroxidase at a
1:20000 dilution in TBST buffer. Finally, Immobilon Western detection reagent
(Millipore, Billerica, MA) was added to the membrane and developed for
chemiluminescence using Fluor-S multiImage (Bio-Red, Hercules, CA).
13
CHAPTER 3: RESULTS
3.1 The examination of MDB2b binding to a fully methylated CpG dinucleotide
To understand how a fully methyl-CpG site on both strands contributes to the
interaction with MBD2b, unmethylated (8uDNA), hemimethylated (8hDNA), and fully
methylated (8mDNA) double-stranded DNA surfaces were generated by coating
annealed oligonucleotides on different flow cells of a streptavidin coated (SA) sensorchip.
8uDNA has no methylated CpG dinucleotides, while 8hDNA is hemi-methylated at
position 8 and 8mDNA is fully methylated at the CpG dinucleotide at position 8 and 9 of
the oligo (Fig. 1, A-C). One surface was left uncoated to be able to correct for
background binding or refractive index changes. MBD2b showed different binding
responses on the three DNA surfaces. It bound more strongly to 8mDNA than to the
8uDNA and 8hDNA surfaces. Interestingly, MBD2b binding to 8hDNA showed a fast-on
and fast-off kinetics, suggesting that each methyl-CpG of both strands in DNA
contributes weakly to the binding of MBD2b (Fig. 2). Together, a fully methyl-CpG site
can lead a more stable binding of MBD2b. Based on this observation, we tried to explore
the mechanism of MBD2b binding to a fully methyl-CpG site.
14
Figure 1. Oligonucleotides used for the study. The sequence of the each biotinylated
oligo is biotinylated on the bottom strand. All oligonucleotide pairs at 0.1 μM were
annealed at 95
o
C for 10 minutes and cooled down to room temperature over the course
of one hour. For single methyl-CpG oligos, non-, one-, and two-strand CpG dinucleotides
were methylated at 8
th
or/and 9
th
nucleotide (A-C). For dimethyl-CpG oligos, 8
th
, 9
th
, 14
th
,
and 15
th
nucleotides or 8
th
, 9
th
, 18
th
, and 19
th
nucleotides were methylated on trans- or cis-
sides of DNA helix (D and E). The short version of 8mDNA has the same sequence of
the regular 8mDNA oligo from 1-14
th
nucleotides.
Figure 2. Interaction of MBD2b with non-methylated, hemimethylated and fully
methylated DNA. Three-fold serial dilutions of MBD2b were injected continuously for 1
minute followed a five-minute dissociation with running buffer flowing over the surfaces.
Each sample was triply and injected in random order over the sensor chip. Signal from
the uncoated surface was used to subtract non-specific responses from the three coated
surfaces. The data (black lines) was fitted globally (red lines) using CLAMP with a 1:1
interaction model with a mass transport correction to obtain association and dissociation
rates.
15
From the studies of Dr. Gerd Pfeifer’s laboratory in 2004, it appears that
MBD3L1, a protein which is highly homologous to MBD2 and MBD3 but without an
N-terminal MBD domain, can stabilize binding of MBD2b to methyl-CpG sites (Fig. 3;
Jiang et al., 2004 and Rauch and Pfrifer 2005). The Pfeifer lab used the yeast two hybrid
system and gel shift assay to demonstrate that MBD3L1 can interact with MBD2b (Fig.
4). To verify this result, we measured binding of MBD2b to methylated DNA in the
absence and presence of MDB3L1 (Fig. 5). First, we tested whether MBD3L1 would
bind to a fully methyl-CpG dinucleotide on its own. As expected, SPR analysis showed
negligible binding (Jiang et al., 2002). When the same concentration of MBD2b and
MBD3L1 were mixed before exposure to the sensor chip, the sensorgram showed no
obvious difference between MBD2b/MBD3L1 complex and MBD2b only on the
8mDNA surface, suggesting that MBD3L1 does not increase MBD2b binding to a single
fully methyl-CpG site. This result was in apparent disagreement with the Pfeifer lab’s
findings. There are several possible reasons to explain this observation. One most
possibility is MBD3L1 associating with MBD2b can not contribute the
MBD2b/MBD3L1 complex binding to methylated CpGs. Because MBD3L1 unable to
bind on the DNA, it apparently can not enhance the interaction between MBD2b and
methylated DNA. However, it does not mean MBD3L1 has no function on increasing
MBD2b binding to methyl-CpGs. One considered model is MBD3L1 can recruit another
factor interacting with DNA. However, Pfeifer lab’s data have shown only MBD2b and
MBD3L1 still can work together to have higher affinity on methylated DNA. Thus,
MBD3L1 in MBD2b/MBD3L1 complex able to recruit another MBD2b lights a potential
direction for the true mechanism. In the future, we will examine binding of MBD2b and
16
MBD3L1 to a series of more complex target containing different numbers of CpGs at
different distances. For this thesis work, the MBD2b and MBD3L1 interaction was not
explored further. Instead I focused on the role of the MBD domain in MBD2b in
mediating binding to methylated DNA. Perhaps through understanding the binding
characteristic of MBD domain, we can have more idea about the mechanism of
MBD2b/MBD3L1 complex increasing the affinity with methylated DNA.
Figure 3. The cartoon shows MBD2b and MBD3L1 domain structures. MBD2b has an
N-terminal MBD and a C-terminal coiled coil in the full length of 262 amino acids.
MBD3L1 has similar sequence with MBD2b but lacks the MBD. (Figure from Jiang et al.,
2004)
17
Figure 4. The yeast two-hybrid system shows the interactions between MBD proteins.
