UNIVERSITY OF CINCINNATI

Date:__09/27/2004______

I, _Katherine L. Lillard______, hereby submit this work as part of the requirements for the degree of: Doctor of Philosophy in: Molecular Genetics, Biochemistry, and Microbiology It is entitled: The BLM helicase functions in alternative lengthening of telomeres.

This work and its defense approved by:

Chair: Joanna Groden______Iain Cartwright______Carolyn Price______James Stringer______

Kathleen Dixon______

THE BLM HELICASE FUNCTIONS IN ALTERNATIVE LENGTHENING OF TELOMERES

A dissertation submitted to the

Division of Research and Advanced Studies Of the University of Cincinnati

in partial fulfillment of the requirements for the degree of

DOCTORATE OF PHILOSOPHY (Ph.D.)

In the Department of Molecular Genetics, Biochemistry & Microbiology Of the College of Medicine

2004

by

Kate Lillard-Wetherell B.S., University of Texas at Austin, 1998

Committee Chair: Joanna Groden, Ph.D. ABSTRACT

Somatic cells from persons with the inherited chromosome breakage syndrome

Bloom syndrome (BS) feature excessive chromosome breakage, intra-and inter- chromosomal homologous exchanges and telomeric associations. The gene mutated in

BS, BLM, encodes a RecQ-like ATP-dependent 3’-to-5’ helicase that presumably functions in some types of DNA transactions. As the absence of BLM is associated with excessive recombination, in vitro experiments have tested the ability of BLM to suppress recombination and/or resolve recombination intermediates. In vitro, BLM promotes branch migration of Holliday junctions, resolves D-loops and unwinds G-quadruplex

DNA. A function for BLM in maintaining telomeres is suggested by the latter, since D- loops and perhaps G-quadruplex structures are thought to be present at telomeres.

In the present study, the association of BLM with telomeres was investigated.

Given the association of BLM with recombination, it was of particular interest to determine the nuclear localization of BLM with respect to telomeres in cells using recombinational pathways for telomere lengthening, termed ALT. Using the telomere repeat protein TRF2 as a telomere marker, we demonstrate that BLM co-localizes with telomeres in cells using ALT, but not in telomerase-positive or primary cells. BLM co- localizes with TRF2 in foci actively synthesizing DNA during late S and G2/M; co- localization is enriched during these phases of the cell cycle when ALT is thought to occur. By immunoprecipitation, BLM associates with telomeres and TRF2 in cells using

ALT. In S. cerevisiae, we demonstrate that BLM expression rescues a defect in recombinational telomere lengthening associated with absence of SGS1. These data

ii establish a spatial and temporal association of BLM with telomere synthesis in cells using ALT and demonstrate conserved function(s) for BLM and SGS1 in ALT.

Additionally, the regulation of BLM activity using telomere substrates was investigated in vitro. We find that TRF1 and TRF2 physically and functionally interact with BLM in vitro. TRF2 stimulates BLM unwinding of telomeric and non-telomeric substrates. Conversely, TRF1 inhibits BLM unwinding of telomeric substrates only.

Neither TRF1 nor TRF2 regulate unwinding activity of the UvrD helicase. Finally, BLM helicase activity is stimulated by TRF2 with equimolar concentrations of TRF1, but not when TRF1 is present in molar excess. Based on these data, we present a model for the coordinated regulation of BLM helicase activity by TRF1 and TRF2 at telomeres in cells using ALT.

iii Acknowledgments

Thank you to members of the Groden lab, past and present. To Joanna for being a wonderful mentor who always had words of encouragement when I needed them most, to Rose for taking care of us all (we would be hopelessly lost without you), to Amod,

Chelsea, the Gregs and all the Fries gang for making me relax and have fun (every now and again), to Bill for keeping me company in the late hours my last few months in the lab, to Kathy and Mary for all your support, guidance and wisdom, and a very special thank you to Al… you are an absolute treasure! I’m not sure I would have made it through without your help and companionship.

Thank you to my grandmother, my parents, and all my siblings for their continual and everlasting support, to Shonnie for being a friend I can always count on, and last but certainly not least, I thank my husband Seel for being a shoulder to cry on, a friend to laugh with and a partner to rely on for the last ten years and the many years to come.

iv

Table of Contents

Abstract ii

Acknowledgements iv

Table of Contents 1

Abbreviations 4

List of Tables 6

List of Figures 6

Chapter One. Literature Review

Introduction 8 Clinical features of Bloom syndrome. 8 Bloom’s syndrome is characterized by a mutator phenotype. 11 Sensitivity of BS cells to DNA damaging agents. 15 BLM encodes a RecQ-like helicase. 16 Protein partners implicate BLM in DNA repair. 18 BLM is a structure-specific helicase. 19 BLM is regulated by the cell cycle. 22 BLM responds to DNA damage. 23 BLM in DSB repair. 24 Repair of stalled replication forks. 27 BLM is a target and regulator of apoptosis. 27 Other human disorders associated with RecQ deficiency. 29 Sgs1 is a functional homolog of BLM. 31 Murine models for BS. 34 Conclusions 36

Chapter Two. Rationale and Research Objectives 37

1 Chapter Three. Functional association of BLM with recombination-mediated telomere lengthening. I. Introduction 38 II. Materials and Methods Cell lines. 44 Cytogenetics. 44 Cell cycle synchronization, BrdU pulse-labeling, and flow cytometry. 45 Immunostaining. 45 Chromatin immunoprecipitations. 46 In vivo protein co-immunoprecipitations. 46 BLM knock-down. 47 Human telomere length analysis. 47 Yeast expression vectors. 48 Yeast strains and crosses. 48 Yeast telomere length analysis. 49 III. Results Analysis of TAs by FISH analysis is transformed BS fibroblasts. 49 Nuclear localization of endogenous BLM and TRF2. 50

BLM and TRF2 foci are enriched during G2/M and undergo DNA synthesis. 52 Association of BLM with telomeric DNA and TRF2 in vivo. 58

Effects of reducing BLM expression on telomere length in cells using ALT. 60 Rescue of type II telomere lengthening pathway by BLM in sgs1 est2 60 mutant S. cerevisiae.

IV. Discussion 64

Chapter Four. Association and regulation of the BLM helicase by TRF1 and TRF2 in vitro. I. Introduction 72 II. Materials and Methods Protein expression and purification. 74 Expression constructs, in vitro transcription and translation (IVTT), 75 and interaction site mapping by in vitro immunoprecipitation.

2 In vitro immunoprecipitation of full length proteins. 77 ELISA assay for detecting TRF1-BLM interaction. 78 DNA substrates. 79 DNase I footprinting. 80 Helicase assays. 80 Gel shift assays. 81 III. Results TRF1 and TRF2 interact with BLM in vitro. 81 BLM unwinds substrates that resemble native telomere conformations. 86 TRF1 and TRF2 oppositely regulate BLM activity on telomeric substrates. 87 TRF2 but not TRF1 affects BLM unwinding of a non-telomeric 3’-overhang. 92 Effects of TRF1 and TRF2 in combination on BLM helicase activity vary 96 with relative concentrations. IV. Discussion 96

Chapter Five. Thesis summary 101

Chapter Six. Bibliography 110

3 Abbreviations aa amino acid ALT alternative lengthening of telomeres APB ALT-associated PML body ATM ataxia telangiectasia mutated protein BIR break-induced replication BLM Bloom’s syndrome protein bp base pair BrdU bromodeoxyuridine BS Bloom’s syndrome BSA bovine serum albumin DSB double-strand break DTT dithiothreitol ECTR extrachromosomal telomeric repeats ELISA enzyme-linked immunosorbent assay FISH fluorescence in situ hybridization GPA glycophorin A HJ Holliday junction HR homologous recombination HPRT hypoxanthine phosphoribosyltransferase HRDC helicase and RNase D C-terminus HU hydroxyurea ICL interstrand cross-links IR ionizing radiation IVTT in vitro transcription and translation LOH loss of heterozygosity MLH1 mutL homologue 1 protein MMC mitomycin C MN micronuclei NBS1 Nijmegen Breakage Syndrome 1 protein

4 ND10 nuclear domain 10 NE nuclear extract NHEJ non-homologous end-joining NLS nuclear localization signal nt nucleotide PD population doubling PML promyelocytic leukemia protein PNB PML nuclear body QR quadriradial RTS Rothmund-Thomson syndrome RMN RAD50-MRE11-NBS1 protein complex RQC RecQ C-terminal domain SCE sister chromatid exchange SDSA synthesis-dependent strand annealing SGS1 slow growth suppressor 1 protein SSA single-strand annealing TA telomeric association TRF terminal repeat fragment TRF1 TTAGGG repeat factor 1 protein TRF2 TTAGGG repeat factor 2 protein UDS unscheduled DNA synthesis UV ultraviolet WRN Werner’s syndrome protein WS Werner’s syndrome

5 LIST OF TABLES Table 1 TAs in metaphase spreads prepared from BS and non-BS lymphoblasts. 14

LIST OF FIGURES

Figure 1 Sun sensitivity in BS male. 10 Figure 2 Cytogenetic abnormalities in metaphase spreads from BS cells. 13 Figure 3 BLM protein structure, protein partners and germline BLM mutations. 20 Figure 4 A role for BLM in repair of DSBs and stalled replication fork. 28 Figure 5 BIR-like mechanisms for ALT . 41 Figure 6 Analysis of telomere associations of homologous chromosome arms. 51 Figure 7 BLM co-localizes with TRF2 in cells using ALT. 53 Figure 8 BLM localizes to APBs. 54 Figure 9 Cell cycle regulated association of BLM with PML and TRF2. 56 Figure 10 BLM and TRF2 co-localize with foci of DNA synthesis during 57 late S/ G2/M. Figure 11 BLM co-immunoprecipitates with telomeric DNA and TRF2 from 59 cells using ALT. Figure 12 Knockdown of BLM and TRF analysis in Saos2 cells. 61 Figure 13 Structure of telomeres is S. cerevisiae type I and type II survivors. 63 Figure 14 BLM rescues the type II pathway in sgs1 est2 yeast mutants. 65 Figure 15 A model for the function of BLM in ALT. 71 Figure 16 Recombinant protein purity. 76 Figure 17 BLM directly interacts with TRF1 and TRF2. 84 Figure 18 Mapping of BLM domains that interact with TRF1 and TRF2. 85 Figure 19 BLM helicase unwinds telomere-like substrates. 88 Figure 20 BLM binding to the telomeric D-loop substrate demonstrated by 89 DNase I footprinting. Figure 21 Effect of TRF1 and TRF2 on BLM unwinding of telomere substrates. 90 Figure 22 TRF1 and TRF2 do not regulate UvrD helicase activity. 91 Figure 23 BLM unwinds the non-telomeric 3’-overhang substrate. 93

6 Figure 24 Gel shift analysis. 94 Figure 25 Effects of TRF1 and TRF2 on unwinding of non-telomeric 95 3’-overhang by BLM. Figure 26 Effect of combined TRF1 and TRF2 on BLM unwinding of telomeric 97 substrate. Figure 27 Model for the coordinated regulation of BLM by TRF1 and TRF2 109 at telomeres in cells using ALT.

