University of Southern California Dissertations and Theses

Molecular mechanisms of recurrent chromosomal translocations in human leukemias and lymphomas

(USC Thesis Other)

Molecular mechanisms of recurrent chromosomal translocations in human leukemias and lymphomas

PDF

Download

Open document

Flip pages

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content

MOLECULAR MECHANISMS OF RECURRENT CHROMOSOMAL
REARRANGEMENTS IN HUMAN LEUKEMIAS AND LYMPHOMAS

by

Albert G Tsai

A Dissertation Presented to the
FACULTY OF THE GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(BIOCHEMISTRY AND MOLECULAR BIOLOGY)

August 2008

Copyright 2008 Albert G Tsai

ii
Table of contents

List of tables iii

List of figures iv

Abstract vi

Chapter 1: Introduction 1
The molecular mechanisms of chromosomal rearrangements 1
The bcl-1 and bcl-2 translocations 4
Mechanisms of double-strand breakage at the immunoglobulin
heavy chain locus
10
DNA structure as a basis for translocation breakpoint clustering 24
Bisulfite as a probe of DNA structure 24

Chapter 2: Human chromosomal translocations at CpG sites and a
theoretical basis for their lineage and stage specificity
26
Summary 26
Introduction 26
Materials and Methods 30
Results 42
Discussion and theory 82

Chapter 3: DNA structure as a basis for translocation breakpoint
clustering at the bcl-1 MTC and bcl-2 MBR
94
Summary 94
The V(D)J recombination assay 95
Attempts to circumvent the expression blocking effect 102
In vitro transcription through the bcl-1 113
Electrophoretic mobility shift assay 117
Chemical probing methods 123
Enzymatic probing and direct labeling methods 128
Hydrolytic deamination assay 135

Chapter 4: Bisulfite as a probe of DNA structure 140
Summary 140
Introduction 140
Materials and Methods 143
Results 146
Discussion 161
Conclusions 165

iii
Chapter 5: Conclusions and future directions

166
Bibliography 172

iv
List of tables

Table 2-1 Summary of translocation breakpoint features. 43

Table 2-2 Statistical analysis of CpG and CAC proximity in various
translocations.
53

v
List of figures

Figure 1-1 General mechanism of chromosomal rearrangements. 2

Figure 1-2 The three types of chromosomal rearrangement junction
structures.
5

Figure 1-3 The bcl-1 translocation. 8

Figure 1-4 Structure of an immunoglobulin and its relation to V(D)J
recombination.
11

Figure 1-5 Molecular mechanism of V(D)J recombination. 14

Figure 1-6 Sequence of immunoglobulin heavy chain
rearrangement.
18

Figure 1-7 The R-loop:deaminase model for class switch
recombination.
22

Figure 2-1 Sequence-level breakpoint diagrams for bcl-2 clusters. 45

Figure 2-2 Sequence-level breakpoint diagrams for the bcl-1 MTC,
E2A, main SCL, and SIL clusters.
47

Figure 2-3 Distributions of breakpoints at various distances from
CpG or CAC in representative translocations.
50

Figure 2-4 Sequence-level diagrams for balanced translocations. 59

Figure 2-5 Proportions of junctions with nucleotide additions in
various translocations.
72

Figure 2-6 The RAGs endonuclease is capable of nicking one-base
mismatches, and can process nicks, gaps, and flaps into
double-strand breaks.
78

Figure 2-7 Theoretical mechanism for CpG-type double-strand
breakage.
85

Figure 2-8 Methylation status of the bcl-2 MBR and bcl-1 MTC. 88

Figure 3-1 The V(D)J recombination assay. 96

vi
Figure 3-2 Expression blocking effect. 100

Figure 3-3 V(D)J recombination assay of pAT90. 103

Figure 3-4 Alternative construct designs attempting to circumvent
the expression blocking effect.
108

Figure 3-5 In vitro transcription through the bcl-1. 114

Figure 3-6 Gel shift of bcl-1 PCR fragment. 119

Figure 3-7 Incomplete extension of PCRs used for gel shift. 124

Figure 3-8 KMnO
4
and OsO
4
chemical probing. 126

Figure 3-9 Direct labeling method of DNA breakage detection. 129

Figure 3-10 P1 and the RAG endonuclease probing of the bcl-1. 131

Figure 3-11 Deamination assay. 137

Figure 4-1 Mechanism of bisulfite-catalyzed cytosine deamination. 142

Figure 4-2 Bisulfite reactivity of individual molecules of
oligonucleotide substrates.
148

Figure 4-3 Bisulfite reactivity on dsDNA correlates broadly with
relative propensity to breathe.
151

Figure 4-4 Bisulfite reactivity on dsDNA correlates broadly with
relative propensity to form B-A-intermediate structure.
156

Figure 4-5 Bisulfite is minimally reactive on DNA-2’-O-methyl RNA
hybrid duplexes.
157

Figure 4-6 Bisulfite reactivity at the bcl-1 MTC. 158

Figure 4-7 Hypothetical model of bisulfite-catalyzed deamination of
cytosines in double-stranded DNA.
163

vii
Abstract
The bcl-2 translocation, t(14;18)(q32;q21), is found in about 50% of all
non-Hodgkin’s lymphomas, and the bcl-1 translocation, t(11;14)(q13;q32), in
>90% of mantle cell lymphomas, making them the most common
chromosomal rearrangements in human lymphoma. They occur when
double-strand breaks on the bcl-2 region, 18q21, or the bcl-1 region, 11q13,
are joined to those from an ongoing V(D)J recombination event on 14q32 via
the nonhomologous end joining (NHEJ) pathway. While the V(D)J and NHEJ
processes have been extensively characterized, the nature of double-strand
breakage at the bcl-1 and bcl-2 regions has remained a mystery since their
discovery almost 25 years ago. Computational statistical analysis of more
than 1,700 breakpoints from human lymphoid and myeloid chromosomal
rearrangements yields an intriguing finding: Translocations occurring at the
pro-B/pre-B stage, including the bcl-1 and bcl-2 translocations, are highly
focused to the dinucleotide sequence CpG. In addition to the bcl-1 major
translocation cluster (MTC), the bcl-2 major breakpoint region (MBR), the bcl-
2 intermediate cluster region (icr), and the bcl-2 minor cluster region (mcr),
CpG hotspots also occur at the E2A cluster region from t(1;19)(q23;p13)
E2A-PBX1 translocations. We do not see such hotspots in rearrangements
in lymphoid-myeloid progenitors, mature B cells, or T cells, including
t(12;21)(p12;q22) TEL-AML1, t(8;21)(q22;q22) AML1-ETO, t(9;22)(q34;q11)
BCR-ABL, t(8;14)(q24;q32), c-myc-IgH switch, and del(1)(p32) SCL-SIL, just

viii
to name a few. Furthermore, we demonstrate a unique breakpoint
distribution around CpGs, consistent with the structure-specific endonuclease
activity of the lymphoid-specific RAG complex acting on bubble, mismatch,
and gap DNA substrates. CpG is of special significance in vertebrates
because of the tendency for cytosines in this context to be methylated at the
C5 position, and deaminated to another endogenous base, thymine. The
resulting T:G mismatches are not always repaired, leading to a genome-wide
CpG depletion over evolutionary time termed “CpG suppression,” as well as
high rates of mutations at CpGs in tumor suppressor genes in many cancers.
However, CpG has not previously been described as a motif for
chromosomal translocation. We propose a lesion-specific double-strand
breakage mechanism involving the RAG complex acting at deaminated
methyl-CpGs to explain the observed stage-specificity and unique breakpoint
distributions at these cluster regions.

1
Chapter 1: Introduction
The molecular mechanisms of chromosomal rearrangements
Genetic mutations are thought to underlie virtually all cancers. One of
the most important types of mutations in lymphomas and leukemias is
chromosomal rearrangement – a process by which a segment of DNA from
one region of the genome is moved to another region or deleted from the
genome altogether (Greaves and Wiemels, 2003; Vega and Medeiros, 2003).
These rearrangements can have various oncogenic effects, for instance,
placing enhancers upstream of oncogenes, disrupting tumor suppressor
genes, or creating chimeric genes (Look, 1997b).
Mechanistically, a rearrangement event is composed of two main
steps: 1. DNA double-strand breakages at two different regions of the
genome, and 2. joining of the wrong ends together (Fig. 1-1). Based on
sequence information from many different rearrangements from many
different cancers, it appears that most, if not all, rearrangement events are
joined using the nonhomologous end joining (NHEJ) pathway and variants
thereof (Ferguson and Alt, 2001b; Greaves and Wiemels, 2003; Lieber et al.,
2006a; Povirk, 2006). NHEJ joins any available broken DNA ends without
preserving the original DNA sequence, and without too much regard for the
cause of breakage (Lieber, 2008; Weinstock et al., 2007; Zarrin et al., 2007).
While work into the more detailed aspects of NHEJ is ongoing, NHEJ has

2

Figure 1-1. General mechanism of chromosomal rearrangements. The thick
black bar represents one region of the genome, and the red bar another.
They may be on the same or different chromosomes. Parallel double lines
indicate the surrounding sequences of DNA. Double-strand breakages occur
on each region, creating four free ends. Rearrangement occurs when ends
from different regions are joined together, i.e. a black end joins to a red end.

double-strand breakages
misjoining

3
been quite extensively characterized and successfully reconstituted in vitro
using the logically requisite enzymatic components identified by both genetic
and biochemical methods (Ma et al., 2004).
The largest remaining questions, then, are 1. how cells prevent
double-strand breakages from progressing to rearrangements, 2. how
chromatin structure and nuclear architecture influence rearrangements, and
3. the causes of double-strand breakage. The main goal of my work has
been to address a special but important case of this third question – the
mechanism(s) of double-strand breakage in the bcl-1 and bcl-2
translocations.
Translocations are one class of rearrangement events where the DNA
of one chromosome is transferred to another chromosome, resulting in a
hybrid chromosome termed a “derivative.” In “balanced” translocations, the
chromosomes essentially switch arms, resulting in two derivatives. The
fused region in a derivative chromosome is termed the “junction,” and the
exact position where the DNA sequence changes significantly from that
before the translocation is termed the “breakpoint.” Naturally, at each
junction there are two breakpoints – one for each chromosome involved.
Junctions can assume one of three structures: 1. the breakpoints from each
chromosome meet at one position in the derivative, 2. the breakpoints from
each chromosome are separated by a number of added nucleotides in the
derivative, or 3. the breakpoints from each chromosome overlap by a number

4
of nucleotides (Fig. 1-2). Translocations are typically detected by metaphase
karyotyping, fluorescent in-situ hybridization (FISH), Southern blotting, and/or
PCR. If the genomic positions of both breakpoints on a particular derivative
can be narrowed down to a sufficient degree, then the junction can be
sequenced and characterized. The cell of origin, breakpoint positions, and
junctional structure for a given rearrangement provide very valuable clues to
the mechanism(s) of double-strand breakage.
The bcl-1 and bcl-2 translocations
Almost 25 years have passed since the discovery of the bcl-1
translocation, t(11;14)(q13;q32), present in greater than 90% of mantle cell
lymphomas (MCL) and the bcl-2 translocation, t(14;18)(q32;q21), present in
80-90% of follicular lymphomas (FL) and 20-30% of diffuse large B-cell
lymphomas (DLBL) (Tsujimoto et al., 1984a; Tsujimoto et al., 1984b; Vega
and Medeiros, 2003). In contrast to the rapid progress of research into the
joining phase and into the nature of double-strand breakage on 14q32, the
mechanism(s) of double-strand breakage on 11q13 and 18q21 have
remained a mystery.
The bcl-1 and bcl-2 translocations are the two most common
translocations in human lymphomas in the United States, with the bcl-1
occurring in about 5% of all lymphomas, and the bcl-2 in about 50% (Jaffe,
2001). Both are balanced translocations found primarily in mature B-cell
neoplasms at a median age of 50-60 years, though the translocations

5
A. ↓
partner 1
aaacagctagccaagcggggggcactcccagcattgcata

||||||||||||||||||||
derivative
aaacagctagccaagcggggaagcacagtgactattacag

||||||||||||||||||||
partner 2
atcgggactactgacccactaagcacagtgactattacag

↑
B. ↓
partner 1
aaacagctagccaagcggggggcactcccagcattgcata

|||||||||||||||||
derivative
aaacagctagccaagcgaaagcacacagtgactattacag

|||||||||||||||||
partner 2
atcgggactactgacccactaagcacagtgactattacag

↑
C. ↓
partner 1
aaacagctagccaagcggggaacactcccagcattgcata

||||||||||||||||||||||
derivative
aaacagctagccaagcggggaagcacagtgactattacag

||||||||||||||||||||||
partner 2
atcgggactactgacccaggaagcacagtgactattacag
↑

Figure 1-2. The three types of chromosomal rearrangement junction
structures. Chromosomal rearrangement junctions can be classified by the
positions of the breakpoints on the partner chromosomes relative to one
another in the junction. In the standard breakpoint diagrams below, the first
line is the sequence of one partner chromosome before rearrangement, and
the third line is the sequence of the other partner chromosome before
rearrangement. Between them is the sequence of the fusion region in the
derivative. Each arrow marks the breakpoint of a partner chromosome, or
the exact position where the DNA sequence changes significantly from that
before the translocation. Vertical bars, “|”, denote identical nucleotide
sequences between the partners and the derivative. (A) In this first type of
junctional structure, the breakpoints meet in the derivative. (B) In this second
type, some nucleotides, underlined, are inserted between the breakpoints.
(C) In this last type, the breakpoints overlap, and the underlined nucleotides
could have come from either chromosome. This is termed “microhomology.”
That in the figure is an example of four-base microhomology.

6
themselves most probably occur during the pro-B/pre-B stage of B-cell
development. Both can be detected at low levels in the B-cells of normal
individuals, indicating that the translocations themselves are insufficient to
induce lymphoma (Bordeleau and Berinstein, 2000; Espinet et al., 2005).
Breakpoints on 14q32 map to the immunoglobulin heavy chain (IgH) locus,
and join to largely uninteresting regions flanking known oncogenes – the
CCND1 gene encoding the cyclin D1 cell cycle progression protein in the
case of bcl-1, and the bcl-2 gene encoding the bcl-2 anti-apoptotic protein in
the case of bcl-2 (Jaffe, 2001). Juxtaposition of these genes with the strong
IgH µ enhancer leads to their overexpression, which is thought to be critical
for lymphomagenesis (Fig. 1-3). In mouse models, overexpression of either
oncogene alone is insufficient to cause lymphoma. As in humans, secondary
genetic alterations and/or additional factors are likely required (Bodrug et al.,
1994; Vaux et al., 1988).
Most intriguingly, while 11q13 breakpoints from bcl-1 translocations
can be found over a roughly 200 kb region centromeric to the CCND1 gene,
30-50% occur at the center of this region in a 150 bp zone termed the major
translocation cluster (MTC) (Bertoni et al., 2004). And while 18q21
breakpoints from bcl-2 translocations can be found over a roughly 30 kb
region in the 3’ UTR of the bcl-2 gene, 50% occur in a 175 bp zone termed
the major breakpoint region (MBR), 13% occur in the 105 bp intermediate

7
Figure 1-3. The bcl-1 translocation. Chromosomes 14 and 11 are
represented as black and red bars, respectively, and diagrammed telomeric
to centromeric, left to right. The immunoglobulin heavy chain locus on
chromosome 14 is organized into V, D, J, and constant segments, with
interspersed enhancers as described in the text. The strong immunoglobulin
heavy chain µ enhancer is denoted E µ, and the sections 3’ to E µ have been
omitted for simplicity. Double-strand breakages at D and J segments on
chromosome 14 join with a double-strand breakage on chromosome 11,
placing E µ upstream of the cyclin D1 oncogene, CCND1. This induces
overexpression of cyclin D1 protein, which is thought to contribute to
lymphomagenesis. The bcl-2 translocation follows the same scheme, except
involving the bcl-2 gene on chromosome 18.

8
Figure 1-3 continued

E µ
chr. 14
chr. 11 cyclin D1 gene
der. 11
der. 14 cyclin D1 gene E µ
E µ

V
H

chr. 14
D
H
J
H

chr. 11 cyclin D1 gene
double-strand breakages
misjoining
oncogene overexpression

9
cluster region (icr) about 18 kb centromeric to the MBR, and 5% occur in the
561 bp minor cluster region (mcr) about 11 kb centromeric to the icr
(Weinberg et al., 2007). As shall be demonstrated in chapter 2, these cluster
regions are highly unique in size and distribution, with important implications
for the mechanism(s) of double-strand breakage.
Mechanisms of double-strand breakage at the immunoglobulin
heavy chain locus
Physiologically, lymphocytes undergo a developmentally-regulated
program of chromosomal deletions to generate diversity in the antibodies –
also termed B-cell receptors (BCR) or immunoglobulins (Ig) – and T-cell
receptors (TCR) which recognize foreign antigens as part of the adaptive
immune response. These deletion events involve the same two steps as
rearrangements in general – double-strand breakage and joining by NHEJ.
Understanding the mechanisms of these physiologic double-strand
breakages provides insight into the mechanisms of the non-physiologic
double-strand breakages that occur in lymphoid cancers.
Structurally, antibodies and TCRs are divided into variable domains,
which sense and confer specificity to the antigen, and constant domains,
which are responsible for the receptor’s other functions (Fig. 1-4). Antibodies
are composed of either two heavy chains and two κ light chains, or two
heavy chains and two λ light chains. TCRs are composed of either two
TCR α chains and two TCR β chains, or one TCR γ chain and one TCR δ chain.

10
Figure 1-4. Structure of an immunoglobulin and its relation to V(D)J
recombination. (A) In this schematic of a rearranged heavy chain locus, the
V, D, J, and constant regions are color-coded to correspond to the amino
acids they encode in the typical Y-shaped structure of an antibody shown in
panels B-E. Arrows and zig-zag lines indicate the V(D)J-rearranged
junctions, which are a major source of variability among immunoglobulins.
These arrows correspond to those in panel E. Antibodies are composed of
four polypeptides: two heavy chains, outlined by dashed red lines in (B); and
two light chains, outlined by dashed red lines in (C). (D) The variable
domains of an antibody correspond to the rearranged V(D)J exons in the
heavy and light chains. The constant domains are responsible for the
effector functions of the antibody. (E) This enlargement of the antigen
receptor binding pocket shows that the junction regions encode many of the
amino acids which confer specificity to the antigen. The inherent variability of
the NHEJ process and the many combinations of V, D, and J segments
which can be used contribute greatly to antibody diversity. T-cell receptors
have a different shape but are organized similarly. The structure was imaged
in Rasmol from coordinates by (Padlan, 1994).

11
Figure 1-4 continued

heavy
chains
light
chains
A.
B. C.
D. E.
constant
domains
variable
domains

12
Each chain forms part of the variable domain and part of the constant
domain, and is located at a particular genomic locus – 14q32.3 for the heavy
chain, 2p12 for κ, 22q11 for λ, 14q11-12 for α, 7q32-33 for β, 7p15 for γ, and
14q11-12 for δ. The variable domains of each chain are encoded in a single
exon created using a process called V(D)J recombination.
During V(D)J recombination, the RAG complex makes double-strand
breaks at recombination signal sequences (RSS) adjacent to receptor gene
segments, and NHEJ joins the gene segments together to form the variable
domain exon of the receptor (Bassing et al., 2002; Dudley et al., 2005b;
Schatz, 2004; Tonegawa, 1983b).
The consensus sequence for RSS is CACAGTG (12/23 bp spacer)
ACAAAAACC (Lieber et al., 2004). The CACAGTG or comparable sequence
is termed the “heptamer” while the ACAAAAACC or comparable sequence is
termed the “nonamer.” Mechanistically, a RAG complex first binds an RSS
and makes a nick 5’ of the heptamer (Fig. 1-5). After synapsis with a partner
RSS, it uses the free hydroxyl group at the nick to attack the opposite strand,
thereby creating a double-strand break – with one hairpin DNA end and one
blunt DNA end – at each RSS. For synapsis to occur, an RSS having a 12
spacer, termed a “12 signal,” must pair with an RSS having a 23 spacer,
termed a “23 signal.” This is termed the “12-23 rule.”

13
Figure 1-5. Molecular mechanism of V(D)J recombination. V, D, and J
segments are flanked by sequence elements called recombination signal
sequences (RSS), represented as triangles. The consensus for RSS is
CACAGTG (12/23 spacer) ACAAAAACC. The RAG complex, composed of
RAG-1, RAG-2, and HMG-1, binds these signals, nicks the position 5’ to
CAC, and after synapsis of a 12 signal (12-spacer RSS) and a 23 signal (23-
spacer RSS), uses the free 3’ hydroxyls to attack the opposite strands,
thereby hairpinning both DNA ends 5’ to the signals. The ends containing
the RSS, called the signal ends, are tightly bound by the RAG complex,
which must be degraded before those ends can be joined without loss by the
ligase IV:XRCC4:XLF complex into the signal joint. The hairpin ends, also
called the coding ends, are bound by Ku, which recruits DNA protein kinase
catalytic subunit (DNA-PK
CS
). The hairpins themselves are sufficient to
activate DNA-PK
CS
, causing it to autophosphorylate and recruit and
phosphorylate Artemis. Binding to activated DNA-PK
CS
activates the hairpin-
opening activity of Artemis, with a preference for the position 2 nt 3’ to the
hairpin tip. Artemis may also recede the end to a limited degree, usually less
than 8 nt. Terminal deoxynucleotidyl transferase (TdT) adds nucleotides
template-independently to the coding ends and the ligase IV:XRCC4:XLF
complex joins them together into the coding joint.

14
Figure 1-5 continued

15
After double-strand breakage, the blunt ends containing the RSS, termed
signal ends, are later joined by ligase IV without loss to form a signal joint.
The hairpin ends, also called coding ends because they form the coding
sequence of the immune receptor, are opened and endonucleolytically
recessed by Artemis:DNA-PK
CS
. Recessing generally follows a distribution
which falls with distance from the free end, and is usually less than eight
nucleotides in human cells (Gauss and Lieber, 1996a). Terminal
deoxynucleotidyl transferase (TdT), polymerase µ, and/or polymerase λ add
nucleotides template-independently and the ligase IV:XRCC4:XLF complex
joins them together.
There are some important caveats to this simplified description. In
vitro, purified, truncated, recombinant RAG complexes can synapse two 12
signals, though not to the same degree as a 12 and a 23 signal (Yu et al.,
2004c). Furthermore, the minimal motif for RAG nicking in vitro appears to
be CAC, albeit at a lower efficiency compared to CACA and much lower
compared to a consensus heptamer-nonamer. In vivo, this minimal CAC
motif is present in certain translocation junctions from lymphoid tumors
(Marculescu et al., 2002b; Raghavan et al., 2001a). Also, coding ends
occasionally recombine with signal ends into what are termed “hybrid joints,”
at about 5% the rate of coding or signal joints, as demonstrated ex vivo in
pre-B cell lines transfected with plasmids bearing RSS (Gauss and Lieber,
1993).

16
Double-strand breakage by the RAG complex appears to be the most
critical and controlled step of the V(D)J recombination process, while NHEJ
just cleans up in the aftermath. It is thought that lymphocytes limit the V(D)J
recombination process to specific antigen receptor loci in a lineage-specific,
stage-specific, locus-specific, and allele-specific manner by restricting the
activity of RAG complexes to certain regions at certain times (Jung et al.,
2006). This is termed the “accessibility” hypothesis, and recent data
suggests that it may be due in part to RAG complexes binding to
trimethylated histone H3, lysine 4 (trimethyl-H3K4), which is heavily enriched
at the antigen receptor loci of pro-B/pre-B and pro-T/pre-T cells (Liu et al.,
2007; Matthews et al., 2007). RAGs are also regulated according to the cell
cycle. Because RAG complex activity is high during the G
0
and G
1
phases
and low during the G
2
, S, and M phases, it is thought that V(D)J
recombination occurs mainly during G
0
and G
1
(Schlissel et al., 1993). This
may be mediated by destruction of RAG2 before S phase by ubiqutin-
dependent degradation (Jiang et al., 2005).
As a prototypical example of a V(D)J locus, the immunoglobulin heavy
chain locus is organized, telomerically to centromerically, into variable (V)
segments; diversity (D) segments; joining (J) segments; the µ intronic
enhancer (E µ); and a number of constant regions and enhancers (C µ, C δ,
C γ3, C γ1, ψC ε, C α1, E α1, ψC γ, C γ2, C γ4, C ε1, C α2, E α2) (Honjo et al.,
2004) (Fig. 1-6). Each V segment is preceded by a promoter and has a 23

17
Figure 1-6. Sequence of immunoglobulin heavy chain rearrangement. The
1,237,987 bp immunoglobulin heavy chain locus on chromosome 14 is
organized into 123 V segments spanning 882,443 bp; 26 D segments
spanning 53,617 bp; 6 J segments spanning 2,263 bp; and 11 constant
regions spanning 278,105 bp and averaging 6 kb each; with 3 interspersed
enhancers. Only 39 of the Vs are functional, averaging 496 bp each, and
each is preceded by a promoter and followed by a 23 signal. Only 20 of the
Ds are functional, averaging 26 bp, and each is flanked on both sides by 12
signals. All 6 Js are functional, averaging 55 bp, are preceded by a 23
signal, and are followed by a splice donor site. Except for C δ, the constant
regions are preceded by switch regions of typically 1-10 kb of ~50 bp G-rich
repeats. Each constant region is preceded by a splice acceptor for the VDJ
exon and followed by a poly(A) signal. During the pro-B/pre-B stage, V(D)J
recombination joins a D and a J to form a D-J joint. Intervening Ds and Js
are deleted. Then, a V is joined to the D-J to form a single VDJ variable
domain exon. The first transcripts splice the VDJ exon to the µ or δ constant
region, producing antibodies of IgM and IgD class or isotype. At the mature
activated stage, a B cell can undergo a process called class switch
recombination (CSR), whereby the switch region preceding C µ is
recombined with the switch region for the new isotype. This results in an
antibody of a different constant domain but the same antigen specificity.

18
Figure 1-6 continued

19
signal on its centromeric end. Each D segment is flanked by 12 signals on
both sides, and each J segment has a 23 signal on its telomeric side. Each
constant region is followed by a poly(A) signal, and except for C δ, each is
preceded by a highly repetitive switch region.
During the pro-B/pre-B stage, the centromeric 12 signal of a D
segment recombines with the 23 signal of a J segment (Honjo et al., 2004).
The intervening D and J segments are lost along with the signal joint. The
telomeric 12 signal of the D segment then recombines with the 23 signal of a
V segment, and the intervening V and D segments are lost. Transcription
runs through the single exon of rearranged V, D, and J segments; through
any remaining J segments; through the intronic enhancer E µ; and through
the constant regions C µ and C δ. Alternative splicing from the VDJ exon to
the C µ or C δ results in the IgM and IgD isotypes of antibody heavy chains,
respectively. If a B cell fails to produce a functional heavy chain from one of
its two IgH loci, or a functional light chain from one of its four IgL loci, it will
die.
After producing a functional complete immunoglobulin receptor, a B
cell can be activated by foreign antigen and T cell stimulation. This causes
some activated B cells to undergo the class switch recombination (CSR)
process, whereby a double-strand break at the C µ switch region recombines
with that at another switch region. This results in expression of a different
constant region, i.e. production of an antibody with a different isotype. CSR

20
does not use the RAG complex to generate double-strand breaks and does
not appear to have precise sequence-specific breakage mechanism
(Lansford et al., 1998). Instead, transcription through the kilobases
extremely G-rich repeats in the switch regions – about 50% G on the
untranscribed strand – results in a structure called an R-loop, where the
transcribed RNA is hybridized to the transcribed DNA, and the untranscribed
DNA strand is displaced and single-stranded (Huang et al., 2006; Roy et al.,
2008; Yu et al., 2003) (Fig. 1-7). This structure is a perfect substrate for an
enzyme called activation-induced deaminase (AID), which is required for
CSR, is induced by B cell activation, and deaminates single-stranded
cytosines to uracils, especially in the context WRC, where W is the
nucleotide A or T and R is the nucleotide A or G (Bransteitter et al., 2003b;
Muramatsu et al., 2000; Pham et al., 2003a; Yu et al., 2004a; Yu et al.,
2005). Deletion of the switch regions in mice reduces the efficiency of both
R-looping and CSR (Huang et al., 2007; Luby et al., 2001). AID-deaminated
cytosines, in conjunction with the base excision repair (BER) pathway
involving uracil DNA glycosylase (UNG) and apurinic endonuclease (APE),
and possibly the mismatch repair (MMR) pathway and other DNA repair
proteins as well, are thought to generate the requisite double-strand
breakages in CSR (Stavnezer et al., 2008).
Based simply on joining partner, then, it is often possible to stage
lymphoid translocations. For instance, translocations to the IgH RSS such as

21

Figure 1-7. The R-loop:deaminase model for class switch recombination. A
putative CSR event to IgG3 is shown. Switch region RNA transcripts (green
lines) were paired with the C-rich DNA template strand to form an R-loop
structure. The entire displaced G-rich DNA strand and part of the C-rich DNA
at the edges of the R-loop are single-stranded, and therefore, serve as
targets for AID. AID deaminates C residues located in the single-stranded
region to convert them to uracil (U). UDG removes uracils in the DNA and
leaves behind an abasic (apyrimidinic) site, which is cleaved by APE. The
sum of nicks on both strands results in double-strand DNA breaks in the
switch region, which are repaired by the NHEJ pathway to complete CSR.
Coding regions (VDJ and constant region exons) are indicated by rectangles.
Vertical lines in the switch region indicate base pairing between the switch
transcript and the C-rich DNA template. Figure and caption from (Yu and
Lieber, 2003).

22
the bcl-1 and bcl-2 likely occur at the pro-B/pre-B stage, while translocations
to the IgH switch regions such as those involving c-myc or bcl-6 likely occur
at the activated B cell stage. Other times, the sequence character of the
breakpoint indicates the stage of the rearrangement. If breakpoints occur
next to RSS-like elements such as those at SCL-SIL deletions, it is likely that
the rearrangement occurred at the pro-B/pre-B or pro-T/pre-T stage.
Additional information can come from nucleotide additions at junctions.
Large numbers of added nucleotides are typical of TdT, which is mostly
specific to pro-T/pre-T or pro-B/pre-T stage cells, though low level expression
of TdT is observed in acute myeloid leukemias (AML) with t(8;21)(q22;q22)
AML1-ETO translocations. The clinical entity sometimes provides a limit as
well. It is unlikely that a translocation in a pre-B cell leukemia occurred at the
mature B stage because it is unlikely that the cell differentiated to that stage
and then de-differentiated back.
Unsurprisingly, the best-understood translocations from lymphoid
tumors are simply recapitulations of V(D)J recombination, involving non-
V(D)J loci. However, it shall be demonstrated that in humans these cases
are confined mostly to pre-T cell leukemias and represent a vast minority of
cases. Nothing about conventional V(D)J biochemistry explains the
breakpoint distributions at the bcl-1 MTC, bcl-2 MBR, bcl-2 icr, bcl-2 mcr, or
E2A clusters. We provide a plausible explanation for these clusters in
chapter 2.

23
DNA structure as a basis for translocation breakpoint clustering
Based on work published in this lab involving the bcl-2 MBR, three
years was spent pursuing the hypothesis that sequence elements at the bcl-1
MTC induce the formation of a non-B-DNA structure, which is directly
cleaved by the RAG complex (Raghavan et al., 2005a; Raghavan et al.,
2004a; Raghavan et al., 2005c; Raghavan et al., 2004b). However, as shall
be seen in chapter 3, this endeavor was fraught with inconsistencies and
technical difficulties and ultimately proved fruitless. Some technical
improvements were made, and further improvements are suggested.
Specific experimental details are omitted for brevity, and left in my
notebooks. While it is impossible to fully rule out DNA structure as a basis
for translocation breakpoint clustering, I provide some general thoughts on
the topic and suggestions for experimental proof.
Bisulfite as a probe of DNA structure
The clearest data supporting the notion of a non-B-DNA structure at
the bcl-1 MTC or bcl-2 MBR came from the structure-specific chemical probe
bisulfite. Bisulfite converts single-stranded, unmethylated cytosines into
uracils, which, in the context of DNA, can be cloned and sequenced (Gough
et al., 1986; Raghavan et al., 2004b; Raghavan et al., 2006; Yu et al., 2003).
While the structural reasons for the observed pattern of bisulfite reactivity at
cruciforms and R-loops are clear, the basis for reactivity at the centromeric
end of the bcl-1 MTC and the centromeric and telomeric ends of the bcl-2

24
MBR is unclear. To determine the true nature of bisulfite reactivity on
presumably double-stranded DNA, we performed bisulfite probing on
oligonucleotides containing sequences previously characterized by other
groups using other methods.
As shall be demonstrated in chapter 4, we observed a large difference
between different sequences of double-stranded oligonucleotides, allowing
us to speculate how, exactly, bisulfite might be reacting at the bcl-1 MTC and
bcl-2 MBR.

25
Chapter 2: Human chromosomal translocations at CpG
sites and a theoretical basis for their lineage and stage
specificity
Summary
The dinucleotide CpG only constitutes 1% of the human genome, yet
accounts for 25% of p53 point mutations in cancer. Here we analyze and
compare the sequences of over 1700 human chromosomal translocation and
rearrangement breakpoints and show that CpG also accounts for about 40%
of breakpoints from the most common translocation cluster regions in human
lymphoma, including the bcl-2 major breakpoint region, bcl-2 intermediate
and minor cluster regions, bcl-1 major translocation cluster, and E2A cluster
– all of which are specific to pro-B/pre-B cells. We do not see CpG hotspots
in rearrangements in lymphoid-myeloid progenitors, mature B cells, or T
cells. A lesion-specific double-strand breakage mechanism involving the
RAG complex acting at deaminated methyl-CpGs can explain the stage-
specificity and unique breakpoint distributions at these cluster regions.
Introduction
The dinucleotide sequence CG, also designated CpG, is statistically
underrepresented in vertebrate genomes, but is preserved in small and
functionally important regions around promoters termed CpG islands
(Antequera et al., 1990; Bestor, 2003; Bird, 1992; Takai and Jones, 2003).

26
CpG to TpG transitions account for a disproportionate number of mutations in
p53, BRCA1, BRCA2, p16, and Rb across a wide range of neoplasms
(Pfeifer, 2006). The common thread between these two observations is that
in vertebrates the cytosines of CpGs have a tendency to be methylated at the
C5 position, and after spontaneous deamination of the N4 amino group are
chemically indistinguishable from another endogenous DNA nucleoside,
thymine. The resulting T:G mismatches are not always repaired, leading to
point mutations seen in cancer and a genome-wide depletion over
evolutionary time referred to as “CpG suppression.” However, CpG has not
previously been described as a motif for chromosomal translocation.
Chromosomal translocations are common in many cancers and may
place enhancers upstream of oncogenes, disrupt tumor suppressor genes, or
create chimeric genes (Look, 1997a). The two basic mechanistic
requirements for a translocation are: (a) double-strand breaks on two
different chromosomes, and (b) joining of the wrong ends together. The vast
majority of recurrent translocations in cancer appear to use the
nonhomologous DNA end joining (NHEJ) pathway (or alternative enzymes
when NHEJ is not present) for the joining phase (Ferguson and Alt, 2001a;
Franco et al., 2008; Greaves and Wiemels, 2003; Lieber et al., 2006b;
Povirk, 2006; Yan et al., 2007). In contrast, the breakage mechanisms vary
from cancer to cancer and locus to locus and are largely undefined.