MBD2b shows weak interaction with itself, but MBD3L1 shows strong interaction with
itself. However, the interaction between MBD2b and MBD3L1 shows different results
may be due to the orientation of fused binding domain and activity domain with MBD2b
and MBD3L1. (Figure from Jiang et al., 2004)
18
Figure 5. Interaction of MBD3L1 with DNA, and effect of MDB3L1 on DNA binding by
MBD2b. The upper figures show MBD3L1 injected over the chip using the same program
as in the Fig. 2 (one-minute sample injection followed five-minute buffer flowing). The
bottom figures show a comparison of MBD2b with or without MBD3L1 in the samples.
The two proteins were mixed in 20 uM reaction buffer for 30 minutes on ice followed a
three-fold serial dilution. The data (black lines) was fitted globally (red lines) using
CLAMP with a 1:1 interaction model with a mass transport correction to obtain
association and dissociation rates.
3.2 MBD domain is important in MBD2b binding to a fully methyl-CpG dinucleotide
MBD2 has an N-terminal glycine-arginine-rich repeat region ((GR)
n
), a MBD
domain, and a C-terminal coiled-coil domain. The MBD domain is highly conserved in
MBD-containing proteins. Due to having two initiation codons, two translated protein
products from MBD2 mRNA have been discovered. The full length protein is referred to
as MBD2a. MBD2b is a shorter version lacking the N-terminal (GR)
n
region (Fig. 6). To
further study how the MBD domain of MBD2b contributes to the binding of methylated
DNA, we created a single MBD domain from MBD2a (1xMBD; MBD2
145-217
) and
examined its binding kinetics in SPR experiments. 1xMBD was injected onto the chip
19
containing 8uDNA, 8hDNA, and 8mDNA surfaces, and the injection time was one
minute followed five-minute dissociation. The sensorgram of 1xMBD on the 8mDNA
surface showed the presence of an unstable complex different from MBD2b (Fig. 4).
MBD2b showed a clearly slower dissociation than 1xMBD, suggesting that the
MBD2b/8mDNA complex is more stable than the 1xMBD/8mDNA complex. In addition,
very little binding is seen when 1xMBD is injected over the 8uDNA and 8hDNA surfaces,
indicating that a fully methyl-CpG dinucleotide is required for MBD domain binding.
Taking the MBD2b and 1xMBD data together, it suggests that MBD domain can mediate
binding to the methylated DNA site, but that the C-terminal region of MBD2b is also
required to make a more stable interaction.
Figure 6. The cartoon shows two types of MBD2: MBD2a and MBD2b. MBD2a and
MBD2b are translated from MBD2 RNA with different start codons. MBD2a has a
glycin/arginine rich region, a MBD domain, and a coiled coil. MBD2b lacks the
glycine/arginine rich region in the N terminus. (Figure from Jiang et al., 2004)
Compared with previous data by Dr. Shirakawa and his colleagues, however, our
result of 1xMBD binding to single methyl-CpG DNA was different (Inomata et al., 2008).
In the Inomata paper, the MBD domain (amino acid 1-75 from MBD1) is shown to
exhibit much slower association and dissociation over a surface coated with single fully
20
methyl-CpG DNA than in our SPR experiments (Fig. 8 and table 2). We further
checked every step in their experiment and surprisingly found that their chip was coated
with high response unit (RU) of methylated DNA (~600RU) instead of our coating
(~50RU) on the chip surface. Figure 5 from the Inomata paper clearly shows that 100 nM
MBD1
1-75
can cause a response of over 150 RU on the methylated DNA surface after
one-minute injection, indicating that there were much more oligonucleotides coated on
their chip than on our chip surface (~15 RU in Fig. 7). To examine the effect of a higher
DNA density, we generated a chip with three different density 8mDNA surfaces. 50 RU,
500 RU, and 1500 RU 8mDNA were coated on the different surfaces of a chip, and a
surface was left as a blank control. The same concentration of 1xMBD flowed over the
four surfaces continuously for one minute followed by a 5-minute buffer flow to measure
dissociation. 1xMBD on the 500 RU 8mDNA DNA surfaces was found to have a
dramatically slower dissociation rate consistent with the result of MBD1
1-75
on the 600
RU methylated DNA surface (Fig. 9). On the 1500 RU DNA surface, a much higher
response unit of MBD/DNA complex and an obvious slower dissociation was shown,
indicating that a high density of DNA on the surface caused tighter MBD-DNA binding.
A rebinding phenomenon may explain the differences between the three surfaces in our
data, and suggests that when too much DNA is coated, the binding kinetics of 1xMBD
becomes dependent on the concentration of methylated DNA on the surface.
21
Figure 7. 1xMBD binding to non-, hemi-, and fully methylated DNA surfaces. The
samples were three-fold diluted and randomly injected as triplication. The slight signal
was observed in the figures of 1xMBD binding to 8uDNA and 8hDNA.
Figure 8. SPR data from the Inomata manuscript. The oligo
5’-biotin-AAAAAAGATCGAC
m
GACGTAC was annealed with its complementary
strand and then coated on the SA sensor chip at 600 RU. The running buffer was
composed of 20 mM potassium phosphate (PH 6.5), 50 mM KCl, and 5 mM DTT. The
MBD1
1-75
samples were diluted from 10-100 nM and injected at a flow rate of 50 μL
min
-1
for one minute. The association and dissociation rates were fitted globally using a
simple 1:1 Langmuir model. (Figure from Inomata et al., 2008)
22
MBD domain
from
k
a
(M
-1
, s
-1
)
k
d
(s-1)
K
D
(M)
MBD1 (4.63+0.17) x 10
4
(3.26+0.03) x 10
-3
(7.04+0.32) x 10
-8
MBD2b (2.05+0.8) x 10
7
(4.99+0.3) x 10
-1
(2.42+0.9) x 10
-8
Table 2. Comparison of kinetic data from Inomata et al., to our data using the MBD from
MDB2b. The k
a
, k
d
, and KD values of MBD1
1-75
are fitted globally using a simple 1:1
Langmuir model (reference from Inomata, K. et al., 2008). The k
a
, k
d
, and KD values of
MBD2b are fitted globally using 1:1 model with mass transport correction (sensorgram
shown in figure 2). The equilibrium dissociation constants (KD) were calculated from k
a
and k
d
(KD = k
d
/ k
a
).