7 CHAPTER ONE. Literature Review

Introduction

Bloom’s syndrome (BS) is a rare autosomal recessive disorder characterized at the cellular level by an extraordinary genomic instability that includes excessive homologous recombination, chromosome breakage and somatic hyper-mutation. The clinical manifestation of this genomic instability is an extraordinary predisposition to cancer of all tissues and cell types. The gene mutated in BS encodes the recQ-like BLM helicase. Since its discovery, studies have focused on how the BLM helicase maintains genomic stability. This review discusses the cellular and molecular aspects of BS, and the known and putative functions for the BLM helicase in DNA recombination, replication and repair.

Clinical features of Bloom syndrome

Initially classified as a sun-sensitive erythema associated with congenital dwarfism

(Bloom, 1954), BS is first and foremost a cancer predisposition syndrome. BS persons are susceptible to the development of all common cancers at a much higher frequency than the general population (German and Passarge, 1989). During the first decade of life there is a slightly higher occurrence of non-Hodgkin lymphoma, acute leukemia and rare tumors, although all tissues may be affected (reviewed in German, 1997). Persons surviving into adulthood display an increased frequency of all types of tumors; skin and colorectal carcinomas are the most common. The median age of cancer onset is 24.7 years, significantly earlier than the general population; almost a third of BS persons are

8 diagnosed with more than one primary cancer in a lifetime (German and Ellis, 2001).

Clinical data compiled in the Bloom’s Syndrome Registry from 168 cases show that ninety-percent of deaths result from cancer and associated complications. This underscores the importance of early diagnosis and close surveillance in the management of BS.

Diagnosis of BS is most often made on the basis of physical characteristics. Most prominent is the small size or proportional dwarfism that is apparent from birth (Bloom,

1954; Keller et al., 1999). As growth deficiency is associated with numerous disorders, BS may not be immediately suspected unless there is a family history or other indications that prompt further investigation (Vanderschueren-Lodeweyckx et al., 1984).

Dermatological changes, including sun-sensitive skin lesions of the face and exposed forearms, and irregular spotty skin pigmentation (distinct from the sun hypersensitivity), are common characteristics that distinguish BS from other diseases involving growth deficiency (German and Passarge, 1989). An observant physician may also recognize a characteristic facial appearance, which has been described as “keel- shaped” (German, 1995) (Figure 1). Excessive diarrhea, vomiting and lack of interest in feeding have been described in some BS infants; “wasting” as a result of malnutrition may occur from birth to 8 years of age (German, 1995; Keller et al., 1999).

Immunodeficiency, first manifested in chronic middle ear infections and respiratory infections, is another clinical characteristic of most BS persons (Mori et al., 1990; German and Ellis, 2001). In fact, chronic pulmonary disease is the second most common cause of

9

Figure 1. Male BS person presenting with sun-sensitive erythema. Also, note unusual facial shape, described as “keel-shaped.” Photo courtesy of James L. German, III.

10 mortality in BS (German and Ellis, 2001). T-cell function appears to be normal; however, a deficiency in secreted IgM levels versus non-secreted IgM levels has been observed in some affecteds, suggesting a defect in the maturation of B-cells from IgM-surface- bearing to IgM-secreting cells (Hutteroth et al., 1975; Kondo et al., 1992; Schoen and

Shearn, 1967; Weemaes et al., 1991). Reduced levels of mu-S mRNA may account for the reduced secretion of IgM (Kondo et al., 1992); however, the underlying cause of reduced expression is unknown.

Most if not all males with BS are sterile. Male hypogonadism from early development results in reduced spermatogenesis or failure to produce any sperm, termed azoospermia (Kauli et al., 1977). Several BS females followed by the Registry have given birth to normal children, although the onset of menopause at an early age

(often by the mid- to late twenties) dramatically reduces fertility of BS females

(Chisholm et al., 2001; German, 1995). Finally, an increased incidence of diabetes is associated with BS (Kondo et al., 1991; Mori et al., 1990). Currently 12% of persons in the BS registry have been diagnosed with non-insulin-dependent diabetes (German,

1995; German and Ellis, 2001).

Bloom’s syndrome is associated with a mutator phenotype.

Our first understanding of the molecular basis of BS came from early cytogenetic analysis, which revealed excessive chromosomal instability and hyper-mutability in cells from persons with BS (Figure 2A-C). Chromosome breaks, gaps and structural rearrangements occur at a frequency many times higher in cells from BS persons than in

11 those from normal persons (German, 1993; Kusunoki et al., 1994; Tachibana et al.,

1996)(Figure 2A). This chromosomal instability is also suggested by a high frequency of micronuclei, small “nuclei” carrying DNA in the cytoplasm that result from lagging or broken chromosomes. (German and Crippa, 1966; Honma et al., 2002). The most notable characteristic of BS cells is excessive homologous intra- and inter-chromosomal recombination. Exchanges between sister chromatids of the same chromosome, commonly called sister chromatid exchanges (SCEs), occur at a 5- to 10-fold excess in BS cells as compared to normal cells (Figure 2B) (Chaganti et al., 1974). Inter-chromosomal exchanges between homologous chromosomes are observable in metaphase preparations as symmetrical four-armed conformations called quadriradials (QRs). QRs are found in 1% of BS cells, but are exceedingly rare in normal cells (Figure 2C)

(German, 1995). Additionally, BS cells have a high frequency of telomeric associations

(TAs) (Figure 2D-F). TAs are present in 0.57% of metaphases from BS blood lymphocytes in short-term culture as compared to 0.1% of non-BS controls, a statistically significant difference by chi-square analysis (p=0.05) (Table 1). Frequently, a non- stained area can be visualized between chromosome ends along with a “string” of stained material (Figure 2E-F). A subset of TAs appear as end-to-end fusions (Figure

2D). Anaphase and telophase bridges occur at a high frequency and may be coincidental to TAs (German and Crippa, 1966).

The genomic instability of BS is associated with a high mutation frequency at specific loci (Groden and German, 1992) and by frequent loss of heterozygosity (LOH)

12

Figure 2. Cytogenetic abnormalities in metaphase spreads from BS cells. A) Chromosome breakage near centromere. B) Characteristic “harlequin” chromosomes seen in metaphases from BS cells. SCEs are visualized by labeling cells for two cell cycles with bromodeoxyuridine (BrdU), resulting in differentially labeled sister chromatids. Each alternating region of light and dark represents a sister chromatid exchange (marked by arrows on one chromosome). C) Quadriradial (QR) between homologous chromosomes. D) TA comprised of chromosome nos. 11. E) TA comprised of chromosome nos. 12. F) TA comprised of chromosome nos. 14. Photos courtesy of James L. German, III and Steven Schonberg.

13 Table 1. TAs in metaphase spreads prepared from BS and non-BS lymphoblasts

Source of metaphases Ave Ave BSL-1a BSL-3 BSL-4 BSL-5 BSL-6 BSL-8 BSL-9 CL-1b CL-2 BSLs CLs # TAs 3201022275 501921 # metaphases 103 1031 2941 6667 3462 1471 500 500 500 16174 1000 c Frequency 2.91 1.94 0.34 0.33 0.78 0.34 1.00 0.00 0.20 0.57 0.10 a metphases prepared from lymphocytes from 7 BS donors (BSL) b metaphases prepared from lymphocytes from 2 non-BS donors (CL) c frequenceis calculated as configurations per 100 metphases scored (Lillard-Wetherell et al., 2004). due to increased somatic recombination (Groden, et al., 1990). Rates of spontaneous mutation at the hypoxanthine phorphoribosyltransferase (HPRT) locus were examined in BS and non-BS lymphoblastoid cells (Tachibana et al., 1996). Mutations occurred at a frequency of 1.39x10-6 in BS cells versus 1.75x10-8 in non-BS cells. Molecular analysis of mutant clones revealed a majority of clones containing large deletions within the HPRT locus and flanking sequences, as well as a minority of cells with point mutations and small deletions. Two studies used erythrocytes from BS persons with MN blood type, determined by heterozygosity at the glycophorin A (GPA) gene locus, to evaluate mutational frequency and type (Kyoizumi et al., 1989; Langlois et al., 1989). Variant erythrocytes, those that expressed only M or N type GPA, occurred between 50- and

100- fold more frequently in BS vs. unrelated non-BS donors or unaffected parental donors (Kyoizumi et al., 1989; Langlois et al., 1989). Homozygous variants, those expressing one allele at twice the expected suggested that this LOH was occurring by somatic crossing-over in BS cells.

14 Sensitivity of BS cells to DNA damaging agents.

The genomic instability of BS cells first suggested a susceptibility to DNA damage or inability to repair it. Sensitivity to DNA damaging agents was measured in

BS cells by examining viability, SCE induction and gross chromosomal abnormalities following exposure. Sensitivity of BS cells to mitomycin C (MMC), a bifunctional alkylating agent that induces bulky adducts and interstrand cross-links (ICLs), was demonstrated. ICLs tether complementary strands of DNA resulting in an inability to unwind DNA. Recent data indicates that ICLs may be converted to double-strand breaks (DSBs) and then repaired by homologous recombination (McKenna et al., 2003;

Niedernhofer et al., 2004). BS cells are also more sensitive than control cells to the effects of MMC on cell viability. (Hook et al., 1984; Shiraishi and Sandberg, 1978) The effect of

MMC on the rate of SCE is less certain. In one study, a 2-fold increase in SCE was reported for BS cells in response to MMC treatment (Shiraishi and Sandberg, 1978).

However, Hook et al. reported normal sensitivity of BS cells to MMC with respect to SCE induction (Hook et al., 1984). Cytogenetic studies by the German laboratory demonstrate an increase in the formation of QRs in BS lymphoblasts following MMC treatment (Schonberg and German, unpublished data); a recent report demonstrates an increase in micronuclei formation (Honma et al., 2002).

Individuals with BS exhibit sun-sensitive lesions on the face and exposed forearms indicating that BS cells may be sensitive to UV irradiation. The earliest study to address UV sensitivity in BS evaluated unscheduled DNA synthesis (UDS) in BS and control cells (Giannelli et al., 1981). While basal rates of UDS were within normal range,

15 the rate of UDS in BS cells was 19-29% higher than normal in cells treated with UV.

Giannelli et al. proposed that the increase in UDS was due to an increase in the number of damaged sites repaired by a DNA synthesis-dependent pathways that repair double- strand breaks. Two additional studies examined the effect of UV on SCE induction in BS cells. In both studies, the rate of SCE (over basal levels) increased substantially more in

BS cells than in control cells following UV treatment (Kurihara et al., 1987; Mamada et al., 1989).

Sensitivity of BS cells to gamma irradiation is cell cycle stage-dependent. BS cells are more sensitive to gamma irradiation during S- and G2-phases of the cell cycle; this is marked by an increase in the frequency of chromosome breaks (Aurias et al.,

1985). In two similar studies, radiosensitivity and/or induction of chromatid breaks and gaps after gamma irradiation were elevated during G2-phase for BS cells, as well as cells derived from individuals with three other familial cancer syndromes, ataxia- telangiectasia, Fanconi anemia, and xeroderma pigmentosum (Parshad et al., 1983;

Parshad et al., 1985).