27
The best-characterized translocations in lymphomas are thought to be
examples of aberrant V(D)J recombination. Physiologically, lymphocytes
utilize V(D)J recombination to generate diversity in the antibodies and T-cell
receptors which recognize foreign antigens as part of the adaptive immune
response. During this process, the RAG complex makes double-strand
breaks at recombination signal sites (RSS) adjacent to receptor gene
segments, and NHEJ joins the gene segments together to form the variable
domain exon of the receptor (Bassing et al., 2002; Dudley et al., 2005a;
Schatz, 2004; Tonegawa, 1983a). Translocations can occur when the RAGs
accidentally create breaks on other chromosomes, and NHEJ then joins the
different chromosomal loci.
Though the consensus recognition sequence (termed recombination
signal sequence or RSS) for RAGs is CACAGTG (12 / 23 spacer)
ACAAAAACC, the minimal motif for RAG-mediated breakage appears to be
CAC (Lewis et al., 1997). If such a motif occurs near a site of recurrent
lymphocyte translocation, and especially if it also joins to the normal loci of
V(D)J recombination – 14q32.3, 2p12, 22q11 in B cells; 14q11-12, 7q32-33,
7p15 in T cells – it is termed a cryptic recombination signal site (cRSS). A
puzzling aspect to this is that CAC is extremely common throughout the
genome. Even more troubling is that cases involving cRSS seem to explain
only a minor fraction of events. The most common recurrent translocations
in non-Hodgkin lymphoma – the bcl-2 MBR translocation

28
t(14;18)(q32;q21.33), present in about a quarter of all non-Hodgkin’s
lymphomas, and the bcl-1 MTC translocation t(11;14)(q13.2;q32), present in
about a third of all mantle cell lymphomas – lack proper cryptic sites, as
demonstrated by sequence analysis and by ex vivo transfection experiments
(Marculescu et al., 2002c; Raghavan et al., 2001b; Raghavan et al., 2004a).
Here we show that breakpoints in translocations involving the bcl-2
MBR, bcl-2 mcr, bcl-2 icr, bcl-1 MTC, and the E2A loci are highly focused to
CpG hotspots and not CACs, featuring a bell-shaped distribution with the
apex at CpG. This pattern is in contrast to a second pattern of translocations
in closely-related lineages, such as BCR-ABL in chronic myelogenous
leukemia (CML); TEL-AML1 and MLL translocations in lymphoid-myeloid
hematopoietic stem cells (LM HSC); and c-myc or bcl-6 to immunoglobulin
heavy chain switch regions in mature B cells exhibit a scattered distribution
over many kilobases. And, as mentioned, a third pattern of translocation
breakpoints arises in T-cell rearrangements such as SCL-SIL, p16 deletions,
and LMO2 or TTG1 to T-cell receptor loci. These typically involve cryptic
RSS sites rather than CpGs, with an asymmetric “spike-and-single-sided tail”
distribution, where the largest breakpoint peak occurs immediately 5’ to CAC
and tails off in one direction, away from the pseudo-signal end and into the
“coding” end, typically for less than 15 bp.
We propose that the basis for CpG hotspots in the most common pro-
B/pre-B stage translocations is the same as that for the enrichment of CpG

29
transition mutations in cancer and CpG suppression during evolution –
namely, deamination of methylcytosines to thymines, coupled with poor
repair. The RAG complex is able to nick various intermediates during or prior
to repair of the T:G mismatch, resulting in a double-strand break. Finally,
aberrant joining to ends from a V(D)J recombination or other double-strand
break event leads to chromosomal translocation.
Materials and methods
Breakpoint References
Junctions and breakpoints were obtained from literature and GenBank
searches and aligned with sequence data from the UCSC genome browser
(Kent et al., 2002) using BLAT (Kent, 2002) and verified by eye.
Breakpoints from bcl-2 translocations, many in Fig. 2-1 A-C, are given
characters according to the reference, as follows: a (Bakhshi et al., 1985), b
(Cleary and Sklar, 1985), c (Tsujimoto et al., 1985a), d (Cleary et al., 1986b),
e (Cleary et al., 1986a), f (Bakhshi et al., 1987), g (Tsujimoto et al., 1988), h
(Crescenzi et al., 1988), i (Gauwerky et al., 1988), j (Ngan et al., 1989), k
(Cotter et al., 1990), l (Eick et al., 1990), m (Aster et al., 1992), n (Limpens et
al., 1991), o (Price et al., 1991), p (Wyatt et al., 1992), q (Liu et al., 1994), r
(Limpens et al., 1995), s (Ji et al., 1995), t (Kneba et al., 1995), u (Galoin et
al., 1996), v (Fuscoe et al., 1996), w (Dolken et al., 1996), x (Matolcsy et al.,
1997), y (Willis et al., 1997), z (Delage et al., 1997), 0 (Luthra et al., 1997), 1
(Wang et al., 1998), 2 (Akasaka et al., 1998), 3 (Rauzy et al., 1998), 4

30
(Soubeyran et al., 1999), 5 (Takacs et al., 2000), 6 (Jager et al., 2000), 7
(Hostein et al., 2001), 8 (Summers et al., 2001), 9 (Albinger-Hegyi et al.,
2002), + (Jenner et al., 2002), − (Marculescu et al., 2002a), # (Roulland et
al., 2003), = (Ladetto et al., 2003), % (Schmitt et al., 2006), $ (Young et al.,
2008).
Breakpoints from SCL-SIL deletions, many in Fig. 2-2 C-D, are
referenced as follows: c (Brown et al., 1990), f (Aplan et al., 1990), i (Bernard
et al., 1991), j (Aplan et al., 1992a), n (Bash et al., 1993), o (Breit et al.,
1993), p (Janssen et al., 1993).
Breakpoints from bcl-1 translocations, many in Fig. 2-2A, are
referenced as follows: a (Tsujimoto et al., 1985b), b (Rabbitts et al., 1988), c
(Meeker et al., 1991), d (Williams et al., 1993), e (Rimokh et al., 1994), f
(Galiegue-Zouitina et al., 1994), g (Segal et al., 1995), h (Willis et al., 1997), i
(Pott et al., 1998), j (Stamatopoulos et al., 1999), k (Degan et al., 2002), l
(Welzel et al., 2001a), m (Nyvold et al., 2007), n (unpublished breakpoints on
GenBank, DQ912821, DQ400340, DQ401134, DQ241758).
Breakpoints from E2A-PBX1 translocations, some shown in Fig. 2-2B,
are all from one reference: a (Wiemels et al., 2002a).
Breakpoints from TEL-AML1 translocations came from the following
references: (Andersen et al., 2001; Ford et al., 1998; McHale et al., 2003;
Pine et al., 2003; Wiemels et al., 2000; Wiemels and Greaves, 1999).

31
Breakpoints from MLL translocations were separated into primary
ALLs, primary AMLs, and therapy-related ALLs and AMLs, from the following
references: (Akao and Isobe, 2000; Bizarro et al., 2007; Borkhardt et al.,
1997; Chervinsky et al., 1995; Felix et al., 1999; Felix et al., 1997; Fu et al.,
2005; Gale et al., 1997; Gillert et al., 1999; Gu et al., 1994; Gu et al., 1992;
Hayette et al., 2000; Kuefer et al., 2003; Langer et al., 2003; Marschalek et
al., 1995; Megonigal et al., 1998; Meyer et al., 2006; Negrini et al., 1993;
Raffini et al., 2002; Reichel et al., 2001; Reichel et al., 1999; Reichel et al.,
1998; Super et al., 1997; Vieira et al., 2006; Wechsler et al., 2003) and
unpublished breakpoints on Genbank, DQ284496 to DQ284498, DQ249868,
AY839727 to AY839754, AY288756 to AY288777.
Breakpoints from AML1-ETO translocations came from the following
references: (Erickson et al., 1992; Wiemels et al., 2002b; Xiao et al., 2001;
Zhang et al., 2002).
Breakpoints from BCR-ABL translocations in CMLs came from the
following references: (Chissoe et al., 1995; de Klein et al., 1986; Heisterkamp
et al., 1985; Jeffs et al., 1998; Koduru et al., 1993; Litz et al., 1993; Mills et
al., 1992; Sowerby et al., 1993; Zhang et al., 1995).
Breakpoints from c-myc-IgH switch translocations in sporadic Burkitt
lymphomas came from the following references: (Apel et al., 1992; Battey et
al., 1983; Busch et al., 2004; Busch et al., 2007; Care et al., 1986; Dyson
and Rabbitts, 1985; Gauwerky et al., 1988; Gelmann et al., 1983; Muller et

32
al., 1995; Murphy et al., 1986; Saito et al., 1983; Showe et al., 1985; Wilda et
al., 2004; Wiman et al., 1984).
Breakpoints from bcl-6-IgH switch translocations came from the
following references: (Akasaka et al., 2000; Baron et al., 1993; Chaganti et
al., 1998; Chen et al., 2006; Kerckaert et al., 1993; Nakamura et al., 1997;
Sanchez-Izquierdo et al., 2001; Schwindt et al., 2006; Yang et al., 2006; Ye
et al., 1995; Yoshida et al., 1999).
Breakpoints from lymphoid p16 deletions came from the following
references: (Cayuela et al., 1997; Duro et al., 1996; Kitagawa et al., 2002).
Breakpoints from SCL translocations came from the following
references: (Aplan et al., 1992b; Begley et al., 1989; Bernard et al., 1992;
Bernard et al., 1993; Bernard et al., 1990; Chen et al., 1990a; Chen et al.,
1990b; Fitzgerald et al., 1991; Jonsson et al., 1991; Xia et al., 1992).
Breakpoints from LMO2-TCR translocations came from the following
references: (Boehm et al., 1988b; Cheng et al., 1990; Dik et al., 2007;
Fischer et al., 2007; Garcia et al., 1991; Marculescu et al., 2002b; Van
Vlierberghe et al., 2006; Yoffe et al., 1989) and unpublished breakpoints on
GenBank, EF450258, EF450768.
Breakpoints from HOX11-TCR translocations came from the following
references: (Kagan et al., 1989; Kagan et al., 1994; Lu et al., 1990; Park et
al., 1992; Salvati et al., 1999; Zutter et al., 1990).

33
Breakpoints from TTG1-TCR translocations came from the following
references: (Boehm et al., 1988a; Boehm et al., 1991; McGuire et al., 1989).
Statistical methods
A breakpoint “at” a CpG is defined as being at any of the three sites 5’
to the C, between the C and G, or 3’ to the G, i.e.
T CG, C
T G, or CG
T .
“Distance” to a motif is calculated as the number of nucleotides between the
breakpoint and the closest motif site.
The goal of using statistics was to establish the proximity of
breakpoints to CpGs. The first and most important test is an intuitive eyeball
evaluation of the data. After that, p-values are introduced to reduce
concerns about optical illusions. It is important to note that p-values
themselves are single numbers funneled down from many complicated
factors, and unfair, unnatural, and/or ad hoc treatment of these factors can
bias p-values.
The first question is, what are some natural ways to measure whether
breakpoints occur next to CpGs? One way is to count the number of
breakpoints occurring at CpGs, the number not occurring at CpGs, and then
compare this to “random.” Another way is to measure the distances between
breakpoints and the closest CpGs, and then compare this to “random.” But
what is “random”? “Random” might be the distribution of breakpoints we
would see if we removed all the CpGs in those regions from a population of
humans and waited a hundred years for them to develop lymphomas.

34
Without such an experiment, the p-values will all be theoretical. Most simply,
“random” is depicted as equal breakage from the most 5’ to the most 3’
breakpoint.
This is not a difficult case to make for the bcl-2 and bcl-1
translocations, which place one of the strongest enhancers next to but not
within their respective oncogenes. Breakpoints are known to occur over tens
or hundreds of kilobases of apparently empty sequence and all result in
essentially the same tumor. However, this assumes that breakpoints do not
naturally cluster around CpG-rich regions for other reasons. To deal with this
issue we further analyze the subclusters. Uniform breakage as a
representation of “random” is not as good an assumption in translocations
which result in fusion genes, but still fair enough. In these cases we worry
about selection issues such as stability of the fusion mRNA, critical protein
binding regions, cryptic splice sites, etc. Top and bottom strands were
treated separately and calculated together and separately, but because all
the examined motifs were symmetrical except CAC, this does not make
much difference.
Given this definition of “random,” the calculation is trivial. The
binomial statistic groups breakpoints into those which occur at CpG and
those that do not (i.e. 1 or 0). The p-value is the probability that one would
see the observed number of events at CpG, or more given random chance,
and is calculated as:

35
p-value =
∑
=
−
−
−
N
k x
x N x
p) (1 p
x)! (N x!
N!
(standard binomial distribution)
N = total number of events
k = observed number of events at motif
p = random probability of an event occurring at the motif, calculated as
the number of sites at motifs divided by the total number of sites
from the most 5’ to the most 3’ breakpoint
A significant p-value indicates that breakpoints occur at CpGs more often
than by random chance. The main weakness of this calculation is the
definition of random chance. Another potential weakness is that breakage
might occur at CpGs but get recessed by Artemis:DNA PK
CS
, in which case
these CpG-caused breakpoints would not score in the binomial statistic.
To deal with this issue, we analyzed the distributions of distance from
the breakpoint to the closest CpG, comparing the actual breakpoints with the
simulated “random”/uniform breakpoints using two different statistical tests.
“Distance” to a CpG was calculated as the number of nucleotides between
the breakpoint and the closest CpG site. Distance distributions for uniformly
or “randomly” distributed breakpoints were generated by traversing the
region from the most 5’ to the most 3’ breakpoint, treating each strand
independently, and calculating the distance to the closest CpG site for each
site along the way.
The Student’s t-test compares the averages of the distributions,
relative to a shared standard deviation. First, all the distances were

36
subjected to a standard Box-Cox log(x+1) transformation to correct for the
non-normality of the distributions. Next, the average and standard deviation
were calculated for each set of breakpoints. Finally, the t statistic was
calculated as:
t-value =
2 1
2 1
x x
s
x x
−
−
(Student’s t-test)
x
1
= average of log(x+1)-transformed actual breakpoint distances
x
2
= average of log(x+1)-transformed random breakpoint distances
s
⎯x1- ⎯x2
=
() ( )
⎟
⎟
⎠
⎞
⎜
⎜
⎝
⎛
+
− +
− + −
2 1 2 1
2
2 2
2
1 1
1 1
2
1 1
n n n n
s n s n

n
1
= number of actual breakpoints
n
2
= number of random breakpoints
s
1
= standard deviation of actual breakpoints
s
2
= standard deviation of random breakpoints
Using Excel, the t-value was converted to a p-value, which is the probability
of seeing the observed average distance to CpG or smaller, given random
chance. A significant p-value indicates that, on average, breakpoints are
significantly closer to CpGs than would be expected by random chance.
However, the t-test works best when comparing two sets of data which are
normally-distributed.
Therefore, we used the Mann-Whitney U-test, which is independent of
the shape of each individual distribution, to compare how actual breakpoints
and random breakpoints are distributed relative to one another. First, the

37
distances from both groups were combined and sorted from least to greatest.
Next, each distance value was ranked from 1 to n
1
+n
2
, where n
1
is the
number of actual breakpoints and n
2
is the number of random breakpoints.
In cases where there were many ranks for a particular distance, the ranks
were averaged. Going back to the original groupings, the ranks for each
distance value in the actual breakpoint group were summed into the value
R
1
. The U statistic was calculated as:
U =
()
2
1
1 1
1
+
−
n n
R
For large samples, a z-value can be derived as:
z =
v
v
m U
σ
−

m
v
=
2
2 1
n n ⋅

σ
v
=
()
12
1
2 1 2 1
+ +n n n n

Excel was used to derive p-values from z-values. The p-value is the
probability of seeing the observed sum rank or lower, given random chance.
A significant p-value indicates that the actual breakpoints are nonrandomly
ordered closer to CpG than the random breakpoints.
At least in our set, these distance statistics behave more or less the
same and have various strengths and limitations. They work best when the
motif being tested is >1% of the breakpoint window and spread evenly.
Otherwise, the distance distributions can be affected by local fluctuations of

38
motif density. If translocation to, say, a CpG island more often results in
deregulation of an oncogene or tumor suppressor, then breakpoints may test
significantly for CpG proximity even though frequency versus distance
graphs show they are mostly far away. When breakpoints are clustered
tightly, neighboring motifs may be pulled along, resulting in statistical
significance, depending on the motif density. In these situations, significant
p-values still indicate that breakpoints are closer to a motif than they should
be given random chance – which is true – but there can be a number of
different reasons for proximity, and we are mostly interested in breakage
directly at those motifs. These calculations are necessarily insensitive to
sequences outside the breakpoint window. If a given window is very dense
for a specific motif and breakpoints gravitate towards but do not occur at the
motif, these algorithms will not detect it. Widening the window will solve this,
but makes the implicit and dangerous assumption that breakpoints could
occur outside the window and still result in the same tumor.
There are a number of things one can do to increase the signal-to-
noise ratio, such as using a binomial with a window of 5 bp, making an upper
limit of 20 bp for distance measurements; while some of these may have a
logical basis, we felt it best to avoid such ad hoc tweaking and arbitrary
cutoffs.

39
RAG Nicking Assays
GST-core RAGs and oligonucleotides were purified as described
previously (Yu et al., 2002). Substrates were prepared by annealing 1.5-fold
excess of cold oligonucleotide(s) to hot oligonucleotide as previously
described. The nicking assay was done as described previously, except that
no competitor oligonucleotide was added. The oligonucleotides used for the
various substrates are as follows, listed as labeled top strand-bottom strand:
12 signal, AT188-AT189; CG duplex from Fig. 2-6A, AT190-AT191; TA
duplex, AT216-AT217; TG mismatch, AT216-AT191; TT mismatch, AT216-
AT193; CT mismatch, AT190-AT193; CA mismatch, AT190-AT217; CG
duplex from Fig. 2-6C, AT191-AT190; nick, AT191-AT218 & AT244; one-
nucleotide gap, AT191-AT218 & AT240; two-nucleotide gap, AT191-AT235 &
AT244; three-nucleotide gap, AT191-AT235 & AT240; one-nucleotide loop,
AT191-AT192; one-nucleotide flap, AT191-AT244 & AT245; three-nucleotide
flap, AT191-AT243 & AT245. The oligonucleotide sequences are as follows:
AT188,
ggtattacgatgcttggactggttattatacccacagtgtctcagagtccaacaaaaacccatccctggg;
AT189,
cccagggatgggtttttgttggactctgagacactgtgggtataataaccagtccaagcatcgtaatacc;
AT190,
ggtattacgatgcttggactggttattataccgtcagtctctcagagtccaagatttacgcatccctggg;
AT191,

40
cccagggatgcgtaaatcttggactctgagagactgacggtataataaccagtccaagcatcgtaatacc;
AT193,
cccagggatgcgtaaatcttggactctgagagactgactgtataataaccagtccaagcatcgtaatacc;
AT216,
ggtattacgatgcttggactggttattatactgtcagtctctcagagtccaagatttacgcatccctggg;
AT217,
cccagggatgcgtaaatcttggactctgagagactgacagtataataaccagtccaagcatcgtaatacc;
AT218, ggtattacgatgcttggactggttattatac; AT235, ggtattacgatgcttggactggttattat;
AT240, gtcagtctctcagagtccaagatttacgcatccctggg; AT243,
accgtcagtctctcagagtccaagatttacgcatccctggg; AT244,
cgtcagtctctcagagtccaagatttacgcatccctggg; AT245,
ggtattacgatgcttggactggttattatacc.
Methylation Analysis
Genomic DNA from FACS-sorted pre-B cells was a generous gift of
Dr. Marcus Muschen. Bisulfite treatment and DNA purification were done
using the Zymo Research EZ DNA Methylation Kit, according to
manufacturer’s directions. PCR was done using AT301,
ggatgttattggttattgaggagt, and AT302, tcaaaaatctaatcattctattcccta, for bcl-2;
and AT297, gggttgtttttaagttttggttatt, and AT298, ccaataccccaaattccctta, for
bcl-1. Conditions were 40 cycles of 95° C, 0:30; 56° C, 0:45; 72° C, 1:00;
with an initial denaturation 95° C, 2:00 and a final extension 72° C, 5:00.

41
Results
Overview of Translocation Breakpoint Distribution and CpG Sites
The first clue to determining the reasons for a given breakage site is
its precise location. Because free DNA ends are often degraded prior to
joining, it can be difficult to know, down to the specific nucleotide, where the
initial break occurred. The best approximation one can make is the
breakpoint, or the exact position in the translocated allele where the
sequence changes significantly compared to that before the translocation.
We first plotted 695 sequence-resolution breakpoints from pro-B/pre-B
stage translocations on bcl-2, bcl-1, and E2A (Table 2-1, Fig. 2-1, Fig. 2-2).
A distinctive pattern of breakpoint clustering around CpGs quickly emerged.
Breakpoints on bcl-2 span 30,435 bp, including 355 CpGs accounting for
1,061 possible breakage sites within or immediately adjacent to CpGs.
Breakpoints on bcl-1 span 126,216 bp, 2,082 CpGs, and 6,188 sites; and
E2A breakpoints span 2,515 bp, 43 CpGs, and 129 sites. In all three cases,
if breakpoints were scattered randomly throughout the region, less than 6%
would occur at CpGs. However, in sequenced translocations from actual
patients, breakpoints at CpGs account for a striking 43% of events at bcl-2,
35% of events at bcl-1, and 53% of events at E2A, and this does not include
breakpoints that are even one base pair away from a CpG. Next, we asked
whether breakpoints are, on average, closer to CpGs than if they were
uniformly distributed from the most 5’ to the most 3’ sequenced

42
Table 2-1. Summary of translocation breakpoint features. Numbers of
breakpoints are listed for rearrangements and clinical entities analyzed in the
text, followed by estimated stage during which the rearrangement occurs and
the apparent types of breakage involved. “Random”-type breakage does not
appear attached to a specific motif, is more regional in nature, and is
presumed to be due to ionizing radiation, reactive oxygen species, or some
other relatively sequence-independent mechanism. Abbreviations are as
follows: follicular lymphoma (FL), diffuse large B-cell lymphoma (DLBL),
mantle cell lymphoma (MCL), acute lymphoblastic leukemia (ALL), acute
myeloid leukemia (AML), chronic myelogenous leukemia (CML), Burkitt
lymphoma (BL), chronic lymphocytic leukemia (CLL), lymphoid-myeloid
hematopoietic stem cell (LM HSC), class switch recombination (CSR).

43
Table 2-1 continued
Name Clinical Entity Rearrangement Breakpoints Stage of
Rearrangement
Breakage Types
bcl-2 FL and DLBL t(14;18)(q32;q21) 551 pro-B/pre-B CpG, V(D)J
bcl-1 MCL t(11;14)(q13;q32) 114 pro-B/pre-B CpG, V(D)J
E2A-PBX1 pre-B / B ALL t(1;19)(q23;p13) 30 pro-B/pre-B CpG, random
TEL-AML1 pre-B / B ALL t(12;21)(p12;q22) 53 LM HSC random
MLL primary ALL t(4;11)(q21;q23)
t(9;11)(p22;q23)
other
254
17
20
LM HSC random
MLL primary AML t(4;11)(q21;q23)
t(9;11)(p22;q23)
other
11
6
7
LM HSC random
MLL therapy-related ALL
and AML
t(4;11)(q21;q23)
t(9;11)(p22;q23)
t(11;16)(q23;p13)
other
5
12
11
7
LM HSC random,
topoisomerase II-
related
AML1-ETO AML t(8;21)(q22;q22) 133 LM HSC random
BCR-ABL CML t(9;22)(q34;q11) 35 LM HSC random
c-myc sporadic BL t(8;14)(q24;q32) 125 activated B CSR, random
bcl-6 mature B lymphomas t(3;14)(q27;q32) 38 activated B CSR, random
SCL-SIL pre-T / T ALL del(1)(p32) 209 pro-T/pre-T V(D)J
p16 deletion pre-T / T ALL
T-CLL
pre-B / B ALL
del(9)(p21) 26
2
13
pro-T/pre-T V(D)J
SCL pre-T / T ALL

pre-B / B ALL
t(1;14)(p32;q11)
t(1;7)(p32;q34)
t(1;3)(p32;q21)
t(1;14)(p32;q11)
14
2
2
1
pro-T/pre-T V(D)J, random
LMO2 pre-T / T ALL t(11;14)(p13;q11)
t(7;11)(q34;p13)
del(11)(p12p13)
27
2
2
pro-T/pre-T V(D)J, random
HOX11 pre-T / T ALL t(10;14)(q24;q11) 13 pro-T/pre-T V(D)J
TTG1 pre-T / T ALL t(11;14)(p15;q11) 6 pro-T/pre-T V(D)J
total 1748

44
Figure 2-1. Sequence-level breakpoint diagrams for bcl-2 clusters.
Breakpoint distributions are plotted for the (A) bcl-2 MBR, (B) bcl-2 icr, and
(C) bcl-2 mcr. Each breakpoint is represented as an individual character
adjoining the breakpoint site, with the top strand sequences running
telomeric to centromeric, der(14) breakpoints above, and der(18) breakpoints
below. Different characters denote different references, as listed in the
materials and methods.

45
Figure 2-1 continued

A.
B.
C.

46
Figure 2-2. Sequence-level breakpoint diagrams for the bcl-1 MTC, E2A,
main SCL, and SIL clusters. Breakpoint distributions are plotted for the (A)
bcl-1 MTC, running telomeric to centromeric; (B) E2A cluster, running
centromeric to telomeric; (C) main SCL cluster, running telomeric to
centromeric; and (D) SIL cluster, running telomeric to centromeric. Each
breakpoint is represented as an individual character adjoining the breakpoint
site. Different characters denote different references, as listed in the
materials and methods. Note the different shapes of the distributions in
panels C and D versus panel A from Fig. 2-1. Breakpoints in the bcl-2 MBR
fall into bell-shaped distributions, with peaks at CpGs and a decline of
breakpoints on both sides. Breakpoints at the main SCL and SIL peaks are
spiked at CAC, and tail off only in the direction 5' of the CAC. This implies
that different mechanism(s) operate at the MBR versus SCL and SIL.

47
Figure 2-2 continued

A.
B.
C.
D.

48
breakpoints (i.e., a “random” distribution throughout the region). On average,
bcl-2 breakpoints occur 7.2 bp away from the closest CpG, versus 59.1 bp if
randomly distributed. The bcl-1 breakpoints average 3.1 bp away, versus
45.3 bp. E2A breakpoints average 5.1 bp away, versus 29.1 bp.
These findings are best illustrated by plots comparing the distributions
of breakpoints at various distance intervals (Fig. 2-3). In actual lymphomas
involving bcl-2, bcl-1, and E2A, the highest percentage of breakpoints occurs
at a distance of 0 bp, which, strikingly, is directly at a CpG. With increasing
distance from CpG, the proportion of breakpoints falls away. In stark
contrast, if the breakpoints were randomly distributed, the proportion would
increase with each distance interval. A similar analysis done for the motif
CAC shows that breakpoints from actual patients are distributed similarly to
random with respect to CAC.
Interestingly, a large proportion of breakpoints in all these
translocations occur within small clusters with high CpG density. In order to
demonstrate that the CpGs themselves are hotpots, and that the proximity of
breakpoints to CpGs goes beyond coincidentally increased CpG density
within these cluster regions, we analyzed each cluster independently using
three statistical measurements (Table 2-2). The binomial statistic groups
breakpoints into those at CpG versus those that are not, then compares this
to the numbers that would be expected if breakpoints were randomly
distributed. We used two statistical measures of breakpoint proximity to

49
Figure 2-3. Distributions of breakpoints at various distances from CpG or
CAC in representative translocations. Breakpoint frequency at distances of 0
bp, 1 to 2 bp, 3 to 4 bp, 5 to 8 bp, and greater than 8 bp from either CpG or
CAC are charted. The distribution for actual lymphoma or leukemia
breakpoints is in dark blue, while that for a random distribution between the
farthest breakpoints is in light blue. If the dark blue and light blue bars
parallel one another, then the patient breakpoints appear random in their
distribution relative to the specified motif. However, when they follow
opposite trends, i.e. the light blue bars rise with increasing distance from the
specified motif while the dark blue bars fall, then the breakage process
appears to concentrate around the motif. Breakpoints at bcl-2, bcl-1, and
E2A appear to concentrate around CpG (shown in panels A-C) but not CAC
(shown in panels D-F). By contrast, distributions for translocations not
occurring in pro-B/pre-B cells do not cluster around CpG (four examples in
panels G-J), and SCL-SIL and lymphoid ∆p16 breakpoints are highly focused
to CAC (shown in panels K-L).

50
Figure 2-3 continued
A. bcl-2, CpG
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
0 1 to 2 3 to 4 5 to 8 >8
distance (bp)
frequency
D. bcl-2, CAC
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
0 1 to 2 3 to 4 5 to 8 >8
distance (bp)
frequency
B. bcl-1, CpG
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0 1 to 2 3 to 4 5 to 8 >8
distance (bp)
frequency
E. bcl-1, CAC
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
0 1 to 2 3 to 4 5 to 8 >8
distance (bp)
frequency
C. E2A, CpG
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0 1 to 2 3 to 4 5 to 8 >8
distance (bp)
frequency
F. E2A, CAC
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0 1 to 2 3 to 4 5 to 8 >8
distance (bp)
frequency

█ actual leukemia or lymphoma breakpoints
█ breakpoints uniformly distributed from most 5’ to most 3’ actual breakpoints

51
Figure 2-3 continued
G. TEL, CpG
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
0 1 to 2 3 to 4 5 to 8 >8
distance (bp)
frequency
H. MLL from ALLs, CpG
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
1
0 1 to 2 3 to 4 5 to 8 >8
distance (bp)
frequency
I. BCR from CMLs, CpG
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0 1 to 2 3 to 4 5 to 8 >8
distance (bp)
frequency
J. c-myc, CpG
0
0.1
0.2
0.3
0.4
0 1 to 2 3 to 4 5 to 8 >8
distance (bp)
frequency
K. SCL-SIL, CAC
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
0 1 to 2 3 to 4 5 to 8 >8
distance (bp)
frequency
L. ∆p16 from lymphoid cell lines, CAC
0
0.1
0.2
0.3
0.4
0.5
0.6
0.7
0.8
0.9
0 1 to 2 3 to 4 5 to 8 >8
distance (bp)
frequency

█ actual leukemia or lymphoma breakpoints
█ breakpoints uniformly distributed from most 5’ to most 3’ actual breakpoints

52
Table 2-2. Statistical analysis of CpG and CAC proximity in various
translocations. Statistical measures of CpG and CAC proximity are
compared. Values related to CpG-type translocations in pre-B stage cells
are highlighted in pink. Values related to V(D)J-type translocations in
immature lymphocytes are highlighted in blue. p-values are calculated with
“random” defined as a uniform breakpoint distribution from the most 5’ to the
most 3’ breakpoint, and depend on the number of breakpoints as well as the
degree of clustering and the regional motif density, therefore precise values
should not be compared between different clusters. In highly-clustered
translocations, i.e. CpG-type or V(D)J-type, nearby motifs may score
significantly but not to the same level. Calculations are described in the
materials and methods.

53
Table 2-2 continued

Locus
Percentage at
CpG
Actual Average
Distance to CpG
(bp)
Random Average
Distance to CpG
(bp)
p-value at CpG
(binomial
distribution)
p-value proximity
to CpG
(Student’s t-test)
p-value proximity
to CpG (Mann-
Whitney U-test)
Percentage at
T CAC
Actual Average
Distance to
T CAC (bp)
Random Average
Distance to
T CAC (bp)
p-value at
T CAC
(binomial
distribution)
p-value proximity
to
T CAC
(Student’s t-test)
p-value proximity
to
T CAC (Mann-
Whitney U-test)
bcl-2 MBR 43% 4.39 11.2 1.8 ⋅ 10
-96
2.3 ⋅ 10
-50
1.2 ⋅ 10
-42
0.21% 21.2 22.1 1.0 0.45 0.27
bcl-2 icr 73% 0.818 21.5 1.5 ⋅ 10
-8
1.2 ⋅ 10
-12
3.4 ⋅ 10
-7
0% 51.4 52.1 1.0 0.79 0.49
bcl-2 mcr 74% 0.632 40.4 5.4 ⋅ 10
-18
2.3 ⋅ 10
-31
7.6 ⋅ 10
-13
0% 10.2 30.4 1.0 5.6 ⋅ 10
-5
5.6 ⋅ 10
-5
bcl-2 unclustered 19% 45.1 62.2 0.0019 0.0016 0.024 0% 44.0 37.9 1.0 0.79 0.82
bcl-1 MTC 37% 2.55 7.80 7.0 ⋅ 10
-9
1.8 ⋅ 10
-13
1.1 ⋅ 10
-12
0% 21.8 22.8 1.0 0.68 0.56
bcl-1 unclustered 20% 8.80 45.3 0.083 0.000480.00090 0% 38.3 32.8 1.0 0.64 0.69
E2A 63% 1.08 3.25 0.00011 0.000200.00052 0% 11.3 17.2 1.0 0.068 0.019
PBX1 0% 93.7 64.3 1.0 0.96 0.96 3.4% 30.1 37.0 0.37 0.21 0.27
TEL in TEL-AML1 1.9% 41.1 41.6 0.86 0.66 0.66 1.9% 38.5 33.4 0.63 0.61 0.55
AML1 in TEL-AML1 1.9% 61.7 61.7 0.85 0.49 0.41 0% 42.7 37.5 1.0 0.87 0.89
MLL in primary ALL 4.7% 79.0 84.9 0.16 0.65 0.58 2.0% 32.2 35.6 0.37 0.26 0.23
MLL in primary AML 0% 64.4 85.4 1.0 0.41 0.47 0% 23.4 33.9 1.0 0.12 0.094
AML1 in AML1-ETO 4.5% 48.6 46.7 0.45 0.61 0.70 0% 38.0 35.7 1.0 0.77 0.67
ETO in AML1-ETO 0% 103 88.9 1.0 0.99 0.99 1.5% 52.8 41.6 0.62 0.99 1.0
BCR in CML 0% 28.7 21.4 1.0 0.97 0.96 4.3% 26.5 26.5 0.27 0.35 0.43
ABL in CML 0% 63.0 50.3 1.0 0.91 0.89 2.9% 30.0 25.1 0.56 0.68 0.72
c-myc 17% 8.85 8.90 0.59 0.71 0.75 0% 39.4 38.7 1.0 0.33 0.36
bcl-6 19% 8.32 10.9 0.28 0.13 0.15 1.7% 41.2 57.3 0.41 0.027 0.019
SCL-SIL 0.48% 16.6 60.9 1.0 8.4 ⋅ 10
-15
5.9 ⋅ 10
-26
35% 2.16 37.3 7.6 ⋅ 10
-75
3.7 ⋅ 10
-185
8.7 ⋅ 10
-108
lymphoid ∆p16 7.3% 63.7 96.7 0.13 0.048 0.11 17% 11.3 46.0 1.5 ⋅ 10
-6
3.4 ⋅ 10
-20
4.4 ⋅ 10
-14
SCL translocations* 20% 32.1 53.5 0.014 0.00029 0.0012 25% 10.2 12.5 0.00070 0.015 0.040
LMO2* 9.7% 26.2 47.1 0.16 0.060 0.036 29% 4.58 16.3 4.8 ⋅ 10
-7
3.6 ⋅ 10
-12
1.6 ⋅ 10
-9

* Mixture of signal and coding joint events, therefore calculated for
T CAC and GTG
T ; see text for details.