Figure 9. Binding of 1xMBD to increased density surfaces. The 1xMBD samples were
three-fold diluted to 800, 267, 89.9, 39.6, and 9.9 nM and triply injected in random order
over the sensor chip. To clearly observe the curve of the low-density DNA surface, we
used a small scale to show the sensorgram. The scales of the middle and right panels are
identical.
In the Inomata manuscript, the authors hypothesized two possible models
describing the reversible interaction of the MBD with DNA. One model to explain this
mobility is a flip model in which the protein turns around while never leaving the DNA
strand. The other is a dissociation/re-association model, in which the protein gets off and
then can get back on in either orientation. In their deliberations, the authors used the k
off
[(3.26+0.03) x 10
-3
S
-1
] from their BIACORE data (which we believe to be slow due to
23
excessive coating of the surface). Since their measured k
off
was more than three orders
of magnitude smaller than the exchange rate k
ex
(4.30+0.51 S
-1
) of strands of the duplex
DNA estimated by chemical exchange 2D NMR spectroscopy, they concluded that the
flip model was more appropriate for explaining the phenomenon. According to our data,
however, if the chip surface had been coated with a lower concentration of methylated
DNA, the phenomenon of stronger binding would be eliminated, and the dissociation
would become faster, so that flip model would probably not be considered correct.
3.3 2xMBDs show stronger binding than 1xMBD to fully methyl-CpG dinucleotides
The MBD domain in MBD2 has been identified that can bind to methylated
DNA but shown a fast dissociation from the complex in our Biacore kinetics analysis.
Although the C-terminal region of MBD2b may improve binding kinetics, the mechanism
is still unknown. The region contains a coiled-coil domain that is thought to mediate
protein/protein interaction. A clone containing only the coiled-coil domain has since been
constructed in the Laird-Offringa lab by a high school student, David Lam, and this
domain will be tested in the future to see if it interacts with methylated DNA, MBD,
MBD2b or the coiled coil domain itself. DNA-binding proteins such as transcription
factors may use multiple DNA-binding domains or form a polymer to increase the
binding activity. These results illustrate that multimerization can increase affinity. Thus,
we considered what the effect might be if MBD2b could homodimerize. The yeast
two-hybrid data from the Pfeifer lab (Fig. 4; Jiang et al., 2004) implied that MBD2b can
weakly interact with itself. Homodimerizetion might increase the binding affinity even
24
though only one MBD domain in MBD2b has a weak binding ability. In addition, the
data showed that MBD3L1, a protein containing a similar sequence with MBD2 but
lacking a MBD domain, could strongly interact with MBD2b and itself. Stimulated by
MBD3L1 interaction, MBD2b could increase its binding affinity for methylated DNA
shown in the gel shift assay, suggesting that MBD3L1 may increase MBD2b by
facilitating the combination of MBD2b proteins in a complex. Dr. Adrian Bird and his
colleagues have generated MBD multimers and examined the binding of 1xMBD and
4xMBD to multi-methylated DNA (Jorgensen et al., 2006). The gel shift assay in the
paper indicated that 4xMBD had more complexes binding to methyl-CpG moieties than
1xMBD did. The calculation of binding constants from the gel shift assay clearly showed
4xMBD had a higher affinity for methylated DNA than 1xMBD (table 3). The data
inspired our thinking to examine 2xMBD engineered from MBD2 binding to one or two
methyl-CpG sites.
Table 3. Binding constants of single 1x and 4xMBDs to DNA targets with 1, 2, or 3
methyl groups. The binding constants were calculated from the gel shift assay data shown
in Dr. Adrian Bird’s lab’s paper published in 2006. The probe (25 fmol) was incubated
with 3-300 nmol 1xMBD or 1-100 nmol 3xMBD in the reaction buffer. (Reference from
Jorgensen et al., 2006)
25
We created three 2xMBD engineering proteins with different linker lengths:
2xMBDn with null linker, 2xMBD
5
with a 5-amino-acid linker, and 2xMBD
10
with a
10-amino-acid linker (Fig. 10). The three different linkers were designed to connect two
MBD domains in order to mimic the optimal spacing between two MBD domains for
binding two methyl-CpG dinucleotides of DNA. In vivo, methylated CpGs are found
spaced at an average distance of 8-10 nucleotides apart. We then examined 2xMBDs on
the chip containing 8uDNA, 8hDNA, and 8mDNA surfaces. Totally different from
1xMBD, 2xMBDs have shown very high affinities to a sully single methylated CpG of
DNA (Fig. 11). The sensorgrams in figure 11 show the affinities of 2xMBDs binding to a
fully single methyl-CpG are much higher than the affinity of 1xMBD (Fig 7 V.S. Fig 11).