BLM encodes a RecQ-like helicase.

The elevated somatic recombination of BS cells facilitated the identification of

BLM, the gene mutated in BS. Analysis of the SCE levels in lymphocytes of selected BS persons revealed the existence of a some cells without the high SCE rates characteristic of BS (Ellis et al., 1995). These “corrected” cells were hypothesized to arise from intragenic recombination within BLM, the gene mutated in BS, resulting in phenotypically wild-type cells with normal SCE. This type of recombination would

16 only be observed in affected persons who carried two unique mutated BLM alleles, and would not be found in those carrying mutant alleles that were identical by descent.

Using cell pairs with high and low SCE, a candidate for the BLM gene was identified by somatic crossover point mapping and was confirmed by the identification of mutations in the BLM gene in other BS persons (Ellis et al., 1995). As the genomic instability of BS indicated a defect relating to some aspect of DNA metabolism, it was not unexpected to find that BLM putatively encoded a 1417-amino acid protein with homology to the RecQ family of helicases (Ellis et al., 1995).

RecQ-like helicases contain seven conserved helicase motifs placing them in a large family of proteins that belong to helicase superfamily II (SF2). RecQ helicase motifs I and II contain “Walker box” sequences characteristic of ATPases and a DEXH sequence (Nakayama et al., 1984). In BLM, the helicase domain is centralized in the protein (amino acids 649-1006) and a conserved RecQ-C-terminal domain (RQC) is found immediately adjacent to the core helicase domain (amino acids 1006-1077). The

N-terminus (amino acids 1-431) facilitates oligomerization of BLM into a putative hexameric ring structure, the active form of the helicase (Karow et al., 1999). The C- terminus contains the nuclear localization signal (NLS) (amino acids 1334-1339)

(Hayakawa et al., 2000) and a putative nucleic acid-binding domain commonly found in

RecQ helicases, called the HRDC domain (helicase and RNase D C-terminal domain)

(amino acids 1212-1292) (Liu et al., 1999).

The majority of BLM mutations identified in persons with BS disrupt the helicase domain indicating that helicase function is essential for the role of BLM in maintaining

17 genomic stability (Ellis et al., 1995) (Figure 3A). Nonsense, frameshift, exon-skipping and exon-deletion mutations encode a truncated BLM protein missing one or more helicase motifs. The frameshift mutation occurring at nucleotide 2281 of the BLM cDNA, commonly referred to as BLMAsh, is carried by at least 1% of Ashkenazi Jews and demonstrates a founder effect in this population. Other nonsense and frameshift mutations may not directly disrupt helicase function, but result in aberrant localization of BLM due to exclusion of the NLS. Missense mutations within the helicase domain may affect ATP-binding and/or hydrolysis, thus disrupting helicase activity (German and Ellis, 2001; Rong et al., 2000). Other point mutations in the C-terminus of BLM are, as yet, poorly characterized, but presumably disrupt sub-cellular localization, DNA- binding or protein interactions.

Protein partners implicate BLM in DNA repair.

BLM is part of two protein super-complexes proposed to function in DNA repair, called “BASC” and “BRAFT.” In addition to BLM and BRCA1, BASC (BRCA1- associated genome surveillance complex) includes the DNA mismatch repair proteins

MSH2, MSH6 and MLH1, the DSB repair proteins ATM, RAD50, MRE11 and NBS1, and a single replication protein RFC (Wang et al., 2000). As the name implies, BASC is thought of as a dynamic “genome-surveillance” complex that senses DNA damage and deploys appropriate proteins to mediate DNA repair.

The BRAFT protein complex was recently identified by in vivo immunoprecipitation of BLM (Meetei et al., 2003). BRAFT is named for its most

18 abundant protein components, including BLM, RPA, Fanconi anemia complementation

(FANC) group proteins, and topoisomerase IIIα (TopoIIIα). Like BASC, BRAFT most likely serves as scaffolding complex for DNA repair proteins. BLM and MLH1 are the sole components common to both BASC and BRAFT; therefore, each complex is likely to respond to unique types of DNA damage or cellular stress cues.

Separate studies have demonstrated that BLM interacts directly with BASC and

BRAFT proteins, including MLH1, MSH6, ATM, RPA, FANCD2, and TopoIIIα (Wu et al., 2000; Pedrazzi et al. 2001; Pedrazzi et al., 2003; Wu et al., 2001; Langland et al., 2001;

Pichierri et al., 2004; Brosh et al., 2000). BLM also interacts directly with p53, RAD51,

RAD51L, FEN1 and WRN (Braybrooke et al., 2003; Sharma et al., 2004; von Kobbe et al.,

2002; Wang et al., 2001; Wu et al., 2001). Some of these interactions have been mapped to specific BLM domains (Figure 3A); more information regarding these interactions with respect to regulation of BLM activity and putative cellular functions will be discussed below.

BLM is a structure-specific helicase.

Similar to other characterized recQ-like helicases, purified recombinant BLM is a

DNA-dependent ATPase and ATP- and Mg2+-dependent helicase that unwinds DNA substrates in a 3’-5’ direction (Karow et al., 1997). Each RecQ family helicase, however, has unique substrate preferences. In vitro analyses of the DNA binding and strand- displacement activities of BLM demonstrate that it acts upon a wide spectrum of substrates (Figure 3B). BLM unwinds partially duplexed substrates with 3’ single-strand tails efficiently, but cannot efficiently unwind duplexes with 5’ overhangs or blunt-

19

Figure 3. BLM protein structure, protein partners, germline BLM mutations and graphical representations of key BLM biochemical substrates. A.) Graphical illustration represents the major features present in the BLM protein. Putative functions, protein interactions and other features are noted (top of figure) for the three major regions of BLM, the N-terminus (aa 1-648), helicase domain (aa 649-1005) and C-terminus (aa 1006- 1417). The green box highlights the helicase domain with 7 conserved helicase motifs from E. coli RecQ; the yellow box highlights the RecQ C-terminal domain (RQC); the blue box highlights the helicase and RNase D C-terminal domain (HRDC); NLS represented the functional nuclear localization signal within the carboxyl-terminal segment. Representative mutations found in these domains are listed (bottom of figure) according to mutation type: frameshift, nonsense or missense. Letters and numbers represent amino acids in the BLM protein, where X denotes a stop codon. Mutational data summarized from information available online from BLMbase, the mutation registry for Bloom syndrome (http://www.uta.fi/imt/bioinfo/BLMbase/) (Rong, et al., 2000). Protein features are summarized from previous reviews (German and Ellis, 2001; Bachrati and Hickson, 2003). B.) Graphical representation of three substrates readily unwound by BLM: the D-loop, X-junction and G-quadruplex (the latter depiction is adapted from Kondo et al, 2004)

20 ended substrates (Karow et al., 1997). BLM cannot unwind long DNA duplexes (>/= 259 bp) indicating low processivity (Brosh et al., 2000). The single-stranded DNA binding protein, replication protein A (RPA), interacts with BLM and increases BLM processivity presumably by stabilizing single-stranded DNA intermediates (Brosh et al., 2000).

BLM can bind and unwind DNA “bubbles,” short regions of unpaired DNA flanked by fully duplexed DNA (Mohaghegh et al., 2001; van Brabant et al., 2000). BLM has a much stronger affinity for binding and melting D-loops, structures that result from pairing of an invading DNA strand to a complementary strand within a bubble structure

(van Brabant et al., 2000). This activity is important for considering BLM function, as there is similarity between this D-loop and the first heteroduplex intermediate formed by RAD51 during recombination. Consistent with a function of BLM as an “anti- recombinase”, BLM can unwind a synthetic X-structure, a four-armed blunt-ended duplex substrate that resembles a Holliday junction (HJ) (Karow et al., 2000; Mohaghegh et al., 2001). This activity is an exception to the binding preference of BLM for substrates containing single-stranded regions. In fact, the X-structure contains no unpaired regions,

BLM binding may rely upon recognition of the crossover portion of the X-structure, since it cannot bind a linear duplexed DNA with the same nucleic acid composition.

BLM partners p53 and the MSH2/MSH6 heterodimer regulate BLM activity on synthetic

HJs (Yang et al., 2004; Yang et al., 2002). While MSH2/MSH6 stimulates HJ resolution, p53 attenuates unwinding by BLM. BLM can also catalyze reverse branch migration of a synthetic HJ formed in a RecA-mediated strand exchange reaction (Karow et al., 2000).

21 BLM disrupts quadruplex or “G4” DNA, an unusually stable secondary structure that forms from non-canonical base pairing between guanine residues, either inter- or intra-strand (Sun et al., 1998). Furthermore, quadruplex-interacting ligands specifically inhibit unwinding of these substrates by BLM (Huber et al., 2002; Li et al., 2001). While quadruplexes form readily in vitro, recent data indicates the existence of these structures in vivo, enriched at telomeres (Chang et al., 2004). This activity, then, has unique implications for BLM function in maintaining the integrity of repetitive, G-rich regions of the genome such as telomeres and ribosomal repeats. Telomere length hyper- variability and reduced rDNA has been described in BS cells (Schawalder et al., 2003), supporting the hypothesis that BLM function to maintain these regions.

BLM is regulated by the cell cycle.

BLM is marker of proliferation. Its expression is highest in proliferating cells, including tumor cells and lymphoid tissues, and is almost undetectable in quiescent tissues. BLM protein levels are regulated during the cell cycle, increasing steadily during S-phase, reaching peak levels that persist during G2/M and declining again as cells enter G1 (Dutertre et al., 2000). Hyper-phosphorylation of BLM occurs during S- phase, suggesting that post-translational modification may regulate BLM stability or activity (Beamish et al., 2002). BLM is predominantly localized to the nucleolus during

S-phase and to subnuclear structures called nuclear domain 10 (ND10) or promyelocytic leukemia (PML) nuclear bodies (PNBs) (so named due to the presence of the PML tumor suppressor) during G2/M (Yankiwski et al., 2000; Yankiwski et al., 2001; Zhong et al.,

22 1999). The C-terminal domain of BLM mediates nucleolar accumulation, while the N- terminus mediates PNB localization. Localization to nucleoli, but not PNBs, is essential for the role of BLM in maintaining genomic stability (Yankiwski et al., 2001). Inducible expression of BLM C-terminal mutants do not localize to nucleoli and fail to correct SCE levels in BS cells; several of these mutants exacerbate SCE levels. Inducible expression of

BLM N-terminal mutants do not localize to PNBs but correct SCE and genomic instability in BS cells (Yankiwski et al., 2001). The localization of BLM to these subnuclear domains may represent differential binding to GC-rich DNA repeat regions that localize there. The N-teminal domain required for PNB localization also facilitates binding to telomere repeats; the C-terminal domain mediates nucleolar localization and binding to ribosomal DNA (Schawalder et al., 2003).

BLM responds to DNA damage.