54
CpG. The Student’s t-test compares the average distances to CpG between
actual lymphoma breakpoints and randomly distributed breakpoints. The
Mann-Whitney U-test compares how actual breakpoints and random
breakpoints are distributed relative to one another.
CpGs in the bcl-2 MBR, icr, and mcr are Hotspots for Translocation
Breakpoints
Three cluster regions have been described for bcl-2 translocations.
Fifty percent occur in the major breakpoint region (MBR); 13% in the
intermediate cluster region (icr) about 18 kb centromeric to the MBR; and 5%
in the minor cluster region (mcr) about 11 kb centromeric to the icr (Weinberg
et al., 2007). Of the bcl-2 translocations that were actually sequenced, 88%
occur in the 175 bp MBR, which is further divided into three breakpoint
peaks, and contains five CpGs (Fig. 2-1A). Strikingly, two CpGs are found at
the tallest spikes of peak I, one CpG at those of peak II, and the last two
CpGs at those of peak III. Of 487 breakpoints in the MBR, 208 (43%) occur
zero base pairs away from (directly at) one of these five CpGs (p<10
-95
).
The average MBR breakpoint is 4.4 bp from the nearest CpG, compared to
11.2 bp if randomly distributed (p<10
-41
).
At the 105 bp icr, there are only two CpGs, spaced 95 bp apart (Fig. 2-
1B). Eight of the 11 breakpoints occur at one of these two CpGs (p<10
-7
),
while the other three are less than 6 bp away. The large region between the

55
CpGs contains no breakpoints. Overall, they average less than one base pair
away (0.82 bp) from a CpG, versus 21.5 bp if randomly distributed (p< 10
-6
).
The 561 bp mcr contains six CpGs, and with the exception of two
CpGs adjacent to one another, they are spaced widely apart (Fig. 2-1C).
However, breakpoints are highly concentrated at the CpGs, while the
intervening regions are completely bare. Of the 19 collected breakpoints, 14
are located directly at CpGs (p<10
-17
), with an overall average 0.63 bp to the
nearest CpG, in contrast to 40.3 bp if randomly distributed (p<10
-12
). CpG is
also the most significant motif among the 27 breakpoints that are not
members of the three cluster regions (MBR, mcr, or icr). Thirteen, or roughly
half, are located less than 10 bp away from a CpG, including five right at
CpGs (p=0.002).
CpGs in the bcl-1 MTC are Hotspots for Translocation Breakpoints
In mantle cell lymphomas, 30% of translocations are detected within
the 150 bp major translocation cluster (MTC) (Bertoni et al., 2004), which
contains 7 CpGs (Fig. 2-2A). The MTC accounts for 104 of 114 breakpoints
in the collection, with 38 occurring at CpGs (p<10
-8
). Overall, the 104
breakpoints average a distance of 2.5 bp from the closest CpG, compared to
7.8 bp if distributed randomly over the 150 bp region (p<10
-11
). Notably,
even among the ten unclustered breakpoints, CpG is still the most significant
motif, averaging 8.8 bp from a CpG, in contrast to 45.3 bp if randomly
distributed (p=0.0009).

56
CpGs in E2A but not PBX1 in E2A-PBX1 Translocations are Breakpoint
Hotspots
Most of the E2A translocations occur within a 23 bp zone containing
two CpGs (Wiemels et al., 2002a) (Fig. 2-2B). Fifteen of the 24 collected
breakpoints occur at CpGs (p=0.0001), overall averaging 1.1 bp to the
closest CpG, versus 3.3 if randomly distributed (p=0.0005). Interestingly,
breakpoints on PBX1 do not have a tendency to occur at or close to CpGs
(p>0.9) and have only a slight tendency towards CAC (p=0.05), though not
directly at CAC (p>0.1).
CpG is the Most Significant Dinucleotide Motif, and is More Significant than
CAC for the bcl-2, bcl-1 and E2A
As controls, we looked at breakpoint proximity to all the other
dinucleotide motifs, six mononucleotide motifs,
T CAC, and
T CAC / GTG
T
(Table 2). While certain motifs actually have lower average distances than
CpG (e.g. the mononucleotide motif S (C or G, of which CpG is a subset)),
this is not significant when compared to random chance. Certain motifs
score significantly because they happen to occur next to CpGs. If a
breakpoint spike occurs at CpG and there happens to be a rare CAC motif a
few base pairs away, then CAC will score significantly – but not to the same
level as CpG, and never across different translocations or clusters. Among
the bcl-2 MBR, bcl-2 icr, bcl-2 mcr, bcl-1 MTC, and E2A breakpoint clusters,

57
and to a lesser degree the unclustered breakpoints, CpG is consistently the
most significant motif in our three statistical tests.
Initial Chromosomal Breakage Likely Occurs Directly at or Near CpGs
While an individual breakpoint may be our best guess as to where
initial breakage occurred, more precise information may be obtained by
examining both derivatives of a balanced translocation. In all translocations,
breakage of the first chromosome generates two free DNA ends, breakage of
the second creates two more, and degradation can occur before one end of
the first chromosome joins to one end of the second. When a translocation is
balanced, the remaining end of the first chromosome joins to the remaining
end of the second, with or without degradation. One of these derivative
chromosome junctions is termed “direct,” and the other is termed reciprocal.
Knowledge of both the direct and the reciprocal breakpoints on the
chromosome of interest provides a window within which the initial breakage
likely occurred.
Both direct and reciprocal breakpoints are available for 46 cases from
bcl-2, 23 cases from bcl-1, and six cases from E2A (Fig. 2-4). Of these, 32
cases from bcl-2, 15 cases from bcl-1, and four cases from E2A have one or
both breakpoints at CpGs. Importantly, one case from bcl-1 and nine cases
from bcl-2, including one case from the bcl-2 mcr, exhibit a breakage window
of zero; that is, there is no nucleotide loss from either junction. All ten are at

58

Figure 2-4. Sequence-level diagrams for balanced translocations. Breakpoint distributions for balanced
translocations are plotted running telomeric to centromeric for the (A) bcl-2 MBR, (B) bcl-2 mcr, and (C) bcl-1
MTC; and running centromeric to telomeric for the (D) E2A cluster. Each breakpoint is represented as an
individual character adjoining the breakpoint site, and has a partner breakpoint on the opposite strand. Different
characters denote different paired breakpoints.

A.
B.
C.
D.

59
CpGs, strongly supporting the notion that initial breakage occurs at or near
CpG edges.
Overview of Recurrent Translocations in Leukemias and Lymphomas
In order to determine whether CpG hotspots are a general
phenomenon, we sought to perform the same statistical analysis on the most
common rearrangements in lineage-related lymphomas and leukemias, with
two examples from each cell type or stage of development (Tables 1 and 2,
Fig. 2-3). Understanding the classification and frequencies of hematologic
malignancies is necessary to obtain a reasonably representative sample.
Knowing which cells from which a tumor arose has clinical importance,
but also tells us which oncogenic processes have occurred or are occurring;
therefore, hematopoietic tumors are classified by lineage and stage (Jaffe,
2001). Myeloid neoplasms are categorized as chronic myeloproliferative
diseases (CMPD), myelodysplastic/myeloproliferative diseases (MDS/MPD),
myelodysplastic syndromes (MDS), and acute myeloid leukemias (AML).
Lymphoid neoplasms are categorized as B-cell, T and NK-cell, and Hodgkin
lymphoma (HL); but for the purposes of this discussion will be divided into
immature (pre-B or pre-T stage) or mature malignancies.
Chronic myelogenous leukemia (CML) and acute myelogenous
leukemia (AML) are the primary myeloid neoplasms, with incidences about 1-
1.5/100,000 people per year and 3/100,000 people per year, respectively
(Jaffe, 2001). Ninety to ninety-five percent of CMLs feature the Philadelphia

60
chromosome, t(9;22)(q34;q11) BCR-ABL translocation, as do some AMLs
and ALLs. CMLs can progress to AMLs through a blast crisis. In AMLs, the
most common translocations are the t(8;21)(q22;q22) AML1-ETO in 15% of
cases, and the t(15;17)(q22;q12) PML-RAR α in 5%. Translocations
involving the MLL locus on 11q23 are less common overall, but occur in 50%
of infant AML.
Pre-B and pre-T cancers usually take the form of acute lymphoblastic
leukemia (ALL)/acute lymphoblastic lymphoma (LBL), with an overall
incidence roughly similar to CML (Jaffe, 2001). They are divided into B-
ALLs, with B-cell origin and surface immunoglobulin; T-ALLs, with a T-cell
origin and surface T-cell receptor; and common ALLs, with a nonmarking
lymphoblast or pre-B origin and rearranged immunoglobulin heavy chain.
The most common rearrangements found in B-ALL and pre-B ALL are the
t(12;21)(p12;q22) TEL-AML1 in 20% of pre-B ALLs, then the t(1;19)(q23;p13)
E2A-PBX1 in 5%, then the various MLL translocations in 5%, which include
t(4;11)(q21;q23) MLL-AF4 and t(9;11)(p22;q23) MLL-AF9. MLL
translocations are especially frequent in infant ALL, present in 85% of cases.
TEL-AML1 and AML1-ETO translocations involve the same AML1 gene on
21q22, but of the sequenced breakpoints, the closest are 28 kb apart. Many
normal T lymphocytes feature inv(7)(p15;q35), an inversion of chromosome 7
between the TCR β and TCR γ loci. Other than this, the most common
rearrangement found in T-ALLs is the SCL-SIL interstitial deletion,

61
del(1)(p32), in about 25% of cases. The interstitial p16 deletion del(9)(p21)
occurs in a small percentage of both T-ALLs and pre-B ALLs / B-ALLs. Less
common are SCL-TCR, LMO2-TCR, HOX11-TCR, and TTG1-TCR
translocations.
Worldwide, mature lymphoid cancers constitute greater than 90% of
lymphoid neoplasms and 4% of all cancers, and greater than 6% of all
cancers in the United States (Jaffe, 2001). They are divided into Hodgkin
lymphoma (HL) and non-Hodgkin lymphomas (NHL), with the latter
composing about 90% of the group. HL appear to be mostly activated B-
cells but recurrent translocations have not been reported. About 85% of NHL
are B-cell lymphomas, the major types with recurrent translocations being
diffuse large B-cell lymphoma (DLBL) in 30.6% of cases; follicular lymphoma
(FL), 22.1%; marginal zone lymphoma (MZL), 9.4%; mantle cell lymphoma,
6.0%; Burkitt lymphoma, 2.5%; and anaplastic large cell lymphoma (ALCL),
2.4%.
The t(14;18)(q32;q21) bcl-2 translocation is found in about half of
NHL, including 80-90% of FL and 20-30% of DLBL (Vega and Medeiros,
2003). About 50% of bcl-2 breakpoints occur in the 175 bp major breakpoint
region (MBR) a few kb telomeric of the bcl-2 oncogene; 13% in the 105 bp
intermediate cluster region (icr) about 18 kb centromeric to the MBR; and 5%
in the 561 bp minor cluster region (mcr) about 11 kb centromeric to the icr
(Weinberg et al., 2007). The t(11;14)(q13;q32) bcl-1 translocation is found in

62
almost all MCL, with about 30% of breakpoints in the 150 bp major
translocation cluster (MTC) about 120 kb telomeric to the bcl-1 oncogene
(Bertoni et al., 2004). The remaining 70% are spread 100 kb on either side
of the MTC, with two minor cluster regions each constituting a few percent.
Both of these translocations join their respective oncogenes to the RSS on
14q32, juxtaposing them with the strong immunoglobulin heavy chain µ
enhancer.
Almost all Burkitt lymphomas contain the t(8;14)(q24;q32) myc
translocation. Ten to twenty percent of DLBLs contain the t(3;14)(q27;q32)
bcl-6 translocation. Both of these translocations join their respective
oncogenes to the repetitive immunoglobulin switch sequences on 14q32. An
additional 5-10% join bcl-6 to various other loci. In MZL, the most common
translocation is t(11;18)(q21;q21) API2-MALT1 in 30% of cases, though few
junctions are available. A quick assessment of them reveals no CpG
proximity and a lack of nucleotide additions. The most common translocation
in T-cell lymphomas is the t(2;5)(p23;q35) ALK-NPM1 in 40-70% of
anaplastic large cell lymphoma (ALCL), for which no translocation junctions
are available (Vega and Medeiros, 2003).
CpG Hotspots are not General to all Translocations
By far, the most common rearrangement known in pre-B
lymphoblastic leukemia is the t(12;21)(p12;q22) TEL-AML1 (ETV6-RUNX1)
translocation. The 53 TEL breakpoints obtained spread across 12.9 kb, only

63
one occurring at a CpG (p>0.8), with an overall average of 41.1 bp to the
nearest CpG, close to the 41.6 bp if randomly distributed (p>0.6). The 53
AML1 breakpoints were more dispersed over a 165 kb region, with only one
occurring at a CpG (p>0.8), and an overall average of 61.7 bp to the closest
CpG, virtually identical to the 61.7 bp if randomly distributed (p>0.4). Thus,
neither TEL nor AML1 breakpoints appear to involve CpG hotspots.
The second most common rearrangement in pre-B lymphoblastic
leukemia is the E2A-PBX1 translocation analyzed previously. Translocations
involving the MLL locus on 11q23 are the third most common, and adjoin
MLL with more than 20 partners across many chromosomes in cases of AML
and ALL. The collection of 350 MLL breakpoints – 24 from primary AML, 291
from primary ALLs, and 35 leukemias secondary to topisomerase II-inhibitor
treatment – are mostly cases of t(4;11)(q21;q23) MLL-AF4 and
t(9;11)(p22;q23) MLL-AF9, and fall within an 8.6 kb window on MLL. Of the
primary AMLs, none occur at CpGs (p=1), averaging 64.4 bp to the nearest
CpG, compared to 85.4 if randomly distributed (p>0.5). Of the primary ALL
breakpoints, 12 occur at CpGs (p>0.1), averaging 79.0 bp to the closest CpG
overall, not too dissimilar to the 84.9 bp if randomly distributed (p>0.4).
Analysis of the B-lineage subset of primary ALLs also yielded no significant
CpG proximity, as did the secondary leukemias. Thus, MLL breakpoints do
not appear to involve CpG hotspots.

64
BCR-ABL translocations t(9;22)(q34;q11) result in Philadelphia
chromosomes seen in chronic myelogenous leukemias (CML). All 35 BCR
breakpoints collected fall into the 2.8 kb major breakpoint cluster region (M-
BCR). Zero occur at CpGs (p=1), with an average of 28.7 bp to the nearest
CpG, compared to 21.4 if distributed randomly (p>0.9). Breakpoints on ABL
scatter over more than a megabase and often fall into repetitive DNA, thus
not all could be located. A 125 kb window containing 25 of the 27 locatable
breakpoints was used for analysis. Again, none of the breakpoints occur at
CpGs (p=1), with an average of 63.0 bp to the closest CpG, versus 50.3 if
randomly distributed (p>0.8). Thus, neither BCR nor ABL breakpoints appear
to involve CpG hotspots.
The t(8;14)(q24;q32) c-myc-IgH switch translocation in sporadic
Burkitt lymphoma is thought to occur as an error of CSR in mature B cells.
The mechanism of double-strand breakage at the highly repetitive switch
regions of 14q32 is an area of intense investigation, while the mechanism at
c-myc is speculated to be similar but unknown. The 125 breakpoints found
on c-myc span 4.1 kb, averaging 8.8 bp to the nearest CpG, compared to 8.9
bp if randomly distributed (p>0.7). It is interesting to note that c-myc has a
higher CpG density (244 CpGs in 4,099 bp, or 16.8 bp/CpG) than either the
bcl-2 MBR (5 CpGs, 150 bp, 35 bp/CpG) or bcl-1 MTC (7 CpGs, 150 bp,
21.4 bp/CpG), yet only 17% (14 breakpoints) occur at CpGs (p>0.5)

65
compared to 43% for the MBR and 37% for the MTC. Thus, c-myc
breakpoints do not appear to involve CpG hotspots.
Similarly, the t(3;14)(q27;q32) bcl-6-IgH switch translocation joins the
first intron of the bcl-6 gene to the switch regions of 14q32 in a variety of
mature B-cell lymphomas. The 2 kb covering the 37 breakpoints obtained is,
like c-myc, very CpG-rich (2045 bp, 102 CpGs = 20.0 bp/CpG). However,
only seven breakpoints occur at CpGs (p>0.2), and the overall average
distance to the nearest CpG is 8.3 bp, as opposed to 10.9 bp if randomly
distributed (p>0.1). Thus, bcl-6 breakpoints do not appear to involve CpG
hotspots.
Graphs of the breakpoint distributions at various distance intervals
confirm these observations (available upon request as a resource). A
summary of these results is shown in Table 2.
T Cell and Some Pre-B Cell Rearrangements Tend to Use a V(D)J-Type
Rather than a CpG-Type Rearrangement Mechanism
In normal V(D)J recombination at the immunoglobulin and TCR loci,
the RAG cmplex binds an RSS and make a nick 5’ of CAC. After synapsis
with a partner RSS, they use the free hydroxyl group at the nick to attack the
opposite strand, thereby creating a double-strand break – with one hairpin
DNA end and one blunt DNA end – at each RSS. The blunt ends containing
the RSS are termed signal ends, later joined by the ligase IV:XRCC4:XLF
complex without loss to form a signal joint. The hairpin ends, also called

66
coding ends because they form the coding sequence of the immune
receptor, are opened and endonucleolytically recessed by Artemis:DNA-
PK
CS
before terminal deoxynucleotidyl transferase (TdT) adds nucleotides
template-independently and the ligase IV:XRCC4:XLF complex joins them
together. Recessing generally follows a distribution which falls with distance
from the free end, and is usually less than eight nucleotides in human cells
(Gauss and Lieber, 1996b). Therefore, the coding joint breakpoint plot
shows a large spike 5’ of CAC which falls further 5’.
The most common T-ALL rearrangements are the intrachromosomal
deletions, del(1)(p32) SCL-SIL and del(9)(p21) ∆p16. p16 deletions also
occur in B-ALLs. While the furthest SCL-SIL breakpoints are 89 kb apart,
204 of 209 breakpoints map to one of three CACs (two shown in Figs. 2-2C
and 2-2D). 35% of SCL-SIL breakpoints occur at CAC (p<10
-74
), with an
average distance of 4.5 bp from the nearest CAC, compared to 40.4 bp if
breakpoints were randomly distributed (p<10
-107
). CpG scores significantly
due to proximity to these CACs (p<10
-14
) but not nearly on the same level.
Individually analyzed, each of the three breakpoint peaks is also most
significant for CAC(A), and all but one of the 209 breakpoints are compatible
with a typical coding joint recombination. Interestingly, the one exception
occurs at a CpG. Thus, SCL-SIL deletions appear to be V(D)J-type events.
More than seven widely-spaced cRSS spanning 231 kb are used in 41
breakpoints from p16 deletions (Kitagawa et al., 2002). They average 11.3

67
bp to the nearest CAC, compared to 46.0 bp if randomly distributed (p<10
-13
),
with seven occurring at a CAC (p<10
-5
). We find three breakpoints at CpG
(p=0.13) and an average distance of 63.7 bp to the nearest CpG, versus 96.7
bp if randomly distributed (p=0.05). Individually analyzed, 28 breakpoints are
compatible with V(D)J-type coding joint events but not CpG-type; three are
compatible with CpG-type but not V(D)J-type; seven with both types; and
three with neither. Thus p16 deletions appear to be V(D)J-type events.
Separate analysis of the B-ALL subset results in the same conclusion.
The SCL locus also translocates to the TCR loci in a small percentage
of T-ALLs, but none of the 20 translocation breakpoints we obtained use the
same sites as SCL-SIL deletions. They span 64 kb, with five at CAC / GTG
(p=0.0007), and average 10.2 bp away, compared to 14.6 bp if randomly
distributed (p=0.04). A small cluster of seven breakpoints centered around a
CACA is likely to be V(D)J-mediated, with two precise signal joints and three
precise coding joints. However, CpG is more significant by measures of
proximity, averaging 32.1 bp away, compared to 51.7 bp if randomly
distributed (p=0.001), but only four breakpoints occur at CpG (p=0.01). This
is complicated by the fact that, at many of the relevant CACs, CpGs are
found close by. A breakpoint-by-breakpoint analysis indicates only nine of the
20 could be V(D)J-type and six are incompatible with either V(D)J-type or
CpG-type breakage. The most significant motif ends up being CACA/TGTG;

68
but as a group, SCL translocations cannot be confidently assigned as any
one type and could be a mixture which includes random-type breakage.
A small percentage of T-ALLs translocate one of the TCR loci to
LMO2, with breakpoints spanning over 40 kb. Of 31 breakpoints, nine occur
at CAC/GTG (p<10
-6
), overall averaging 4.58 bp to the closest CAC / GTG,
versus 16.3 bp if randomly distributed (p<10
-8
). CpG is somewhat significant,
with breakpoints averaging 26.2 bp to the nearest CpG, compared to 47.1 bp
if randomly distributed (p=0.04), but only three occur at CpG (p=0.16).
Individual breakpoint analysis indicates 7 breakpoints compatible with signal
joints but not CpG, 11 with coding joints but not CpG, seven with CpG but not
V(D)J, three with both, and three with neither. On the whole, LMO2
translocations appear to be V(D)J-type recombination events, though a third
or more could be other types.
The t(10;14)(q24;q11) HOX11-TCR and t(11;14)(p15;q11) TTG1-TCR
translocations appear to be other examples of cRSS recombination. Directly
5’ of the main HOX11 CACAG is the sequence CGCG, complicating
analysis; but in TTG1, the closest CpG to the CACAGTG is more than 20 bp
away.
Graphs of the breakpoint distributions at various distance intervals
confirm these observations (available upon request as a resource). A
summary of these results is shown in Table 2.

69
CpG Hotspots are Specific to Pro-B/Pre-B Stage Translocations
Especially in lymphoid lineages, the stage of development during
which a translocation occurs provides important clues as to the mechanism
of breakage. Many translocations can be staged by joining partner.
Specifically, bcl-2 and bcl-1 join to the heavy chain RSS during the pro-B
stage; c-myc and bcl-6 to the heavy chain switch regions during the activated
B-cell stage; LMO2, HOX11, TTG1, and SCL translocations to the TCR RSS
during the pro-T or pre-T stage. Staging for other translocations, however,
requires other means. Additional evidence can be found in nucleotide
additions at breakpoint junctions.
Less than 35% of TEL-AML1, MLL, BCR-ABL, bcl-6, and c-myc
junctions contain nucleotide additions, with microhomologous joining
constituting the largest proportion of junctions. The average lengths of
additions range from 2 to 7.9 bp, with medians from 1 to 4 bp. Aberrant
expression of TdT in some AMLs may account for the comparatively high
50% of breakpoints from AML1-ETO with additions (Jaffe, 2001). However,
this is low in comparison to translocation junctional addition frequencies from
pro-B/pre-B or pro-T/pre-T cells.
Greater than 90% of bcl-2, bcl-1, E2A, and SCL-SIL rearrangements
feature junctional additions, as do more than 80% of lymphoid p16 deletions
(Fig. 2-5). Junctional additions at bcl-2, bcl-1, and E2A are especially long,
with averages from 11.9 to 16.3 bp and medians from 10 to 14 bp.

70
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
TEL-AML1
MLL ALL
bcl-2
bcl-1
E2A-PBX1
bcl-6
c-myc
AML1-ETO
MLL AML
BCR-ABL
SCL-SIL
p16 deletion
Nucleotide Addition No Addition and No Microhomology Microhomology
LM HSC pro-B / pre-B mature B myeloid pre-T

Figure 2-5. Proportions of junctions with nucleotide additions in various
translocations. Stacked percentages of each type of junction are plotted for
various translocations. Proportion of junctions with nucleotide additions,
microhomology, or neither are shown in dark blue, light blue, and light green,
respectively. LM HSC is lymphoid-myeloid hematopoietic stem cell.
Junctions from pro-B/pre-B and pro-T/pre-T rearrangements, i.e. bcl-1, bcl-2,
E2A-PBX1, SCL-SIL, and p16 deletions, have a high degree of nucleotide
additions due to strong expression of TdT. In contrast, junctions from non-
lymphoid rearrangements tend to use microhomology. Interestingly, some
AMLs with AML1-ETO translocations aberrantly express low levels of TdT,
which may account for their relatively high rate of nucleotide additions.

71
Those at SCL-SIL and p16 deletion are about 7 bp. The high frequency and
length of junctional addition is likely a reflection of the robust activity of TdT in
pro-B/pre-B and pro-T/pre-T stage cells. These junctional additions are also
seen in normal immunoglobulin and TCR junctions, and are thought to be
coexpressed with RAGs as a means of increasing receptor diversity.
Although TEL-AML1, MLL, and E2A-PBX1 translocations all occur in
pre-B or B-ALLs and may have rearranged immunoglobulin loci, there is a
critical difference. TEL-AML1 and MLL translocations themselves likely
occur before the pro-B stage, and therefore prior to TdT and RAG
expression, consistent with what has been inferred biologically (Jansen et al.,
2007; Pine et al., 2003). Cells with these translocations may later express
TdT and RAGs and rearrange their receptor loci. However, E2A-PBX1
translocations appear to occur at the pro-B or pre-B stage, during which TdT
is expressed. This conclusion was previously reached by Wiemels et al.
(Wiemels et al., 2002a).
Taken together with the lack of CpG hotspots in pre-T ALLs, AMLs,
CMLs, c-myc, and bcl-6, CpG-type translocations appear to be exclusive to
pro-B or pre-B stage cells which express TdT, and therefore likely RAGs as
well.
Coexistence of Multiple Distinct Breakage Mechanisms
p16 deletions in pre-B ALLs seem to use a cryptic V(D)J-type
mechanism, like their pre-T counterparts, indicating that these mechanisms

72
coexist with, but are dwarfed by the CpG-type mechanism. V(D)J-type and
CpG-type hotspots appear to be very special cases where an active
breakage mechanism creates double-strand breaks within small windows of
specific regions in the genome.
Even within one translocation, the break on one chromosome may
occur by one mechanism and the break on the other chromosome by another
mechanism. This is evidenced by CpG-type breakages at bcl-1 and bcl-2
joining to V(D)J-type breakages at the heavy chain locus, and experimentally
by I-SceI restriction cuts with V(D)J-type and class switch-type ex vivo and in
vivo, respectively (Weinstock et al., 2007; Zarrin et al., 2007).
Unlike the interstitial deletions at SCL-SIL and p16, a large proportion
of the SCL and LMO2 translocations in pre-T cells do not use V(D)J-type
breakage – even if they join to the RSS at the TCR loci. CpG was at least
marginally significant in these two rearrangements, thus CpG-type breakage
might account for some of them. However, a remaining fraction escape
either the V(D)J-type or CpG-type classification and for now are assumed to
be random-type breakage.
We postulate that this “any-break-to-any-break” joining also occurs in
the E2A-PBX translocation, where breakpoints at the E2A cluster at two CpG
hotspots within a 17 bp region of a 3.2 kb intron, but those on PBX1 are
distributed across 107 kb without CpG or CAC propensity, presumably due to
random-type breakage.

73
The Widths and Shapes of CpG-type, V(D)J-type, and “Random”-type
Breakpoint Distributions Provide Insight into the Mechanisms of Breakage
The null hypothesis to a sequence-specific breakage mechanism is
that breakage is sequence-independent and occurs equally everywhere (e.g.,
by reactive oxygen species or ionizing radiation). Any clustering of
breakpoints, then, would be due to random chance, selective growth
advantage, and/or broad chromatin accessibility effects. It is easy to imagine
that open chromatin is more susceptible to translocation, or that breakpoints
in certain regions provide growth advantages over other regions for reasons
such as stability of the fusion mRNA, cryptic splice sites, or protein binding
requirements. This seems to be the situation for non-recurrent
rearrangements, as well as translocations such as BCR-ABL and TEL-AML1,
where breakpoint clusters extend over kilobases and different leukemias
rarely share the same breakpoint.
By contrast, breakpoints from V(D)J-type and CpG-type
rearrangements can occur over equally large or larger regions, but strongly
prefer small windows and even specific nucleotide positions, such as occurs
at SIL or the bcl-2 MBR. It is difficult to imagine growth advantage and
accessibility effects as the basis for clustering down to 150 bp or so as in the
MBR and MTC, let alone a 20-fold preference for one site over another site
five or ten bases away. This extremely high degree of clustering, down to
the level of a few base pairs, suggests a site-specific breakage mechanism.

74
For V(D)J recombination, we know it to be the RAG endonuclease acting at
RSS or cRSS.
However, the different shapes of the breakpoint distributions for
V(D)J-type and CpG-type peaks suggest they operate by different
mechanisms. The reason that cRSS have a “spike-and-single-sided tail”
distribution is that a site-specific break occurs first, followed by limited
recessing only into the pseudo-coding end side of the break (Fig. 1D and
1E). However, the breakpoint distribution of a CpG hotspot is more bell-
shaped, with breakpoints on either side of the tallest CpG spike; therefore,
initial breakage likely occurs a little heterogeneously about the CpG (Fig. 1A).
Additionally, the majority of der(14) or “top strand” breakpoints tend to
be somewhat 5’ of the CpG(s). What is likely occurring is that an initial
breakage might occur around a CpG, and both the telomeric end (the left end
in Fig. 1A), and the centromeric end (the right end in Fig. 1A) are recessed
away from the CpG. The telomeric end becomes part of der(14), and the
centromeric end part of der(18). However, the top strand only represents
der(14) breakpoints, and therefore the top strand only shows the recessing
on the telomeric or left end. If there were an equal number of der(18) bottom
strand breakpoints, and they were reflected onto the top strand, then the
distribution would be a symmetrical hill centered at the CpG, rather than
shifted slightly off.

75
Heterogeneity in nick sites is not typical of sequence-specific
processes such as V(D)J recombination, or sequence-specific nucleases
such as restriction enzymes; it is, however, consistent with structure-specific
nucleases.
The RAG Endonuclease is Capable of Nicking One-Base Mismatches, and
can Process Nicks, Gaps, and Flaps into Double-Strand Breaks
We have previously shown that the RAG complex can nick large
structural deviations such as three to ten-base bubbles and heteroduplexes,
but not one-base bubbles (Raghavan et al., 2007). After lengthening the
oligonucleotide substrates, we find nicking at one-base bubbles (Fig. 2-6 A-
B). Moreover, nicking efficiency is inversely proportional to the expected
stability of the mismatch, as estimated previously (Peyret et al., 1999). The
most unstable mismatches, C:T and C:A, nicked at about 6% the efficiency of
a very good 12 signal. The more stable mismatches T:T and T:G nicked six-
fold less at about 1%, while the duplexes were not visibly nicked. This
further demonstrates the structure-specificity of the RAG endonuclease.
Nicks, gaps, and flaps are cut at a similar or higher rate than the
unstable mismatches (Fig. 2-6C). Importantly, the nick, gap, and flap
substrates display a heterogeneous nicking pattern about the lesion, similar
to the pattern of breakpoints seen about CpG at the bcl-2 MBR.

76
Figure 2-6. The RAGs endonuclease is capable of nicking one-base
mismatches, and can process nicks, gaps, and flaps into double-strand
breaks. The diagrammed 12% PAGE-purified, 5’-labeled substrates were
annealed by boiling and slow cooling, then incubated with purified GST-core
RAGs and HMG1, deproteinized, and run on an 8% denaturing PAGE. (A)
One-base bubbles having arms of 31 bp and 38 bp and lacking CAC were
nicked at a low level. For comparison, duplex DNA is not visibly nicked, and
a good 12 signal is nicked efficiently. Lanes without RAGs were incubated
with HMG1 alone and do not show nicking. The diagrammed structure on
the left bottom margin of the gel depicts RAG nicking on the wrong side of
the heptamer adjacent to the spacer. (B) Quantitation shows that nicking
efficiency increases with the ∆G of the mismatch (Peyret et al., 1999). That
is, the RAG nicking is more efficient for pairs (matched or mismatched base
pairs) that have a higher (more positive) ∆G, and RAG nicking is less efficient
for pairs that have a lower (more negative) ∆G. (C) Nick, gap, and flap
substrates are cut at a similar level or higher (relative to the mismatches in
panel A) at heterogeneous positions around the lesion. The 12 signal and
duplex DNA controls are similar to those in (A). The diagrammed structures
on the right margin depict the two alternative flap configurations that result
after annealing the flap oligonucleotides.

77
Figure 2-6 continued

12 signal
top bottom
− + − +
CG duplex
top bottom
− + − +
TA duplex
top bottom
− + − +
TG mismatch
top bottom
− + − +
TT mismatch
top bottom
− + − +
CT mismatch
top bottom
− + − +
CA mismatch
top bottom
− + − +
substrate
strand
RAGs
*
*
*
*
*
*
*
31 bp 38 bp
A.

78
Figure 2-6 continued

B.
Substrate Nicking Efficiency
12 signal 66.7%
cacTgtg 1.8%
T-G mismatch 0.7%
T-T mismatch 0.8%
C-T mismatch 4.1%
C-A mismatch 4.0%

C
G
T
G
T
A
T
T
C
T
C
A
Peyret et al., 1999

79
Figure 2-6 continued

*
CG
duplex
− +

12 signal
− +
substrate
RAGs

nick
− +

1 nt gap
− +

2 nt gap
− +

3 nt gap
− +

1 nt loop
− +

1 nt flap
− +

3 nt flap
− +
*
*
C.