One possibility to explain the observation is the two MBD domains linked in 2xMBDs
may alternatively bind to the methyl-CpG site of 8mDNA (Fig. 12). If 1xMBD
dissociates from the methyl-CpG site, it would flow away from the surface in the running
buffer. However, if the 2xMBD protein contains two MBD domains, one MBD domain
could quickly replace the binding to a methylated CpG when the other MBD domain
dissociated, which could form an alternative binding and increases the apparent affinity
to the DNA surface. Therefore, the presence of two MBD domains within the 2xMBD
proteins may have increased the affinity of 2xMBDs binding to single methyl-CpG site.
1xMBD would have no such ability. Otherwise, the flip model from Dr. Shirakawa’s
discussion is not able to explain our finding because 1xMBD had evidently weaker
affinity to a fully methylated CpG than 2xMBDs. If the flip model is truly applied to the
MBD domains in 1xMBD and 2xMBDs, they should have similar affinities to 8mDNA.
26
Figure 10. 1xMBD and 2xMBDs shown in the cartoon. Red is an MBD domain, and
green is a linker. 1xMBD has a five-amino-acid linker. I used 2xMBD constructs with a
linker of zero, five or ten amino acids. The linker connects the C terminus of the first
MBD domain to the N terminus of the second MBD domain.
27
Figure 11. Binding of 2xMBDs to non-methylated, hemi-methylated or singly fully
methylated DNA. Each column of oligo coated on the chip surface shown from left to
right is 8uDNA, 8hDNA, and 8mDNA. Each row of examined sample shown from top to
bottom is 2xMBDn, 2xMBD
5
, and 2xMBD
10
. Each protein was three-fold diluted, and
randomly and triply injected on the surface for one minute. The sensorgram of 2xMBDn
binding to 8uDNA shows a slight response perhaps due to weakly non-specific
interaction.
Figure 12. The alternative binding model. Two MBD shown in red may alternatively
bind to the same methyl-CpG of DNA and increase the affinity.
28
However, when we looked 2xMBDs binding to the un- and hemi-methylated
DNA, apparent responses were still able to be detected (10-30 RU).According to the 3D
structural studies of MBD domain from MBD1, an MBD domain not only interacts with
a fully methyl-CpG, its loop L1 and helix α1 also has other electrostatic binding or
hydrogen bond to interact with DNA bases, deoxyribose sugars, and phosphates (Fig. 13;
Ohki et al., 2001). Therefore, one possible alternative explanation for 2xMBDs having
stronger binding ability to a fully methylated CpG might be its second MBD domain that
can weakly interact with non-methylated sequence of the DNA. Due to the enhancement
of interaction from the second MBD domain, the affinity of 2xMBDs binding to the
methylated CpG may be improved. In addition, the other MBD domain may quickly
rebind to the methyl-CpG because of the promotion from the second MBD binding close
to the methylated CpG (Fig. 14). Thus, instead of two linked MBD domains trading
places on the methylated CpG, one MBD domain would bind to the methylated site while
the other one would facilitate the rebinding of that domain. In order for this latter model
to be true, the DNA target must be long enough to fit the two linked MBD domains.
Therefore, how many nucleotides of DNA that a MBD domain needs to occupy became a
key question. The structural studies pointed that it is a remarkably small area (Ohki et al.,
2001). Ohki and his colleages showed that the MBD domain only invaded DNA in a
narrow area of the major groove and that no any contact was involved in other parts of
DNA. In their studies, only six nucleotides are occupied and involved in MBD domain
binding (Fig. 15). Thus, our original 30-nt target 8mDNA would likely be able to
accommodate the second MBD domain of 2xMBD on the unmethylated part of the oligo,
and this could indeed stabilize the interaction. To examine the possibility, we then
29
explored whether the length of the oligonucleotide would affect 1xMBD and 2xMBD
binding. We designed a short version of 8mDNA (S8mDNA) that had 14 nucleotides
with a methyl-CpG dinucleotide at 8
th
and 9
th
nucleotide. Because of the half length of
the oligo lacking enough space for two MBD domains binding but still allowing one
MBD domain binding, the second MBD domain in 2xMBDs no longer bound to the oligo
(Fig. 16). The S8mDNA was coated with 8mDNA and 8m18mDNA on the different
surfaces of a chip. 1xMBD and 2xMBDs on the S8mDNA surface showed similar shapes
of the binding curve compared to the binding on the 8mDNA surface (Fig. 17 compared
by Fig. 11). Therefore, our experiments support the idea that 2xMBDs used both MBD
domains to alternately bind to a methyl-CpG site. However, even the alternative binding
model is more preferential for 2xMBDs binding to a fully methylated CpG, the rebinding
model still has weak effect shown in our SPR experiments because the dissociations seem
to become faster in figure 17 compared to figure 11.
30
Figure 13. Schematic summary of protein-DNA contacts. MBD binding to DNA
represents interfacial hydrophobic (yellow arrow), hydrogen bond (red arrow), and
electrostatic (blue) interactions with DNA bases, deoxyribose sugars, and phosphates,
suggesting that MBD domain not only binds to methylated CpGs but other components
of DNA. (Figure from Ohki et al., 2001)
Figure 14. The rebinding model. The second MBD binding to the non-methylated
sequence may promote the other MBD to rebind to the same methylated CpG of DNA
and increase the affinity.
31
Figure 15. 3D structural study of MBD domain binding to methylated DNA. The
modeling structure shows MBD contact a very small interface invading in the major
groove. The interface takes about six nucleotides length of DNA. (Figure from Ohki et al.,
2001)
Figure 16. The short version of 8mDNA may reduce the affinity of 2xMBDs bind to a
fully single methylated CpG of DNA. The second MBD loses it binding site and no longer
interacts with the DNA, which decrease the affinity of the other MBD binding to the
methylated CpG.
32
Figure 17. Binding of 1xMBD and 2xMBD to a shorter singly methylated DNA target.