BLM expression levels and localization have been evaluated following treatment with certain DNA damaging agents, particularly those that induce DSBs. BLM protein accumulates in response to ionizing radiation (IR) and treatment with the DSB-inducing chemicals etoposide and bleomycin (Ababou et al., 2000; Bischof et al., 2001). BLM moves into IR-induced foci with protein partners RAD51, RPA and PML at sites of singled-stranded DNA that are presumed to represent sites of homologous recombination (HR) repair (Bischof et al., 2001). Phosphorylation of BLM occurs simultaneously with the accumulation and translocation after IR treatment. BLM interacts directly with ATM, a PI-3 kinase that is activated in response to DSBs to

23 phosphorylate key proteins involved in DNA damage repair (Beamish et al., 2002;

Shiloh and Kastan, 2001). ATM phosphorylates BLM during mitosis and in response to

IR (Ababou et al., 2000; Beamish et al., 2002). Importantly, expression of BLM protein with mutated ATM phosphorylation sites fails to correct radiation-induced damage in

BS cells indicating this phosphorylation is essential for the function of BLM in repair of

IR-induced DSBs (Beamish et al., 2002).

In response to the treatment of cells with agents that create DSBs during replication, including hydroxyurea (HU) which stalls replication forks and UV-C which creates nicks that are converted to DSBs during replication, BLM protein levels are reduced. Additionally, BLM is phosphorylated in an ATM-independent manner. HU- induced phosphorylation is also dependent on another DNA damage-responsive kinase, the ATM and RAD3-related protein (ATR) (Franchitto and Pichierri, 2002). Similar to

IR-induced foci, BLM forms foci with RAD51, as well as p53 and the DSB proteins,

RAD50-MRE11-NBS1 (RMN complex) (Franchitto and Pichierri, 2002; Sengupta et al.,

2003). Intriguingly, BLM is essential for translocation of p53 and the RMN complex to

HU-induced foci, suggesting that BLM is a “sensor” for DSBs occuring during replication (Franchitto and Pichierri, 2002; Sengupta et al., 2003).

BLM in DSB repair.

Replication-coupled HR may facilitate repair of DSBs wihout necessitating strand exchange (crossing over). The increase in crossovers found in BS cells may indicate a role for BLM in promoting these HR repair pathways. Three models are presented in

24 Figure 4. In all models, BLM functions to 1) promote the formation of a D-loop that is productive for replication and 2) facilitate resolution of HJs by branch migration, thus suppressing crossover resolution. When a double HJ is created, BLM-mediated resolution may require TopoIIIα to decatenate strands (Figure 4A). In support of this model, recent in vitro evidence demonstrates that BLM and TopoIIIα can resolve a catenated double-HJ substrate (Wu and Hickson, 2003). Importantly, interaction between BLM and TopoIIIα is essential for full correction of SCE in BS cells; fewer than normal TopoIIIα foci are found in cells without BLM than in normal cells, suggesting that BLM recruits TopoIIIα to sites of DSBs (Hu et al., 2001). BLM is capable of recruiting TopoIIIα to its DNA substrate in vitro resulting in increased TopoIIIα activity

(Wu and Hickson, 2002).

Richardson and Jasin (2000) have reported that replication-coupled HR repair is commonly initiated from only one end of a DSB, a process termed break-induced replication (BIR). This type of repair would result in the formation of only one HJ, essentially a “migrating D-loop”, (Figure 4B-C). When one side of the break is not available for re-joining, the missing chromosome arm may be completely replaced by replication (Figure 4B). BIR is thought to facilitate recombinational pathways for telomere lengthening in yeast and some human cells lacking telomerase (Kraus et al.,

2001; Lundblad, 2002; Reddel et al., 2001). When both sides of the break are present, one end may be extended by replication and then re-joined to the recipient chromosome arm directly by non-homologous end-joining (NHEJ) or by single-strand annealing (SSA) of homologous sequences followed by end-joining (Figure 4C). The latter process, termed

25 synthesis-dependent strand annealing (SDSA), is the primary mechanism by which

DSBs are repaired in Drosophila melanogaster. Importantly, mutation of the BLM homolog in Drosophila, DmBLM, results in a failure to synthesize new DNA during

SDSA repair of P-element induced DSBs, such that SDSA is aborted and DSBs are repaired by error-prone NHEJ (Adams et al., 2003).

SSA may also be used without HR to facilitate end-joining. In this case, ends are processed to expose complementary single-strand regions of direct repeats or microhomologies that anneal and re-ligate, deleting the DNA between the repeats.

Using linearized plasmids to simulate a DSB, Langland et al. found that BS cell extracts did not effectively utilize sites of microhomology for end-joining resulting in small deletions (Langland et al., 2002). Using a similar assay, Gaymes et al. reported preferential utilization of distant sites of microhomology by BS cell extracts resulting in large deletions (Gaymes et al., 2002). This error-prone end-joining in BS cells was dependent on the activity of the Ku70/86 heterodimer, key proteins in NHEJ. Recent data also demonstrated an association of BLM with Ku70/80 and the associated kinase,

DNA-PK in vitro (Onclercq-Delic et al., 2003). In the absence of ATP, BLM associated with DNA-PK and Ku70/80; however, addition of ATP resulted in phosphorylation of

BLM by DNA-PK and release of BLM from DNA. The authors suggest that BLM stays at the site of a DSB while it is being repaired by NHEJ, and is removed after successful repair and DNA-PK phosphorylation. If NHEJ is unsuccessful, BLM would remain at the DSB and facilitate BIR repair.

26 Repair of stalled replication forks.

Replication forks stall when they encounter DNA lesions. An inability to bypass these lesions can collapse the fork into a DSB. Collapsed forks may then be repaired by

HR mechanisms and deleterious sister chromatid exchange (for a review, see Helleday,

2003). Thus, BLM may facilitate bypass mechanisms for the repair of stalled forks that does not involve HR. Of particular interest is the ‘chickenfoot model’ for replication fork repair (Figure 4D). In this model, a replication fork encounters damage and arrests; the leading and lagging strands regress to form an HJ-like structure that resembles a chickenfoot (reviewed in Cox, 2001). This daughter-daughter pairing allows the daughter strand that encountered the damage to synthesize past the lesion using the other daughter as a template. BLM could then reverse migrate the chickenfoot such that the replication fork is re-established without recombination. BS cells display a retarded rate of nascent DNA chain elongation and accumulate abnormal daughter-daughter replication intermediates (Hand and German, 1975; Waters et al., 1978). The latter may represent ‘chickenfeet’ that are unresolved in the absence of BLM and that must be resolved subsequently by HR mechanisms. Such functions for BLM could account for the excessive SCE in cells lacking BLM.

BLM is a target and effector of apoptosis.

BLM is cleaved prior to apoptosis induced by HU, UV-C, anti-Fas and, other apoptosis-inducing agents (Ababou et al., 2002; Bischof et al., 2001; Freire et al., 2001).

This cleavage is mediated by caspase-3, creating a BLM fragment lacking the first 436

27 D

A

B C

BLM

or BIR-SSA/NHEJ BIR

crossover no crossover chickenfoot model

Figure 4. A role for BLM in repair of DSBs and stalled replication fork. (A) BLM may function with TopoIIIa to resolve double HJ substrates (right) to prevent crossover resolution (left) (Depiction adapted from Ira et al., 2003) (B) BLM may promote branch migration of the single HJ formed by BIR in which replication proceeds to the end of the chromosome. (C) Alternatively, BIR may re-replicate only a portion of the homologous chromosome which is then re-joined with its recipient chromosome end by SSA or NHEJ. (D) BLM may function in a bypass mechanism to repair stalled replication forks. In this case the stalled fork regresses such that the daughter strands are paired to form a ‘chickenfoot’ The daughter that encountered the damage can use its sister as a template for replication past the lesion. After replication occurs, BLM may promote reverse migration of the fork such that the fork is re-established and the damage is bypassed. In all diagrams, dashed lines represent newly replicated DNA.

28 amino acids but with an intact helicase domain. Cleaved BLM is helicase-proficient, but cannot bind to TopoIIIα (Freire et al., 2001), an interaction essential for restoring SCE to normal levels in BS cells. Freire et al. (2001) proposed that disruption of the TopoIIIα-

BLM interaction by caspase-3 cleavage prevents activation of BLM-mediated repair in cells committed to apoptosis.

BLM may also participate in apoptosis. It interacts directly with p53, a key tumor suppressor that controls the cellular switch that either arrests cells to allow for repair of damaged DNA or directs apoptotic processes. While BS cell lines have normal

Fas-induced apoptosis, p53-depedent apoptosis in response to genotoxic agents is attenuated (Wang et al., 2001). Furthermore, BLM over-expression exerts an anti- proliferative effect on p53-positive, but not p53-negative cells (Garkavtsev et al., 2001).

These data suggest that BLM cooperates in p53-mediated growth arrest and apoptosis, and disruption of these signals in BS cells may contribute to tumorigenesis.

Other human disorders associated with RecQ deficiency.

Higher eukaryotes have multiple RecQ-like helicases. Five RecQ-like helicases are currently known in humans and include RecQ1, BLM, WRN, RECQ4 and RECQ5. In addition to BS, defects in two other RecQ-like helicases are associated with inherited disorders, including Rothmund-Thomson syndrome (RTS) and Werner’s syndrome

(WS). Although clinically distinct, the chromosomal instability associated with deficiency of RecQ-like helicases indicates a conserved function for these helicases in maintaining genomic stability.

29 RECQ4 (aka RECQL4 or RTS) encodes a protein of 1208-amino acids that is mutated in Rothmund-Thomson syndrome (RTS) (Kitao et al., 1999; Ohhata et al., 2000).

RTS is characterized by skeletal abnormalities, dwarfism (postnatal), and skin abnormalities (Baro et al., 1989; Shuttleworth and Marks, 1987). Individuals with RTS have an unusually high predisposition to osteosarcoma and skin malignancies (Baro et al., 1989; Cumin et al., 1996; el-Khoury et al., 1997; Nishijo et al., 2004; Piquero-Casals et al., 2002; Varughese et al., 1992; Wang et al., 2003; Wang et al., 2001). As with BS, cytogenetic analysis revealed chromosome instability, with a high frequency of translocations, deletions and rearrangements (Der Kaloustian et al., 1990; Miozzo et al.,

1998; Ying et al., 1990). RTS cells are sensitive to UV and gamma irradiation; however unlike BS cells, RTS deficiency does not result in MMC sensitivity (Lindor et al., 2000;

Prache-de-Carrere et al., 1996; Smith and Paterson, 1982). RecQ4 has yet to be characterized biochemically and its cellular functions remain largely unknown.

WRN encodes a protein of 1432-amino acid that is mutated in Werner’s syndrome (WS) (Yu et al., 1996). Werner’s syndrome (WS) is characterized as a premature aging disorder; those diagnosed with WS display many of the same phenotypes associated with normal aging, such as cataracts, osteoporosis and atherosclerosis (reviewed in Goto, 2000). Affected individuals are also susceptible to cancer, particularly soft-tissue sarcomas (Goto et al., 1996). Cells from WS individuals undergo premature senescence, are prone to chromosomal translocations and are hyper- mutable (Beadle et al., 1978; Hoehn et al., 1975; Kyoizumi et al., 1998). The WRN

30 helicase is unique from other RecQ-like helicases in that it also functions as an exonuclease (Huang et al., 1998). BLM and WRN have similar helicase substrate preferences, including HJs, D-loops, quadruplex DNA and bubbles (Huang et al., 1998;

Mohaghegh et al., 2001). Like BLM, WRN has low processivity and is stimulated by

RPA (Shen et al., 1998). Furthermore, BLM and WRN share many of the same protein partners, and may interact directly with one another (von Kobbe et al., 2002).