80
Discussion and Theory
Mechanistic and Developmental Model for Classifying Translocations and the
Role of the CpG Motif
Numerous parallels between the translocations at the bcl-2 MBR and
the bcl-1 MTC have suggested that they share a common mechanism.
However, mechanistic insight into the nature of breakage and the precise
breakpoint distribution has been lacking (Bertoni et al., 2004; Jaeger et al.,
2000; Tsujimoto et al., 1988; Welzel et al., 2001b).
Here we have detailed an additional key observation that translocation
breakpoints within the bcl-2 MBR on chromosome 18 and bcl-1 MTC on
chromosome 11 are highly focused to CpG hotspots, and that this also
extends to the bcl-2 icr about 18 kb downstream of the MBR, the bcl-2 mcr
about 28 kb downstream of the MBR, and the E2A cluster on chromosome
19. We demonstrate that this phenomenon is prominent in pro-B/pre-B stage
cells, but not in myeloid cells, lymphoid-myeloid precursors, mature B cells,
or T-cells. Moreover, at the MBR, the fine structure of the three-peak pattern
is consistent with slightly heterogeneous nick sites about the CpGs followed
by limited 5’ recessing. No known single factor can account for all of these
observations.
Why might CpGs be targets for translocations? The most apparent
feature of CpG is its tendency to be methylated and then deaminated to a
T:G mismatch. While such events are virtually invisible on the time scales in

81
which biochemical experiments are normally done – having a half-life of
>30,000 years in double-stranded DNA, not even accounting for repair – they
are conspicuously frequent in cancers and have remodeled vertebrate
genomes over evolutionary time (Pfeifer, 2006). Unmethylated cytosines
also deaminate, but at a five-fold lower rate, and result in uracils.
Normally, base-excision repair corrects these defects. Glycosylases
first remove these damaged bases, creating abasic sites. Uracil DNA
glycoslyase (UDG), though, is >2,000-fold more efficient at processing
uracils than thymine DNA glycosylase (TDG) or methyl-CpG binding domain
protein 4 (MBD4) are at processing T:G mismatches (Walsh and Xu, 2006).
This is thought to be the basis for preferential mutagenesis at methylated
cytosines over unmethylated ones. Following glycosylase activity, APE1 or
spontaneous β-elimination create single-strand gaps, which are filled-in by
polymerase β and ligated by ligase III:XRCC1 (Friedberg et al., 2006; Lindahl
et al., 1997; Nash et al., 1997). However, the intermediate mismatches and
gaps are vulnerable lesions at which structure-specific nucleases may act.
We have previously shown that RAGs can nick heteroduplexes as
small as two or three base pairs (Raghavan et al., 2007), and here, we have
demonstrated nicking of single-base mismatches and across from nicks; one,
two, and three-nucleotide gaps; and one and three-nucleotide flaps as well.
Nicking of one-base mismatches would create a single-strand break, which
unlike normal V(D)J-type nicks at CACs, would not be ligated readily by

82
ligase I due to the mismatch at the point of ligation (Tomkinson et al., 2006).
Nicking across from nicks and gaps creates double-strand breaks, and we
observe that these nick sites are heterogeneous, which could explain the
heterogeneity in breakpoints at CpG hotspots. Thus, for mechanistic
reasons, RAGs are a reasonable candidate for an endonuclease mechanism;
and for stage-specificity reasons, would go a long way to explain why
translocation breakpoints in lymphoid-myeloid precursors and mature B cells
do not occur at CpG hotspots.
However, T cells also express RAGs but do not appear to translocate
to CpG hotspots. The recent finding of activation-induced deaminase (AID)
expression in early B cells may be able to account for this lineage-specificity
(Feldhahn et al., 2007; Mao et al., 2004). The major activity of AID is thought
to be deamination of single-stranded cytosines in mature B cells during class
switch recombination and somatic hypermutation. The finding of AID in pre-
B cells is consistent with earlier work showing that pre-B cell lines can
undergo class switch recombination (Alt et al., 1982; Burrows et al., 1983).
Lower but still significant rates of deamination exist for genomic regions
which are thought to be mostly double-stranded, such as c-myc and bcl-6
(Liu et al., 2008). Therefore, AID is a reasonable explanation for lineage-
specificity.
Based on this evidence, we propose the model diagrammed in Figure
2-7. Some level of AID expression in early B cells leads to an increased rate

83

Figure 2-7. Theoretical mechanism for CpG-type double-strand breakage.
Deamination at a methylcytosine within a CpG creates a T:G mismatch which
persists due to catalytic inefficiency of methyl-CpG binding domain protein 4
(MBD4) and thymine DNA glycosylase (TDG) in cleaving the thymine
glycosidic bond, leaving an abasic site ( ). Single-strand breaks are
generated either by the normal base excision repair pathway of glycosylase
and AP endonuclease (APE) activity, or by RAGs nicking the mismatch
directly, resulting in a poorly-ligatable flap. Heterogeneous RAG nicking of
the remaining strand creates a double-strand break close to the original site
of the T:G mismatch. Nicking by the RAG complex at one-base mismatches
and across from nicks, gaps, and flaps is shown in the Fig. 2-6. Nucleolytic
resection of the ends before aberrant joining results in a breakpoint
distribution similar to that seen in Fig. 2-1.

RAGs MBD4, TDG
RAGs
deamination
APE
RAGs
me
CG
GC
me

T
G
G
C
me

G
G
C
me

84
of cytosine deaminations, which are adequately repaired at unmethylated
cytosines but persist at methylated cytosines as T:G mismatches. These
sites are then converted to double-strand breaks by RAGs, before or during
attempted repair by T:G mismatch repair enzymes. The resulting DNA ends
join to ends from a V(D)J recombination or other double-strand break event
and lead to the chromosomal translocations observed so frequently in human
B-cell lymphomas.
Basis for Translocation Clustering and the Relationship to Translocation
Hotspots
Clearly, if this hypothesis is correct, other requirements exist for
translocation targeting to the 175 bp bcl-2 MBR and the 150 bp bcl-1 MTC,
i.e. to these CpGs as opposed to all the others in the surrounding tens or
hundreds of kilobases. Possibilities can be grouped into those which
promote breakage and recombination (such as increased methylation,
selective AID or targeting of the RAG complex, slow repair rate,
chromosomal positioning, and failure to sequester free ends at certain loci)
and those which prevent recombination events from appearing in cancers
(such as lack of a growth advantage). None of these various possibilities
preclude one another, and it could be that there are multiple contributory
mechanisms and/or different mechanisms operating at different loci. As the
other possibilities are discussed elsewhere, only the first two will be
discussed below.

85
Though unlikely, the most obvious explanation for these 20-150 bp
clusterings would be selective hypermethylation within those zones.
Methylation at the bcl-2 MBR and bcl-1 MTC differed greatly within and
between FACS-sorted pre-B cells from two individuals of the same age and
race. CpGs outside the cluster zones were methylated at the same or higher
rate (Fig. 2-8). Hence, while there is no hypermethylation, there are still
ample targets for deamination-based translocations (Fig. 2-7).
Hydrolytic deamination of methylcytosine is extremely slow in duplex
DNA, but is accelerated 500 to 1000-fold when the base is unpaired or
single-stranded (Frederico et al., 1990). Moreover, AID is much more active
on unpaired cytosines than those which are base paired (Bransteitter et al.,
2003a; Pham et al., 2003b; Yu et al., 2004b). Therefore, one would expect
CpGs in more single-stranded regions to be more susceptible to
deamination. We have previously shown that the bcl-2 MBR is reactive with
the single-strand-specific chemical probe bisulfite, suggesting some degree
of single-stranded character in this region (Raghavan et al., 2004a). The bcl-
1 MTC also shows a high degree of bisulfite reactivity, but it is elevated
mostly at the centromeric end of the 150 bp MTC (AT & MRL, unpublished).
This naturally leads to the hypothesis that AID may be drawn to the cytosines
within these regions, and the deamination products of methylcytosines
persist while those of unmethylated cytosines are repaired. Such a scenario

86
Figure 2-8. Methylation status of the bcl-2 MBR and bcl-1 MTC. FACS-
sorted pre-B cells from two patients were bisulfite sequenced to determine
the methylation pattern at the (A) bcl-2 MBR and (B) bcl-1 MTC. Each line of
circles represents the CpGs from a single bisulfite sequenced molecule,
where a filled circle represents a methylated CpG and an empty circle
represents an unmethylated CpG. The five CpGs within the MBR are
represented by the second through sixth circles, with the first circle being the
closest CpG telomeric of the MBR. The seven CpGs within the MTC are
represented by the first through seventh circles, with the eighth circle being
the closest CpG centromeric of the MTC.

87
Figure 2-8 continued
A. bcl-2 MBR.

Molecule Patient A Patient B
1
|||||| ||||||
2
|||||| ||||||
3
|||||| ||||||
4
|||||| |||z||
5
|||||| |||z||
6
z||||| |||z||
7
z||||| |||z||
8
z||||| zz||||
9
z|z||| zzz|||
10
zzz||z |z||zz
11
zzzz|z zzz|z|
12
z|zzzz zz|zzz
13
z|zzzz zzz|zz
14
zzzzzz zzz|zz
15
zzzzzz zzz|zz
16
zzzzzz zzzzzz

B. bcl-1 MTC.

Molecule Patient A Patient B
1
|||||||| |z|z||||
2
|||||||| |z|z||||
3
|||||z|| |z|z|||z
4
|||||||z |z|z|||z
5
|||||||z |z|z|||z
6
|z|||||z z|zzz|||
7
|z|||||z |zzz|zzz
8
||||||zz |zzz|zzz
9
||||||zz zzzz||zz
10
|z|z|||z zzzz|z|z
11
|z|z|||z zzzz|z|z
12
|z|z|||z zzzzzz|z
13
|||zz||z zzzzz|zz
14
|||zz||z zzzzz|zz
15
z|z||zz| zzzzzzzz
16
|z|z||zz zzzzzzzz

88
would explain the breakpoint hotspots at the CpGs within the bcl-2 MBR and
bcl-1 MTC.
Regions of increased single-stranded character might also be caused
by slippage at repeat sequences, induced by a passing RNA or DNA
polymerase. Such a mechanism has been suggested based on sequence
data from a yeast LYS2 reporter system (Kim et al., 2007). The formation of
cruciforms at palindromic repeats appears to drive the t(11;22), the most
common constitutional translocation in humans (Kurahashi et al., 2006). At
the bcl-2 mcr, there is an eight base pair direct repeat on each side of one
CpG hotspot. A slippage event between the top and bottom strand would
place the CpG within a loop, thereby making it vulnerable to hydrolytic
deamination or to AID. Therefore, there may be multiple mechanisms by
which single-stranded character could be generated.
Current Lack of Mouse Models of Translocation-Based Lymphoma
We eagerly searched the literature for mouse models to prove or
disprove the theory. According to the proposed mechanism, AID-
overexpressing mice would have increased rates of translocation in both B
and T lineages. Similarly, we would expect RAG-overexpressing, MBD4-
knockout, UNG-knockout, and MSH2-knockout mice to develop more
translocations in germinal center B-cells.
Unfortunately, AID-overexpressors developed only T-cell lymphomas,
and without clonal tranloscations (Okazaki et al., 2003). In general,

89
lymphomas in mice mainly seem to be of T-cell origin, including p53
-/-
and
p53
+/-
mice (Bassing et al., 2008; Liao et al., 1998). Presumably, point
mutations of critical genes are sufficient to drive T-cell lymphomagenesis at
early ages in mice. These may mask any latent or slower-developing
tumors. Experimental demonstration for this comes from a study where mice
constitutively expressing c-myc developed only T-cell lymphomas, while
RAG1-knockouts constitutively expressing c-myc developed a range of
hematological malignancies, including B-LBLs, pre-T cell lymphomas,
myeloid progenitor tumors, and cutaneous macrophage tumors (Smith et al.,
2005).
Mouse B cell follicular lymphomas and not T cell lymphomas
developed in UNG
-/-
mice after 12 months, but these were never
characterized for translocations (Nilsen et al., 2003). It is not unlikely that, in
some initiating germinal center B cells, AID either created irreparable p53
mutations or lesions at c-myc leading to a CSR-type translocation to the
switch regions. Interestingly, MSH2
-/-
mice developed B and T lymphoblastic
lymphomas, and these were not examined for translocations either (Reitmair
et al., 1995). MBD4 knockout mice are not especially susceptible to tumors
(Wong et al., 2002). RAG1/2-overexpressing mice are small and die early
(Barreto et al., 2001).
As far as the issue has been examined, c-myc appears to be the only
recurrent translocation target in mouse B-cell lymphomas, aberrantly joining

90
to the immunoglobulin heavy chain switch regions as part of an aborted CSR
attempt. This translocation in humans, t(8;14), results in Burkitt lymphoma
most commonly; but, in mice, the equivalent t(12;15) translocation results in
plasmacytomas (Muller et al., 1995).
More fundamentally, even in mice with genome-wide instability, many
translocations simply have no effect at all. H2AX
-/-
mice develop many
different translocations, but do not develop tumors (Bassing et al., 2008).
Overexpression of the bcl-2 or bcl-1 oncogenes in mice can cause
hyperplasia but does not lead to lymphoma (Lovec et al., 1994; McDonnell et
al., 1989). As is the case with the many asymptomatic humans who harbor B
cells with t(14;18) translocations, additional mutations may be required, and it
conceivable that mice do not live long enough for secondary mutations to
drive these translocation-bearing B cells to malignancy.
In humans, B cell lymphomas are greater than five-fold more frequent
than T-cell lymphomas, with a median age around 50-60 years, and p53
mutations are not found all that frequently among hematopoietic cancers
(Jaffe, 2001; Wada et al., 1993). As mice do not live past three years, even
calorie-restricted, their lifespan may simply be too short to reveal slower-
developing tumors.
While physiologic lymphoid processes in mice and humans appear to
be very similar and many components are likely interchangeable (e.g. V(D)J
recombination, CSR, somatic hypermutation) small but important

91
mechanistic and phenotypic differences exist. However, mice and humans
appear to be very different with regard to rare and non-physiologic processes
such as lymphomagenesis, of which translocations seem to play a large part
in humans but not in mice.
These factors present technical challenges in developing a mouse
model for translocation-based lymphomas as occurs in humans.

92
Chapter 3: DNA structure as a basis for translocation
breakpoint clustering at the bcl-1 MTC and bcl-2 MBR
Summary
In order to detect the presence of a non-B-DNA structure at the bcl-1
MTC, we employed a combination of chemical, physical, and enzymatic
methods. An ex vivo transfection assay attempted to recapitulate the
translocation process on plasmid, and especially the breakpoint focusing to
the bcl-1 MTC. However, such efforts were ultimately unsuccessful for a
variety of technical reasons. Progress has been made to overcome some of
these technical difficulties, and suggestions are made regarding proper
design of these experiments.
Introduction
Following the initial publication of a non-B-DNA structure directly
cleaved by the RAG complex as the cause of double-strand breaks in bcl-2
MBR translocations, efforts to identify and characterize the structure led to
the conclusion that the bcl-2 MBR was forming an intramolecular triplex
(Raghavan et al., 2005a; Raghavan et al., 2004a; Raghavan et al., 2005b;
Raghavan et al., 2005c; Raghavan et al., 2004b). This was based on ex vivo
transfection assays, RAG nicking of bcl-2 MBR sequence on plasmid as
detected by primer extension, bisulfite chemical probing, potassium
permanganate (KMnO
4
) and osmium tetroxide (OsO
4
) chemical probing,

93
electrophoretic mobility shift assay (EMSA), circular dichroism (CD), and
electron microscopy (EM). Attempts to apply some of these approaches to
the bcl-1 MTC are detailed below.
The V(D)J recombination assay
The V(D)J recombination assay attempts to recapitulate
recombination processes on plasmids transfected into cell lines such as
293T and Reh. A construct bearing a tryptophan operon promoter (P
trp
) and
the chloramphenicol acetyltransferase (cat) antibiotic resistance gene
separated by recombinogenic sequences with an intervening oop
transcriptional terminator (oop) is transiently transfected into the cell line (Fig.
3-1). After two days, substrates are harvested out by alkaline lysis,
transformed into competent DH10B E. coli, and plated onto LB-agar
containing chloramphenicol. If recombination occurs on a particular plasmid
molecule during the two days, the terminator is removed, transcription runs
through the cat gene, and a colony bearing the plasmid grows on the plate.
If recombination does not occur, the terminator is intact, transcription stops
before reaching the cat gene, and E. coli transformed with the plasmid do not
grow.
While the substrates carry an SV40 origin of replication and an SV40
large T antigen, not all of them replicate within the transfected cells.
However, only the replicated fraction appear to be accessible for
recombination (Lieber et al., 1987). Replicated substrates lose their bacterial

94
Figure 3-1. The V(D)J recombination assay. A construct bearing a
tryptophan operon promoter (P
trp
) and the chloramphenicol acetyltransferase
antibiotic resistance gene (cat) separated by recombinogenic sequences with
an intervening oop transcriptional terminator (oop) is transfected into a RAG-
expressing cell line. After two days, substrates are harvested out by alkaline
lysis, transformed into competent DH10B E. coli, and plated onto LB-agar
containing chloramphenicol. If recombination occurred on a particular
plasmid molecule during the two days, the terminator is removed,
transcription (diagrammed as the green dotted line) runs through the cat
gene, and a colony bearing the plasmid grows on the plate. If recombination
does not occur, the terminator is intact, transcription stops before reaching
the cat gene, and E. coli transformed with it do not grow. (A) When good
consensus recombination signal sequences (RSS) are used, recombination
efficiency is about 10% in the Reh cell line. (B) When bcl-1 sequence is
used in place of a signal, recombination drops more than 100-fold into
background levels.

95
Figure 3-1 continued

Transfect into RAG
+
cells
48 h, 37 ° C
bcl-1
P
trp

oop cat gene V(D)J
P
trp

>99.995%

bcl-1
P
trp

oop cat gene V(D)J
cat gene
<0.005%
☺
☺
☺
☺

B.
Recover, transform
into E. coli, plate
onto amp-cam
RAG
+
cell nucleus
Transfect into RAG
+
cells
48 h, 37 ° C
V(D)J
P
trp

oop cat gene
Recover, transform
into E. coli, plate
onto amp-cam
V(D)J
P
trp

90%

V(D)J
P
trp

oop cat gene V(D)J
cat gene
10%
☺
☺
☺
☺

A.
RAG
+
cell nucleus

96
methylation pattern and are therefore resistant to DpnI restriction digestion,
which cuts sites with bacterial dam methylation, i.e. methylated adenines in
the context GATC. For that reason, recombination efficiency is measured as
the number of recombined colonies divided by the estimated number of DpnI-
resistant colonies screened. The number of DpnI-resistant colonies divided
by the total number of colonies is a rough indication of transfection efficiency,
and varies widely, typically from 5-25% even when transfections are done
seconds apart. However, this does not seem to affect the recombination
efficiency, which, in my hands, varied less than two-fold for pGG51 in Reh (to
be described). The main advantage of this method over PCR-based
detection methods is that it can give reasonable quantitation of
recombination efficiency.
As a positive control for the assay, I use the plasmid pGG51, which
has a perfect 12 signal on one side of the terminator and a perfect 23 signal
on the other. A recombination efficiency significantly lower than 10% in Reh
may indicate a problem with the transfection or the cell culture. Reh are a
pre-B cell line containing a TEL-AML1 translocation and naturally express the
RAG proteins which recombine signals. If using 293T, one must cotransfect
RAG expression constructs as well. Transfection into RAG-deficient or
NHEJ-deficient cell lines results in a recombination efficiency <10
-5
.
We attempted to recapitulate the bcl-1 MTC translocation by cloning a
DNA fragment containing the MTC on one side of the terminator and a pair of

97
RSS on the other. If a recombination process occurred on the plasmid
similarly to the translocations seen in mantle cell lymphomas, the MTC would
join to an RSS, the terminator would be deleted, and a colony harboring the
plasmid would grow on the antibiotic plate (Fig. 3-1B). However, initial
attempts to perform this assay using kilobase or larger pieces of the bcl-1
were hindered by an effect I termed “expression blocking” (Fig. 3-2). I
found that, when the transcriptional terminator had been deleted from a given
substrate, it failed to confer antibiotic resistance to transformed bacteria.
Further deletion of the bcl-1 fragment conferred resistance, demonstrating
that the problem was not due to the cat gene or the antibiotic plate. Thus the
bcl-1 sequence appears to “block” expression of the gene, and even if the
MTC was recombining with the RSS in some plasmids, we would never see
them because the transformed bacteria would die regardless.
After cloning a large number of bcl-1 fragments into such test
substrates, I found this effect to be both length-dependent and sequence-
dependent (Fig. 3-2 B-C). Certain fragments would block expression when
cloned in one orientation but not the other. If a particular fragment did not
block expression, a shorter subfragment of it also would not block
expression. If a particular fragment did block expression, a longer piece
containing that fragment would also block expression.
We found that the fragment used by another group to do the assay
was also affected (Marculescu et al., 2002b). They reported no breakpoint

98
Figure 3-2. Expression blocking effect. (A) Even with the transcriptional
terminator removed, substrates with large bcl-1 sequences between the
bacterial tryptophan operon promoter (P
trp
) and the chloamphenicol
acetyltransferase resistance gene (cat) fail to confer resistance to
transformed DH10B E. coli. Bacteria transformed with the pAT3 ClaI (pAT7)
substrate, for example, grow many large colonies LB-agar containing
ampicillin (amp) but none on LB-agar containing both ampicillin and
chloramphenicol (amp-cam). In the substrate pAT11, where the bcl-1
sequence is additionally deleted from pAT7, many large colonies grow on
amp-cam plates, indicating that the gene itself is functional and the plates are
not defective (not shown). Thus the bcl-1 sequence appears to “block”
expression of the downstream gene. Whether this occurs at the level of
transcription (e.g. the bcl-1 DNA terminates the RNA transcript), RNA
stability (e.g. the bcl-1 RNA destabilizes the RNA), or translation (e.g. the bcl-
1 RNA prevents translation of the gene) is not known. Various fragments of
the bcl-1 containing the MTC were tested for blocking effect in the (B)
telomeric-to-centromeric or “forward” orientation, or the (C) centromeric-to-
telomeric or “reverse” orientation. Lengths are drawn to scale and the boxed
region denotes the 100 bp MTC. Listed to the left are the construct numbers
and the oligos used to clone them (5’ oligo, then 3’ oligo). Fragments colored
red exhibited blocking effect and did not grow on amp-cam plates.
Fragments colored green did grow.

99
Figure 3-2 continued
A.

P
trp

bcl-1 cat
amp-cam amp

100

B. bcl-1 forward orientation
C. bcl-1 reverse orientation
pAT32, 34, 46, 48 (SH1, SH4)

pAT36 (SH11, SH12)

pAT5, 7, 56, 58 (AT29, AT30)

pAT38 (SH15, SH16)

pAT50 (AT29, AT90)

pAT52 (AT89, AT30)

pAT54 (AT29, AT91)

pAT75 (AT89, AT91)

pAT82 (AT29, AT113)

pAT84 (AT114, AT30)

pAT86 (AT112, AT91)

pAT33, 35, 47, 49 (SH4, SH1)

pAT37 (SH12, SH11)

pAT6, 8, 57, 59 (AT30, AT29)

pAT39 (SH16, SH15)

pAT51 (AT90, AT29)

pAT53 (AT30, AT89)

pAT55 (AT91, AT29)

pAT76 (AT91, AT89)

pAT83 (AT113, AT29)

pAT85 (AT30, AT114)

pAT87 (AT91, AT112)
Figure 3-2 continued

101
focusing into the MTC. At least a few of the bcl-2 substrates are also
affected (Raghavan et al., 2005b; Raghavan et al., 2004b).
We therefore proceeded to do the assay using a substrate with a
suitable fragment of the bcl-1 (Fig. 3-3). Out of an estimated 2.45 ⋅ 10
6
DpnI-
resistant colonies screened, we obtained 85 unique recombinants. 21
occurred in the 150 bp MTC, and 64 in the eligible surrounding ~450 bp,
which is about what is expected for random breakage throughout the region
(Fig. 3-3A). 13 of the 85 occurred within 30 bp 5’ of the RSS heptamer, and
thus these breaks at the RSS side could be V(D)J-mediated (Fig. 3-3B). Of
the 13, 5 were also at the MTC but none of them at CpGs; 7 were in the
vector, outside the cloned bcl-1 fragment (Fig. 3-3C). This is far below the
focusing or the efficiency reported for the bcl-2 MBR (Fig. 3-3D). Because
the statistical significance is marginal, p-value hovering around 0.05, and the
recombination efficiency is ten-fold lower than at the bcl-2 – i.e. the signal is
drowned in the background – it is probably not possible or practical to make
a convincing argument from doing the same experiments that were done for
the bcl-2 MBR, e.g. testing the RAG dependence or comparing joining to
signal versus coding ends.
Attempts to circumvent the expression blocking effect
While testing the blocking effect of the various bcl-1 fragments, we
concurrently tried to circumvent the problem altogether by using an alternate
construct design, and by improving the components of the current design.

102
Figure 3-3. V(D)J recombination assay of pAT90. The junctions of the 85
assay recombinants were sequenced and organized by 5’ breakpoint (A) and
3’ breakpoint (B). Each line represents a recombinant, where the red lines
denote intact sequence, and the end of the red line indicates the position of
the breakpoint. The black breakpoint lines are markers denoting the
boundaries of the MTC and the 3’ RSS. Breakpoints do not occur at the 5’
RSS for reasons of expression blocking by sequences between the two RSS.
This effect was also seen in the prior study of the bcl-2 MBR (Raghavan et
al., 2005b). The positions of the breakpoints correspond roughly to the
substrate structure is shown at the top of each diagram. (C) 13 of the 85
occurred within 30 bp of the 3’ RSS, thus double-strand breakage at the 3’
breakpoint is possibly mediated by the RAG complex. 5 of these have a 5’
breakpoint within the 150 bp MTC but none occur at CpGs; 7 have a 5’
breakpoint in the vector, outside the cloned bcl-1 fragment. (D) The
efficiency of recombination and the degree of focusing to the cluster region
are much lower than those reported for the bcl-2 MBR (Raghavan et al.,
2005b).

103

A.
MTC
bcl-1 reverse orientation
P
trp

oop V(D)J cat gene
Figure 3-3 continued

104

B.
V(D)J
bcl-1 reverse orientation
P
trp

oop V(D)J cat gene
Figure 3-3 continued

105
Figure 3-3 continued

D.
substrate DpnI-resistant
colonies
screened
number of
recombinants
5’ breakpoint
within cluster
region (percent
of
recombinants)
3’ breakpoint at
RSS (percent
of
recombinants)
breakage
efficiency at
cluster region
breakage
efficiency at
RSS
number and
breakage
efficiency at
cluster region
and RSS
pSCR71* 238,000 33 23 (70%) 11** (33%) 0.000097 0.000046 10 (0.000042)
pSCR72* 294,800 29 20 (69%) 5** (17%) 0.000068 0.000017 3 (0.000010)
pAT90 2,450,000 85 21 (25%) 13*** (15%) 0.0000086 0.0000053 5 (0.0000020)

* data from (Raghavan et al., 2005b)
** breakpoints are less than 14 bp 5’ of the heptamer
*** breakpoints are less than 30 bp 5’ of the heptamer

C.
bcl-1 MTC

106
First, we tried to use deletion of the ccdB bacterial suicide gene as a
selectable marker for recombination (Fig. 3-4A). When recombination occurs
between the two sequences of interest, the gene is deleted, and E. coli
transformed with recombined plasmids can grow into colonies on plates.
Unrecombined plasmids retain the gene, thus E. coli transformed with them
are killed by the ccdB protein product and unable to grow into colonies.
There are a number of complications with this design. First, to clone
such a substrate in the first place, one must inhibit expression or activity of
the ccdB. Since it is controlled by the lac promoter, one can clone it using
lac
Iq
E. coli, which overexpress the lac repressor. However, the amount of
plasmid recoverable by this method is relatively low, possibly due to residual
expression killing some of the bacteria during culture. Alternatively, one may
clone it using bacteria with a stably integrated ccdA gene. ccdA is the
natural inhibitor of ccdB. Secondly, calculating the recombination efficiency
can be complicated. Normally, one compares bacterial counts between
DpnI-digested plasmid plated onto ampicillin plates to plasmid plated onto
ampicillin-chloramphenicol plates, using the same batch of competent cells.
Thus in theory, transformation efficiency should be similar. However, to
calculate recombination efficiency in the ccdB system, one must compare
colony counts from DpnI-digested plasmid transformed into ccdB-resistant
bacteria with plasmid transformed with ccdB-susceptible bacteria. If the
resistant bacteria are one strain, e.g. lac
Iq
or stably-integrated ccdA, and the

107
Figure 3-4. Alternative construct designs attempting to circumvent the
expression blocking effect. (A) The ccdB bacterial suicide gene was taken
from pZero-2 as the lac promoter (P
lac
) followed by a fusion of the lacZ α with
the ccdB gene, and blunt cloned between the bcl-1 and V(D)J RSS. In
theory, a recombination event between the bcl-1 and RSS deletes out the
suicide gene and allows colony growth when the recombined plasmid is
transformed into E. coli and plated onto LB-agar containing ampicillin.
Unrecombined plasmids retain the ccdB gene, which, after transformation
into E. coli, kills the bacteria and prevents colony growth. In practice, an
unacceptable number of colonies still grew despite apparently retaining the
ccdB gene. This could be due to point mutation, resistance, incomplete
expression or penetrance, etc. (B) pAT211 is the most sophisticated
construct design tested. pAT211C, diagrammed, was not actually
constructed but is simply pAT211 with the oop terminator inserted into the
ClaI site. Most of the modifications made to the original pGG51-derived
substrates had no apparent effect. The most beneficial effect seemed to be
obtained by modifying the sequences directly upstream of the antibiotic
resistance gene. See text for details.

108
Figure 3-4 continued

Transfect into RAG
+
cells
48 h, 37 ° C
bcl-1
P
lac

Recover, transform
into E. coli, plate
onto amp
V(D)J
☺
☺
☺
☺

A.
RAG
+
cell nucleus
lacZ-ccdB
bcl-1
P
lac

V(D)J lacZ-ccdB

B.
bcl-1
P
tac

oop cat
human
J
H
6
human
D
H
3
T7 M13r M13f
stop
codons
hammerhead
translational
enhancer
better
Shine-Dalgarno
SalI
BamHI XmaI
ClaI
MluI
SpeI XbaI NheI
pAT211C
XhoI
E µ
EcoRV EcoRI

109
susceptible bacteria are another, e.g. DH10B, then the transformation
efficiencies must be taken into account. Competencies may be inconsistent.
A better, but more expensive, approach is to use a lac
Iq
strain plated onto
ampicillin for the the resistant case, and the same batch of lac
Iq
competent
cells plated onto ampicillin and IPTG as the susceptible case. IPTG will
release the lac repressor protein from P
lac
, allowing expression of the ccdB
gene.
Unfortunately, most of this discussion is academic because, after
transfection, all of the resistant colonies sequenced apparently retained the
ccdB gene. It is possible that the transfection process introduced small
inactivating mutations into the ccdB gene, and/or an unacceptably high
proportion of E. coli are stochastically or naturally resistant to the gene.
Another approach for overcoming the expression blocking effect was
to improve the components of the existing design (Fig. 3-4B). I tested many
of these changes both sequentially and in combination by colony counts with
or without deletion of the oop terminator. Ideally, a perfect substrate would
allow growth of all colonies without the terminator and no colonies with the
terminator. Because cloning into and sequencing of the existing substrate
was cumbersome, I also introduced standard sequencing primer sites and
unique restriction sites around each of the components.
First, I replaced P
trp
with the tac promoter (P
tac
), which supposedly
induces transcription three-fold better than P
trp
(de Boer et al., 1983). The

110
idea behind this was that an increased overall transcription rate might
increase the number of transcripts pushing through the presumed blockage.
However, it had no effect. Next, I tried replacing the cat gene with the
aminoglycoside phosphotransferase (aph) gene, which confers resistance to
kanamycin, neomycin, and other aminoglycosides. I found it more difficult to
control the kanamycin concentration in the LB-agar plates to both allow
growth of recombined plasmids and prevent growth of unrecombined
plasmids, thus this came out actually worse than the cat gene. Presuming
that the bcl-1 RNA itself might be the cause of the expression blocking effect,
I cloned a hammerhead self-cleaving ribozyme between the bcl-1 sequence
and the antibiotic resistance gene. Supposedly, this ribozyme is 100%
efficient (Samarsky et al., 1999). The intention was that the ribozyme would
cleave the resistance gene RNA away from the inhibiting bcl-1 RNA.
However, it did not have any effect. A later publication reported that lack of a
5’ pyrophosphate actually accelerated degradation of the RNA in E. coli
(Celesnik et al., 2007).
Next, I tried inserting stop codons and a translational enhancer, and
modifying the Shine-Dalgarno sequence upstream of the antibiotic resistance
gene. The stop codons were intended to terminate any inadvertent upstream
translation that might interfere with translation of the antibiotic resistance
gene. A quick glance at operon literature actually showed that stop codons
for an upstream gene in an operon could actually be 3’ of the start codon for

111
a downstream gene. Modifying the sequences directly upstream of the start
codon proved to have large effects. The first translational enhancer and
Shine-Dalgarno sequence I tried, the g10-L ribosomal binding site from T7
(oligos AT248, AT250), actually conferred antibiotic resistance even when
P
tac
was deleted (Olins and Rangwala, 1989). I concluded that it might have
promoter activity. The second one I tried, from the E. coli atpE gene (oligos
AT260, AT261), was dependent on P
tac
, and permitted use of a bcl-1
fragment that previously blocked expression, the forward orientation of the
366 bp piece from pSH1-4. This is the most current design.
In an effort to improve recombination efficiency, I cloned in the human
RSS most frequently used in mantle cell lymphoma translocations, as well as
the immunoglobulin heavy chain core enhancer. The hope was that those
RSS might recombine better with the MTC, and that the enhancer might
enhance recombination in general. While not tested extensively, neither
appeared to have much effect. An idea I did not try was using I-SceI to
cause breakage instead of two RSS. In theory, the MTC break is
independent of the RSS break, and the I-SceI enzyme is likely more efficient
than the RAG complex. However, one would have to cotransfect I-SceI
enzyme expression vector and would also lose specificity to the G
0
and G
1

phases. Replication-induced breaks could be a problem.
I did not get around to two interesting modifications. The first was
improving the oop terminator to a double-terminator. The double-terminator

112
is reported to be 98.4% efficient in the correct orientation (Conboy, 2003).
More efficient termination would allow use of a lower concentration of
antibiotic, which would kill fewer of the E. coli actually transformed with
recombined plasmids. The second modification was adding a stabilizing
RNA 5’ to the bcl-1 sequence, e.g. the E. coli ompA (Belasco et al., 1986;
Celesnik et al., 2007; Hansen et al., 1994). If the presence of the bcl-1 in the
5’ UTR is destabilizing the resistance gene transcript, perhaps the ompA
RNA would counteract the effect.
In vitro transcription through the bcl-1
Examination of the sequences involved in the expression blocking
effect did not reveal any obvious additional start codons with nearby
upstream Shine-Dalgarno sequences, or any reading frame effect. In the
hopes that expression blocking could be a readout for something interesting
occurring at the bcl-1, we performed experiments to test various hypotheses.
The first hypothesis that came to mind was that the bcl-1 sequence might be
terminating the transcript, e.g. a non-B-DNA structure might impede the
progression of the RNA polymerase. To this ends, I performed in vitro
transcriptions on plasmids containing a 733 bp piece of the bcl-1 including
the MTC (Fig. 3-5A). We did not see any bands indicative of transcription
termination or pausing on supercoiled plasmid. A similar result was seen on
linearized plasmid, and as a positive control, a restriction cut within the bcl-1
MTC showed a clear band where transcripts terminated (not shown). We did

113
Figure 3-5. In vitro transcription through the bcl-1. (A) A 733 bp fragment of
the bcl-1 encompassing the MTC was cloned into the multicloning site of
pBluescript (pBS) in the forward (pAT13) or reverse (pAT14) orientations.
Large-scale preparations of these supercoiled plasmids were purified by
CsCl density gradient centrifugation; equal µg amounts were transcribed with
radiolabeled UTP and either T7, T3, or E. coli RNA polyermases according to
manufacturer’s directions; and the products were run on denaturing PAGE.
The overall intensity of the T7 and T3 lanes is greater than that of the E. coli
RNA polymerase lane, likely due to higher transcription rate of those
polymerases. The band observed in the E. coli polymerase lanes is
unrelated to the bcl-1 because it is present in the pBS lane. It is possible that
it is related transcription in the ColE1 origin elsewhere in the plasmid. The
same plasmids were used for R-looping studies. (B) After transcription with
radiolabeled UTP and T7 RNA polymerase, and treatment with RNaseA or
both RNaseA and RNaseH, products were run on agarose and imaged by
staining with ethidium bromide and UV transillumination. The second lane in
each set of four shows the RNA smear expected from transcription. In the
third lanes, RNaseA has degraded away the single-stranded RNA. A tail
behind the supercoiled position of pAT14 remains (indicated by the yellow
arrow). In the fourth lanes, RNaseH has been added in addition to RNaseA,
and the tail from the transcribed, RNAseA-treated pAT14 has disappeared.
RNaseH specifically degrades RNA that is hybridized to DNA, thus this would

114
suggest the presence of an RNA-DNA hybrid. (C) After drying the gel and
exposing it on a phosphorimager screen, the radioactive profile of the gel
reveals no radiolabeled uridine at the position of the tail of transcribed,
RNaseA-treated pAT14. This suggests that either there are too few uridines
incorporated into the RNA at the putative RNA-DNA hybrid to be detected, or
that there is simply no RNA-DNA hybrid there. If there truly is no RNA-DNA
hybrid, then it is strange that the tail disappears with RNaseH.