S8mDNA is the short version of 8mDNA containing 14 nucleotides, which is half the
size of 8mDNA. It was coated together with 8mDNA and 8m18mDNA on the different
surfaces of a chip. 1xMBD and 2xMBDs were three-fold diluted and examined by
triplicate injections in random order.
To investigate how 2xMBDs bind to double methyl-CpGs of DNA, we also
designed two dimethyl-CpG oligos, 8m14mDNA and 8m18mDNA. 8m14mDNA has two
methyl-CpGs relatively positioned in trans on the DNA helix, and 8m18mDNA has two
methyl-CpGs relatively positioned in cis on the DNA helix (Fig. 1 D and E). 8mDNA,
8m14mDNA, and 8m18mDNA were coated on the different surfaces of a chip, and a
blank surface was left as a control. Each 2xMBD was examined by random order
triplicate injections over the chip surfaces for one minute followed a five-minute
dissociation, and all experiments were carried out three times. The results showed that
2xMBDs had higher affinities to dimethylated DNA than single methylated DNA (Fig.
18). To look insight of 2xMBDs binding to dimethylated DNA. Slower dissociation of
2xMBDs binding to cis-dimethylated CpGs than trans-dimethylated CpGs was observed,
33
suggesting that 2xMBDs bound more stably to a cis-methylated rather than a
trans-methylated DNA helix due to the constraint of the linking lengths (Fig. 18). In
addition, 2xMBD
10
exhibited a slight but clearly slower dissociation from the complex
with trans-dimethylated DNA than 2xMBDn and 2xMBD
5
, suggesting that the longer
flexible linker allowed the two linked MBD domains to contact the methyl-CpG sites on
opposite sides of the helix. Together, 2xMBDs had higher affinity binding to
dimethyl-CpG DNA than 1xMBD binding to single methyl-CpG DNA, and the spacing
between two MBD domains could affect their ability to bind to methyl groups on the
same or opposite sides of DNA helix.
34
Figure 18. Binding of 2xMBDs to singly methylated, trans-dimethylated or
cis-dimethylated DNA. Each column of oligo coated on the chip surface shown from left
to right is 8mDNA, 8m14mDNA, and 8m18mDNA. Each row of examined sample
shown from top to bottom is 2xMBDn, 2xMBD
5
, and 2xMBD
10
. Each protein was
three-fold diluted, and randomly and triply injected on the surface for one minute.
Looking at our data, we observed an interesting phenomenon. 2xMBDs on the
trans-dimethylated DNA surface had higher response units during equilibrium than on the
single methylated DNA and cis-dumethylated DNA surfaces (Fig. 18). The data
suggested that the trans-positioned dimethyl-CpG sites were possibly occupied by two
proteins and therefore increase the response. This effect might be most amplified by the
2xMBDs with null or a short linker, which may be disabled to allow two MBD domains
to contact opposite-side methyl-CpGs. In contrast, 2xMBD proteins showed lower
response units of complex but more stable binding in the sensorgrams to cis-side
35
dimethyl-CpG sites, suggesting that a single 2xMBD protein could easily occupy the
two same-side methyl-CpGs. We next examined the binding of MBD2b to the single and
dimethylated surfaces (Fig. 19). The strong response (~80 RU) on the trans-dimethylated
DNA surface indicated that two MBD2b proteins can independently occupy either
methyl-CpG sites of trans-dimethylated DNA without steric interference. Meanwhile,
MBD2b binding to the cis-dimethylated DNA showed the lower response (~60RU) as
well as binding to the single methylated DNA. These data suggested that the other parts
of MBD2b may mask the other methylated CpG of DNA, making it unaccessible to
proteins. We did not observe any stabilization of MBD2b binding to one V.S. two methyl
groups on the DNA target, indicating that our hypothesis that the C-terminal domain of
MDB2b allows dimerization is incorrect, or that the length between the two methyl
groups is too close to provide enough space for two MBD2b binding on it.
Figure 19. Binding of MBD2b to singly methylated or dimethylated DNA. MBD2b diluted
from 100 to 1.2 nM, and these samples were randomly and triply tested on the 8mDNA,
9m14mDNA, and 8m18mDNA surface. Even though the 8mDNA was coated on the
different chip in different time, the shape of curves is similar when MBD2b was
examined on the two chips (one chip was coated by 8uDNA, 8hDNA, and 8mDNA, and
the other one was coated by 8mDNA, 8m14mDNA, and 8m18mDNA).
36
To more clearly visualize how many 2xMBDs might be binding to the
dimethyl-CpG sites of the cis- and trans-dimethylated DNA, gel shift assays were carried
out. 2xMBD
5
was incubated with 10 uM biotinylated DNA probes 8mDNA,
8m14mDNA, and 8m18mDNA for 30 minutes. The complex with 8mDNA was first
examined. One clear shifted band indicated that 8mDNA can interact with one 2xMBD
(Fig. 20).The complex with trans-dimethylated DNA showed two bands of similar
intensity with increasing concentrations of 2xMBD, suggesting that each methyl-CpG
site of the trans-dimethylated DNA can form a complex independently. However,
2xMBD binding to cis-dimethylated DNA showed predominantly a single strong shifted
band. The second band appeared only at high 2xMBD concentration, suggesting that the
cis-dimethylated DNA is preferentially bound by one 2xMBD protein (Fig. 21). Together,
the results agreed our analysis of SPR data that the spacing between two linked MBD
domains determines the binding model of 2xMBDs. The short linker allowed 2xMBD to
interact with the two cis-methylated CpGs but constrained the contact of trans-methylated
CpGs.
37
Figure 20. Gel shift of 2xMBD with a singly methylated DNA target (right DNA helix).