Association of BLM with WRN inhibits WRN exonuclease activity (von Kobbe et al.,

2002). These data have led many to speculate that BLM and WRN are involved in overlapping DNA repair pathways.

SGS1 is a functional homolog of BLM.

While higher eukaryotes encode multiple RecQ-like helicases, the genomes of prokaryotes and simple eukaryotes generally have only one RecQ-like encoding gene.

Examples include RecQ in E. coli, the founding member of the RecQ family, and the budding and fission yeast, SGS1 in S. cerevisiae and Rqh1 from S. pombe. Much of what is known for mammalian RecQ helicases was initially studied by genetic and biochemical analysis of RecQ helicases in lower organisms. Sgs1 is of particular interest as it may be considered a functional homolog of BLM and WRN, although it lacks exonuclease function.

The S. cerevisiae Sgs1 (slow growth suppressor 1) protein is closely aligned with

BLM in terms of size, sequence and mutant phenotype. SGS1 mutation results in premature senescence, slow growth, chromosome mis-segregation, increased

31 chromosomal rearrangement, and elevated homologous and illegitimate recombination

(Myung et al., 2001; Sinclair and Guarente, 1997; Sinclair et al., 1997; Watt et al., 1996). sgs1 mutants have a high frequency of SCE, mimicking the BS phenotype, (Onoda et al.,

2000).

Premature senescence of sgs1 mutants may be associated with increased recombination and instability within the rDNA repeat regions (Sinclair and Guarente,

1997). Premature senescence is also associated with deficiency of WRN, it was surprising to find that BLM, not WRN, suppressed premature senescence, rDNA instability and accumulation of extrachromosomal rDNA circles in sgs1 mutants (Heo et al., 1999). Versini et al. (2003) reported that sgs1 mutants replicate DNA more quickly, but replication forks are slowed significantly through the rDNA arrays. The authors hypothesize that frequent stalling of replication forks leads to hyper-recombination within the rDNA repeats of sgs1 mutants. In support of the latter, Sgs1 localizes frequently to replication foci and is involved in an intra-S phase DNA damage checkpoint (Frei and Gasser, 2000; Myung and Kolodner, 2002). Intriguingly, Sgs1 is also essential for RNA polymerase I-transcription. In the absence of both Sgs1 and another DNA helicase, Srs2, ribosomal RNA synthesis decreases by 10-fold. These data suggest dual functions for Sgs1 in facilitating rDNA replication and transcription (Lee et al., 1999).

Sgs1 associates with Rad51 and topoisomerase III (Top3). Disruption of RAD51 restores growth and homologous recombination to normal levels in sgs1 mutants indicating epistasis of SGS1 and RAD51 in recombination. Mutation of sgs1 suppresses

32 the slow growth phenotype of top3 suggesting that Sgs1 generates DNA structures, such as recombination intermediates, that must be acted upon by Top3 (Gangloff et al., 1994).

Expression of BLM in double sgs1top3 mutants restores the slow growth phenotype suggesting overlapping function of BLM and Sgs1 in conjunction with Top3 (Heo et al.,

1999). Additionally, Sgs1 interacts directly with Top3 and expression of Sgs1 lacking the

Top3 interaction domain in sgs1 null strains fails to suppress the hyper-recombination defect in these cells (Bennett et al., 2000; Bennett and Wang, 2001; Ui et al., 2001). Based on these data, Ira et al. (2003) hypothesized that Sgs1-Top3 promote non-cross-over resolution of HJs during Rad51-dependent HR repair. Using an inducible HO endonuclease system to create a site specific DSB, Ira et al. recently demonstrated an increase in cross-overs during Rad51-dependent HR repair in sgs1 and top3 mutants.

Rescue of this phenotype in sgs1 mutants by Sgs1 expression required an intact Top3 interaction domain. These data support a conserved function of BLM-TOPIIIα and Sgs1-

Top3 in promoting resolution of HJs during HR repair.

In the absence of telomerase, yeast can utilize two recombination-mediated pathways to maintain telomeres requiring either Rad50 (type II) or Rad51 (type I). sgs1 mutants fail to give rise to survivors utilizing the Rad50-dependent pathway indicating a requirement for Sgs1 in type II recombination (Cohen and Sinclair, 2001; Huang et al.,

2001; Johnson et al., 2001). Both type I and type II pathways may utilize a BIR-type mechanism, but differentially amplify sub-telomeric and telomeric repeat regions (Ira and Haber, 2002). These pathways will be discussed in detail in Chapter Three.

33 Murine models for BS.

Three murine models for BS have been generated in separate laboratories. Each model was created using a distinct gene targeting strategy; each model has a distinct phenotype. The first targeting strategy deleted a 180-bp segment of the Blm gene upstream of the helicase domain which was replaced by the PGK-NEO cassette (Chester et al., 1998). The homozygous knockout is lethal by embryonic day 13.5 and is severely anemic compared to littermates at the time of death. As might be expected for a BS murine model, knockout embryos display severe growth retardation. Significantly, these mice also demonstrate the hallmark genomic instability of BS, including increased

SCE and micronuclei formation.

The second targeting strategy created two alleles, one containing 4 copies of exon

3 and the PGK-NEO cassette (Blmm2) and the other, created by Cre-mediated deletion of

PGK-NEO, retains two copies of exon 3 (Blmm3). Both alleles are null and/or produce truncated protein products (Luo et al., 2000). Analysis of Blmm3/m3 mice revealed normal meiotic recombination, but increased somatic recombination. When crossed to a mouse model of intestinal cancer, the Blmm2/m3 mice displayed LOH and increased tumor susceptibility as a result of increased somatic recombination. While the apparent genomic instability is consistent with BS, this mouse does not display the smaller size or other phenotypes found in the human disorder or the other knockouts. It has been demonstrated subsequently that the Blmm3 allele is hypomorphic and produces BLM

34 protein at low levels, and truncated, amino-terminal containing protein (McDaniel et al.,

2003).

Another murine model created by Goss et al. created an allele (BlmCin) to simulate the human Ashkenazi allele, BLMAsh (Goss et al., 2002). In contrast to the two mouse models of BS previously reported (Chester et al., 1998; Luo et al., 2000), gene-targeting replaced exons 10, 11 and 12 of Blm with an Hprt cassette to create the mutant allele.

Although BlmCin/Cin mice do not survive embryogenesis (J. Groden, unpublished data), tissues from BlmCin/+ had approximately a 50% reduction in BLM protein in comparison with Blm+/+ tissues, allowing examination of Blm haploinsufficiency in vivo, ie. reduction in wild-type Blm gene dosage and its gene product. Mouse cells heterozygous for this allele have a subtle increase in genomic instability, as indicated by an increase in micronuclei. Heterozygous mice developed lymphoma earlier than wild-type littermates in response to challenge with murine leukemia virus, and, when crossed with mice carrying a mutation in the Apc tumor suppressor, developed twice the number of intestinal tumors as their normal littermates. Characterization of these intestinal tumors indicated that some BlmCin/+ tumors were more dysplastic histologically and had lost the normal Apc allele by somatic recombination. Therefore, mutation of one allele of Blm has measurable consequences on the phenotype of murine somatic cells and tumor susceptibility. Highlighting the importance BLM in tumorigenesis, the effects of BLM haploinsufficency apparently apply to human populations. A recent epidemiological study demonstrated that human carriers of BLMAsh have a nearly three-fold increase in the occurrence of colorectal cancer (Gruber et al., 2002).

35 Conclusions

The absence of BLM is associated with excessive chromatid and chromosome exchange. Functional studies of BLM have explored its roles in recombination and DNA repair, and suggest functions in suppressing recombination by promoting non- recombinogenic pathways for DNA repair and resolution of HJs without crossing over.

The biochemical properties of BLM may also be important in pathways outside the boundaries of DNA repair. In this thesis, the function of BLM in telomere lengthening will be explored.

36 CHAPTER TWO. Thesis rationale and research objectives.

Early cytogenetic studies of BS cells discovered a high frequency of telomere associations (TAs) between homologous chromosome arms. More recent studies have demonstrated telomere length hyper-variability in BS cells (Schawalder et al., 2003).

These data indicate instability of telomeric repeats in BS and suggest a function for BLM in maintaining telomere structure and/or length.

While telomerase is the primary mechanism of telomere lengthening in humans and yeast, telomeres also can be lengthened by telomerase-independent pathways mediated by recombination, referred to as ALT (alternative lengthening of telomeres).

In S. cerevisiae, two ALT-related pathways have been identified that require either Rad50 or Rad51 (Le et al., 1999). Sgs1, the S.cerevisiae BLM homolog, is required for the Rad50- dependent telomere lengthening such that sgs1 mutants utilize only the Rad51- dependent pathway (Cohen and Sinclair, 2001; Huang et al., 2001; Johnson et al., 2001).

Mammalian ALT cells have long terminal telomeric repeats which resemble those found in yeast utilizing the Rad50-dependent pathway (Bryan et al., 1995).

Based on these studies, we hypothesis that BLM functions in telomere lengthening pathways that are mediated by recombination by promoting resolution of recombinant telomeres. This hypothesis was tested by cytogenetic, cell biological, genetic and biochemical methods to investigate the in vivo association of BLM with telomeres and telomeric proteins, the conserved functions of BLM and Sgs1 in ALT, and the in vitro association and regulation of BLM by telomeric proteins.

37 CHAPTER THREE. Functional association of BLM with recombination-mediated telomere lengthening.

I. Introduction

The ends of linear chromosomes are organized into nucleoprotein complexes termed telomeres. Telomeric DNA in vertebrate cells is composed of repeating units of the hexa-nucleotide sequence (TTAGGG)n/(CCCTAA)n that are duplexed with the exception of a 3’-overhang composed of the G-rich strand. The 3’-overhang is protected from degradation, end-joining and recombination by insertion into the proceeding duplexed repeats, forming a structure termed the T-loop (Griffith et al., 1999). The 3’- overhang results from incomplete replication at the end of the chromosome or “the end replication problem” (Olovnikov, 1973). Lagging strand DNA synthesis requires the addition of an occasional RNA primer from which new DNA can be extended to form an Okazaki fragment. The RNA is removed and the remaining DNA is ligated to form a contiguous DNA strand. At chromosome ends, there is no mechanism by which the distal-most RNA primer can be replaced with DNA; therefore, removal of this primer results in a 3’-overhang and net loss of DNA with each subsequent round of replication

(Harley et al., 1990).