115
Figure 3-5 continued
A.

pBS

pAT13

pAT14

pBS

pAT13

pAT14

pBS

pAT13

pAT14
T7 E. coli T3

116
Figure 3-5 continued

B. C.
pBS

pBS + T7

pBS + T7 + RNAse A

pBS + T7 + RNAse A +
H

pAT13

pAT13 + T7

pAT13 + T7 + RNAseA

pAT13 + T7 + RNAseA
+ H

pAT14
pBS

pBS + T7

pBS + T7 + RNAse A

pBS + T7 + RNAse A +
H

pAT13

pAT13 + T7

pAT13 + T7 + RNAseA

pAT13 + T7 + RNAseA
+ H

pAT14

117
see some R-loop-like behavior after transcription through the bcl-1 in the
centromeric-to-telomeric orientation, but this is common with sequences
having a high guanine content on the untranscribed strand (Fig. 3-5B).
Strangely, the expected band was not observed in the radioactive profile of
the gel when using radiolabeled UTP to label the nascent RNA. It may be of
some value to cut out the band and examine its structure using bisulfite or P1
nuclease probing. If there truly is no RNA there, it would be interesting to
know the structure of the DNA created during the transcription process.
Other hypotheses for the blocking effect include termination
dependent on E. coli proteins such as rho, the bcl-1 transcript destabilizing
the RNA transcript, or the bcl-1 sequence preventing translation.
We additionally attempted to see if the bcl-1 terminated transcription in
eukaryotic cells by transfecting 293T cells with a construct where the bcl-1
sequence was cloned into the intron of a GFP gene. If the bcl-1 terminated
transcription, the mRNA would be terminated before reaching the second
exon, functional GFP would not be expressed, and the cells would not glow
green under UV light. However, the cells did glow green, therefore the bcl-1
was not terminating transcription in those constructs in eukaryotic cells. A
luciferase plasmid, pCLH22, was used as a transfection control.
Electrophoretic mobility shift assay
Certain non-B-DNA structures have altered mobility when
eletrophoresed through a gel. Curved DNA electrophoreses slower than

118
expected, as do structures bulkier than normal B-DNA such as quadruplexes
and triplexes (Sinden, 1994). EMSA experiments formed a part of the basis
for the bcl-2 MBR triplex hypothesis (Raghavan et al., 2005a; Raghavan et
al., 2004a).
Early on, we observed an interesting “shifted” band distinct from the
expected main band when a 354 bp piece of the bcl-1 containing the MTC
was PCRed and run on agarose or acrylamide (Fig. 3-6). We believed that a
small percentage of the DNA was stochastically adopting a non-B-DNA
structure which was resolving on gel. The shifted band was finicky but had a
few strange but consistent properties. The band moved faster than expected
when ethidium was present in the buffer and gel, and slower without (Fig. 3-
6A). We thought that ethidium did not intercalate well into the structure, thus
all the duplex DNA, including the ladder, were slowed down in ethidium.
Different temperature incubation schemes had various effects, so we thought
that temperature was affecting the structure (Fig. 3-6B). Most convincingly,
the shifted band was cleavable by P1 nuclease, which is a single-strand-
specific DNA endonuclease (Fig. 3-6C). This latter result was very much like
that observed for the bcl-2, though the bcl-2 gel shifts were all done on 23 cm
native acrylamide gels run slowly to keep the temperature down while the
bcl-1 were done on agarose or 5 cm mini-PAGE (Raghavan et al., 2004a).
Attempted bcl-1 gel shifts on the 23 cm gels was never successful. We
presumed that the temperature and running time were dissipating the band.

119
Figure 3-6. Gel shift of bcl-1 PCR fragment. A shifted band distinct from the
main band is present upon electrophoresis. Its position is marked with red
arrows. It is likely an artifact of incomplete extension. (A) When run with
ethidium, it is faster than the main band. (B) When run without ethidium, it is
slower than the main band. In panels A and B, samples marked 1 contain 10
mM EDTA and have been incubated at 80°C for 10 minutes, while those
marked 2 have been incubated on ice for 10 minutes. (C) The quantity of
the shifted band changes with various temperature incubation schemes.
Note that the lanes are equally loaded, but much of the DNA is left in the gel
wells when incubated at 100° C. (D) P1 nuclease cleaves the shifted band
into smaller species marked with green arrows. With increasing amounts of
P1, the shifted band disappears and the smaller species increase in intensity.

120

344 bp
A.
C.
344 bp
run with 0.5 µg ethidium bromide in
gel and running buffer

1 2
no ethidium bromide during run,
poststained

1 1 2
10’ incubation at

then

25’ incubation at
-20 ° C -20 ° C

ice ice

22 ° C 22 ° C

22 ° C 37 ° C

22 ° C 55 ° C

22 ° C 72 ° C

22 ° C 80 ° C

22 ° C 90 ° C

22 ° C 100 ° C

100 ° C 100 ° C
no ethidium bromide during run, poststained
B.
Figure 3-6 continued

121
Figure 3-6 continued

D.
units of P1 nuclease
1 kb ladder

0 U

0.00001 U?

0.0001 U

0.001 U

0.01 U

0.1 U

0.4 U

0.6 U

122
Many variations were tried, including phenol-chloroform extraction,
incubation with KCl; SssI-methylation; dialysis to pH 7.2, 7.7, or 8.0; running
it in gels at pH 7.2, 7.7 or 8.0; etc. They had little or no effect. Various
fragment sizes had the shifted band, including 243 bp, 310 bp, 415 bp, and
528 bp fragments in addition to the 354 bp; the shift was most apparent in
the 354 bp and 410 bp. MgCl
2
in the gel and buffer seemed to have a similar
effect as ethidium; in mini-PAGE, the shift would move faster than the main
band when 10 mM MgCl
2
was added, and was slower than the main band
without. Direct TA cloning of the band sequenced as full-length product, but
it is likely that there was contamination from the main band.
We were able to show that the shift was not single-stranded DNA by
two methods. First, ssDNA of each strand was generated by primer
extension of the PCR product, running it on a 23 cm denaturing PAGE gel,
imaging by x-ray film, cutting out the correct length band, and running it side
by side with the normally-shifted PCR product. The ssDNA ran faster than
both the main band and the shifted band. Second, each primer was
separately labeled and PCRed with the other primer cold, then run on gel.
The radioactive profile of both reactions showed both a main band and a
shift, demonstrating that both primers, and therefore both strands, were
present at both the main band and the shifted band.
However, when we ran an almost identical fragment of the bcl-1
restriction-cut out of plasmid rather than PCR-amplified, there was no shifted

123
band. This led us on the path that the shift might be an artifact of PCR rather
than a stochastic non-B-DNA structure. The experiment which should have
been done all along was running radiolabeled PCR product on denaturing
PAGE, which showed multiple incompletely extended products by two
different forward primers (Fig. 3-7). The amount roughly corresponds to that
seen in the shifted band, and explains the P1 sensitivity of the shifted band.
It is likely that P1 was simply degrading the single-stranded region of the
incompletely extended product, leaving shorter double-stranded regions.
This incomplete extension phenomenon is partly related to the
polymerase, as we did not see it with any of the RNA polymerases in the
transcription gels (Fig. 3-5). Most of the time we used a standard Taq
polymerase, but we also observed it with Vent exo
-
polymerase.
More definitive evidence would come from doing PCR with
radiolabeled primers, cutting out the shifted band from the native PAGE gel,
then running it on denaturing PAGE. If the band were entirely composed of
shortened products, that would confirm its identity.
Chemical probing methods
The original report of a non-B-DNA structure at the bcl-2 MBR used
three chemical probes as evidence – sodium bisulfite (NaHSO
3
), potassium
permanganate (KMnO
4
), and osmium tetroxide (OsO
4
) (Raghavan et al.,
2004b). As shall be demonstrated in chapter 4, we do not really understand
the nature of bisulfite reactivity in double-stranded DNA, and the nature of

124

Figure 3-7. Incomplete extension of PCRs used for gel shift. Equal-size
fragments of the bcl-1 and the ampicillin resistance gene (amp
r
) were PCRed
from the same plasmid, using combinations of 5’ radiolabeled and cold
primers, and run on denaturing PAGE. Paused or incompletely extended
products are observed in the lanes where the bcl-1 forward primer is
radiolabeled. Vertical lines indicate the position of the MTC for the forward
primer.
354 bp amp
r
both

354 bp bcl-1 both

354 bp bcl-1 forward

354 bp bcl-1 reverse

410 bp amp
r
both

410 bp bcl-1 both

410 bp bcl-1 forward

410 bp bcl-1 reverse
lableled primer(s)
sequence
length

125
KMnO
4
and OsO
4
reactivity may not be fully understood either. Technical
issues also hamper these methods.
The asymmetry of bisulfite reactivity between the two strands at the
bcl-2 MBR was used in support of an intramolecular triplex structure
(Raghavan et al., 2005a). However, bisulfite probing data for the bcl-1 MTC
reflects symmetrical reactivity. When bisulfite reactivity at the bcl-2 MBR was
originally compared to other regions, a cutoff of seven out of ten consecutive
cytosine deaminations was used to show that the MBR is reactive while other
regions are not. This cutoff is artificial and has no structural meaning, thus it
is not used in any subsequent analyses. In chapter 4, we show that
consecutive cytosines have a high propensity to react for unidentified
reasons. Background reactivity in switch regions approaches 25%, and
mostly occurs at stretches of consecutive cytosines (Dr. Kefei Yu,
unpublished observation).
KMnO
4
and OsO
4
oxidize the 5, 6 double bond of single-stranded
thymines, resulting in piperidine-cleavable adducts detectable on gel (Fig. 3-
8). They react at very low concentrations and neutral pH, and can be
diffused into cells. However, the true nature of their reactivity has not been
investigated much further than “single-stranded thymines.” The conditions
are typically not standardized to the specific structure or its expected
frequency, thus analysis can be ad hoc, especially if insufficient controls are
performed. The simple interpretation “it reacts therefore it is single-stranded

126

Figure 3-8. KMnO
4
and OsO
4
chemical probing. KMnO
4
and OsO
4
oxidize
the 5, 6 double bond of single-stranded thymines, resulting in piperidine-
cleavable adducts detectable on gel.

127
and it does not react therefore it is double-stranded” may not be correct. If
the concentration, time, temperature, or buffer are modified slightly, the
cleavage pattern can change; and if piperidine treatment is done too long,
extra cleavages may occur. To detect T:G mismatches, 100 mM KMnO
4
is
used at 25° C for 10 minutes with strong indications not to go beyond
(Tabone et al., 2006). Cruciforms were cleaved at 1.6-3.2 mM OsO
4
at 25° C
for 3-30 minutes (Lilley and Palecek, 1984). The bcl-2 MBR was cleaved by
40 mM KMnO
4
at 4° C for 30 seconds or 5 mM OsO
4
at room temperature for
15 minutes at room temperature, and the cleavage patterns by KMnO
4
and
OsO
4
were extremely different (Raghavan et al., 2004b). Another major
problem was that no positive or negative control DNA structures were done
to verify the conditions used. In my hands, it was difficult to get KMnO
4
to
behave the same way twice, but this could have been due to the detection
methods we used, primer extension and LM-PCR.
Difficulties in performing successful primer extensions and LM-PCRs
prevented proper detection of chemical probe breakage at the bcl-1. The
incomplete extension problem is a major problem of primer extension (Fig. 3-
5). In genomic DNA, there are probably hundreds of priming sites, so one
cannot do a simple primer extension to detect breakage. One solution is to
use LM-PCR to increase the specificity. However, this increases the
complexity of the system and requires much more optimization of primers,
cycling temperatures, and polymerases (Ausubel, 1987; Vigneault and

128
Drouin, 2005). When doing restriction-cut genomic DNA controls, I always
found unexpected bands throughout the gel. Moreover, even a restriction-cut
control cannot account for all mispriming. Without sequencing all the bands,
one cannot be sure that they are all from the bcl-1 MTC. A mispriming region
might not show up on the restriction-cut control because it is too far from a
restriction site. However, when KMnO
4
or OsO
4
is added, they react all over
the genome and mispriming regions show up on the gel.
Enzymatic probing and direct labeling methods
Another cleavage method for detecting non-B-DNA structure is to use
structure-specific enzymes. I found them to be more consistent and specific
than chemical probes. P1 nuclease cleaves single-stranded DNA at neutral
pH. The structure-specificity of the RAG complex has not been fully tested,
but because it is expressed in pre-B cells at the same time as the
translocations of interest, it is the most relevant enzyme to test.
In order to avoid primer extension and LM-PCR artifacts, I decided to
use the direct labeling method (Fig. 3-9). This limited analyses to plasmid
DNA, but gave much clearer results. For these experiments, I mostly used
pAT13, which was constructed by cloning a 733 bp piece of the bcl-1 into the
SalI site of pBluescript. In order, the relevant restriction sites are EcoRI,
HindIII, SalI containing bcl-1, XhoI, then KpnI. One first cuts at one of the
inner restriction sites, HindIII or XhoI, dephosphorylates the ends with SAP,
and labels either the top strand by T4 PNK or the bottom strand by Klenow

129

Figure 3-9. Direct labeling method of DNA breakage detection. The inner
restriction digest opens the plasmid at one site, exposing two ends for 3’ end-
fill labeling by Klenow exo
-
, or 5’ labeling by T4 PNK after dephosphorylation.
Radiolabel is denoted by the green dots. The outer digest cleaves away one
of the labels as a short fragment which runs off the gel. The remaining label
allows one to measure distance to a nick or cut within the bcl-1 sequence.
To measure from the other side of the bcl-1, one uses XhoI and KpnI. This
method circumvents the artifacts associated with primer extension and LM-
PCR and allows for accurate quantitation.

EcoRI
HindIII
bcl-1
EcoRI
HindIII
KpnI
XhoI
bcl-1
HindIII
KpnI
XhoI
EcoRI
HindIII
bcl-1
HindIII
KpnI
XhoI
inner restriction digest,
(dephosphorylation), labeling
outer restriction

130
exo
-
end-filling. Then, one cuts at the relevant outer restriction site, EcoRI for
HindIII and KpnI for XhoI, to cleave away one of the labels as a short
fragment which runs off the gel. The remaining label allows one to measure
distance to a nick or cut within the bcl-1 sequence.
We first probed supercoiled pAT13 with P1 nuclease, then did direct
labeling and ran the products on denaturing PAGE (Fig. 3-10A). Many bands
appeared but none particularly strong, relevant to the MTC, and consistent
between the two labelings. Because each strand is represented twice on the
gel – the 5’ end by one cut site with one labeling method, and the 3’ end by
the other cut site with the other labeling method – any specific nicks or cuts
within the bcl-1 should be represented in both sets of lanes. We concluded
that P1 was nicking all over the plasmid and those nicks were being labeled
and showing up on the gel as background. A better way to handle this would
have been to cut the products with the opposite outer restriction site, run
them on agarose gel, cut out the correct size band, and then run on
denaturing PAGE. This would at least ensure that the bands were from the
fragment of interest. One would lose the plasmids with double-strand cuts
within the bcl-1 but the nicked species would remain.
To avoid the excess background, we decided to direct label before
nicking with P1, but in the process losing the supercoiling (Fig. 3-10B). The
human genome is not heavily supercoiled so this is not much of a loss. As
expected, the background nicking reduced but still there was nothing

131
Figure 3-10. P1 and the RAG endonuclease probing of the bcl-1. The black
vertical lines in the gel denote the position of the MTC for that set of lanes.
(A) Supercoiled pAT13 was treated with various amounts of P1 nuclease,
then direct labeled and run on gel. (B) pAT13 was direct labeled, then
treated with various amounts of P1 nuclease. (C) Direct-labeled pAT13 was
treated with various amounts of RAG complexes.

132
Figure 3-10 continued

A.
units of P1 strand end

0

0.00625

0.0125

0.05

0.1

0.5

1
0

0.00625

0.0125

0.05

0.1

0.5

1
0

0.00625

0.0125

0.05

0.1

0.5

1
0

0.00625

0.0125

0.05

0.1

0.5

1
HindIII

top (T4 PNK)
HindIII

bottom (Klenow)
XhoI

bottom (T4 PNK)
XhoI

top (Klenow)
pAT 13, supercoiled, P1 nuclease

133
Figure 3-10 continued

B.
units of P1 strand end

0

0.00625

0.0125

0.05

0.1

0.5

1
0

0.00625

0.0125

0.05

0.1

0.5

1
0

0.00625

0.0125

0.05

0.1

0.5

1
0

0.00625

0.0125

0.05

0.1

0.5

1
HindIII

top (T4 PNK)
HindIII

bottom (Klenow)
XhoI

bottom (T4 PNK)
XhoI

top (Klenow)
pAT 13, linearized, P1 nuclease

134
Figure 3-10 continued

C.
end
strand
RAG complex species
concentration (ng/ µL)
0

2.5

5

10
XhoI

bottom
(T4 PNK)

human
XhoI

bottom
(T4 PNK)

mouse
HindIII

top
(T4 PNK)

human
HindIII

top
(T4 PNK)

mouse
pAT 13, linearized, RAG endonuclease
0

2.5

5

10
0

2.5

5

10
0

2.5
1
5

10

135
interesting to note. The overall lane intensities fell with increasing
concentrations of P1, presumably due to P1 clipping away the label at frayed
ends. The effect was more pronounced for the T4 PNK label because it
labels directly at the terminus, while Klenow labeled a few bases in from the
ends.
Similar gels involving the RAG complex also yielded little (Fig. 3-10C).
A specific band did appear in the bottom strand of the MTC, but more
consistent with a cryptic site nick than a non-B-DNA structure. The RAG
complex turned out to be active in transcription buffer, so we tried concurrent
T7 transcription through the bcl-1. It did not improve nicking; in fact, it
seemed to reduce it. Similar experiments on plasmids with the bcl-2 MBR
did not show nicking either, with or without transcription.
Hydrolytic deamination assay
In an effort to link non-B-DNA structure to translocations, and with the
revelation that breakpoints were occurring at CpGs, we hypothesized that
DNA structure was causing the CpGs at the bcl-1 MTC and bcl-2 MBR to
hydrolytically deaminate faster than neighboring regions. This connected
with the bisulfite data suggesting that strings of consecutive cytosines might
be breathing faster than other sequences. The bcl-1 MTC is adjacent to
such a region while two peaks of the bcl-2 MBR fit the description.
To test this, we decided to use an in vitro deamination assay whereby
a cytosine deamination event on plasmid would confer kanamycin resistance

136
to E. coli transformed with it (Fig. 3-11). We hoped that placing the bcl-1 and
bcl-2 sequences next to a specific cytosine would increase its deamination
rate. Prior assays detected deaminations at sites within the coding region of
the gene (Chen and Shaw, 1993; Frederico et al., 1990; Frederico et al.,
1993; Shen et al., 1994). We, on the other hand, wished to control the
sequence around a specific deamination site of interest. Our chosen site
was at the start codon, and the bcl-1 and bcl-2 sequences would be right
next to it, encoding the first ten amino acids of the gene. Like previous
groups, we used the strain NR8052, which is ung
-
and therefore cannot
repair the uracils which result from cytosine deamination.
The first problem encountered was that initial constructs did not confer
kanamycin resistance, even with a functional start codon. After struggling
with various modifications for weeks, an accidental in-frame 10 amino acid
deletion was created which conferred resistance. The next problem was
that, even with the start codon mutated to GTG, the bacteria were still
growing on kanamycin at a 0.3-1% rate. This is far too high a background to
be detecting events expected at a 0.0001% rate or less. Increasing the
kanamycin concentration was ineffective, as it killed a high proportion of
those with functional start codons. The solution was to mutate the start
codon to ACG instead. The background reduced to less than 0.01%. We
believe that the high background with GTG but not ACG is due to the fact
that GTG is actually a rare start codon. At a 0.01% background, one must

137
A.

B.
substrate colonies screened colonies with
deaminated
substrate
deamination rate at
37° C over 14 days
lacZ (Frederico et al., 1990) – – 8 ⋅ 10
-7

ACGCGCGC 3,000,000 0 <3 ⋅ 10
-7
ACGGGCCC 2,500,000 2 8 ⋅ 10
-7

ACG-bcl-1 MTC downstream 4,600,000 1 2 ⋅ 10
-7

ACG-bcl-2 MBR peak 1 3,700,000 2 5 ⋅ 10
-7

Figure 3-11. Deamination assay. (A) Deamination of the cytosine within the
mutation start codon, ACG, results in AUG. In ung
-
bacteria, this is not
repaired and after replication results in a functional start codon, ATG.
Bacteria transformed with ACG substrates fail to grow on kanamycin while
those with ATG substrates grow into colonies. (B) The four substrates tested
did not differ significantly in their deamination rates. For comparison, the first
reported rate is 7 ⋅ 10
-13
/s for double-stranded DNA, which translates to 8 ⋅
10
-7
over 14 days (Frederico et al., 1990). They found that various sites
differed by less than three-fold.

ACG
P
tac

bcl-1 kan
r
gene
☺
transform into NR8052
(ung
-
), plate onto kan
AUG
P
tac

bcl-1 kan
r
gene

138
still screen about a hundred colonies to see one real event. By adjusting the
kanamycin concentration, limiting the incubation time, and only picking the
largest colonies, this can be reduced to screening about twenty colonies per
actual event. If screening large numbers, restreaking is probably the best
strategy.
In addition to the bcl-2 MBR peak 1 and the downstream region of the
bcl-2 MTC, we tested the deamination rates for substrates with the sequence
ggccc and cgcgc directly downstream of the start codon. These are similar
to our bisulfite substrates in chapter 4. Substrates were purified by cesium
chloride density gradient centrifugation, resuspended in 100 mM KCl, 50 mM
HEPES-KOH, 10 mM MgCl
2
, 1 mM EDTA, pH 7.4, allowed to sit for 14 days
at 37° C, transformed into electrocompetent NR8052 E. coli, plated onto
ampicillin or kanamycin, and the colony counts compared. Kanamycin-
resistant colonies were sequenced to confirm the deamination event.
Positive controls with functional start codons were done to compare growth
efficiencies between the different constructs. They were all roughly similar.
The results are summarized in Fig. 3-11B. Deamination rates for all our
substrates were not significantly different from the previously-reported rate
for double-stranded DNA (Frederico et al., 1990). It appears, then that the
hypothesis is incorrect. Further screening to determine more precise rates is
unnecessary, as the 100-fold increased translocation rate at the bcl-1 and

139
bcl-2 cannot be explained by small differences in hydrolytic deamination
rates.

140
Chapter 4. Bisulfite as a probe for DNA structure
Summary
Bisulfite is the only known sequenceable probe for nucleic acid
structure, and while it is reactive at the bcl-1 major translocation cluster
(MTC) – a zone of 150 bp responsible for 30% of the bcl-1 translocations that
occur in virtually all mantle cell lymphomas (MCL) – the structural reasons
underlying this reactivity are unclear. Here we report a range of bisulfite
reactivity on duplex oligonucleotide sequences structurally characterized by
other groups using other methods. We find that bisulfite reactivity correlates
broadly with relative propensity to breathe and with relative propensity to
form a B-A-intermediate structure. The significance for analysis of duplex
DNA structural deformations is discussed.
Introduction
While 11q13 breakpoints from t(11;14)(q13;q32) translocations in
mantle cell lymphomas (MCL) can occur over a >200 kb region called bcl-1,
30% occur in a 150 bp zone termed the major translocation cluster (MTC)
(Bertoni et al., 2004). Based on previous work involving a very similar
translocation region, the bcl-2 major breakpoint region (MBR) in
t(14;18)(q32;q21) from follicular and diffuse large B-cell lymphomas, we
hypothesized that this 100-fold enrichment for translocations in the MTC is
due to formation of a non-B DNA structure with increased propensity to

141
double-strand breakage (Raghavan et al., 2004b). To determine the
presence or absence of such a structure, we employed bisulfite probing.
The structure of a particular sequence of DNA is often inferred from its
pattern of reactivity with structure-specific chemical or enzymatic probes.
After treatment with such a probe, the DNA is typically cleaved at reactive
sites, and the positions of these cleaved sites are calculated from the
fragment lengths observed using gel-based detection methods such as direct
labeling, primer extension, and ligation-mediated PCR.
Unlike cleavage-based methods, bisulfite converts single-stranded,
unmethylated cytosines into uracils, which, in the context of DNA, can be
cloned and sequenced (Gough et al., 1986; Raghavan et al., 2004b;
Raghavan et al., 2006; Yu et al., 2003) (Fig. 4-1). This is technically much
easier to perform and reproduce, ensures specificity to the sequence of
interest, and allows one to examine specific clones and thereby resolve
multiple structures. Previous work has demonstrated the utility of bisulfite in
analyzing conceptually simple, stable structures such as R-loops and
cruciforms. For these structures, the reasons for the observed pattern of
bisulfite reactivity are very clear: they contain completely unpaired cytosines,
and mainly those cytosines react.
We also observe bisulfite reactivity at the bcl-1 MTC, but it does not
seem to form any simple structures with completely unpaired cytosines.
Reactivity is largely symmetrical between the two strands, increases with

142

Figure 4-1. Mechanism of bisulfite-catalyzed cytosine deamination. The
primary reactive species is thought to be the sulfite ion, which attacks the
5,6-double bond of a single-stranded, unmethylated cytosine and adds to the
6 position. A hydrogen atom is added to the 5 position as the ring loses its
aromaticity. Equilibrium is established quickly, estimated to be within 5
minutes. After protonation of N3, water can deaminate the amino group on
the 4 position. This is the rate-limiting step of the reaction. Substrates are
typically treated with 300 mM NaOH to remove the sulfonate group before
PCR and sequencing.

143
supercoiling, and does not depend on the source of the DNA. Therefore the
reasons for reactivity in this region are unclear, and to determine the true
nature of bisulfite reactivity on presumably double-stranded DNA, we
performed bisulfite probing on oligonucleotides containing sequences
previously characterized by other groups using other methods.
Materials and methods
Bisulfite on oligonucleotides
DNA oligos were ordered from Operon Biotechnologies and oligos
containing 2’-O-methyl RNA were ordered from IDT. All oligos were purified
on denaturing PAGE by standard protocol (Ausubel, 1987). 650 pmol of gel-
purified long strand oligo was mixed with 1,300 pmol of gel-purified short
strand oligo in 65 µL TE with 100 mM NaCl in a screw-cap tube, and
annealed by boiling the tube in 1 L water for 5 minutes, then allowing to cool
at room temperature overnight. Bisulfite mixture was prepared by mixing 0.5
g NaHSO
3
(Sigma S-8890) with 0.525 mL of ddH
2
O and 0.2626 mL of 2 M
NaOH, then mixing 457.5 µL of the resulting solution with 12.5 µL of 20 mM
hydroquinone. 15 µL of annealed oligo was mixed with 235 µL of bisulfite
mixture and put into an air incubator at 37° C for 16 hours. After treatment,
oligos were serially precipitated with ethanol until the pellet could be
dissolved in 20 µL TE, and 5 µL was run on 5% native mini-PAGE along with
small amounts of component single-stranded oligos and untreated substrate
more than one lane away. The gel was stained with 0.25 µg/mL ethidium

144
bromide in deionized water for 20 minutes with light shaking, transferred to a
UV transilluminator overlain with clean plastic wrap, and the double-stranded
form cut out with a clean razor blade. The gel slice was treated 4 times with
0.3 M NaOH for 5 minutes at room temperature, washed 4 times with TE for
5 minutes at room temperature, then crushed and soaked in 200 µL PAGE
diffusion buffer (0.5 M ammonium acetate, 10 mM magnesium acetate, 0.1%
SDS, 1 mM EDTA, pH 8.0) at 37° C, 250 rpm for >8 hours. The supernatant
was precipitated with ethanol and resuspended in 20 µL TE. 0.5 µL was
used for a 10 µL PCR with primers AT131 and AT27 using NEB Taq DNA
polymerase using 30 cycles of 94° C, 0:30; 55° C, 0:30; 72° C, 0:30; with an
initial denaturation 94° C, 2:00 and final extension 72°C, 2:00. Primers are
designed to amplify only the long strand and do not anneal to any cytosines
on the long strand, preventing bisulfite conversion PCR bias. PCR products
were checked by running on 5% native PAGE, Invitrogen topo TA cloned,
and individual molecules sequenced on a Li-Cor IR
2
DNA analyzer using
M13 forward primer according to manufacturers’ directions. Molecules with
bisulfite conversions at every or all but one cytosine were interpreted as
having become single-stranded either before or during bisulfite treatment and
not used for analysis. 16 molecules of data were obtained for each PCR, two
PCRs were done for each bisulfite treatment, and two bisulfite treatments
were done for each substrate, giving 64 molecules of data per substrate.
Substrates are listed long strand-short strand as follows: GGGCCC, AT160-

145
AT161; GGCGCC, AT162-AT163; GGCC, AT164-AT165; GCGC, AT166-
AT167; CGCG, AT168-AT169; GCGCGC, AT183-AT184; 2’-O-methyl-
GGGCCC, AT183-AT259; 2’-O-methyl-GCGCGC, AT160-AT264. Oligo
sequences are as follows: AT160,
gtggggttattgtgggtgtacctgcgttcatgggcccatgcgatccttgaaggaatttggagagaggggt;
AT161, tcaaggatcgcatgggcccatgaacgcaggt; AT162,
gtggggttattgtgggtgtacctgcgttcatggcgccatgcgatccttgaaggaatttggagagaggggt;
AT163, tcaaggatcgcatggcgccatgaacgcaggt; AT164,
gtggggttattgtgggtgtacctgcgttcatggccatgcgatccttgaaggaatttggagagaggggt;
AT165, tcaaggatcgcatggccatgaacgcaggt; AT166,
gtggggttattgtgggtgtacctgcgttcatgcgcatgcgatccttgaaggaatttggagagaggggt;
AT167, tcaaggatcgcatgcgcatgaacgcaggt; AT168,
gtggggttattgtgggtgtacctgcgttcatcgcgatgcgatccttgaaggaatttggagagaggggt;
AT169, tcaaggatcgcatcgcgatgaacgcaggt; AT183,
gtggggttattgtgggtgtacctgcgttcatgcgcgcatgcgatccttgaaggaatttggagagaggggt;
AT184, tcaaggatcgcatgcgcgcatgaacgcaggt; AT259,
tcaaggatcg(caugcgcgcaug)
2’-O-methyl RNA
aacgcaggt; AT264,
tcaaggatcg(caugggcccaug)
2’-O-methyl RNA
aacgcaggt; AT27,
acccctctctccaaattcct; AT131, gtggggttattgtgggtgt.
Bisulfite on genomic DNA and pSH9
Genomic DNA from Reh cells was probed with bisulfite as described
previously, using the same bisulfite mixture as above (Raghavan et al.,

146
2004b). PCR was done using primers SH5, ataaggctgctgtacacatc, and SH6,
ggaggaacgctaaccaagcc, using 35 cycles of 94° C, 1:00; 58° C, 0:45; 72° C,
1:30; with an initial denaturation 95° C, 5:00 and final extension 72° C, 4:00.
The 7.6 kb plasmid pSH9 was constructed by cloning a 1,340 bp piece
of the bcl-1 from Reh cells into the SalI site of pGG51, using the PCR
primers SH26, agctgtcgacaccgcggctcaacccttcacct, and SH27,
agctgtcgacggtggtgcttgaagttgagatg, and purifying by cesium chloride density
gradient centrifugation. 1.5 µg was used for bisulfite probing performed
similarly to that for genomic DNA. Linearized plasmid was obtained by SfiI
digestion according to New England Biolabs instructions, followed by running
on 0.8% low-melt agarose, cutting out the linearized band, extracting with
phenol twice and chloroform once, and precipitating.
Results
Bisulfite reactivity correlates broadly with relative propensity to breathe.
Our first thought was that bisulfite reactivity may be an indication of
breathing, as bisulfite is considered to be a single-strand specific probe.
Breathing refers to the transient opening and closing, or unpairing and
repairing, of individual base pairs. Based on
1
H NMR measurements of
imino proton exchange, Dornberger et al. reported base pair lifetimes and
equilibrium constants for C:G pairs in a matched set of palindromic 10-mer
and 12-mer sequences: catGGGCCCatg, catGGCGCCatg, catGGCCatg,
catGCGCatg, and catCGCGatg (Dornberger et al., 1999).