The protein samples (2xMBD
5
) were incubated with 2 μL 10 uM biotinylated 8mDNA in
20 μL binding buffer (10 mM Tris, pH 8, 50 mM NaCl, 5 mM MgCl
2
, 0.1 mM EDTA,
0.1 % NP40, 1 mM DTT, and 5 % glycerol) for 30 minutes on ice. The first lane shows
biotinlyated 8mDNA with no protein as a negative control. The shifted bands showed
increased intensity from left to right dependent on the growing concentration of protein.
A single shifted band indicates that only one complex is formed between 2xMBD and
8mDNA.
38
Figure 21. Gel shift assay of 2xMBD with trans- or cis-dimethylated DNA targets (the
position of methylated CpGs of DNA helix shown below). The procedure of the gel shift
assay is described in the Methods section and in Figure 11. On the left panel, 2xMBD
5
is
shown to associate with biotinylated 8m14mDNA. The first lane is oligo only as a
negative control. The amount of the complex increases dependent on the growing
concentration of protein. In contrast to the 8mDNA, there are two shifted bands equally
increasing, suggesting that the two methyl-CpG sites on the oligo are each associating
with a protein. In the right panel, 2xMBD
5
is shown to associate with 8m18mDNA. In
contrast to the left panel, only one shifted band was observed when 2xMBD5 is in low
concentration. An upper shifted band was slightly visible in the highest concentration of
2xMBD5.
39
CHAPTER 4: CONCLUSION AND DISCUSSION
Our SPR experiments have provided the insight into the mechanism of binding
of an MBD domain to a fully methyl-CpG dinucleotide. A published report of an
engineered 4xMBD protein made from MBD1 successfully demonstrated to obtain a
higher affinity than 1xMBD, indicating that the multimerization of MBD domains can
increase their affinity for methylated DNA (Jorgensen et al., 2006). To investigate the
kinetics of multimerization, we engineered 2xMBDs and 1xMBD and examined their
binding to methyl- and dimethyl-CpGs. The SPR experiments have shown that 2xMBDs
had higher affinities than 1xMBD no matter binding to the single or dimethylated CpGs.
The most critical improvement of 2xMBDs’ binding affinities is to apparently reduce the
dissociation rates instead of to gain a faster association. In contrast, 1xMBD quickly
dissociates from the CpG-methylated DNA.
In the sensorgrams, the curves of the 2xMBDs indicated that their protein/DNA
complexes were more stable than 1xMBD, which evidences adding another MBD domain
able to much improve the affinity. This effect is even visible when the target DNA carries
only one fully methylated CpG. On such a target, the reduced dissociation rate is
probably a consequence of alternate binding by the two MBD domains, which allows
rebinding. This idea is supported by our observation that interactions of 1xMBD with a
higher density surface also show an apparently slower dissociation. When 2xMBDs were
examined on a dimethyl-CpG surface, interesting things were observed. 2xMBDs
generated a more stable complex with dimethyl-CpGs than with single methyl-CpG, no
matter whether the two CpGs were opposite (trans) or adjacent (cis). In addition, we
40
found that the length of the linker between two MBD domains can affects their binding.
For the 2xMBDs with a short linker, the two MBD domains could bind to the same-side
dimethyl-CpGs but not to opposite-side dimethyl-CpGs. Only two MBD domains with a
longer linker appeared to be able to contact dimethyl-CpGs on opposite sides of the DNA
helix. Our gel shift experiment showed that the short-linker 2xMBD (2xMBD
5
) could
form two complexes: one with one protein bound to the same-side dimethylate CpGs of
the DNA helix, and another with two proteins bound to the two opposite methylated
CpGs of the DNA helix. The data suggests that the ability of two linked MBD domains
interacting with methyl-CpGs depends on the spacing between two domains.
Based on our observations with multimerization and the fact that MBD2b can
stably bind to methylated DNA in the presence of MBD3L1, we hypothesize that MBD2b
may somehow multimerize via MBD3L1. The presence of C-terminal coiled coil
domains in MBD2b and MBD3L1 suggested that they may potentially associate to form a
complex through this motif which is involved in protein-protein interactions. Yeast
two-hybrid studies indicate that MBD2b and MBD3L1 can make MBD2b-MBD3L1,
MBD2b-MBD2b and MBD3L1-MBD3L1 interactions (Fig. 4; Jiang et al., 2004). In the
same paper, MBD3L1 was shown to increase MBD2b binding to methylated DNA (the
basis for the methylated-CpG island recovery assay, MIRA). However, our SPR
experiment did not show an increase in affinity upon mixing MBD2b and MBD3L1. The
exact reason is still unclear, but we think it may be related to the spacing. On the one
hand, the N-terminal region (amino acid 1-103) of MBD3L1 was found to be required for
interaction with MBD2b by in vitro protein binding assay. On the other hand, the
41
C-terminal region (amino acid 201-262) of MBD2b was capable of binding to
MBD3L1 in vitro (Jiang et al., 2004).The observations from Pfeifer lab may result in a
complex formed by the association of C-terminal region of MBD2b. In addition,
MBD3L1 may use its coiled coil to promote the dimerization. Thus, if two
MBD2b/MBD3L1 complexes can interact with each other through the C-terminal
coiled-coil domain of MBD3L1, then two separate MBD2b might be brought together to
bind to methylated DNA and increase the affinity (Fig. 22). That is why the spacing can
explain no apparent increase appeared in MBD2b binding affinity in the
MBD2b/MBD3L1 mixed solution. We have observed that MBD2b may mask the other
methylated CpG when it was binding to a methylated CpG of cis-methylated DNA. Thus,
one MBD2b/MBD3L1 complex would have the same masking and interfere with the
other complex binding to the same DNA. However, if we could provide long enough
DNA, the multimerization of MBD2b proteins may show its improvement on the DNA
binding. To further study the mechanism of the MBD2b/MBD3L1 interaction, we will
examine whether the C-terminal part of MBD2b can interact with MBD3L1 in an SPR
experiment. Next, we will create a multi-methylated oligo to determine whether it will
allow the hypothetical MBD2b/MBD3L1 multimer to bind on it.