The “end replication problem” results in cells with a finite lifespan due to telomere shortening. In order to bypass replicative senescence, must activate specific mechanisms to replace lost telomeric sequences. The multi-subunit ribonucleoprotein reverse transciptase called telomerase is required for the primary mechanism of telomere length maintenance in humans. Telomerase is active in embryonic tissues,

38 germ cells and stem cells, but not in cells with a finite lifespan (Shay and Wright, 1996;

Wright et al., 1996). Telomerase-independent mechanisms for telomere lengthening were first discovered in human cells when a small percentage of transformed, immortalized cell lines were found without telomerase activity (Bryan et al., 1997; Bryan et al., 1995;

Dunham et al., 2000; Lundblad and Blackburn, 1993). These telomerase-independent pathway(s) were termed alternative lengthening of telomeres, or ALT. ALT can be used by some immortalized transformed human cells, some human tumors and cells derived from telomerase-null mouse lines, but does not occur in any normal tissues or cell types

(Henson et al., 2002). While telomere length in telomerase-positive human immortalized cell lines are maintained at a steady, homogenous length averaging 10 kb, telomeres in human cells using ALT average 20 kb and are highly heterogeneous, varying from 3 to

50 kb (Bryan et al., 1995).

The rapid erosion and elongation of telomeres in cells using ALT suggested a recombinational pathway for telomere lengthening (Bryan et al., 1995). Dunham et al. were the first to demonstrate that recombination occurs between telomeric repeats in cells using ALT. In this study, a plasmid with a NEO-cassette was integrated into telomeric DNA in a human cell line using ALT. Two cell clones derived from integration of this plasmid contained 2 and 3 chromosomes with tagged telomeres at 23rd population doubling; the number of tagged telomeres in each clonal population increased to 10 after 40 population doublings. Integration into a subtelomeric region did not result in transfer or copying to any other chromosomes demonstrating telomere repeat-specific recombination. Additional studies have demonstrated that homologous

39 recombination is upregulated at telomeres in cells using ALT, while the rate of recombination outside of the telomere remains unchanged (Bailey et al., 2004; Bechter et al., 2004; Bechter et al., 2003; Londono-Vallejo et al., 2004).

Transfer of a telomere tag to multiple chromosomes is likely accomplished by

BIR in which chromosomes copy telomeric sequences from one chromosome to another by replication (Figure 5A). However, telomere-telomere replication may not be the only mechanism for ALT. Intra-chromosomal recombination may also allow replication to be initiated from the T-loop, potentially allowing continuous replication by a rolling circle type mechanism (Figure 5B). Linear and circular extrachromosomal telomeric repeat

DNA is present at higher levels in cells using ALT (Tokutake et al., 1998); such DNAs may represent byproducts of telomeric recombination or may serve as additional templates for replication. Circular telomeric DNAs could facilitate telomere synthesis by rolling circle replication (Figure 5C), as in K. lactis (Natarajan and McEachern, 2002).

A consistent feature of human cells using ALT is the presence of ALT-associated

PML bodies (APBs) (Grobelny et al., 2000; Yeager et al., 1999). PML nuclear bodies

(PNBs) are nuclear matrix-associated subnuclear structures that exist in most cell types, so named due to the presence of the promyelocytic leukemia or PML protein, a putative tumor suppressor. PNBs are highly dynamic structures and protein components vary with cell type, cell cycle, cell health and external stimuli. APBs compose a subset of

PNBs present only in cells using ALT that contain unique telomeric components, including telomeric DNA (chromosomal and/or extrachromosomal) and telomere (or

TTAGGG) repeat binding factors, TRF1 and TRF2, as well as aggregates of proteins

40

telomere of separate circular ECTR Chromosome or linear ECTR t-loop

A B C

Figure 5. Mechanisms of ALT. (A) ALT may occur by BIR using a separate chromosome or linear extra-chromosomal telomeric repeat (ECTR) DNA as a template. (B) Alternatively, replication may be initiated directly from the T-loop or (C) by using circular ECTR DNA as a template. Both B and C may facilitate continuous, rolling circle synthesis of new telomeric repeats.

41 involved in HR repair, including RAD51, RAD52, RPA and the RMN complex (Grobelny et al., 2000; Wu et al., 2000; Yeager et al., 1999). Analysis of the appearance of APBs during the cell cycle reveals that they are enriched during G2/M (Grobelny et al., 2000).

Importantly, BrdU incorporates into foci that co-localize with APBs during late S-phase and G2 (Wu et al., 2000). Thus, APBs may be scaffolding sites that facilitate telomere lengthening by bringing together telomeres and proteins required for ALT during late S and G2/M after most replication is complete. Alternative recombinational pathways for telomere lengthening also exist in S. cerevisiae in which telomerase has been inactivated

(Lundblad and Szostak, 1989; Lundblad and Blackburn, 1993). Telomerase-null

“survivors” are grouped into two classes based on the structure of their telomeric sequences (Chen et al., 2001; Le et al., 1999; Teng and Zakian, 1999). In the type I group, the terminal telomeric repeats are considerably shorter than their telomerase-positive counterparts, but the shortened length appears to be maintained over time.

Additionally, type I survivors have amplified Y’-elements, subtelomeric long repeat tracts. In contrast, the type II survivors have normal numbers of Y’-elements, but have heterogeneous, often lengthy terminal telomeric repeats. The two groups may arise during different times in crisis. Survivors that activate recombinational telomere lengthening early during crisis may amplify the terminal telomeric repeats (type II), while those that activate later, after most terminal repeats tracts have been lost, may amplify the sub-telomeric Y’-elements (type I). Each survivor pathway requires a distinct set of protein components. Type I telomere lengthening requires Rad51, Rad54 and Rad57, while type II telomere lengthening requires Rad59, Rad50 and associated

42 proteins, Mre11 and Xrs2 (Chen et al., 2001). The separation in protein components required for type I and type II pathways mimics the separation in proteins required for the two known BIR pathways for DSB repair (Ira and Haber, 2002) . Thus, both human and yeast ALT likely occur by BIR. Importantly, the type II pathway additionally requires Sgs1, the BLM homolog and sole RecQ-like helicase in S. cerevisiae (Cohen and

Sinclair, 2001; Huang et al., 2001; Johnson et al., 2001).

BLM associates with PML bodies and interacts directly with at least two HR repair proteins that reside within APBs, RAD51 and RPA (Grobelny et al., 2000).

Indirect evidence also suggests a connection between BLM and the APB-associated

RMN complex (Franchitto and Pichierri, 2002). This, together with the finding that the

BLM homolog Sgs1 is required for type II recombination-mediated telomere lengthening in S. cerevisiae (Cohen and Sinclair, 2001; Huang et al., 2001; Johnson et al., 2001), led us to investigate the association of BLM with telomeres was further investigated.

Fluorescence in situ hybridization with sub-telomeric and pan-telomeric probes confirmed the presence of TAs in SV40-transformed BS fibroblasts previously identified by the German laboratory in lymphoblastoid cell lines by cytogenetic analysis.

Immunostaining with antibodies specific for BLM, TRF2, and PML analyzed the localization of BLM in primary cell lines and in immortalized cell lines using ALT or telomerase. BrdU pulse-labeling determined that BLM and TRF2 foci are associated with sites of DNA synthesis. Finally, the function of BLM in ALT was tested by knock- down of BLM, telomere length analysis, and complementation analysis of type II telomere lengthening by BLM expression in telomerase-null sgs1 mutant yeast.

43 II. Methods

Cell lines. Two immortalized and telomerase-negative human cell lines, Saos2 and WI-

38-VA13/2RA, two immortalized and telomerase-positive human cell lines, MCF7 and

Hela, and one non-immortalized primary cell line of diploid human fibroblasts, WI-38, were obtained from ATCC (Yeager et al., 1999). The SV40-transformed BS fibroblast cell line GM08505 was obtained from the Coriell Cell Repository. MCF7 and Saos2 cells were cultured in Dulbecco’s Modified Eagle Medium (Gibco-BRL) containing 10% heat inactivated fetal bovine serum (FBS) (Hyclone), and WI-38 and WI-38-VA13/2RA in

Minimal Essential Media (Gibco-BRL) containing 10% FBS; all cell lines were grown at

37°C and in 5% CO2.

Cytogenetics. Telomeric fluorescence in situ hybridization (FISH) was performed to analyze telomeric associations in BS cells as follows. Prior to hybridization, metaphase spreads prepared on glass slides were fixed in 4% paraformaldehyde, treated with

1mg/ml pepsin in 20 mM Glycine (pH 2) and dehydrated in a serial ethanol series (70,

90, 100%). A FISH hybridization mixture (10 mM Tris, 70% Formamide, 0.5% blocking agent (Roche), 0.5 ug/mL Biotin-labeled (CCCTAA)3 probe) was applied to the slide, heated to 80°C for 3 minutes and hybridized overnight in a wet chamber at room temperature. Excess hybridization mixture was removed by washing twice in 70% formamide in 10mM Tris-HCl (pH 7.2) and TBS with 0.1% Tween-20 and slides were then blocked for 30 minutes in PBG (PBS with 0.2% fish gelatin (Sigma) and 0.5% BSA).

Slides were incubated for 30 minutes at 37°C in fluorescein Avidin D (Vector

44 Laboratories) diluted 1:100 in PBG followed by three washes at 45°C in PBS with 0.1%

Tween-20. Slides were next incubated for 30 minutes at 37°C with FITC-conjugated anti-

Avidin antibody (Vector Laboratories) diluted 1:100 in PBG followed by three washes at

45°C in PBS with 0.1% Tween-20. DAPI was added to final wash to stain DNA. Slides were mounted in Fluoromount and coverslips were sealed with nail polish.

Subtelomeric FISH was performed as indicated by manufacturer (Cytocell). Images were captured on a Hamamatsu digital camera using QED imaging software.

Cell cycle synchronization, bromodeoxyuridine (BrdU) pulse-labeling and flow cytometry. Cells were blocked at G1/S by aphidicolin treatment for 16 to 18 hours

(Sigma, Cat. A0781) (1µg/mL). Cells were released by removal of the aphidicolin- containing media and fixed at indicated time points post-release for either immunofluorescence (1:1 methanol-acetone fixative) or flow cytometry (70% ethanol).

BrdU pulse-labeling was performed as reported by Wu et al. (2000). Cell cycle analyses were performed on propidium iodide-labeled cells using a Coulter Epics XL flow cytometer.

Immunostaining. The rabbit anti-BLM antibody was generated, purified and used for immunofluorescence as described by Yankiwiski et al. (33). An anti-PML monoclonal antibody (Santa Cruz, Cat. PG-M3), an anti-TRF2 monoclonal antibody (Oncogene, Cat.

OP129) and a rat monoclonal anti-BrdU antibody (Accurate Scientific, Cat. OBT0030) were used for immunofluorescence. Secondary antibodies included FITC-conjugated donkey anti-rabbit, FITC-conjugated goat anti-mouse, rhodamine-conjugated goat anti-

45 mouse, and rhodamine-conjugated donkey anti-rat (Jackson ImmunoResearch

Laboratories). Fluorescence microscopy was performed using an Axioplan 2 Zeiss microscope. Images were captured on a Hamamatsu digital camera using QED imaging software.