147
We probed these sequences and another matched sequence,
catGCGCGCatg, with bisulfite. As required for bisulfite probing, we
lengthened the ends to improve stability at 37° C and lengthened one strand
to provide a place for PCR primers. As a result, each substrate is composed
of a long strand and a short strand, but only the long strand can be PCR-
amplified and sequenced, and thus only long strand C to U conversions can
be detected. The single-stranded arms contain no cytosines in order to
prevent any PCR bias towards less-reacted molecules, and this arrangement
also allows us to see reactivity at the ends of the double-stranded portion.
Cytosines within individual molecules were tallied before summation into
larger plots (Fig. 4-2).
Bisulfite reactivity is greatest at the edges of the duplex and
decreases further internally – consistent with single-strandedness due to end
fraying, i.e. breathing at free DNA ends (Fig. 4-3).
Among the 12-mers, GGGCCC is most reactive, followed by
GGCGCC, then GCGCGC; and among the 10-mers, GGCC is more reactive
than GCGC or CGCG (Fig. 4-3). Clearly, reactivity does not correlate with
the base pair lifetime τ
op
, which is inversely related to the forward kinetic
base pair opening rate k
op
.
However, overall reactivity does appear to correlate with overall αK
d
,
the equilibrium constant for base pair opening. Over the five substrates, they
follow the same trend, with GGGCCC having the highest average αK
d
and

148
Figure 4-2. Bisulfite reactivity of individual molecules of oligonucleotide
substrates. Each row of circles represents an oligonucleotide molecule
treated with bisulfite, cloned, and sequenced. Filled circles denote cytosines
which deaminated to uracils, while empty circles denote cytosines which
remained as cytosines and did not deaminate. The “*” indicates that one of
the oligonucleotides in the substrate is a 2’-O-methyl RNA. The substrate
structures are shown in Figs. 4-3 and 4-5.

149
Figure 4-2 continued

GGGCCC
AT160-AT161

GGCGCC
AT162-AT163

GGCC
AT164-AT165

GCGC
AT166-AT167

150
Figure 4-2 continued

CGCG
AT168-AT169

GCGCGC
AT183-AT184

GGGCCC*
AT183-AT259

GCGCGC*
AT160-AT264

151
Figure 4-3. Bisulfite reactivity on dsDNA correlates broadly with relative
propensity to breathe. Bisulfite reactivities for each cytosine on each duplex
DNA substrate are shown in A-F. Boxed in red are the sequences of interest,
and each black circle denotes a bisulfite-catalyzed C to U conversion
detected after cloning and sequencing. 30 nt of the long strand sequence has
been truncated for space. (G) Base pair lifetimes and equilibrium constants
as determined experimentally by Dornberger et al. (Dornberger et al., 1999).
In cases where two positions are listed for one τ
op
and αK
d
, NMR peaks for
the individual component pairs could not be resolved.

152
Figure 4-3 continued

A. catGGGCCCatg

B. catGGCGCCatg C. catGCGCGCatg
D. catGGCCatg E. catGCGCatg F. catCGCGatg

153
Figure 4-3 continued

G.
Sequence Position(s) Correspon-
ding Cs
τ
op
(ms) average
τ
op
(ms)
αK
d
( ⋅ 10
7
) average
αK
d
( ⋅ 10
7
)
catGGGCCCatg G4 GGGCCC 3.9 ± 0.3 3.6 ± 0.7 14.0 ± 0.1 11.2 ± 0.3
G5 / G6 GGGCCC 3.5 ± 0.2 9.8 ± 0.1
catGGCGCCatg G4 / G5 GGCGCC 1.4 ± 0.8 3.2 ± 3.3 7.5 ± 0.3 6.7 ± 1.3
G6 GGCGCC 6.8 ± 1.7 4.8 ± 0.7
catGGCCatg G4 GGCC 4.6 ± 2.5 4.2 ± 3.9 5.3 ± 0.8 5.2 ± 1.3
G5 GGCC 3.7 ± 1.4 5.0 ± 0.5
catGCGCatg G4 GCGC 6.4 ± 1.9 9.3 ± 3.9 5.1 ± 0.3 4.2 ± 0.5
G5 GCGC 12.2 ± 2.0 3.3 ± 0.2
catCGCGatg G4 CGCG 4.3 ± 1.5 4.2 ± 4.2 6.5 ± 0.4 4.4 ± 1.4
G5 CGCG 4.0 ± 2.7 2.3 ± 1.0
(Dornberger et al., 1999)

154
GCGC / CGCG the lowest. Reactivity and αK
d
do not entirely correspond
when comparing individual base pairs rather than substrates as a whole,
though some of this could be due to the extra sequence around the bisulfite
substrates as well as the inability to resolve certain NMR peaks. Because
αK
d
accounts for the reverse kinetic base pair closing rate k
cl
as well as k
op
,
i.e. αK
d
= k
op
/k
cl
, it is a measure of the steady-state frequency at which the
bases are unpaired.
This naturally leads to the hypothesis that bisulfite is detecting
breathed-open bases, and cytosines with a higher propensity to be
transiently unpaired react more with bisulfite.
Bisulfite reactivity correlates with relative propensity to form B-A-intermediate
structure.
The x-ray crystal structures for catGGGCCCatg and catGGCCatg, as
well as an NMR solution structure for catGGCCatg, have been reported
(Dornberger et al., 1998; Goodsell et al., 1993; Ng et al., 2000).
Interestingly, GGGCCC forms a B-A intermediate type structure, while
GGCC has altered stacking characteristics that result in a bend.
In order to explore the possibility that bisulfite could react with B-A
intermediate structures, we obtained circular dichroism (CD) spectra for the
matched 12-mers (Fig. 4-4). CGCATATATGCG is a typical B-form DNA,
while A
16
U
16
is a typical A-form RNA. As one goes from catGCGCGCatg to

155
catGGCGCCatg to catGGGCCCatg, the structure acquires more and more
A-form character.
We became interested in whether bisulfite would react with purely A-
form DNA, and thus we probed hybrid duplex DNA-2’-O-methyl RNA
substrates (Fig. 4-5). The bulky 2’-O-methyl groups force the duplex into A-
form. Bisulfite reactivity on both the catGGGCCCatg and catGCGCGCatg
substrates vanished, thus fully A-form structure is impervious to bisulfite,
consistent with previous data for RNA-DNA hybrids in R-loops (Roy et al.,
2008; Yu et al., 2003).
Bisulfite reactivity at the bcl-1 MTC.
Returning to an earlier question, we re-examined bisulfite reactivity at
and around the bcl-1 MTC. Bisulfite reactivity occurs primarily at the
downstream edge of the bcl-1 MTC (Fig. 4-6). Two peaks of reactivity are
observed. The first and smaller peak occurs at the semi-palindromic
sequence cgaggggaagcccctcc. The second, broader, and larger peak occurs
at the sequence ccctctaagccccctctccccgtcacatccccccgaccctgccc. Note the
stretches of consecutive Cs within both sequences.
The pattern of reactivity is very similar between the different
substrates – extracted genomic DNA, supercoiled plasmid, or linearized
plasmid. However, overall level of bisulfite reactivity is higher on supercoiled
plasmid than on linearized plasmid or genomic DNA. Supercoiling

156
A. 100 mM NaCl

B. 5 M NaCl

Figure 4-4. Bisulfite reactivity on dsDNA correlates broadly with relative
propensity to form B-A-intermediate structure. Circular dichroism spectra for
catGGGCCCatg, catGGCGCCatg, and catGCGCGCatg, along with
appropriate controls, at (A) 100 mM NaCl and (B) 5 M NaCl are shown.

157
A. catGGGCCCatg

B. catGGCGCCatg

Figure 4-5. Bisulfite is minimally reactive on DNA-2’-O-methyl RNA hybrid
duplexes. Bisulfite reactivities for each cytosine on each hybrid duplex DNA-
2’-O-methyl RNA substrate are shown. Notations are as in Fig. 4-3.

158
Figure 4-6. Bisulfite reactivity at the bcl-1 MTC. Bisulfite reactivities for
cytosines in bcl-1 MTC sequences from various substrates are shown. The
bottom line of each plot denotes all the cytosines (bases along the top strand
which can be deaminated) and guanines (bases along the bottom strand
which can be deaminated) in the sequence. Black marks above those
positions denote top strand cytosine deaminations and red marks bottom
strand cytosine deaminations. For each molecule sequenced, deaminations
can only be detected on one strand. Numbers sequenced for each strand
are noted, as well as overall conversion frequency.

159

Figure 4-6 continued

Extracted Reh Genomic DNA
7.8% total conversion frequency
44 top strands
20 bottom strands

pSH 9 Supercoiled
19.4% total conversion frequency
19 top strands
8 bottom strands

pSH 9 Sfil-Cut
11.48% total conversion frequency
23 top strands
9 bottom strands
bcl-1 MTC

160
destabilizes the duplex, accelerating DNA breathing and allowing it to more
easily form non-B-DNA structures.
Because the top and bottom strands of the original bisulfite-probed
DNA molecule are separated during PCR and separately cloned, each
sequence obtained gives information for deamination events on only one of
the two original strands. When the reactivity is rescaled to balance the top
and bottom strand representation, it is roughly symmetrical between the two
strands.
To rule out our primers biasing against binding to converted regions,
we used a second set of outer primers to confirm that few conversions occur
at the binding sites of the first set. A safer strategy – designing two sets of
primers, one set insensitive to top strand conversions and the other
insensitive to those on the bottom – was also performed. These show no
difference in pattern (data not shown).
The pattern does not change if bisulfite treatment is done on intact
Reh cells embedded in 1% low-melt agarose, or intact E. coli embedded in
agarose after transformation with plasmid containing the bcl-1 sequence.
Any proteins that might bind to the DNA are likely denatured by the 5 M salt
concentration of the bisulfite solution. The pattern also does not change if
the bisulfite solution is adjusted to pH 6, though overall reactivity is
decreased about three-fold – consistent with the three-fold lower efficiency of
the bisulfite-catalyzed deamination reaction itself at pH 6 compared to pH 5.2

161
(Sono et al., 1973). Reactivity increases proportionally if DNA is treated for
24 hours rather than the usual 16 hours, but again the pattern of reactivity is
still the same. Neither does it change if the two peaks are separated and
cloned into different plasmids, or if the plasmid or genomic DNA are
pretreated with RNaseH, or if plasmid is extracted from E. coli by a non-
denaturing method rather than alkaline lysis (data not shown).
Therefore the specific pattern of bisulfite reactivity appears dependent
mainly upon the DNA sequence.
Discussion
Rarely are probes of DNA structure investigated beyond the notion
that they are, for instance, “single-strand specific.” While single-stranded
DNA might react much more than double-stranded under the specific
conditions and substrates tested – which are often not uniform between
papers and labs – one cannot be certain that the probe never reacts with any
structures considered double-stranded, or that it always reacts with
structures considered single-stranded. “Single-stranded” typically refers to
loss of standard Watson-Crick base pairing, but structurally might mean
several different things – e.g. strand separation or unwinding, bases flipping
out, partial hydrogen bonding, syn base conformation, etc. For a given
probe, which structures react and which do not react is usually not
extensively tested. In short, reactivity does not always mean single-
strandedness, lack of reactivity does not always mean lack of single-

162
strandedness, and the definition of single-strandedness is itself structurally
vague.
Our probing of oligonucleotides containing sequences of known
structure is the first foray into these issues for bisulfite. We find large
differences in reactivity between substrates with single base pair changes.
Differential bisulfite reactivities between different cytosine positions indicate
that structural differences exist between the cytosine environments. When
cytosines at the edges of a duplex react more than those at the center, or
when those on supercoiled DNA react more than those on linearized, there
are simple structural explanations. However, the structural reasons
underlying differential reactivities between cytosines within double-stranded
regions – or more generally, the specific pattern of bisulfite reactivity within
dsDNA – are more mysterious.
There are thought to be three requirements for bisulfite-catalyzed
cytosine deamination: 1. addition of the sulfonate group to the C6 position, 2.
protonation of the N3 position, and 3. water or hydroxyl ion attack of the C4
position with displacement of the amino group (Shapiro et al., 1974).
Increase in reactivity could occur at any one or several of them (Fig. 4-7).
One explanation for differential cytosine reactivities might be
breathing. Unpairing of cytosine from guanine and increased exposure to
solvent would allow the sulfonate group to be added more easily, and is likely
required for protonation. Consistent with this idea, increased breathing

163

Figure 4-7. Hypothetical model of bisulfite-catalyzed deamination of
cytosines in double-stranded DNA. Bisulfite-catalyzed deamination appears
to occur only when a cytosine is protonated, single-stranded, and sulfonated
(H-
+
C
ss
-SO
3
). It may reach that state by a number of paths. Abbreviations
are as follows: cytosine, C; double-stranded, ds; single-stranded, ss;
protonated, H-
+
; sulfonated, -SO
3
; uracil, U.

C
ds
C
ss

C
ds
-SO
3
C
ss
-SO
3
H-
+
C
ss
-SO
3
H-
+
C
ss
U
ss
-SO
3

164
correlates with increased reactivity when comparing consecutive cytosines to
alternating guanine-cytosine dinucleotides. We see the bulk of reactivity in
the bcl-1 at stretches of consecutive cytosines.
However, if bisulfite reactivity were really an indicator of breathing, we
would expect the highest reactivity at AT-rich regions of the genome, rather
than the GC-rich regions where it is typically found. Additionally, the level of
reactivity is far beyond what is expected from the equilibrium constants of
breathing (on the order of 10
-6
) and the pseudo-first-order rate kinetics of the
bisulfite reaction on free cytosine bases, i.e. fully single-stranded cytosines,
(4.6 ⋅ 10
-4
/s for 5 M). Based on these numbers, one would expect cytosines
to react less than 1% of the time over 16 hours. In fact, we observe
reactivities in excess of 20% for certain cytosine positions. Therefore other
structural factors must account for this excess reactivity. Among many
speculative possibilities, one may be that sulfonation and/or protonation
stabilize the single-stranded state and prevent return to the double-stranded
state (Fig. 4-7). Another could be that, for structural reasons, sulfonation
might occur on certain cytosines that are still mostly base-paired (Fig. 4-7).
These are other points at which sequence-specific effects might influence
bisulfite reactivity.
Another interesting observation is that, especially in the genomic and
plasmid sequences, reactivity is quite heterogeneous from molecule to
molecule, as evident from previously published data (Raghavan et al.,

165
2004b). That some molecules react more or in different regions than others
suggests that either the reaction itself is somewhat stochastic in nature
and/or the DNA substrates are structurally heterogeneous to begin with. To
resolve this issue bisulfite probing might have to be done on substrates with
a known range of structural hetereogeneity, i.e. one would have to compare
the hetereogeneity of reactivity between substrates known to have low
structural heterogeneity and those known to have high structural
heterogeneity, while controlling for overall reactivity. While the algorithm has
already been written to quantitate heterogeneity in reactivity, and overall
reactivity can be scaled, structural heterogeneity remains more or less
theoretical rather than experimentally proven.
Conclusion
The observation of bisulfite-catalyzed cytosine deamination at
presumably double-stranded regions led us to question the true nature of
bisulfite reactivity. Here we have demonstrated that bisulfite reactivity
correlates with structural/dynamic characteristics such as propensity to
breathe and/or propensity to form B-A intermediate structure. Questions
regarding the precise structural mechanism(s) of breathing and the structural
basis of chemical probe reactivities in general are raised.

166
Chapter 7. Conclusion and future directions
In chapter 2, we analyzed over 1,700 human chromosomal
translocation breakpoints grouped according to tissue type, lineage, and
developmental stage. Based on this, we elucidated a new type of double-
strand breakage motif, which appears to shape the breakpoint distributions
within the bcl-2 MBR, bcl-2 icr, bcl-2 mcr, bcl-1 MTC, and E2A cluster
regions. This CpG-type breakage is prominent in pro-B/pre-B cells and
apparently absent or heavily diminished in other cell types. We proposed
that AID-deaminated methyl-CpGs within cluster regions are intercepted by
RAGs before repair can be completed, generating the requisite double-strand
breaks in several of the most common translocations in human lymphoma.
We have labeled this as a theory because definitive proof is not yet
available, and may not be achievable with existing model systems. To
demonstrate AID or RAG complex dependence, one must have a working
translocation model system. However, the translocation process as it occurs
in humans has not been observed in mice. The only translocation model in
mouse is c-myc joining to the immunoglobulin heavy chain switch region.
The V(D)J recombination assay in human cell lines has far too high a
background to detect unrepaired deamination events. However, it may still
be worth attempting anyway just to prove that point. One could methylate
one of the newer substrates I’ve developed, and cotransfect with an AID
expression vector into Reh, or 293 with RAG expression vectors.

167
Certain details of the theory could also be bolstered by more
experimental evidence. We showed that core RAG complexes could nick
one-base mismatches and across from nicks, gaps, and flaps, but we did not
test full-length RAG complexes. The substrates we used were all based on
one artificial sequence context. We did not show that mismatches, nicks,
gaps, or flaps within the bcl-1 MTC or bcl-2 MBR sequence contexts could be
nicked by RAG complexes. The expression of AID by pre-B cells is still in
some doubt. It would be valuable to show AID-dependent deaminations in
such cells, especially within the MTC or MBR. We are attempting to show
AID action at the MTC and MBR in Daudi and Raji, mature B cell lines which
strongly express AID, using a high-throughput sequencing approach.
One of the questions left open is why translocations occur at CpGs
within the MTC, MBR, or E2A clusters versus the hundreds of other CpGs
throughout the bcl-1, bcl-2, or E2A regions. Each step in the proposed
mechanism is an opportunity to explain this clustering.
One possibility is that AID targets these regions due to inherent non-
B-DNA structure. Under this hypothesis, AID would have enhanced activity
at the MTC and MBR in the deamination assay compared to other
sequences of DNA. While it is possible that AID might target some regions
of double-stranded DNA better than others, though not as well as single-
stranded DNA, the basis for this would, with little doubt, be due to the
structure of AID and not of the DNA. As I described in chapter 3, nothing we

168
have observed about the MTC suggests it is much different from any other
DNA. Moreover, there are plenty of adjacent regions which have sequence
elements more suitable for supporting non-B-DNA structures. In the wrong
hands, structural methods can be artefacted, misinterpreted, and/or
overinterpreted. I find the selective application of various structures under
very narrow conditions and the explanations for the all the inconsistencies to
be artfully ad hoc.
In order to put these claims to rest, one could use the breakpoint data
I have accumulated to examine whether breakpoints occur at such putative
non-B-DNA structures more often than expected. I am quite certain that
breakpoints will be random with respect to the proposed structural elements.
The statistics take into account the frequency of the element throughout the
region, and include all the breakpoints – not just the one or two cherrypicked
to demonstrate the fantasy. Using the data from MLL translocations in AMLs
secondary to topoisomerase II inhibitor treatment, I found no clustering of
breakpoints around topoisomerase II consensus sites. Others have noticed
this lack of correlation as well, though the prevailing theory is that breakage
occurs around these sites (Greaves and Wiemels, 2003).
A way to experimentally test whether non-B-DNA structures undely
double-strand breaks might be to do the V(D)J recombination assay using
known cruciform, Z-DNA, quadruplex, triplex, etc. sequences.

169
Another possibility for clustering is that repair in the MTC or MBR is
especially poor or slow. It is possible that TDG or MBD4 do not cleave T:G
mismatches at the region as effectively as at others. A simple experiment
would be to purify TDG and MBD4 and compare their glycosylase activities
on T:G mismatches in the MTC or MBR to mismatches in other sequence
contexts. If oligonucleotides were deemed too short, one could clone the
C →T mutation onto a plasmid, and anneal mutant and wild type restriction
fragments. In theory, one could adjust the ratios of the plasmids such that
the radiolabeled strands are mostly mismatched.
One of the most ignored aspects of DNA structure studies is
chromatin. The RAG complex is known to bind trimethylated histone H3
lysine 4 (trimethyl H3K4) (Liu et al., 2007; Matthews et al., 2007). No
trimethyl H3K4 ChIP has been done for pre-B cells, but data for T cells is
available (Barski et al., 2007). An interesting puzzle in the field has been
why SCL-SIL deletions are so common, yet they are not nicked efficiently by
RAG complexes in vitro or recombined well ex vivo (Marculescu et al.,
2002b; Raghavan et al., 2001a; Zhang and Swanson, 2008). Using the
published resource, I found that the highest peaks of trimethyl H3K4 occur
right at those two loci. We have further plans to assess, computationally, the
relationship between cRSS and trimethyl H3K4. It would be valuable to do a
ChIP experiment to see if a high level of trimethyl H3K4 is present at the
MTC or MBR, but the difficulty is in obtaining enough cells to do it. It is

170
known that mice lacking H2AX have genome-wide instability in all cells, and
it has been proposed that H2AX sequesters double-strand breaks close to
one another in order to prevent chromosomal rearrangement (Bassing et al.,
2003; Franco et al., 2006). It is possible that, when double-strand breaks
occur at the MTC or MBR, they do not bind H2AX well or activate such a
“sequestering” response effectively.
The issue of chromatin on V(D)J recombination assay substrates with
the bcl-2 MBR was not investigated either. It might be useful to place
nucleosomal positioning sequences around the MBR and see how they affect
recombination.
While the proteins logically required for the enzymatic breakage and
rejoining steps of V(D)J recombination have been identified by genetic and
biochemical methods and the process reconstituted in vitro, it is quite likely
that there are many other proteins involved to facilitate the V(D)J process
and prevent translocations. The roles of proteins such as ATM/ATR,
BRCA1/BRCA2, and RAD50/Nbs1/Mre11 in double-strand break repair are
not entirely clear. These could also have roles in clustering at the MTC or
MBR.
The identification of CpG as a motif for translocation in pre-B cells
appears to be a major step forward in the understanding of the mechanisms
of chromosomal rearrangement. We provide a plausible theory for this
observation, and future experiments will either refute it or bear it out.

171
Bibliography
Akao, Y., and Isobe, M. (2000). Molecular analysis of the rearranged genome
and chimeric mRNAs caused by the t(6;11)(q27;q23) chromosome
translocation involving MLL in an infant acute monocytic leukemia. Genes
Chromosomes Cancer 27, 412-417.
Akasaka, H., Akasaka, T., Kurata, M., Ueda, C., Shimizu, A., Uchiyama, T.,
and Ohno, H. (2000). Molecular anatomy of BCL6 translocations revealed by
long-distance polymerase chain reaction-based assays. Cancer Res 60,
2335-2341.
Akasaka, T., Akasaka, H., Yonetani, N., Ohno, H., Yamabe, H., Fukuhara,
S., and Okuma, M. (1998). Refinement of the BCL2/immunoglobulin heavy
chain fusion gene in t(14;18)(q32;q21) by polymerase chain reaction
amplification for long targets. Genes Chromosomes Cancer 21, 17-29.
Albinger-Hegyi, A., Hochreutener, B., Abdou, M. T., Hegyi, I., Dours-
Zimmermann, M. T., Kurrer, M. O., Heitz, P. U., and Zimmermann, D. R.
(2002). High frequency of t(14;18)-translocation breakpoints outside of major
breakpoint and minor cluster regions in follicular lymphomas: improved
polymerase chain reaction protocols for their detection. Am J Pathol 160,
823-832.
Alt, F. W., Rosenberg, N., Casanova, R. J., Thomas, E., and Baltimore, D.
(1982). Immunoglobulin heavy-chain expression and class switching in a
murine leukaemia cell line. Nature 296, 325-331.
Andersen, M. T., Nordentoft, I., Hjalgrim, L. L., Christiansen, C. L., Jakobsen,
V. D., Hjalgrim, H., Pallisgaard, N., Madsen, H. O., Christiansen, M., Ryder,
L. P., et al. (2001). Characterization of t(12;21) breakpoint junctions in acute
lymphoblastic leukemia. Leukemia 15, 858-859.
Antequera, F., Boyes, J., and Bird, A. P. (1990). High levels of de novo
methylation and altered chromatin structure at CpG islands in cell lines. Cell
62, 503-514.
Apel, T. W., Mautner, J., Polack, A., Bornkamm, G. W., and Eick, D. (1992).
Two antisense promoters in the immunoglobulin mu-switch region drive
expression of c-myc in the Burkitt's lymphoma cell line BL67. Oncogene 7,
1267-1271.
Aplan, P. D., Lombardi, D. P., Ginsberg, A. M., Cossman, J., Bertness, V. L.,
and Kirsch, I. R. (1990). Disruption of the human SCL locus by "illegitimate"
V-(D)-J recombinase activity. Science 250, 1426-1429.

172
Aplan, P. D., Lombardi, D. P., Reaman, G. H., Sather, H. N., Hammond, G.
D., and Kirsch, I. R. (1992a). Involvement of the putative hematopoietic
transcription factor SCL in T-cell acute lymphoblastic leukemia. Blood 79,
1327-1333.
Aplan, P. D., Raimondi, S. C., and Kirsch, I. R. (1992b). Disruption of the
SCL gene by a t(1;3) translocation in a patient with T cell acute lymphoblastic
leukemia. J Exp Med 176, 1303-1310.
Aster, J. C., Kobayashi, Y., Shiota, M., Mori, S., and Sklar, J. (1992).
Detection of the t(14;18) at similar frequencies in hyperplastic lymphoid
tissues from American and Japanese patients. Am J Pathol 141, 291-299.
Ausubel, F. M. (1987). Current protocols in molecular biology (New York:
Published by Greene Pub. Associates and Wiley-Interscience: J. Wiley).
Bakhshi, A., Jensen, J. P., Goldman, P., Wright, J. J., McBride, O. W.,
Epstein, A. L., and Korsmeyer, S. J. (1985). Cloning the chromosomal
breakpoint of t(14;18) human lymphomas: clustering around JH on
chromosome 14 and near a transcriptional unit on 18. Cell 41, 899-906.
Bakhshi, A., Wright, J. J., Graninger, W., Seto, M., Owens, J., Cossman, J.,
Jensen, J. P., Goldman, P., and Korsmeyer, S. J. (1987). Mechanism of the
t(14;18) chromosomal translocation: structural analysis of both derivative 14
and 18 reciprocal partners. Proc Natl Acad Sci U S A 84, 2396-2400.
Baron, B. W., Nucifora, G., McCabe, N., Espinosa, R., 3rd, Le Beau, M. M.,
and McKeithan, T. W. (1993). Identification of the gene associated with the
recurring chromosomal translocations t(3;14)(q27;q32) and t(3;22)(q27;q11)
in B-cell lymphomas. Proc Natl Acad Sci U S A 90, 5262-5266.
Barreto, V., Marques, R., and Demengeot, J. (2001). Early death and severe
lymphopenia caused by ubiquitous expression of the Rag1 and Rag2 genes
in mice. Eur J Immunol 31, 3763-3772.
Barski, A., Cuddapah, S., Cui, K., Roh, T. Y., Schones, D. E., Wang, Z., Wei,
G., Chepelev, I., and Zhao, K. (2007). High-resolution profiling of histone
methylations in the human genome. Cell 129, 823-837.
Bash, R. O., Crist, W. M., Shuster, J. J., Link, M. P., Amylon, M., Pullen, J.,
Carroll, A. J., Buchanan, G. R., Smith, R. G., and Baer, R. (1993). Clinical
features and outcome of T-cell acute lymphoblastic leukemia in childhood
with respect to alterations at the TAL1 locus: a Pediatric Oncology Group
study. Blood 81, 2110-2117.

173
Bassing, C. H., Ranganath, S., Murphy, M., Savic, V., Gleason, M., and Alt,
F. W. (2008). Aberrant V(D)J recombination is not required for rapid
development of H2ax/p53-deficient thymic lymphomas with clonal
translocations. Blood 111, 2163-2169.
Bassing, C. H., Suh, H., Ferguson, D. O., Chua, K. F., Manis, J., Eckersdorff,
M., Gleason, M., Bronson, R., Lee, C., and Alt, F. W. (2003). Histone H2AX:
a dosage-dependent suppressor of oncogenic translocations and tumors.
Cell 114, 359-370.
Bassing, C. H., Swat, W., and Alt, F. W. (2002). The mechanism and
regulation of chromosomal V(D)J recombination. Cell 109 Suppl, S45-55.
Battey, J., Moulding, C., Taub, R., Murphy, W., Stewart, T., Potter, H., Lenoir,
G., and Leder, P. (1983). The human c-myc oncogene: structural
consequences of translocation into the IgH locus in Burkitt lymphoma. Cell
34, 779-787.
Begley, C. G., Aplan, P. D., Davey, M. P., Nakahara, K., Tchorz, K.,
Kurtzberg, J., Hershfield, M. S., Haynes, B. F., Cohen, D. I., Waldmann, T.
A., and et al. (1989). Chromosomal translocation in a human leukemic stem-
cell line disrupts the T-cell antigen receptor delta-chain diversity region and
results in a previously unreported fusion transcript. Proc Natl Acad Sci U S A
86, 2031-2035.
Belasco, J. G., Nilsson, G., von Gabain, A., and Cohen, S. N. (1986). The
stability of E. coli gene transcripts is dependent on determinants localized to
specific mRNA segments. Cell 46, 245-251.
Bernard, O., Azogui, O., Lecointe, N., Mugneret, F., Berger, R., Larsen, C. J.,
and Mathieu-Mahul, D. (1992). A third tal-1 promoter is specifically used in
human T cell leukemias. J Exp Med 176, 919-925.
Bernard, O., Barin, C., Charrin, C., Mathieu-Mahul, D., and Berger, R.
(1993). Characterization of translocation t(1;14)(p32;q11) in a T and in a B
acute leukemia. Leukemia 7, 1509-1513.
Bernard, O., Guglielmi, P., Jonveaux, P., Cherif, D., Gisselbrecht, S.,
Mauchauffe, M., Berger, R., Larsen, C. J., and Mathieu-Mahul, D. (1990).
Two distinct mechanisms for the SCL gene activation in the t(1;14)
translocation of T-cell leukemias. Genes Chromosomes Cancer 1, 194-208.
Bernard, O., Lecointe, N., Jonveaux, P., Souyri, M., Mauchauffe, M., Berger,
R., Larsen, C. J., and Mathieu-Mahul, D. (1991). Two site-specific deletions
and t(1;14) translocation restricted to human T-cell acute leukemias disrupt
the 5' part of the tal-1 gene. Oncogene 6, 1477-1488.

174
Bertoni, F., Zucca, E., and Cotter, F. E. (2004). Molecular basis of mantle cell
lymphoma. Br J Haematol 124, 130-140.
Bestor, T. H. (2003). Unanswered questions about the role of promoter
methylation in carcinogenesis. Ann N Y Acad Sci 983, 22-27.
Bird, A. (1992). The essentials of DNA methylation. Cell 70, 5-8.
Bizarro, S., Cerveira, N., Correia, C., Lisboa, S., Peixoto, A., Norton, L., and
Teixeira, M. R. (2007). Molecular characterization of a rare MLL-AF4 (MLL-
AFF1) fusion rearrangement in infant leukemia. Cancer Genet Cytogenet
178, 61-64.
Bodrug, S. E., Warner, B. J., Bath, M. L., Lindeman, G. J., Harris, A. W., and
Adams, J. M. (1994). Cyclin D1 transgene impedes lymphocyte maturation
and collaborates in lymphomagenesis with the myc gene. Embo J 13, 2124-
2130.
Boehm, T., Baer, R., Lavenir, I., Forster, A., Waters, J. J., Nacheva, E., and
Rabbitts, T. H. (1988a). The mechanism of chromosomal translocation
t(11;14) involving the T-cell receptor C delta locus on human chromosome
14q11 and a transcribed region of chromosome 11p15. Embo J 7, 385-394.
Boehm, T., Buluwela, L., Williams, D., White, L., and Rabbitts, T. H. (1988b).
A cluster of chromosome 11p13 translocations found via distinct D-D and D-
D-J rearrangements of the human T cell receptor delta chain gene. Embo J
7, 2011-2017.
Boehm, T., Foroni, L., Kaneko, Y., Perutz, M. F., and Rabbitts, T. H. (1991).
The rhombotin family of cysteine-rich LIM-domain oncogenes: distinct
members are involved in T-cell translocations to human chromosomes 11p15
and 11p13. Proc Natl Acad Sci U S A 88, 4367-4371.
Bordeleau, L., and Berinstein, N. L. (2000). Molecular diagnostics in follicular
non-Hodgkin's lymphoma: a review. Semin Oncol 27, 42-52.
Borkhardt, A., Repp, R., Haas, O. A., Leis, T., Harbott, J., Kreuder, J.,
Hammermann, J., Henn, T., and Lampert, F. (1997). Cloning and
characterization of AFX, the gene that fuses to MLL in acute leukemias with
a t(X;11)(q13;q23). Oncogene 14, 195-202.
Bransteitter, R., Pham, P., Scharff, M. D., and Goodman, M. F. (2003a).
Activation-induced cytidine deaminase deaminates deoxycytidine on single-
stranded DNA but requires the action of RNase. Proc Natl Acad Sci 100,
4102-4107.

175
Bransteitter, R., Pham, P., Scharff, M. D., and Goodman, M. F. (2003b).
Activation-induced cytidine deaminase deaminates deoxycytidine on single-
stranded DNA but requires the action of RNase. Proc Natl Acad Sci U S A
100, 4102-4107.
Breit, T. M., Mol, E. J., Wolvers-Tettero, I. L., Ludwig, W. D., van Wering, E.
R., and van Dongen, J. J. (1993). Site-specific deletions involving the tal-1
and sil genes are restricted to cells of the T cell receptor alpha/beta lineage:
T cell receptor delta gene deletion mechanism affects multiple genes. J Exp
Med 177, 965-977.
Brown, L., Cheng, J. T., Chen, Q., Siciliano, M. J., Crist, W., Buchanan, G.,
and Baer, R. (1990). Site-specific recombination of the tal-1 gene is a
common occurrence in human T cell leukemia. Embo J 9, 3343-3351.
Burrows, P. D., Beck-Engeser, G. B., and Wabl, M. R. (1983).
Immunoglubulin Heavy-Chain Class Switching in a pre-B Cell Line is
Accompanied by DNA Rearrangement. Nature 306, 243-246.
Busch, K., Borkhardt, A., Wossmann, W., Reiter, A., and Harbott, J. (2004).
Combined polymerase chain reaction methods to detect c-myc/IgH
rearrangement in childhood Burkitt's lymphoma for minimal residual disease
analysis. Haematologica 89, 818-825.
Busch, K., Keller, T., Fuchs, U., Yeh, R. F., Harbott, J., Klose, I., Wiemels, J.,
Novosel, A., Reiter, A., and Borkhardt, A. (2007). Identification of two distinct
MYC breakpoint clusters and their association with various IGH breakpoint
regions in the t(8;14) translocations in sporadic Burkitt-lymphoma. Leukemia
21, 1739-1751.
Care, A., Cianetti, L., Giampaolo, A., Sposi, N. M., Zappavigna, V., Mavilio,
F., Alimena, G., Amadori, S., Mandelli, F., and Peschle, C. (1986).
Translocation of c-myc into the immunoglobulin heavy-chain locus in human
acute B-cell leukemia. A molecular analysis. Embo J 5, 905-911.
Cayuela, J. M., Gardie, B., and Sigaux, F. (1997). Disruption of the multiple
tumor suppressor gene MTS1/p16(INK4a)/CDKN2 by illegitimate V(D)J
recombinase activity in T-cell acute lymphoblastic leukemias. Blood 90,
3720-3726.
Celesnik, H., Deana, A., and Belasco, J. G. (2007). Initiation of RNA decay in
Escherichia coli by 5' pyrophosphate removal. Mol Cell 27, 79-90.