42
Figure 22. A cartoon showing the model we propose to explain the ability of the
MBD2b/MBD3L1 complex to bind to methylated DNA. DNA is shown orange, MBD2b is
shown light blue, and MBD3L1 is shown dark blue. The distance between two binding
sites is unknown. In addition, whether the complex can bind to cis- or trans-side located
methyl groups on the DNA helix is also unclear.
43
BIBLIOGRAPHY
Baylin, S.B., Herman, J.G., Graff, J.R., Vertino, P.M., and Issa, J.P. (1998). Alterations in
DNA methylation: a fundamental aspect of neoplasia. Adv. Cancer. Res. 72, 141–196.
Bird, A.P. and Wolffe, A.P. (1999). Methylation-induced repression belts, braces, and
chromatin. Cell 99, 451–454.
Costello, J.F., Fruhwald, M.C., Smiraglia, D.J., Rush, L.J., Robertson, G.P., Gao, X.,
Wright, F.A., Feramisco, J.D., Peltomaki, P., Lang, J.C., et al. (2000). Aberrant
CpG-island methylation has non-random and tumour-type-specific patterns. Nat. Genet.
24, 132–138.
Fujita, N., Shimotake, N., Ohki, I., Chiba, T., Saya, H., Shirakawa, M., and Nakao, M.
(2000). Mechanism of transcriptional regulation by methyl-CpG binding protein MBD1.
Mol. Cell. Biol. 20, 5107–5118.
Hendrich, B., Hardeland, U., Ng, H.H., Jiricny, J., and Bird, A. (1999). The thymine
glycosylase MBD4 can bind to the product of deamination at methylated CpG sites. Nature
401, 301–304.
Inomata, K., Ohki, I., Tochio, H., Fujiwara, K., Hiroaki, H., and Shirakawa, M. (2008).
Kinetic and thermodynamic evidence for flipping of a methyl-CpG binding domain on
methylated DNA. Biochemistry 47, 3266–3271.
Jiang, C.L., Jin, S.G., Lee, D.H., Lan, Z.J., Xu, X., O’Connor, T.R., Szabo, P.E., Mann,
J.R., Cooney, A.J., and Pfeifer, G.P. (2002). MBD3L1 and MBD3L2, two new proteins
homologous to the methyl-CpG-binding proteins MBD2 and MBD3: characterization of
MBD3L1 as a testis-specific transcriptional repressor. Genomics 80, 621–629
Jiang, C.L., Jin, S.G., Pfeifer, G.P. (2004). MBD3L1 is a transcriptional repressor that
interacts with MBD2 and components of the NuRD complex. J. Biol. Chem. 279,
52456–64.
Jones, P.A. and Takai, D. (2001). The role of DNA methylation in mammalian
epigenetics. Science 298, 1068-1070
Jorgensen, H.F., Adie, K., Chaubert, P., and Bird A.P. (2006). Engineering a high-affinity
methyl-CpG-binding protein. Nucleic Acids Research 34, e96.
Myszka, D.G. and Morton, T.A. (1998). CLAMP©: a biosensor kinetic data analysis
program. Trends Biochem Sci. 23, 149-50.
Myszka, D.G. (1999). Improving biosensor analysis. J. Mol. Recognit. 12, 279–284.
44
Nan, X., Ng, H.H., Johnson, C.A., Laherty, C.D., Turner, B.M., Eisenman, R.N., and
Bird, A. (1998). Transcripional repression by the methyl-CpG-binding protein MeCP2
involves a histone deacetylase complex. Nature 393, 386–389.
Ng, H.H., Zhang, Y., Hendrich, B., Johnson, C.A., Turner, B.M., Erdjument-Bromage, H.,
Tempst, P., Reinberg, D., and Bird, A. (1999). MBD2 is a transcriptional repressor
belonging to the MeCP1 histone deacetylase complex. Nat. Genet. 23, 58–61.
Ng, H.H., Jeppesen, P., and Bird, A. (2000). Active repression of methylated genes by the
chromosomal protein MBD1. Mol. Cell. Biol. 20, 1394–1406.
Ohki, I., Shimotake, N., Fujita, N., Jee, J.G., Ikegami, T., Nakao, M., and Shirakawa, M.
(2001). Solution Structure of the Methyl-CpG Binding Domain of Human MBD1 in
Complex with Methylated DNA. Cell, 105, 487–497.
Rauch, T. and Pfeifer, G.P. (2005). Methylated-CpG island recovery assay: a new
technique for the rapid detection of methylated-CpG islands in cancer. Lab. Invest. 85,
1172–80.
Rich, R.L. and Myszka, D.G. (2001). BIACORE J: a new platform for routine
biomolecular interaction analysis. J. Mol. Recognit. 14, 223–228.
Wade, P.A., Gegonne, A., Jones, P.L., Ballestar, E., Aubry, F., and Wolffe, A.P. (1999).
Mi-2 complex couples DNA methylation to chromatin remodelling and histone
deacetylation. Nat. Genet. 23, 62–66.