Chromatin immunoprecipitations. In vivo cross-linking and immunoprecipitations were performed similarly to those described by Hsu et al. (Hsu et al., 1999). Briefly, cells were harvested and washed in PBS. Fixation and cross-linking were carried out in buffer containing 1% formaldehyde/1% methanol. DNA was sheared by sonication and immunoprecipitations performed using non-specific IgG (negative background control), anti-TRF2 antibody (Oncogene, Cat. OP129) and anti-BLM antibody (Santa Cruz, Cat. sc-

7789 and 7790). Cross-links were reversed and DNA purified using two phenol/chloroform/isoamyl alcohol (25:24:1, vol/vol) extractions and ethanol precipitation. DNA was analyzed by slot-blotting onto Zeta-probe membrane (Bio-Rad), and subsequent hybridization to [32P]-labeled telomere repeat probe, (TTAGGG)10. Blots were analyzed using a Molecular Dynamics Phosphorimager and band densities were determined using ImageQuant software.

In vivo protein co-immunoprecipitations. In vivo immunoprecipitations from nuclear extracts using goat anti-BLM antibodies (Santa Cruz, Cat. sc-7789 and 7790) were performed as previously described (Langland et al., 2001). Nuclear extracts were pre- treated with DNase I (10 µg/mL) to eliminate interactions mediated by DNA tethering.

Monoclonal anti-TRF2 antibodies (Oncogene, Cat. OP129) were used for western blot

46 analysis. 50 to 100 µg of nuclear extract were loaded for input lanes on each gel. To confirm that immunoprecipitations were successful, western blots were re-probed using anti-MLH1 antibodies (PharMingen, Cat. 13271A) and anti-BLM antibodies (Novus, Cat.

NB 100-161).

BLM knock-down. BLM expression was knocked-down by treatment of cells with antisense BLM morpholinos designed and purchased from GeneTools, LLC (Philomath,

OR). Sequences for morpholinos used in these experiments are as follows, BLM antisense morpholinos (GATTATTTTGAGGAACAGCAGCCAT) and negative control, inverse BLM antisense morpholinos (TACCGACGACAAGGAGTTTTATTAG). Cells were grown to sub-confluence (60-70%) in 6 well plates overnight in media containing no antibiotics. Morpholinos were delivered to cells using EPEI delivery reagent according to manufacturers directions. BLM expression was determined in treated cells by western blot analysis with anti-BLM antibody (Novus, Cat. NB 100-161).

Human telomere length analysis. Genomic DNA was collected from cells and 10 ug

DNA was digested with RsaI and HinfI restriction enzymes overnight. Digested DNA was electrophoresed on a 0.8% agarose gel in 1xTBE, transferred overnight to positively- charged nylon membrane (Immobilion-NY+). DNA was cross-linked to membrane using a Stratalinker and Southern blots performed according to standard procedure using a [32P]-labeled telomere-specific probe (CCCTAA)3. Blots were incubated with probe in hybridization buffer overnight at room temperature, washed three times for 15

47 minutes at room temperature, and exposed to Biomax MS autoradiographic film for 2 days at -80°C for analysis.

Expression vectors. pYES-BLM expression vector (pJK1) was provided by I. Hickson

(University of Oxford, Oxford, UK) (Karow et al., 1997). Mutation of the BLM in pYES to create pYES-K695E BLM was performed by site-directed mutagenesis using PCR.

Presence of the mutation was confirmed by sequencing. Helicase assays with purified proteins (also expressed from pYES vector and purified from yeast) confirmed unwinding activity of wild-type BLM and lack of unwinding with BLMK795E (data not shown).

Yeast strains and crosses. sgs1::HIS3; URA3 yeast (RDKY3813) was provided by

R.Kolodner (University of California-San Diego, La Jolla, CA) (Myung et al., 2001). The

URA3 marker present in this strain was eliminated by selection on 5-FOA, confirmed by

PCR and lack of growth on media lacking uracil, to create sgs1::HIS3 which was used for all subsequent experiments. est2::KANMX4/EST2 yeast (BY4743-YLR318W) were obtained from ATCC Yeast Genetic Stock Center. The latter was streaked onto standard sporulation media to obtain α; est2::KANMX4 spores which were mated to a; sgs1::HIS3 to produce a double heterozygous diploid strain that was subsequently transformed with pYES-BLM (aka pJK1) (Karow et al., 1997), pYES-K695E BLM, or pYES vector

(Invitrogen, Cat. V825-20). The diploid strain was sporulated and tetrads dissected.

Individual spores with genotypes SGS1 EST2 (wild-type), sgs1::HIS3 EST2, SGS1 est2::KANMX4 and sgs1::HIS3 est2::KANMX4 were selected on the appropriate media and passaged on uracil dropout plates with minimal base containing 2% galactose at

48 30ºC (Clontech, Cat. 8607-1, 8611-1). After re-streaking 10 times (approximately 250 generation times), single colonies were selected and grown in standard liquid uracil dropout medium with minimal base containing 2% galactose (Clontech, Cat. 8607-1,

8611-1). Yeast were diluted 1:1000 every 24 to 72 hours into fresh media for an additional

10 passages (approximately 100 generation times). Yeast were collected daily for telomere length analysis.

Yeast telomere length analysis. For genomic DNA isolation, pelleted yeast were resuspended in 200 µL Buffer A and glass beads, and lysed by vortexing for 2 minutes.

DNA was extracted with phenol:chloroform and ethanol-precipitated. 7.5 µg of genomic

DNA was digested with XhoI (New England Biolabs) at 37ºC overnight and resolved by electrophoresis on a 1.0% TBE agarose gel. Digested DNA was transferred onto a positively charged Hybond-N+ nylon membrane (Amersham, Cat. RPN303B) and probed by standard Southern blotting procedures. A probe for the C1-3A/TG1-3 terminal repeats was created by random-priming using Poly(dA-dC)·Poly(dG-dT) (Amersham,

Cat. 27-7940-01) as described by others (Le et al., 1999).

III. Results

Analysis of TAs by FISH analysis in transformed BS fibroblasts.

Cytogenetic evaluation of BS cells revealed an excess of various chromatid lesions, including the quadriradial (QR) configuration signifying the hyper- recombination associated with BLM mutation (German, 1993). In addition, a significant increase in TAs between homologous chromosome arms at metaphase were observed in

49 BS lymphoblasts as compared to non-BS lymphoblasts (Figure 2D-F and Table 1). TAs may also be found in cells from normal persons, although in cultured cells TAs are most frequently found in cells defective in telomere lengthening, presumably due to recombination and end-joining between shortened telomeres (Blasco et al., 1997;

MacEachern and Iyer, 2001).

Using fluorescence in situ hybridization (FISH) with sub-telomeric and telomeric repeat probes, TAs between homologous chromosome arms were also observed in metaphases prepared from immortalized BS fibroblast cells GM08505 (Figure 6A).

Telomeric sequences were present at the junction of TAs demonstrating that TAs in BS cells do not occur exclusively between critically shortened telomeres (Figure 6B). These cytogenetic data confirmed the presence of TAs in BS cells, and suggested a function for

BLM in telomere protection. Alternatively, TAs may have resulted from intertwined or recombinant chromosome ends that persisted in the absence of the BLM helicase.

Nuclear localization of endogenous BLM and TRF2.

TAs in BS cells suggested that BLM is a telomere-associated protein. Given the proposed functions for BLM in HR, it was of interest to compare localization of BLM with respect to telomeres in telomerase-negative immortalized cells using ALT (Saos2 and VA13/2RA), as well as telomerase-positive immortalized cells (MCF7 and Hela) and primary cells (WI-38). Dual immunostaining was performed with antibodies specific for BLM and TRF2, a telomere repeat binding protein involved in telomere length regulation and protection (Broccoli et al., 1997; Griffith et al., 1999; Karlseder et al., 2002;

Smogorzewska et al., 2000; Stansel et al., 2001). BLM co-localized with TRF2 in both cell

50 A.

B.

Figure 6. Analysis of telomere associations (TAs) of homologous chromosome arms. A.) FISH analysis of metaphases prepared from BS fibroblasts (GM08505) with a chromosome arm-specific subtelomeric probe demonstrates a TA between q-arms of chromosome nos. 14 (white arrow). B.) FISH analysis of metaphases prepared from BS fibroblasts (GM08505) with a telomeric repeat probe shows telomeric signals at the junction of a TA (white arrow).

51 lines using ALT, but did not co-localize with TRF2 in either telomerase-positive or primary cell lines (Figure 7). These data demonstrate a spatial relationship between

BLM and telomeres in cells using ALT, consistent with co-localization of BLM with TRF1 previously described in the VA13/2RA cells (Yankiwski et al., 2000).

Telomeric components, including TRF1 and TRF2, aggregate within APBs specifically in cells using ALT (Yeager et al., 1999). APBs additionally contain two BLM protein partners, RAD51 and RPA. Co-localization of BLM with TRF2 would suggest that BLM is an APB-associated protein. Immunofluorescence analysis was used to determine the localization of BLM and TRF2 with respect to PML in cells using ALT

(Figure 8). As anticipated, both BLM and TRF2 co-localized with PML in cells using

ALT demonstrating that BLM is a component of APBs.

BLM and TRF2 foci are enriched during G2/M and undergo DNA synthesis.

Cells that use the ALT pathway are enriched with APBs in G2/M, suggesting that recombination-mediated telomere lengthening is cell cycle-regulated (Grobelny et al.,

2000). If BLM is involved in ALT, its association with TRF2 and PML is predicted to increase as cells progress into G2/M. To test this, VA13/2RA and Saos2 cells were blocked at the G1/S boundary and analyzed at 4 and 10 hours post-release to compare co-localization of BLM with TRF2 during G1/early S or late S/G2/M, respectively (Figure

9). Cell cycle distribution was confirmed by flow cytometry (Figure 9A). BLM co- localized with TRF2 throughout the cell cycle in both cell lines; however, the number of cells with overlapping foci increased approximately 2-3 fold as cells progressed into late

52 ALT Telomerase Primary

Figure 7. BLM co-localizes with TRF2 in cells using ALT, but not in telomerase-positive or primary cells. Cells that use ALT, Saos2 (A-C) and VA13/2RA (D-F), telomerase- positive cells, MCF7 (G-I) and HeLa (J-L), and the telomerase-negative mortal cell line WI-38 (M-O) were fixed and immunostained with anti-BLM rabbit antibody (FITC- labeled secondary antibody) and anti-TRF2 mouse monoclonal antibody (Rhodamine- labeled secondary antibody). BLM foci (top panel) and TRF2 foci (middle panel) are present in all cell lines analyzed. BLM and TRF2 co-localize in ALT cells (C, F), but not in telomerase-positive cells (I, L) or telomerase-negative mortal cells (O). Antibody controls are performed with anti-BLM plus mIgG (instead of mouse anti-TRF2) (P) and anti-TRF2 plus rIgG (instead of rabbit anti-BLM) (Q).