176
Chaganti, S. R., Rao, P. H., Chen, W., Dyomin, V., Jhanwar, S. C., Parsa, N.
Z., Dalla-Favera, R., and Chaganti, R. S. (1998). Deregulation of BCL6 in
non-Hodgkin lymphoma by insertion of IGH sequences in complex
translocations involving band 3q27. Genes Chromosomes Cancer 23, 328-
336.
Chen, H., and Shaw, B. R. (1993). Kinetics of bisulfite-induced cytosine
deamination in single-stranded DNA. Biochemistry 32, 3535-3539.
Chen, Q., Cheng, J. T., Tasi, L. H., Schneider, N., Buchanan, G., Carroll, A.,
Crist, W., Ozanne, B., Siciliano, M. J., and Baer, R. (1990a). The tal gene
undergoes chromosome translocation in T cell leukemia and potentially
encodes a helix-loop-helix protein. Embo J 9, 415-424.
Chen, Q., Yang, C. Y., Tsan, J. T., Xia, Y., Ragab, A. H., Peiper, S. C.,
Carroll, A., and Baer, R. (1990b). Coding sequences of the tal-1 gene are
disrupted by chromosome translocation in human T cell leukemia. J Exp Med
172, 1403-1408.
Chen, Y. W., Hu, X. T., Liang, A. C., Au, W. Y., So, C. C., Wong, M. L., Shen,
L., Tao, Q., Chu, K. M., Kwong, Y. L., et al. (2006). High BCL6 expression
predicts better prognosis, independent of BCL6 translocation status,
translocation partner, or BCL6-deregulating mutations, in gastric lymphoma.
Blood 108, 2373-2383.
Cheng, J. T., Yang, C. Y., Hernandez, J., Embrey, J., and Baer, R. (1990).
The chromosome translocation (11;14)(p13;q11) associated with T cell acute
leukemia. Asymmetric diversification of the translocational junctions. J Exp
Med 171, 489-501.
Chervinsky, D. S., Sait, S. N., Nowak, N. J., Shows, T. B., and Aplan, P. D.
(1995). Complex MLL rearrangement in a patient with T-cell acute
lymphoblastic leukemia. Genes Chromosomes Cancer 14, 76-84.
Chissoe, S. L., Bodenteich, A., Wang, Y. F., Wang, Y. P., Burian, D., Clifton,
S. W., Crabtree, J., Freeman, A., Iyer, K., Jian, L., and et al. (1995).
Sequence and analysis of the human ABL gene, the BCR gene, and regions
involved in the Philadelphia chromosomal translocation. Genomics 27, 67-82.
Cleary, M. L., Galili, N., and Sklar, J. (1986a). Detection of a second t(14;18)
breakpoint cluster region in human follicular lymphomas. J Exp Med 164,
315-320.

177
Cleary, M. L., and Sklar, J. (1985). Nucleotide sequence of a t(14;18)
chromosomal breakpoint in follicular lymphoma and demonstration of a
breakpoint-cluster region near a transcriptionally active locus on
chromosome 18. Proc Natl Acad Sci U S A 82, 7439-7443.
Cleary, M. L., Smith, S. D., and Sklar, J. (1986b). Cloning and structural
analysis of cDNAs for bcl-2 and a hybrid bcl-2/immunoglobulin transcript
resulting from the t(14;18) translocation. Cell 47, 19-28.
Conboy, C. (2003). Cconboy:Terminator Characterization/Results, pp.
Cconboy:Terminator Characterization/Results.
Cotter, F., Price, C., Zucca, E., and Young, B. D. (1990). Direct sequence
analysis of the 14q+ and 18q- chromosome junctions in follicular lymphoma.
Blood 76, 131-135.
Crescenzi, M., Seto, M., Herzig, G. P., Weiss, P. D., Griffith, R. C., and
Korsmeyer, S. J. (1988). Thermostable DNA polymerase chain amplification
of t(14;18) chromosome breakpoints and detection of minimal residual
disease. Proc Natl Acad Sci U S A 85, 4869-4873.
de Boer, H. A., Comstock, L. J., and Vasser, M. (1983). The tac promoter: a
functional hybrid derived from the trp and lac promoters. Proc Natl Acad Sci
U S A 80, 21-25.
de Klein, A., van Agthoven, T., Groffen, C., Heisterkamp, N., Groffen, J., and
Grosveld, G. (1986). Molecular analysis of both translocation products of a
Philadelphia-positive CML patient. Nucleic Acids Res 14, 7071-7082.
Degan, M., Doliana, R., Gloghini, A., Di Francia, R., Aldinucci, D., Mazzocut-
Zecchin, L., Colombatti, A., Attadia, V., Carbone, A., and Gattei, V. (2002). A
novel bcl-1/JH breakpoint from a patient affected by mantle cell lymphoma
extends the major translocation cluster. J Pathol 197, 256-263.
Delage, R., Roy, J., Jacques, L., Bernier, V., Delage, J. M., and Darveau, A.
(1997). Multiple bcl-2/Ig gene rearrangements in persistent polyclonal B-cell
lymphocytosis. Br J Haematol 97, 589-595.
Dik, W. A., Nadel, B., Przybylski, G. K., Asnafi, V., Grabarczyk, P., Navarro,
J. M., Verhaaf, B., Schmidt, C. A., Macintyre, E. A., van Dongen, J. J., and
Langerak, A. W. (2007). Different chromosomal breakpoints impact the level
of LMO2 expression in T-ALL. Blood 110, 388-392.
Dolken, G., Illerhaus, G., Hirt, C., and Mertelsmann, R. (1996). BCL-2/JH
rearrangements in circulating B cells of healthy blood donors and patients
with nonmalignant diseases. J Clin Oncol 14, 1333-1344.

178
Dornberger, U., Flemming, J., and Fritzsche, H. (1998). Structure
determination and analysis of helix parameters in the DNA decamer
d(CATGGCCATG)2 comparison of results from NMR and crystallography. J
Mol Biol 284, 1453-1463.
Dornberger, U., Leijon, M., and Fritzsche, H. (1999). High base pair opening
rates in tracts of GC base pairs. J Biol Chem 274, 6957-6962.
Dudley, D. D., Chaudhuri, J., Bassing, C. H., and Alt, F. W. (2005a).
Mechanism and control of V(D)J recombination versus class switch
recombination: similarities and differences. Adv Immunol 86, 43-112.
Dudley, D. D., Chaudhuri, J., Bassing, C. H., and Alt, F. W. (2005b).
Mechanism and control of V(D)J recombination versus class switch
recombination: similarities and differences. Adv Immunol 86, 43-112.
Duro, D., Bernard, O., Della Valle, V., Leblanc, T., Berger, R., and Larsen, C.
J. (1996). Inactivation of the P16INK4/MTS1 gene by a chromosome
translocation t(9;14)(p21-22;q11) in an acute lymphoblastic leukemia of B-
cell type. Cancer Res 56, 848-854.
Dyson, P. J., and Rabbitts, T. H. (1985). Chromatin structure around the c-
myc gene in Burkitt lymphomas with upstream and downstream translocation
points. Proc Natl Acad Sci U S A 82, 1984-1988.
Eick, S., Krieger, G., Bolz, I., and Kneba, M. (1990). Sequence analysis of
amplified t(14;18) chromosomal breakpoints in B-cell lymphomas. J Pathol
162, 127-133.
Erickson, P., Gao, J., Chang, K. S., Look, T., Whisenant, E., Raimondi, S.,
Lasher, R., Trujillo, J., Rowley, J., and Drabkin, H. (1992). Identification of
breakpoints in t(8;21) acute myelogenous leukemia and isolation of a fusion
transcript, AML1/ETO, with similarity to Drosophila segmentation gene, runt.
Blood 80, 1825-1831.
Espinet, B., Sole, F., Pedro, C., Garcia, M., Bellosillo, B., Salido, M.,
Florensa, L., Camacho, F. I., Baro, T., Lloreta, J., and Serrano, S. (2005).
Clonal proliferation of cyclin D1-positive mantle lymphocytes in an
asymptomatic patient: an early-stage event in the development or an indolent
form of a mantle cell lymphoma? Hum Pathol 36, 1232-1237.
Feldhahn, N., Henke, N., Melchior, K., Duy, C., Soh, B. N., Klein, F., von
Levetzow, G., Giebel, B., Li, A., Hofmann, W. K., et al. (2007). Activation-
induced cytidine deaminase acts as a mutator in BCR-ABL1-transformed
acute lymphoblastic leukemia cells. J Exp Med 204, 1157-1166.

179
Felix, C. A., Hosler, M. R., Slater, D. J., Megonigal, M. D., Lovett, B. D.,
Williams, T. M., Nowell, P. C., Spinner, N. B., Owens, N. L., Hoxie, J., et al.
(1999). Duplicated regions of AF-4 intron 4 at t(4;11) translocation
breakpoints. Mol Diagn 4, 269-283.
Felix, C. A., Kim, C. S., Megonigal, M. D., Slater, D. J., Jones, D. H., Spinner,
N. B., Stump, T., Hosler, M. R., Nowell, P. C., Lange, B. J., and Rappaport,
E. F. (1997). Panhandle polymerase chain reaction amplifies MLL genomic
translocation breakpoint involving unknown partner gene. Blood 90, 4679-
4686.
Ferguson, D. O., and Alt, F. W. (2001a). DNA double-strand break repair and
chromosomal translocations: lessons from animal models. Oncogene 20,
5572-5579.
Ferguson, D. O., and Alt, F. W. (2001b). DNA double strand break repair and
chromosomal translocation: lessons from animal models. Oncogene 20,
5572-5579.
Fischer, S., Mann, G., Konrad, M., Metzler, M., Ebetsberger, G., Jones, N.,
Nadel, B., Bodamer, O., Haas, O. A., Schmitt, K., and Panzer-Grumayer, E.
R. (2007). Screening for leukemia- and clone-specific markers at birth in
children with T-cell precursor ALL suggests a predominantly postnatal origin.
Blood 110, 3036-3038.
Fitzgerald, T. J., Neale, G. A., Raimondi, S. C., and Goorha, R. M. (1991). c-
tal, a helix-loop-helix protein, is juxtaposed to the T-cell receptor-beta chain
gene by a reciprocal chromosomal translocation: t(1;7)(p32;q35). Blood 78,
2686-2695.
Ford, A. M., Bennett, C. A., Price, C. M., Bruin, M. C., Van Wering, E. R., and
Greaves, M. (1998). Fetal origins of the TEL-AML1 fusion gene in identical
twins with leukemia. Proc Natl Acad Sci U S A 95, 4584-4588.
Franco, S., Gostissa, M., Zha, S., Lombard, D. B., Murphy, M. M., Zarrin, A.
A., Yan, C., Tepsuporn, S., Morales, J. C., Adams, M. M., et al. (2006). H2AX
prevents DNA breaks from progressing to chromosome breaks and
translocations. Mol Cell 21, 201-214.
Franco, S., Murphy, M. M., Li, G., Borjeson, T., Boboila, C., and Alt, F. W.
(2008). DNA-PKcs and Artemis function in the end-joining phase of
immunoglobulin heavy chain class switch recombination. J Exp Med 205,
557-564.

180
Frederico, L. A., Kunkel, T. A., and Shaw, B. R. (1990). A sensitive genetic
assay for the detection of cytosine deamination: determination of rate
constants and the activation energy. Biochemistry 29, 2532-2537.
Frederico, L. A., Kunkel, T. A., and Shaw, B. R. (1993). Cytosine
deamination in mismatched base pairs. Biochemistry 32, 6523-6530.
Friedberg, E. C., Walker, G. C., Siede, W., Wood, R. D., Schultz, R. A., and
Ellenberger, T. (2006). DNA Repair and Mutagenesis, 2nd edn (Washington
D.C.: ASM Press).
Fu, J. F., Hsu, H. C., and Shih, L. Y. (2005). MLL is fused to EB1 (MAPRE1),
which encodes a microtubule-associated protein, in a patient with acute
lymphoblastic leukemia. Genes Chromosomes Cancer 43, 206-210.
Fuscoe, J. C., Setzer, R. W., Collard, D. D., and Moore, M. M. (1996).
Quantification of t(14;18) in the lymphocytes of healthy adult humans as a
possible biomarker for environmental exposures to carcinogens.
Carcinogenesis 17, 1013-1020.
Gale, K. B., Ford, A. M., Repp, R., Borkhardt, A., Keller, C., Eden, O. B., and
Greaves, M. F. (1997). Backtracking leukemia to birth: identification of
clonotypic gene fusion sequences in neonatal blood spots. Proc Natl Acad
Sci U S A 94, 13950-13954.
Galiegue-Zouitina, S., Collyn-d'Hooghe, M., Denis, C., Mainardi, A.,
Hildebrand, M. P., Tilly, H., Bastard, C., and Kerckaert, J. P. (1994).
Molecular cloning of a t(11;14)(q13;q32) translocation breakpoint centromeric
to the BCL1-MTC. Genes Chromosomes Cancer 11, 246-255.
Galoin, S., al Saati, T., Schlaifer, D., Huynh, A., Attal, M., and Delsol, G.
(1996). Oligonucleotide clonospecific probes directed against the junctional
sequence of t(14;18): a new tool for the assessment of minimal residual
disease in follicular lymphomas. Br J Haematol 94, 676-684.
Garcia, I. S., Kaneko, Y., Gonzalez-Sarmiento, R., Campbell, K., White, L.,
Boehm, T., and Rabbitts, T. H. (1991). A study of chromosome 11p13
translocations involving TCR beta and TCR delta in human T cell leukaemia.
Oncogene 6, 577-582.
Gauss, G. H., and Lieber, M. R. (1993). Unequal signal and coding joint
formation in human V(D)J recombination. Mol Cell Biol 13, 3900-3906.
Gauss, G. H., and Lieber, M. R. (1996a). Mechanistic constraints on diversity
in human V(D)J recombination. Mol Cell Biol 16, 258-269.

181
Gauss, G. H., and Lieber, M. R. (1996b). Mechanistic constraints on diversity
in human V(D)J recombination. Mol Cell Biol 16, 258-269.
Gauwerky, C. E., Haluska, F. G., Tsujimoto, Y., Nowell, P. C., and Croce, C.
M. (1988). Evolution of B-cell malignancy: pre-B-cell leukemia resulting from
MYC activation in a B-cell neoplasm with a rearranged BCL2 gene. Proc Natl
Acad Sci U S A 85, 8548-8552.
Gelmann, E. P., Psallidopoulos, M. C., Papas, T. S., and Dalla-Favera, R.
(1983). Identification of reciprocal translocation sites within the c-myc
oncogene and immunoglobulin mu locus in a Burkitt lymphoma. Nature 306,
799-803.
Gillert, E., Leis, T., Repp, R., Reichel, M., Hosch, A., Breitenlohner, I.,
Angermuller, S., Borkhardt, A., Harbott, J., Lampert, F., et al. (1999). A DNA
damage repair mechanism is involved in the origin of chromosomal
translocations t(4;11) in primary leukemic cells. Oncogene 18, 4663-4671.
Goodsell, D. S., Kopka, M. L., Cascio, D., and Dickerson, R. E. (1993).
Crystal structure of CATGGCCATG and its implications for A-tract bending
models. Proc Natl Acad Sci U S A 90, 2930-2934.
Gough, G. W., Sullivan, K. M., and Lilley, D. M. (1986). The structure of
cruciforms in supercoiled DNA: probing the single-stranded character of
nucleotide bases with bisulphite. Embo J 5, 191-196.
Greaves, M. F., and Wiemels, J. (2003). Origins of chromosome
translocations in childhood leukaemia. Nat Rev Cancer 3, 639-649.
Gu, Y., Alder, H., Nakamura, T., Schichman, S. A., Prasad, R., Canaani, O.,
Saito, H., Croce, C. M., and Canaani, E. (1994). Sequence analysis of the
breakpoint cluster region in the ALL-1 gene involved in acute leukemia.
Cancer Res 54, 2327-2330.
Gu, Y., Cimino, G., Alder, H., Nakamura, T., Prasad, R., Canaani, O., Moir,
D. T., Jones, C., Nowell, P. C., Croce, C. M., and et al. (1992). The
(4;11)(q21;q23) chromosome translocations in acute leukemias involve the
VDJ recombinase. Proc Natl Acad Sci U S A 89, 10464-10468.
Hansen, M. J., Chen, L. H., Fejzo, M. L., and Belasco, J. G. (1994). The
ompA 5' untranslated region impedes a major pathway for mRNA
degradation in Escherichia coli. Mol Microbiol 12, 707-716.

182
Hayette, S., Tigaud, I., Vanier, A., Martel, S., Corbo, L., Charrin, C., Beillard,
E., Deleage, G., Magaud, J. P., and Rimokh, R. (2000). AF15q14, a novel
partner gene fused to the MLL gene in an acute myeloid leukaemia with a
t(11;15)(q23;q14). Oncogene 19, 4446-4450.
Heisterkamp, N., Stam, K., Groffen, J., de Klein, A., and Grosveld, G. (1985).
Structural organization of the bcr gene and its role in the Ph' translocation.
Nature 315, 758-761.
Honjo, T., Alt, F. W., and Neuberger, M. S. (2004). Molecular biology of B
cells (Amsterdam; Boston: Elsevier).
Hostein, I., Menard, A., Soubeyran, I., Eghbali, H., Debled, M., Gastaldello,
B., and Soubeyran, P. (2001). A 1-kb Bcl-2-PCR fragment detection in a
patient with follicular lymphoma and development of a new diagnostic PCR
method. Diagn Mol Pathol 10, 89-94.
Huang, F. T., Yu, K., Balter, B. B., Selsing, E., Oruc, Z., Khamlichi, A. A.,
Hsieh, C. L., and Lieber, M. R. (2007). Sequence dependence of
chromosomal R-loops at the immunoglobulin heavy-chain Smu class switch
region. Mol Cell Biol 27, 5921-5932.
Huang, F. T., Yu, K., Hsieh, C. L., and Lieber, M. R. (2006). Downstream
boundary of chromosomal R-loops at murine switch regions: implications for
the mechanism of class switch recombination. Proc Natl Acad Sci U S A 103,
5030-5035.
Jaeger, U., Bocskor, S., Le, T., Mitterbauer, G., Bolz, I., Chott, A., Kneba, A.,
Mannhalter, C., and Nadel, B. (2000). Follicular lymphomas BCL-2/IgH
junctions contain templated nucleotide insertions: novel insights into the
mechanism of t(14;18) translocation. Blood 95, 3520-3529.
Jaffe, E. S. (2001). Pathology and genetics of tumours of haematopoietic and
lymphoid tissues (Lyon; Washington, D.C.: IARC Press).
Jager, U., Bocskor, S., Le, T., Mitterbauer, G., Bolz, I., Chott, A., Kneba, M.,
Mannhalter, C., and Nadel, B. (2000). Follicular lymphomas' BCL-2/IgH
junctions contain templated nucleotide insertions: novel insights into the
mechanism of t(14;18) translocation. Blood 95, 3520-3529.
Jansen, M. W., Corral, L., van der Velden, V. H., Panzer-Grumayer, R.,
Schrappe, M., Schrauder, A., Marschalek, R., Meyer, C., den Boer, M. L.,
Hop, W. J., et al. (2007). Immunobiological diversity in infant acute
lymphoblastic leukemia is related to the occurrence and type of MLL gene
rearrangement. Leukemia 21, 633-641.

183
Janssen, J. W., Ludwig, W. D., Sterry, W., and Bartram, C. R. (1993). SIL-
TAL1 deletion in T-cell acute lymphoblastic leukemia. Leukemia 7, 1204-
1210.
Jeffs, A. R., Benjes, S. M., Smith, T. L., Sowerby, S. J., and Morris, C. M.
(1998). The BCR gene recombines preferentially with Alu elements in
complex BCR-ABL translocations of chronic myeloid leukaemia. Hum Mol
Genet 7, 767-776.
Jenner, M. J., Summers, K. E., Norton, A. J., Amess, J. A., Arch, R. S.,
Young, B. D., Lister, T. A., Fitzgibbon, J., and Goff, L. K. (2002). JH probe
real-time quantitative polymerase chain reaction assay for Bcl-2/IgH
rearrangements. Br J Haematol 118, 550-558.
Ji, W., Qu, G. Z., Ye, P., Zhang, X. Y., Halabi, S., and Ehrlich, M. (1995).
Frequent detection of bcl-2/JH translocations in human blood and organ
samples by a quantitative polymerase chain reaction assay. Cancer Res 55,
2876-2882.
Jiang, H., Chang, F. C., Ross, A. E., Lee, J., Nakayama, K., Nakayama, K.,
and Desiderio, S. (2005). Ubiquitylation of RAG-2 by Skp2-SCF links
destruction of the V(D)J recombinase to the cell cycle. Mol Cell 18, 699-709.
Jonsson, O. G., Kitchens, R. L., Baer, R. J., Buchanan, G. R., and Smith, R.
G. (1991). Rearrangements of the tal-1 locus as clonal markers for T cell
acute lymphoblastic leukemia. J Clin Invest 87, 2029-2035.
Jung, D., Giallourakis, C., Mostoslavsky, R., and Alt, F. W. (2006).
Mechanism and control of V(D)J recombination at the immunoglobulin heavy
chain locus. Annu Rev Immunol 24, 541-570.
Kagan, J., Finger, L. R., Letofsky, J., Finan, J., Nowell, P. C., and Croce, C.
M. (1989). Clustering of breakpoints on chromosome 10 in acute T-cell
leukemias with the t(10;14) chromosome translocation. Proc Natl Acad Sci U
S A 86, 4161-4165.
Kagan, J., Joe, Y. S., and Freireich, E. J. (1994). Joining of recombination
signals on the der 14q- chromosome in T-cell acute leukemia with t(10;14)
chromosome translocation. Cancer Res 54, 226-230.
Kent, W. J. (2002). BLAT--the BLAST-like alignment tool. Genome Res 12,
656-664.
Kent, W. J., Sugnet, C. W., Furey, T. S., Roskin, K. M., Pringle, T. H., Zahler,
A. M., and Haussler, D. (2002). The human genome browser at UCSC.
Genome Res 12, 996-1006.

184
Kerckaert, J. P., Deweindt, C., Tilly, H., Quief, S., Lecocq, G., and Bastard,
C. (1993). LAZ3, a novel zinc-finger encoding gene, is disrupted by recurring
chromosome 3q27 translocations in human lymphomas. Nat Genet 5, 66-70.
Kim, N., Abdulovic, A. L., Gealy, R., Lippert, M. J., and Jinks-Robertson, S.
(2007). Transcription-associated mutagenesis in yeast is directly proportional
to the level of gene expression and influenced by the direction of DNA
replication. DNA Repair (Amst) 6, 1285-1296.
Kitagawa, Y., Inoue, K., Sasaki, S., Hayashi, Y., Matsuo, Y., Lieber, M. R.,
Mizoguchi, H., Yokota, J., and Kohno, T. (2002). Prevalent involvement of
illegitimate V(D)J recombination in chromosome 9p21 deletions in lymphoid
leukemia. J Biol Chem 277, 46289-46297.
Kneba, M., Eick, S., Herbst, H., Pott, C., Bolz, I., Dallenbach, F., Hiddemann,
W., and Stein, H. (1995). Low incidence of mbr bcl-2/JH fusion genes in
Hodgkin's disease. J Pathol 175, 381-389.
Koduru, P. R., Goh, J. C., Pergolizzi, R. G., Lichtman, S. M., and Broome, J.
D. (1993). Molecular characterization of a variant Ph1 translocation t(9;22;11)
(q34;q11;q13) in chronic myelogenous leukemia (CML) reveals the
translocation of the 3'-part of BCR gene to the chromosome band 11q13.
Oncogene 8, 3239-3247.
Kuefer, M. U., Chinwalla, V., Zeleznik-Le, N. J., Behm, F. G., Naeve, C. W.,
Rakestraw, K. M., Mukatira, S. T., Raimondi, S. C., and Morris, S. W. (2003).
Characterization of the MLL partner gene AF15q14 involved in
t(11;15)(q23;q14). Oncogene 22, 1418-1424.
Kurahashi, H., Inagaki, H., Ohye, T., Kogo, H., Kato, T., and Emanuel, B. S.
(2006). Palindrome-mediated chromosomal translocations in humans. DNA
Repair (Amst) 5, 1136-1145.
Ladetto, M., Mantoan, B., Ricca, I., Astolfi, M., Drandi, D., Compagno, M.,
Vallet, S., dell'Aquila, M., Alfarano, A., Rossatto, P., et al. (2003). Recurrence
of Bcl-2/IgH polymerase chain reaction positivity following a prolonged
molecular remission can be unrelated to the original follicular lymphoma
clone. Exp Hematol 31, 784-788.
Langer, T., Metzler, M., Reinhardt, D., Viehmann, S., Borkhardt, A., Reichel,
M., Stanulla, M., Schrappe, M., Creutzig, U., Ritter, J., et al. (2003). Analysis
of t(9;11) chromosomal breakpoint sequences in childhood acute leukemia:
almost identical MLL breakpoints in therapy-related AML after treatment
without etoposides. Genes Chromosomes Cancer 36, 393-401.

185
Lansford, R., Manis, J. P., Sonoda, E., Rajewsky, K., and Alt, F. W. (1998).
Ig heavy chain class switching in Rag-deficient mice. Int Immunol 10, 325-
332.
Lewis, S. M., Agard, E., Suh, S., and Czyzyk, L. (1997). Cryptic signals and
the fidelity of V(D)J Joining. Mol Cell Biol 17, 3125-3136.
Liao, M. J., Zhang, X. X., Hill, R., Gao, J., Qumsiyeh, M. B., Nichols, W., and
Van Dyke, T. (1998). No requirement for V(D)J recombination in p53-
deficient thymic lymphoma. Mol Cell Biol 18, 3495-3501.
Lieber, M. R. (2008). The mechanism of human nonhomologous DNA end
joining. J Biol Chem 283, 1-5.
Lieber, M. R., Hesse, J. E., Mizuuchi, K., and Gellert, M. (1987).
Developmental stage specificity of the lymphoid V(D)J recombination activity.
Genes Dev 1, 751-761.
Lieber, M. R., Ma, Y., Pannicke, U., and Schwarz, K. (2004). The mechanism
of vertebrate nonhomologous DNA end joining and its role in V(D)J
recombination. DNA Repair (Amst) 3, 817-826.
Lieber, M. R., Yu, K., and Raghavan, S. C. (2006a). Roles of nonhomologous
DNA end joining, V(D)J recombination, and class switch recombination in
chromosomal translocations. DNA Repair (Amst) 5, 1234-1245.
Lieber, M. R., Yu, K., and Raghavan, S. C. (2006b). Roles of nonhomologous
DNA end joining, V(D)J recombination, and class switch recombination in
chromosomal translocations. DNA Repair 5, 1234-1245.
Lilley, D. M., and Palecek, E. (1984). The supercoil-stabilised cruciform of
ColE1 is hyper-reactive to osmium tetroxide. Embo J 3, 1187-1192.
Limpens, J., de Jong, D., van Krieken, J. H., Price, C. G., Young, B. D., van
Ommen, G. J., and Kluin, P. M. (1991). Bcl-2/JH rearrangements in benign
lymphoid tissues with follicular hyperplasia. Oncogene 6, 2271-2276.
Limpens, J., Stad, R., Vos, C., de Vlaam, C., de Jong, D., van Ommen, G. J.,
Schuuring, E., and Kluin, P. M. (1995). Lymphoma-associated translocation
t(14;18) in blood B cells of normal individuals. Blood 85, 2528-2536.
Lindahl, T., Karran, P., and Wood, R. D. (1997). DNA excision repair
pathways. Curr Opin Genet Dev 7, 158-169.

186
Litz, C. E., McClure, J. S., Copenhaver, C. M., and Brunning, R. D. (1993).
Duplication of small segments within the major breakpoint cluster region in
chronic myelogenous leukemia. Blood 81, 1567-1572.
Liu, M., Duke, J. L., Richter, D. J., Vinuesa, C. G., Goodnow, C. C.,
Kleinstein, S. H., and Schatz, D. G. (2008). Two levels of protection for the B
cell genome during somatic hypermutation. Nature 451, 841-845.
Liu, Y., Hernandez, A. M., Shibata, D., and Cortopassi, G. A. (1994). BCL2
translocation frequency rises with age in humans. Proc Natl Acad Sci U S A
91, 8910-8914.
Liu, Y., Subrahmanyam, R., Chakraborty, T., Sen, R., and Desiderio, S.
(2007). A plant homeodomain in RAG-2 that binds Hypermethylated lysine 4
of histone H3 is necessary for efficient antigen-receptor-gene rearrangement.
Immunity 27, 561-571.
Look, A. T. (1997a). Oncogenic Transcription Factors in Human Acute
Leukemias. Science 278.
Look, A. T. (1997b). Oncogenic transcription factors in the human acute
leukemias. Science 278, 1059-1064.
Lovec, H., Grzeschiczek, A., Kowalski, M. B., and Moroy, T. (1994). Cyclin
D1/bcl-1 cooperates with myc genes in the generation of B-cell lymphoma in
transgenic mice. Embo J 13, 3487-3495.
Lu, M., Dube, I., Raimondi, S., Carroll, A., Zhao, Y., Minden, M., and
Sutherland, P. (1990). Molecular characterization of the t(10;14) translocation
breakpoints in T-cell acute lymphoblastic leukemia: further evidence for
illegitimate physiological recombination. Genes Chromosomes Cancer 2,
217-222.
Luby, T. M., Schrader, C. E., Stavnezer, J., and Selsing, E. (2001). The mu
switch region tandem repeats are important, but not required, for antibody
class switch recombination. J Exp Med 193, 159-168.
Luthra, R., McBride, J. A., Hai, S., Cabanillas, F., and Pugh, W. C. (1997).
The application of fluorescence-based PCR and PCR-SSCP to monitor the
clonal relationship of cells bearing the t(14;18)(q32;q21) in sequential biopsy
specimens from patients with follicle center cell lymphoma. Diagn Mol Pathol
6, 71-77.
Ma, Y., Lu, H., Tippin, B., Goodman, M. F., Shimazaki, N., Koiwai, O., Hsieh,
C. L., Schwarz, K., and Lieber, M. R. (2004). A biochemically defined system
for mammalian nonhomologous DNA end joining. Mol Cell 16, 701-713.

187
Mao, C., Jiang, L., Melo-Jorge, M., Puthenveetil, M., Zhang, X., Carroll, M.
C., and Imanishi-Kari, T. (2004). T cell-independent somatic hypermutation in
murine B cells with an immature phenotype. Immunity 20, 133-144.
Marculescu, R., Le, T., Bocskor, S., Mitterbauer, G., Chott, A., Mannhalter,
C., Jaeger, U., and Nadel, B. (2002a). Alternative end-joining in follicular
lymphomas' t(14;18) translocation. Leukemia 16, 120-126.
Marculescu, R., Le, T., Simon, P., Jaeger, U., and Nadel, B. (2002b). V(D)J-
mediated translocations in lymphoid neoplasms: a functional assessment of
genomic instability by cryptic sites. J Exp Med 195, 85-98.
Marculescu, R., Le, T., Simon, P., Jaeger, U., and Nadel, B. (2002c). V(D)J-
mediated translocations in lymphoid neoplasms: a functional assessment of
genomic instability by cryptic sites. J Exp Med 195, 85-98.
Marschalek, R., Greil, J., Lochner, K., Nilson, I., Siegler, G., Zweckbronner,
I., Beck, J. D., and Fey, G. H. (1995). Molecular analysis of the chromosomal
breakpoint and fusion transcripts in the acute lymphoblastic SEM cell line
with chromosomal translocation t(4;11). Br J Haematol 90, 308-320.
Matolcsy, A., Warnke, R. A., and Knowles, D. M. (1997). Somatic mutations
of the translocated bcl-2 gene are associated with morphologic
transformation of follicular lymphoma to diffuse large-cell lymphoma. Ann
Oncol 8 Suppl 2, 119-122.
Matthews, A. G., Kuo, A. J., Ramon-Maiques, S., Han, S., Champagne, K.
S., Ivanov, D., Gallardo, M., Carney, D., Cheung, P., Ciccone, D. N., et al.
(2007). RAG2 PHD finger couples histone H3 lysine 4 trimethylation with
V(D)J recombination. Nature 450, 1106-1110.
McDonnell, T. J., Deane, N., Platt, F. M., Nunez, G., Jaeger, U., McKearn, J.
P., and Korsmeyer, S. J. (1989). bcl-2-immunoglobulin transgenic mice
demonstrate extended B cell survival and follicular lymphoproliferation. Cell
57, 79-88.
McGuire, E. A., Hockett, R. D., Pollock, K. M., Bartholdi, M. F., O'Brien, S. J.,
and Korsmeyer, S. J. (1989). The t(11;14)(p15;q11) in a T-cell acute
lymphoblastic leukemia cell line activates multiple transcripts, including Ttg-1,
a gene encoding a potential zinc finger protein. Mol Cell Biol 9, 2124-2132.
McHale, C. M., Wiemels, J. L., Zhang, L., Ma, X., Buffler, P. A., Guo, W.,
Loh, M. L., and Smith, M. T. (2003). Prenatal origin of TEL-AML1-positive
acute lymphoblastic leukemia in children born in California. Genes
Chromosomes Cancer 37, 36-43.

188
Meeker, T. C., Sellers, W., Harvey, R., Withers, D., Carey, K., Xiao, H.,
Block, A. M., Dadey, B., and Han, T. (1991). Cloning of the t(11;14)(q13;q32)
translocation breakpoints from two human leukemia cell lines. Leukemia 5,
733-737.
Megonigal, M. D., Rappaport, E. F., Jones, D. H., Williams, T. M., Lovett, B.
D., Kelly, K. M., Lerou, P. H., Moulton, T., Budarf, M. L., and Felix, C. A.
(1998). t(11;22)(q23;q11.2) In acute myeloid leukemia of infant twins fuses
MLL with hCDCrel, a cell division cycle gene in the genomic region of
deletion in DiGeorge and velocardiofacial syndromes. Proc Natl Acad Sci U
S A 95, 6413-6418.
Meyer, C., Schneider, B., Jakob, S., Strehl, S., Attarbaschi, A., Schnittger, S.,
Schoch, C., Jansen, M. W., van Dongen, J. J., den Boer, M. L., et al. (2006).
The MLL recombinome of acute leukemias. Leukemia 20, 777-784.
Mills, K. I., Sproul, A. M., Ogilvie, D., Elvin, P., Leibowitz, D., and Burnett, A.
K. (1992). Amplification and sequencing of genomic breakpoints located
within the M-bcr region by Vectorette-mediated polymerase chain reaction.
Leukemia 6, 481-483.
Muller, J. R., Janz, S., and Potter, M. (1995). Differences between Burkitt's
lymphomas and mouse plasmacytomas in the immunoglobulin heavy
chain/c-myc recombinations that occur in their chromosomal translocations.
Cancer Res 55, 5012-5018.
Muramatsu, M., Kinoshita, K., Fagarasan, S., Yamada, S., Shinkai, Y., and
Honjo, T. (2000). Class switch recombination and hypermutation require
activation-induced cytidine deaminase (AID), a potential RNA editing
enzyme. Cell 102, 553-563.
Murphy, W., Sarid, J., Taub, R., Vasicek, T., Battey, J., Lenoir, G., and
Leder, P. (1986). A translocated human c-myc oncogene is altered in a
conserved coding sequence. Proc Natl Acad Sci U S A 83, 2939-2943.
Nakamura, Y., Miki, T., Kawamata, N., Ohashi, K., Hirosawa, S., Kobayashi,
H., Maseki, N., Kaneko, Y., Saito, K., Enokihara, H., and Furusawa, S.
(1997). Two Burkitt-type lymphoma/leukemia-derived cell lines presenting
3q27 translocations and immunoglobulin/BCL6 chimeric transcripts.
Leukemia 11, 1993-1994.
Nash, R. A., Caldecott, K. W., Barnes, D. E., and Lindahl, T. (1997). XRCC1
protein interacts with one of two distinct forms of DNA ligase III. Biochemistry
36, 5207-5211.