Zemach, A. and Grafi, G. (2007). Methyl-CpG-binding domain proteins in plants:
interpreters of DNA methylation. Trends Plant Sci. 12, 80-85
Zemach, A., Gaspan, O., and Grafi, G. (2008). The three methyl-CpG-binding domains of
AtMBD7 control its subnuclear localization and mobility. J. Biol. Chem. 283, 8406-8411
Zhang, Y., Ng, H.H., Erdjument-Bromage, H., Tempst, P., Bird, A., and Reinberg, D.
(1999). Analysis of the NuRD subunits reveals a histone deacetylase core complex and a
connection with DNA methylation. Genes Dev. 13, 1924–1935.
Abstract (if available)
Abstract
Mehtyl-CpG binding domain protein 2 and 3 (MBD2 and MBD3) contain methyl-CpG binding domains (MBD) and belong to a family of MBD proteins. Methyl-CpG binding domain protein 3-like-1 (MBD3L1) is a protein with highly homology to MBD2 and MBD3, but it lacks the MBD domain in the N-terminus. A new method, methylated-CpG island recovery assay (MIRA), was developed to purify methylated DNA by using the mixture of MBD2b and MBD3L1. MBD3L1 was found to be able to interact with MBD2b and increase the affinity of MBD2b binding to methylated DNA. However, the mechanism is unknown. Our goal is to explore the mechanism of MBD3L1 involved in MBD2b binding to methylated DNA. We created the engineered 1xMBD and 2xMBDs from MBD2. 2xMBDs showed higher affinities than 1xMBD binding to the single and double fully methylated DNA. The alternative binding model and the rebinding model had been hypothesized to increase the affinities of 2xMBDs, but we observed that the alternative binding model was more preferential for 2xMBDs binding to a fully single methyl-CpG by further studies. In addition, the two linked MBDs were able to contact both methyl-CpGs of cis-methylated DNA, which increased a much more stable binding than trans-methylated DNA. The orientation of 2xMBDs binding suggested the spacing between two MBDs may be an important factor involved in the binding mechanism of MBD2b/MBd3L1 complex. Our findings reveal essential properties of MBDs in binding to methylated DNA. By these results, we can depict a clearer mechanism of how an MBD3b/MBD3L1 complex increases the affinity and binds to methylated DNA.
Linked assets
University of Southern California Dissertations and Theses
Conceptually similar
PDF
Understanding protein–DNA recognition in the context of DNA methylation
PDF
Decoding protein-DNA binding determinants mediated through DNA shape readout
PDF
The relationship between DNA methylation and transcription factor binding in colon cancer cells
PDF
CpG poor promoter SULT1C2 regulated by DNA methylation and is induced by cigarette smoke condensate in lung cell lines
PDF
Identification of DNA methylation markers in diffuse large B-cell lymphoma
PDF
DNA methylation markers for blood-based detection of small cell lung cancer in mouse models
PDF
Genome-wide studies reveal the function and evolution of DNA shape
PDF
DNA methylation inhibitors and epigenetic regulation of microRNA expression
PDF
The function of Rpd3 in balancing the replicaton initiation of different genomic regions
PDF
Biochemical characterization and structural analysis of two hexameric helicases for eukaryotic DNA replication
PDF
Developing a method for measuring kinetic rates of RNA/protein interaction using switchSENSE® technology on the DRX2
PDF
In vivo study with a novel liposomal formulation of a recombinant ADAM disintegrin-like domain showing both anti-angiogenic and tumor growth inhibition activities
PDF
Mechanistic basis for chromosomal translocations at the E2A gene
PDF
Biochemical studies of APOBEC protein family
PDF
Investigating the function and epigenetic regulation of ABCA3, a novel LUAD tumor suppressor gene
PDF
Understanding DNA methylation and nucleosome organization in cancer cells using single molecule sequencing
PDF
DNA methylation and gene expression profiles in Vidaza treated cultured cancer cells
PDF
Structural and biochemical studies of large T antigen: the SV40 replicative helicase
PDF
Impacts of post-translational modifications on interactions between G9a and its N-terminus binding partners
PDF
Nonlinear modeling of the relationship between smoking and DNA methylation in the multi-ethnic cohort
Asset Metadata
Creator
Chen, Po-Han
(author)
Core Title
The kinetic study of engineered MBD domain interactions with methylated DNA: insight into binding of methylated DNA by MBD2b
School
Keck School of Medicine
Degree
Master of Science
Degree Program
Biochemistry and Molecular Biology
Degree Conferral Date
2009-12
Publication Date
05/24/2011
Defense Date
10/28/2009
Publisher
University of Southern California
(original),
University of Southern California. Libraries
(digital)
Tag
MBD protein,methyl-CpG,OAI-PMH Harvest
Language
English
Contributor
Electronically uploaded by the author
(provenance)
Advisor
Laird-Offringa, Ite A. (
committee chair
), Haworth, Ian S. (
committee member
), Tokes, Zoltan A. (
committee member
)
Creator Email
pohanche@gmail.com,pohanche@usc.edu
Permanent Link (DOI)
https://doi.org/10.25549/usctheses-m2759
Unique identifier
UC1167402
Identifier
etd-Chen-3394 (filename),usctheses-m40 (legacy collection record id),usctheses-c127-278087 (legacy record id),usctheses-m2759 (legacy record id)
Legacy Identifier
etd-Chen-3394.pdf
Dmrecord
278087
Document Type
Thesis
Rights
Chen, Po-Han
Type
texts
Source
University of Southern California
(contributing entity),
University of Southern California Dissertations and Theses
(collection)
Repository Name
Libraries, University of Southern California
Repository Location
Los Angeles, California
Repository Email
cisadmin@lib.usc.edu
Tags
MBD protein
methyl-CpG