189
Negrini, M., Felix, C. A., Martin, C., Lange, B. J., Nakamura, T., Canaani, E.,
and Croce, C. M. (1993). Potential topoisomerase II DNA-binding sites at the
breakpoints of a t(9;11) chromosome translocation in acute myeloid
leukemia. Cancer Res 53, 4489-4492.
Ng, H. L., Kopka, M. L., and Dickerson, R. E. (2000). The structure of a
stable intermediate in the A <--> B DNA helix transition. Proc Natl Acad Sci U
S A 97, 2035-2039.
Ngan, B. Y., Nourse, J., and Cleary, M. L. (1989). Detection of chromosomal
translocation t(14;18) within the minor cluster region of bcl-2 by polymerase
chain reaction and direct genomic sequencing of the enzymatically amplified
DNA in follicular lymphomas. Blood 73, 1759-1762.
Nilsen, H., Stamp, G., Andersen, S., Hrivnak, G., Krokan, H. E., Lindahl, T.,
and Barnes, D. E. (2003). Gene-targeted mice lacking the Ung uracil-DNA
glycosylase develop B-cell lymphomas. Oncogene 22, 5381-5386.
Nyvold, C. G., Bendix, K., Brandsborg, M., Pulczynski, S., Silkjaer, T., and
Hokland, P. (2007). Multiplex PCR for the detection of BCL-1/IGH and BCL-
2/IGH gene rearrangements--clinical validation in a prospective study of
blood and bone marrow in 258 patients with or suspected of non-Hodgkin's
lymphoma. Acta Oncol 46, 21-30.
Okazaki, I. M., Hiai, H., Kakazu, N., Yamada, S., Muramatsu, M., Kinoshita,
K., and Honjo, T. (2003). Constitutive expression of AID leads to
tumorigenesis. J Exp Med 197, 1173-1181.
Olins, P. O., and Rangwala, S. H. (1989). A novel sequence element derived
from bacteriophage T7 mRNA acts as an enhancer of translation of the lacZ
gene in Escherichia coli. J Biol Chem 264, 16973-16976.
Padlan, E. A. (1994). Anatomy of the antibody molecule. Mol Immunol 31,
169-217.
Park, J. K., Le Beau, M. M., Shows, T. B., Rowley, J. D., and Diaz, M. O.
(1992). A complex genetic rearrangement in a t(10;14)(q24;q11) associated
with T-cell acute lymphoblastic leukemia. Genes Chromosomes Cancer 4,
32-40.
Peyret, N., Seneviratne, P. A., Allawi, H. T., and SantaLucia, J., Jr. (1999).
Nearest-neighbor thermodynamics and NMR of DNA sequences with internal
A.A, C.C, G.G, and T.T mismatches. Biochemistry 38, 3468-3477.
Pfeifer, G. P. (2006). Mutagenesis at methylated CpG sequences. Curr Top
Microbiol Immunol 301, 259-281.

190
Pham, P., Bransteitter, R., Petruska, J., and Goodman, M. F. (2003a).
Processive AID-catalysed cytosine deamination on single-stranded DNA
simulates somatic hypermutation. Nature 424, 103-107.
Pham, P., Bransteitter, R., Petruska, J., and Goodman, M. F. (2003b).
Processive AID-catalyzed cytosine deamination on single-stranded DNA
stimulates somatic hypermutation. Nature 424, 103-107.
Pine, S. R., Wiemels, J. L., Jayabose, S., and Sandoval, C. (2003). TEL-
AML1 fusion precedes differentiation to pre-B cells in childhood acute
lymphoblastic leukemia. Leuk Res 27, 155-164.
Pott, C., Tiemann, M., Linke, B., Ott, M. M., von Hofen, M., Bolz, I.,
Hiddemann, W., Parwaresch, R., and Kneba, M. (1998). Structure of Bcl-1
and IgH-CDR3 rearrangements as clonal markers in mantle cell lymphomas.
Leukemia 12, 1630-1637.
Povirk, L. F. (2006). Biochemical mechanisms of chromosomal translocations
resulting from DNA double-strand breaks. DNA Repair (Amst) 5, 1199-1212.
Price, C. G., Tuszynski, A., Watt, S. M., Murdoch, S. J., Lister, T. A., and
Young, B. D. (1991). Detection of additional JH/BCL2 translocations in
follicular lymphoma. Leukemia 5, 548-554.
Rabbitts, P. H., Douglas, J., Fischer, P., Nacheva, E., Karpas, A., Catovsky,
D., Melo, J. V., Baer, R., Stinson, M. A., and Rabbitts, T. H. (1988).
Chromosome abnormalities at 11q13 in B cell tumours. Oncogene 3, 99-103.
Raffini, L. J., Slater, D. J., Rappaport, E. F., Lo Nigro, L., Cheung, N. K.,
Biegel, J. A., Nowell, P. C., Lange, B. J., and Felix, C. A. (2002). Panhandle
and reverse-panhandle PCR enable cloning of der(11) and der(other)
genomic breakpoint junctions of MLL translocations and identify complex
translocation of MLL, AF-4, and CDK6. Proc Natl Acad Sci U S A 99, 4568-
4573.
Raghavan, S. C., Chastain, P., Lee, J. S., Hegde, B. G., Houston, S.,
Langen, R., Hsieh, C. L., Haworth, I. S., and Lieber, M. R. (2005a). Evidence
for a triplex DNA conformation at the bcl-2 major breakpoint region of the
t(14;18) translocation. J Biol Chem 280, 22749-22760.
Raghavan, S. C., Gu, J., Swanson, P. C., and Lieber, M. R. (2007). The
structure-specific nicking of small heteroduplexes by the RAG complex:
implications for lymphoid chromosomal translocations. DNA Repair (Amst) 6,
751-759.

191
Raghavan, S. C., Houston, S., Hegde, B. G., Langen, R., Haworth, I. S., and
Lieber, M. R. (2004a). Stability and strand asymmetry in the non-B DNA
structure at the bcl-2 major breakpoint region. J Biol Chem 279, 46213-
46225.
Raghavan, S. C., Hsieh, C. L., and Lieber, M. R. (2005b). Both V(D)J coding
ends but neither signal end can recombine at the bcl-2 major breakpoint
region, and the rejoining is ligase IV dependent. Mol Cell Biol 25, 6475-6484.
Raghavan, S. C., Kirsch, I. R., and Lieber, M. R. (2001a). Analysis of the
V(D)J recombination efficiency at lymphoid chromosomal translocation
breakpoints. J Biol Chem 276, 29126-29133.
Raghavan, S. C., Kirsch, I. R., and Lieber, M. R. (2001b). Analysis of the
V(D)J recombination efficiency at lymphoid chromosomal translocation
breakpoints. J Biol Chem 276, 29126-29133.
Raghavan, S. C., Swanson, P. C., Ma, Y., and Lieber, M. R. (2005c). Double-
strand break formation by the RAG complex at the bcl-2 major breakpoint
region and at other non-B DNA structures in vitro. Mol Cell Biol 25, 5904-
5919.
Raghavan, S. C., Swanson, P. C., Wu, X., Hsieh, C.-L., and Lieber, M. R.
(2004a). A non-B-DNA structure at the bcl-2 major break point region is
cleaved by the RAG complex. Nature 428, 88-93.
Raghavan, S. C., Swanson, P. C., Wu, X., Hsieh, C. L., and Lieber, M. R.
(2004b). A non-B-DNA structure at the Bcl-2 major breakpoint region is
cleaved by the RAG complex. Nature 428, 88-93.
Raghavan, S. C., Tsai, A., Hsieh, C. L., and Lieber, M. R. (2006). Analysis of
non-B DNA structure at chromosomal sites in the mammalian genome.
Methods Enzymol 409, 301-316.
Rauzy, O., Galoin, S., Chale, J. J., Adoue, D., Albarede, J. L., Delsol, G., and
al Saati, T. (1998). Detection of t(14;18) carrying cells in bone marrow and
peripheral blood from patients affected by non-lymphoid diseases. Mol Pathol
51, 333-338.
Reichel, M., Gillert, E., Angermuller, S., Hensel, J. P., Heidel, F., Lode, M.,
Leis, T., Biondi, A., Haas, O. A., Strehl, S., et al. (2001). Biased distribution
of chromosomal breakpoints involving the MLL gene in infants versus
children and adults with t(4;11) ALL. Oncogene 20, 2900-2907.

192
Reichel, M., Gillert, E., Breitenlohner, I., Repp, R., Greil, J., Beck, J. D., Fey,
G. H., and Marschalek, R. (1999). Rapid isolation of chromosomal
breakpoints from patients with t(4;11) acute lymphoblastic leukemia:
implications for basic and clinical research. Cancer Res 59, 3357-3362.
Reichel, M., Gillert, E., Nilson, I., Siegler, G., Greil, J., Fey, G. H., and
Marschalek, R. (1998). Fine structure of translocation breakpoints in
leukemic blasts with chromosomal translocation t(4;11): the DNA damage-
repair model of translocation. Oncogene 17, 3035-3044.
Reitmair, A. H., Schmits, R., Ewel, A., Bapat, B., Redston, M., Mitri, A.,
Waterhouse, P., Mittrucker, H. W., Wakeham, A., Liu, B., and et al. (1995).
MSH2 deficient mice are viable and susceptible to lymphoid tumours. Nat
Genet 11, 64-70.
Rimokh, R., Berger, F., Delsol, G., Digonnet, I., Rouault, J. P., Tigaud, J. D.,
Gadoux, M., Coiffier, B., Bryon, P. A., and Magaud, J. P. (1994). Detection of
the chromosomal translocation t(11;14) by polymerase chain reaction in
mantle cell lymphomas. Blood 83, 1871-1875.
Roulland, S., Lebailly, P., and Gauduchon, P. (2003). Correspondence re:
Welzel et al, Cancer Res, 61: 1629-1636. Cancer Res 63, 1722-1723.
Roy, D., Yu, K., and Lieber, M. R. (2008). Mechanism of R-loop formation at
immunoglobulin class switch sequences. Mol Cell Biol 28, 50-60.
Saito, H., Hayday, A. C., Wiman, K., Hayward, W. S., and Tonegawa, S.
(1983). Activation of the c-myc gene by translocation: a model for
translational control. Proc Natl Acad Sci U S A 80, 7476-7480.
Salvati, P. D., Watt, P. M., Thomas, W. R., and Kees, U. R. (1999). Molecular
characterization of a complex chromosomal translocation breakpoint t(10;14)
including the HOX11 oncogene locus. Leukemia 13, 975-979.
Samarsky, D. A., Ferbeyre, G., Bertrand, E., Singer, R. H., Cedergren, R.,
and Fournier, M. J. (1999). A small nucleolar RNA:ribozyme hybrid cleaves a
nucleolar RNA target in vivo with near-perfect efficiency. Proc Natl Acad Sci
U S A 96, 6609-6614.
Sanchez-Izquierdo, D., Siebert, R., Harder, L., Marugan, I., Gozzetti, A.,
Price, H. P., Gesk, S., Hernandez-Rivas, J. M., Benet, I., Sole, F., et al.
(2001). Detection of translocations affecting the BCL6 locus in B cell non-
Hodgkin's lymphoma by interphase fluorescence in situ hybridization.
Leukemia 15, 1475-1484.
Schatz, D. G. (2004). V(D)J recombination. Immunol Rev 200, 5-11.

193
Schlissel, M., Constantinescu, A., Morrow, T., Baxter, M., and Peng, A.
(1993). Double-strand signal sequence breaks in V(D)J recombination are
blunt, 5'-phosphorylated, RAG-dependent, and cell cycle regulated. Genes
Dev 7, 2520-2532.
Schmitt, C., Balogh, B., Grundt, A., Buchholtz, C., Leo, A., Benner, A.,
Hensel, M., Ho, A. D., and Leo, E. (2006). The bcl-2/IgH rearrangement in a
population of 204 healthy individuals: occurrence, age and gender
distribution, breakpoints, and detection method validity. Leuk Res 30, 745-
750.
Schwindt, H., Akasaka, T., Zuhlke-Jenisch, R., Hans, V., Schaller, C.,
Klapper, W., Dyer, M. J., Siebert, R., and Deckert, M. (2006). Chromosomal
translocations fusing the BCL6 gene to different partner loci are recurrent in
primary central nervous system lymphoma and may be associated with
aberrant somatic hypermutation or defective class switch recombination. J
Neuropathol Exp Neurol 65, 776-782.
Segal, G. H., Masih, A. S., Fox, A. C., Jorgensen, T., Scott, M., and Braylan,
R. C. (1995). CD5-expressing B-cell non-Hodgkin's lymphomas with bcl-1
gene rearrangement have a relatively homogeneous immunophenotype and
are associated with an overall poor prognosis. Blood 85, 1570-1579.
Shapiro, R., DiFate, V., and Welcher, M. (1974). Deamination of cytosine
derivatives by bisulfite. Mechanism of the reaction. J Am Chem Soc 96, 906-
912.
Shen, J. C., Rideout, W. M., 3rd, and Jones, P. A. (1994). The rate of
hydrolytic deamination of 5-methylcytosine in double-stranded DNA. Nucleic
Acids Res 22, 972-976.
Showe, L. C., Ballantine, M., Nishikura, K., Erikson, J., Kaji, H., and Croce,
C. M. (1985). Cloning and sequencing of a c-myc oncogene in a Burkitt's
lymphoma cell line that is translocated to a germ line alpha switch region.
Mol Cell Biol 5, 501-509.
Sinden, R. R. (1994). DNA structure and function (San Diego: Academic
Press).
Smith, D. P., Bath, M. L., Harris, A. W., and Cory, S. (2005). T-cell
lymphomas mask slower developing B-lymphoid and myeloid tumours in
transgenic mice with broad haemopoietic expression of MYC. Oncogene 24,
3544-3553.

194
Sono, M., Wataya, Y., and Hayatsu, H. (1973). Role of bisulfite in the
deamination and the hydrogen isotope exchange of cytidylic acid. J Am
Chem Soc 95, 4745-4749.
Soubeyran, P., Hostein, I., Debled, M., Eghbali, H., Soubeyran, I., Bonichon,
F., Astier-Gin, T., and Hoerni, B. (1999). Evolution of BCL-2/IgH hybrid gene
RNA expression during treatment of T(14;18)-bearing follicular lymphomas.
Br J Cancer 81, 860-869.
Sowerby, S. J., Kennedy, M. A., Fitzgerald, P. H., and Morris, C. M. (1993).
DNA sequence analysis of the major breakpoint cluster region of the BCR
gene rearranged in Philadelphia-positive human leukemias. Oncogene 8,
1679-1683.
Stamatopoulos, K., Kosmas, C., Belessi, C., Kyriazopoulos, P., Papadaki, T.,
Anagnostou, D., and Loukopoulos, D. (1999). Molecular analysis of bcl-1/IgH
junctional sequences in mantle cell lymphoma: potential mechanism of the
t(11;14) chromosomal translocation. Br J Haematol 105, 190-197.
Stavnezer, J., Guikema, J. E., and Schrader, C. E. (2008). Mechanism and
Regulation of Class Switch Recombination. Annu Rev Immunol 26, 261-292.
Summers, K. E., Goff, L. K., Wilson, A. G., Gupta, R. K., Lister, T. A., and
Fitzgibbon, J. (2001). Frequency of the Bcl-2/IgH rearrangement in normal
individuals: implications for the monitoring of disease in patients with follicular
lymphoma. J Clin Oncol 19, 420-424.
Super, H. G., Strissel, P. L., Sobulo, O. M., Burian, D., Reshmi, S. C., Roe,
B., Zeleznik-Le, N. J., Diaz, M. O., and Rowley, J. D. (1997). Identification of
complex genomic breakpoint junctions in the t(9;11) MLL-AF9 fusion gene in
acute leukemia. Genes Chromosomes Cancer 20, 185-195.
Tabone, T., Sallmann, G., Chiotis, M., Law, M., and Cotton, R. (2006).
Chemical cleavage of mismatch (CCM) to locate base mismatches in
heteroduplex DNA. Nat Protoc 1, 2297-2304.
Takacs, I., Zeher, M., Urban, L., Szegedi, G., and Semsei, I. (2000).
Diagnostic value of the detection of t(14;18) chromosome translocation in
malignant hematological and immunopathological diseases using
polymerase chain reaction. Acta Med Okayama 54, 185-192.
Takai, D., and Jones, P. A. (2003). The CpG island searcher: a new WWW
resource. In Silico Biol 3, 235-240.

195
Tomkinson, A. E., Vijayakumar, S., Pascal, J. M., and Ellenberger, T. (2006).
DNA ligases: structure, reaction mechanism, and function. Chem Rev 106,
687-699.
Tonegawa, S. (1983a). Somatic generation of antibody diversity. Nature 302,
575-581.
Tonegawa, S. (1983b). Somatic generation of antibody diversity. Nature 302,
575-581.
Tsujimoto, Y., Finger, L. R., Yunis, J., Nowell, P. C., and Croce, C. M.
(1984a). Cloning of the chromosome breakpoint of neoplastic B cells with the
t(14;18) chromosome translocation. Science 226, 1097-1099.
Tsujimoto, Y., Gorham, J., Cossman, J., Jaffe, E., and Croce, C. M. (1985a).
The t(14;18) chromosome translocations involved in B-cell neoplasms result
from mistakes in VDJ joining. Science 229, 1390-1393.
Tsujimoto, Y., Jaffe, E., Cossman, J., Gorham, J., Nowell, P. C., and Croce,
C. M. (1985b). Clustering of breakpoints on chromosome 11 in human B-cell
neoplasms with the t(11;14) chromosome translocation. Nature 315, 340-
343.
Tsujimoto, Y., Louie, E., Bashir, M. M., and Croce, C. M. (1988). The
reciprocal partners of both the t(14; 18) and the t(11; 14) translocations
involved in B-cell neoplasms are rearranged by the same mechanism.
Oncogene 2, 347-351.
Tsujimoto, Y., Yunis, J., Onorato-Showe, L., Erikson, J., Nowell, P. C., and
Croce, C. M. (1984b). Molecular cloning of the chromosomal breakpoint of B-
cell lymphomas and leukemias with the t(11;14) chromosome translocation.
Science 224, 1403-1406.
Van Vlierberghe, P., van Grotel, M., Beverloo, H. B., Lee, C., Helgason, T.,
Buijs-Gladdines, J., Passier, M., van Wering, E. R., Veerman, A. J., Kamps,
W. A., et al. (2006). The cryptic chromosomal deletion del(11)(p12p13) as a
new activation mechanism of LMO2 in pediatric T-cell acute lymphoblastic
leukemia. Blood 108, 3520-3529.
Vaux, D. L., Cory, S., and Adams, J. M. (1988). Bcl-2 gene promotes
haemopoietic cell survival and cooperates with c-myc to immortalize pre-B
cells. Nature 335, 440-442.
Vega, F., and Medeiros, L. J. (2003). Chromosomal translocations involved
in non-Hodgkin lymphomas. Arch Pathol Lab Med 127, 1148-1160.

196
Vieira, L., Sousa, A. C., Matos, P., Marques, B., Alaiz, H., Ribeiro, M. J.,
Braga, P., da Silva, M. G., and Jordan, P. (2006). Three-way translocation
involves MLL, MLLT3, and a novel cell cycle control gene, FLJ10374, in the
pathogenesis of acute myeloid leukemia with t(9;11;19)(p22;q23;p13.3).
Genes Chromosomes Cancer 45, 455-469.
Vigneault, F., and Drouin, R. (2005). Optimal conditions and specific
characteristics of Vent exo- DNA polymerase in ligation-mediated
polymerase chain reaction protocols. Biochem Cell Biol 83, 147-165.
Wada, M., Bartram, C. R., Nakamura, H., Hachiya, M., Chen, D. L.,
Borenstein, J., Miller, C. W., Ludwig, L., Hansen-Hagge, T. E., Ludwig, W.
D., and et al. (1993). Analysis of p53 mutations in a large series of lymphoid
hematologic malignancies of childhood. Blood 82, 3163-3169.
Walsh, C. P., and Xu, G. L. (2006). Cytosine methylation and DNA repair.
Curr Top Microbiol Immunol 301, 283-315.
Wang, Y. L., Addya, K., Edwards, R. H., Rennert, H., Dodson, L., Leonard,
D. G., and Wilson, R. B. (1998). Novel bcl-2 breakpoints in patients with
follicular lymphoma. Diagn Mol Pathol 7, 85-89.
Wechsler, D. S., Engstrom, L. D., Alexander, B. M., Motto, D. G., and
Roulston, D. (2003). A novel chromosomal inversion at 11q23 in infant acute
myeloid leukemia fuses MLL to CALM, a gene that encodes a clathrin
assembly protein. Genes Chromosomes Cancer 36, 26-36.
Weinberg, O. K., Ai, W. Z., Mariappan, M. R., Shum, C., Levy, R., and Arber,
D. A. (2007). ''Minor'' BCL2 breakpoints in follicular lymphoma: frequency
and correlation with grade and disease presentation in 236 cases. J Mol
Diagn 9, 530-537.
Weinstock, D. M., Brunet, E., and Jasin, M. (2007). Formation of NHEJ-
derived reciprocal chromosomal translocations does not require Ku70. Nat
Cell Biol 9, 978-981.
Welzel, N., Le, T., Marculescu, R., Mitterbauer, G., Chott, A., Pott, C., Kneba,
M., Du, M. Q., Kusec, R., Drach, J., et al. (2001a). Templated nucleotide
addition and immunoglobulin JH-gene utilization in t(11;14) junctions:
implications for the mechanism of translocation and the origin of mantle cell
lymphoma. Cancer Res 61, 1629-1636.

197
Welzel, N., T, T. L., Marculescu, R., Mitterbauer, G., Chott, A., Pott, C.,
Kneba, M., Du, M. Q., Kusec, R., Drach, J., et al. (2001b). Templated
nucleotide addition and immunoglobulin JH-gene utilization in t(11;14)
junctions: implications for the mechanism of translocation and the origin of
mantle cell lymphoma. Cancer Res, 1629-1636.
Wiemels, J. L., Alexander, F. E., Cazzaniga, G., Biondi, A., Mayer, S. P., and
Greaves, M. (2000). Microclustering of TEL-AML1 translocation breakpoints
in childhood acute lymphoblastic leukemia. Genes Chromosomes Cancer 29,
219-228.
Wiemels, J. L., and Greaves, M. (1999). Structure and possible mechanisms
of TEL-AML1 gene fusions in childhood acute lymphoblastic leukemia.
Cancer Res 59, 4075-4082.
Wiemels, J. L., Leonard, B. C., Wang, Y., Segal, M. R., Hunger, S. P., Smith,
M. T., Crouse, V., Ma, X., Buffler, P. A., and Pine, S. R. (2002a). Site-specific
translocation and evidence of postnatal origin of the t(1;19) E2A-PBX1 fusion
in childhood acute lymphoblastic leukemia. Proc Natl Acad Sci U S A 99,
15101-15106.
Wiemels, J. L., Xiao, Z., Buffler, P. A., Maia, A. T., Ma, X., Dicks, B. M.,
Smith, M. T., Zhang, L., Feusner, J., Wiencke, J., et al. (2002b). In utero
origin of t(8;21) AML1-ETO translocations in childhood acute myeloid
leukemia. Blood 99, 3801-3805.
Wilda, M., Busch, K., Klose, I., Keller, T., Woessmann, W., Kreuder, J.,
Harbott, J., and Borkhardt, A. (2004). Level of MYC overexpression in
pediatric Burkitt's lymphoma is strongly dependent on genomic breakpoint
location within the MYC locus. Genes Chromosomes Cancer 41, 178-182.
Williams, M. E., Swerdlow, S. H., and Meeker, T. C. (1993). Chromosome
t(11;14)(q13;q32) breakpoints in centrocytic lymphoma are highly localized at
the bcl-1 major translocation cluster. Leukemia 7, 1437-1440.
Willis, T. G., Jadayel, D. M., Coignet, L. J., Abdul-Rauf, M., Treleaven, J. G.,
Catovsky, D., and Dyer, M. J. (1997). Rapid molecular cloning of
rearrangements of the IGHJ locus using long-distance inverse polymerase
chain reaction. Blood 90, 2456-2464.
Wiman, K. G., Clarkson, B., Hayday, A. C., Saito, H., Tonegawa, S., and
Hayward, W. S. (1984). Activation of a translocated c-myc gene: role of
structural alterations in the upstream region. Proc Natl Acad Sci U S A 81,
6798-6802.

198
Wong, E., Yang, K., Kuraguchi, M., Werling, U., Avdievich, E., Fan, K.,
Fazzari, M., Jin, B., Brown, A. M., Lipkin, M., and Edelmann, W. (2002).
Mbd4 inactivation increases Cright-arrowT transition mutations and promotes
gastrointestinal tumor formation. Proc Natl Acad Sci U S A 99, 14937-14942.
Wyatt, R. T., Rudders, R. A., Zelenetz, A., Delellis, R. A., and Krontiris, T. G.
(1992). BCL2 oncogene translocation is mediated by a chi-like consensus. J
Exp Med 175, 1575-1588.
Xia, Y., Brown, L., Tsan, J. T., Yang, C. Y., Siciliano, M. J., Crist, W. M.,
Carroll, A. J., and Baer, R. (1992). The translocation (1;14)(p34;q11) in
human T-cell leukemia: chromosome breakage 25 kilobase pairs
downstream of the TAL1 protooncogene. Genes Chromosomes Cancer 4,
211-216.
Xiao, Z., Greaves, M. F., Buffler, P., Smith, M. T., Segal, M. R., Dicks, B. M.,
Wiencke, J. K., and Wiemels, J. L. (2001). Molecular characterization of
genomic AML1-ETO fusions in childhood leukemia. Leukemia 15, 1906-
1913.
Yan, C. T., Boboila, C., Souza, E. K., Franco, S., Hickernell, T. R., Murphy,
M., Gumaste, S., Geyer, M., Zarrin, A. A., Manis, J. P., et al. (2007). IgH
class switching and translocations use a robust non-classical end-joining
pathway. Nature 449, 478-482.
Yang, X., Lee, K., Said, J., Gong, X., and Zhang, K. (2006). Association of
Ig/BCL6 translocations with germinal center B lymphocytes in human
lymphoid tissues: implications for malignant transformation. Blood 108, 2006-
2012.
Ye, B. H., Chaganti, S., Chang, C. C., Niu, H., Corradini, P., Chaganti, R. S.,
and Dalla-Favera, R. (1995). Chromosomal translocations cause deregulated
BCL6 expression by promoter substitution in B cell lymphoma. Embo J 14,
6209-6217.
Yoffe, G., Schneider, N., Van Dyk, L., Yang, C. Y., Siciliano, M., Buchanan,
G., Capra, J. D., and Baer, R. (1989). The chromosome translocation
(11;14)(p13;q11) associated with T-cell acute lymphocytic leukemia: an
11p13 breakpoint cluster region. Blood 74, 374-379.
Yoshida, S., Kaneita, Y., Aoki, Y., Seto, M., Mori, S., and Moriyama, M.
(1999). Identification of heterologous translocation partner genes fused to the
BCL6 gene in diffuse large B-cell lymphomas: 5'-RACE and LA - PCR
analyses of biopsy samples. Oncogene 18, 7994-7999.

199
Young, K. H., Xie, Q., Zhou, G., Eickhoff, J. C., Sanger, W. G., Aoun, P., and
Chan, W. C. (2008). Transformation of follicular lymphoma to precursor B-
cell lymphoblastic lymphoma with c-myc gene rearrangement as a critical
event. Am J Clin Pathol 129, 157-166.
Yu, K., Chedin, F., Hsieh, C. L., Wilson, T. E., and Lieber, M. R. (2003). R-
loops at immunoglobulin class switch regions in the chromosomes of
stimulated B cells. Nat Immunol 4, 442-451.
Yu, K., Huang, F. T., and Lieber, M. R. (2004a). DNA substrate length and
surrounding sequence affect the activation-induced deaminase activity at
cytidine. J Biol Chem 279, 6496-6500.
Yu, K., Huang, F. T., and Lieber, M. R. (2004b). DNA substrate length and
surrounding sequence affect the activation induced deaminase activity at
cytidine. J Biol Chem 279, 6496-6500.
Yu, K., and Lieber, M. R. (2003). Nucleic acid structures and enzymes in the
immunoglobulin class switch recombination mechanism. DNA Repair (Amst)
2, 1163-1174.
Yu, K., Roy, D., Bayramyan, M., Haworth, I. S., and Lieber, M. R. (2005).
Fine-structure analysis of activation-induced deaminase accessibility to class
switch region R-loops. Mol Cell Biol 25, 1730-1736.
Yu, K., Taghva, A., and Lieber, M. R. (2002). The cleavage efficiency of the
human immunoglobulin heavy chain VH elements by the RAG complex:
implications for the immune repertoire. J Biol Chem 277, 5040-5046.
Yu, K., Taghva, A., Ma, Y., and Lieber, M. R. (2004c). Kinetic analysis of the
nicking and hairpin formation steps in V(D)J recombination. DNA Repair
(Amst) 3, 67-75.
Zarrin, A. A., Del Vecchio, C., Tseng, E., Gleason, M., Zarin, P., Tian, M.,
and Alt, F. W. (2007). Antibody class switching mediated by yeast
endonuclease-generated DNA breaks. Science 315, 377-381.
Zhang, J. G., Goldman, J. M., and Cross, N. C. (1995). Characterization of
genomic BCR-ABL breakpoints in chronic myeloid leukaemia by PCR. Br J
Haematol 90, 138-146.
Zhang, M., and Swanson, P. C. (2008). V(D)J recombinase binding and
cleavage of cryptic recombination signal sequences identified from lymphoid
malignancies. J Biol Chem 283, 6717-6727.

200
Zhang, Y., Strissel, P., Strick, R., Chen, J., Nucifora, G., Le Beau, M. M.,
Larson, R. A., and Rowley, J. D. (2002). Genomic DNA breakpoints in
AML1/RUNX1 and ETO cluster with topoisomerase II DNA cleavage and
DNase I hypersensitive sites in t(8;21) leukemia. Proc Natl Acad Sci U S A
99, 3070-3075.
Zutter, M., Hockett, R. D., Roberts, C. W., McGuire, E. A., Bloomstone, J.,
Morton, C. C., Deaven, L. L., Crist, W. M., Carroll, A. J., and Korsmeyer, S. J.
(1990). The t(10;14)(q24;q11) of T-cell acute lymphoblastic leukemia
juxtaposes the delta T-cell receptor with TCL3, a conserved and activated
locus at 10q24. Proc Natl Acad Sci U S A 87, 3161-3165.

Abstract (if available)

Abstract The bcl-2 translocation, t(14

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Molecular genetic analysis of human genetic lesions

PDF

The mechanism of mammalian immunoglobulin class switch recombination: R-loop structures and activation-induced deaminase site preferences

PDF

Mechanistic basis for chromosomal translocations at the E2A gene

PDF

Biochemistry and reconstitution of V(D)J recombination in a purified system

PDF

Identification of DNA methylation markers in diffuse large B-cell lymphoma

PDF

The mechanism of R-loop formation in mammalian immunoglobulin class switch recombination

PDF

Studies on the role of Artemis in non-homologous DNA end-joining to understand the mechanism and discover therapies

PDF

Mechanisms of nucleases in non-homologous DNA end joining

PDF

Generation and characterization of fully human anti-CD19 chimeric antigen receptor T (CAR-T) cells for the treatment of hematologic malignancies

PDF

The mechanism of mammalian immunoglobulin class switch recombination

PDF

Molecular elucidation of nonhomologous DNA end-joining in the context of nucleosome core particles

PDF

Enhanced Burkitt 's lymphoma cell killing by the combination treatments of bortezomib with celecoxib and 2,5-dimethyl-celecoxib (DMC)

PDF

Generation and characterization of humanized anti-CD19 chimeric antigen receptor T (CAR-T) cells for the treatment of hematologic malignancies

PDF

Molecular mechanisms of chemoresistance in breast cancer

PDF

Investigating the role of SASH1 gene located on chromosome 6 in ovarian cancer

PDF

Role of TLR4 and AID in lymphomagenesis induced by obesity and hepatitis C virus

PDF

Genetic interaction between androgen receptor and Lef1 in bone mass control

PDF

The kinetic study of engineered MBD domain interactions with methylated DNA: insight into binding of methylated DNA by MBD2b

PDF

CpG methylation profiling in lung cancer cell lines, tumors and non-tumors

PDF

Induction of hypersignaling as a therapeutic approach for treatment of BCR-ABL1 positive Acute Lymphoblastic Leukemia (ALL) cells

Asset Metadata

Creator Tsai, Albert G. (author)

Core Title Molecular mechanisms of recurrent chromosomal translocations in human leukemias and lymphomas

School Keck School of Medicine

Degree Doctor of Philosophy

Degree Program Biochemistry and Molecular Biology

Degree Conferral Date 2008-08

Publication Date 07/21/2008

Defense Date 06/05/2008

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag chromosomal rearrangement,chromosomal translocation,DNA double-strand break repair,non-Hodgkin's lymphoma,OAI-PMH Harvest

Language English

Advisor Lieber, Michael R. (committee chair), Hacia, Joseph G. (committee member), Haworth, Ian S. (committee member), Hsieh, Chih-Lin (committee member), Wang, Clay C. C. (committee member)

Creator Email albertts@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-m1365

Unique identifier UC1133439

Identifier etd-Tsai-20080721 (filename),usctheses-m40 (legacy collection record id),usctheses-c127-196021 (legacy record id),usctheses-m1365 (legacy record id)

Legacy Identifier etd-Tsai-20080721.pdf

Dmrecord 196021

Document Type Dissertation

Rights Tsai, Albert G.

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