53

Figure 8. BLM localizes to APBs. Saos2 cells were fixed and immunostained for PML (Rhodamine-labeled secondary antibody) and BLM (FITC-labeled secondary antibody) or TRF2 (FITC-labeled secondary antibody). Both BLM (a-c) and TRF2 (d-f) co-localize in PML nuclear bodies in Saos2 cells. Antibody controls were performed with anti-TRF2 plus mIgG (instead of mouse anti-PML)(g), anti-PML plus mIgG (instead of mouse anti- TRF2)(h), and anti-PML plus rIgG (instead of rabbit anti-BLM)(i).

54 S and G2/M (Figure 9B). During G1 and early S, foci in which BLM and TRF2 co-localize are present in 15% of VA13/2RA cells and 21% of Saos2 cells; during late S and G2/M, the number of cells with BLM/TRF2 foci increases to 38% and 54%, respectively.

Immunostaining with BLM and PML antibodies demonstrated that co-localization of

BLM with PML is regulated in a similar manner with respect to the cell cycle (Figure

9C). Increased localization in APBs could be due to either translocation of BLM to APBs or to an increase in the number of APBs during late S and G2/M.

Proteins implicated in ALT, including NBS1 and TRF1, co-localize with foci of

BrdU-incorporation during late S and G2 (Wu et al., 2000). Because APBs are enriched during G2/M (Grobelny et al., 2000), it is probable that these foci represent centers of telomere lengthening in cells using ALT. Thus, immunofluorescence was used to determine whether BLM and TRF2 similarly co-localized with active centers of DNA incorporation during late S and G2/M. VA13/2RA cells (ALT) were blocked in G1/S and pulse-labeled with BrdU at 10 hours post-release from block at which time cells were in late S and G2/M as determined by flow cytometry (see Figure 9A). Cells were evaluated with antibodies specific for BrdU as well as for BLM or TRF2. During late S and G2/M, both BLM and TRF2 immunostaining overlapped with BrdU in distinct foci (Figure 10).

In contrast, few distinct BrdU foci were distinguishable in telomerase-positive cells

(MCF7) during G2/M; co-localization with BLM or TRF2 was not observed. Our results indicate that BLM, TRF2 and PML co-localize with sites of ongoing DNA synthesis in

APBs during late S and G2/M.

55

A)

G1

B) C)

Figure 9. Cell cycle-regulated association of BLM with TRF2 and PML. Cells using ALT (Saos2 and VA13/2RA) were blocked in G1/S and co-stained with BLM and PML or TRF2 antibodies at 4 and 10 hours post-release from block. Cells are in G1/early S and late S/G2/M at these time points respectively, as determined by flow cytometry (A). Percentage co-localization of BLM with TRF2 (B) and BLM with PML (C) was scored by analyzing digital photographs for co-localizing signal. Data are presented as a percentage of total cells showing co-localization. The experiment was repeated three times for each cell line to obtain averages and standard deviations.

56

Figure 10. BLM and TRF2 co-localize with foci of DNA synthesis during late S/G2/M. VA13/2RA cells were synchronized in G1/S by aphidicolin block. Cells shown were pulse-labeled with BrdU at 10h post-release from aphidicolin block and subsequently fixed for immunostaining. Flow cytometry was used to confirm cell cycle stage (see Fig 9A). Cells were fixed and immunostained with BrdU antibodies (b,e) (rhodamine- labeled secondary) and either BLM (a) or TRF2 (d) antibodies (FITC-labeled secondary). Antibody controls are performed with anti-BLM and rat IgG (g) and anti-TRF2 and rat IgG (h); rat IgG used instead of anti-BrdU antibody for both controls.

57 Association of BLM with telomeric DNA and TRF2 in vivo.

If BLM functions at telomeres in cells using ALT, it should associate with telomeric DNA in these cells. Therefore, chromatin immunoprecipitation experiments were performed comparing Saos2 (ALT) and MCF7 (telomerase-positive) cells. Proteins were cross-linked to DNA and immunoprecipitated with either normal goat IgG, antibodies specific for BLM or antibodies specific for TRF2. Immunoprecipitated DNA was hybridized with a radiolabeled telomere-specific probe (Figure 11A-B). While TRF2 immunoprecipitated approximately equal amounts of telomeric DNA from cells using

ALT and telomerase, BLM immunoprecipitated significantly higher amounts of telomeric DNA from cells using ALT. These data are consistent with a direct association of BLM with telomeric repeats in cells using ALT.

Other proteins implicated in ALT, including members of the RMN complex, associate with TRF2 and TRF1 in vivo (Wu et al., 2000; Zhu et al., 2000).

Immunoprecipitations were performed to determine whether BLM and TRF2 are components of the same telomere-specific protein complex. BLM immunoprecipitations from nuclear extracts prepared from cells using ALT (Saos2 and VA13/2RA) and telomerase (MCF7 and HeLa) were analyzed by SDS-PAGE and western blot analysis using anti-TRF2 antibodies. TRF2 co-immunoprecipitated with BLM in nuclear extracts from cells using ALT but not from telomerase-positive cells (Figure 11C). DNase treatment did not interfere with immunoprecipitation, suggesting that the interaction is independent of DNA tethering. These data confirm an in vivo association between BLM

58

Figure 11. BLM co-immunoprecipitates with telomeric DNA and TRF2 from cells using ALT. A.) Saos2 (ALT) and MCF7 (telomerase-positive) cells were treated with formaldehyde to crosslink protein to DNA. DNA was sheared by sonication and proteins immunoprecipitated with normal mouse and goat IgG (negative controls), mouse monoclonal TRF2 antibody (positive control), or goat polyclonal BLM antibody. Following rigorous washing, cross-links were reversed and DNA purified by phenol: chloroform extraction. Pure DNA was denatured, slot-blotted on nitrocellulose and evaluated with a [32P]-labeled telomere probe. B.) Quantitation of telomeric DNA immunoprecipitated with BLM. Density of telomeric DNA signal shown in 11A was determined by densitometry. Percentage density was calculated by subtracting density of band in negative control (IgG) from density of band in BLM immunoprecipitations, divided by density of total input DNA. Averages for three experiments are shown with standard deviations. C.) BLM was immunoprecipitated from 500mg of nuclear extracts (NE) prepared from cells using ALT (Saos2 and VA13/2RA) or telomerase-positive cells (MCF7 and Hela). Immunoprecipitations were performed using goat polyclonal BLM antibody or normal goat IgG (negative control). Immunoprecipitated proteins were separated by SDS-PAGE and analyzed by western blot using an anti-TRF2 antibody.

59 and TRF2 and suggest these proteins may be part of a larger telomere-associated protein complex that regulates ALT.

Effects of reducing BLM expression on telomere length in cells using ALT.

To determine if disruption of BLM expression disrupts telomere elongation, BLM expression was repressed by treatment of cells with BLM antisense morpholinos; inverse antisense morpholinos were used as negative controls. BLM expression in treated cells was determined by western blot analysis (Figure 12A). Saos2 cells were treated with morpholinos for the indicated amount of time, collected and analyzed for telomere length by standard telomere restriction fragment (TRF) Southern blotting

(Figure 12B). No clear difference in telomere length could be established between cells treated with BLM antisense morpholinos and inverse antisense morpholinos.

Complementation of type II telomere lengthening by BLM expression in est2 sgs1 mutants

Yeast lacking the catalytic subunit of telomerase EST2 stop dividing; although most est2- do not survive, a subset continues to grow using a Rad51-dependent or a

Rad50-dependent recombination-mediated telomere lengthening, also known as the type I and type II pathway respectively (Le et al., 1999; Lundblad and Blackburn, 1993;

Teng and Zakian, 1999). The RecQ-like helicase Sgs1 is required for the type II pathway

(Cohen and Sinclair, 2001; Huang et al., 2001; Johnson et al., 2001). Terminal telomere repeats in type II survivors are long and heterogeneous, while telomeres in type I

60

Figure 12. Knockdown of BLM and telomere length analysis in Saos2 cells. A) Saos2 cells were collected 48 hours after treatment with antisense or reverse antisense (negative control) BLM morpholinos. Proteins were isolated and equal quantities were analyzed by SDS-PAGE and western blot with anti-BLM antibody. HeLa nuclear extract (NE) is also included as a positive control for western. B) Cells were treated with antisense BLM morpholino (A) and reverse antisense BLM morpholino (RA) for 3 to 6 days. Cells were collected and DNA was analyzed by Southern blot analysis with telomere probe. The telomere restriction fragment (TRF) is unchanged from negative controls at both time points. Size markers are shown to the left.

61 survivors have amplified sub-telomeric Y’-elements and very short terminal repeat tracts (Figure 13A).

To test whether BLM and Sgs1 have conserved functions in recombination- mediated telomere lengthening, rescue of the type II pathway by BLM expression was tested in an sgs1est2 mutated strain. An sgs1::HIS3/SGS1; est2::KAN2/EST2 diploid strain was transformed with a galactose-inducible BLM expression vector (pYES-BLM) or vector alone (pYES) and subsequently sporulated; tetrads were dissected to identify individual sgs1 est2, SGS1 est2, and wild-type spores. Spores were re-streaked and grown in liquid galactose-containing media for an estimated 350 generations, sufficient time to generate stable survivor clones. DNA was isolated, digested with XhoI and analyzed by Southern blot hybridization with a C1-3A/TG1-3 probe (Figure 13B). The telomere probe identified both terminal repeat fragments, as well as Y’ elements due to the presence of short stretches of variable repeats found to separate Y’ elements. A total of 11 cultures of each genotype were analyzed; DNA from 8 of each genotype are shown in Figure 14. SGS1 est2 yeast transformed with BLM expression vector gave rise to both type I survivors (9 of 11), identifiable by sharp Y’ bands and a very short terminal repeat fragment (Y’ TRF), and type II survivors (2 of 11 cultures), identified by the presence of terminal restriction fragments with heterogeneous distribution (Teng and Zakian, 1999).

As predicted by previous studies (Cohen and Sinclair, 2001; Huang et al., 2001; Johnson et al., 2001), the sgs1 est2 survivors transformed with vector only gave rise to only type I pathway survivors (Teng and Zakian, 1999). Three of 11 sgs1 est2 survivors transformed

62

A.

B.

Figure 13. Structure of telomeres is S. cerevisiae type I and type II survivors. A) Expected telomere composition in telomerase-negative survivors utilizing the type I and type II pathways. Telomeres in type I survivors are typified by amplification of Y’-elements and a very short terminal repeat fragment (Y’-TRF). Telomeres in type II survivors have a normal number of Y’-elements and heterogeneous, often very long, TRF length (type II-TRF). B) Structure of a wild-type yeast telomere. 5.2 kb and 6.7 kb Y’-elements (0-4 copies) are present internal to the terminal C1-3A/TG1-3 repeats (~ 1 kb). Y’-elements are separated by short stretches of C1-3A/TG1-3 repeats. XhoI restriction sites are present within Y’-elements as indicated. Digestion with XhoI releases individual Y’- elements and a terminal repeat fragment (TRF) (adapted from Yamada et al., 1998). Other sub-telomeric elements may be present but are not shown.

63 with the BLM expression vector used the type II pathway indicating that BLM can rescue the telomere defect associated with sgs1 mutation.

Rescue of the type II pathway by Sgs1 requires helicase activity indicating that

Sgs1 functions to unwind a specific structure unique to type II recombination (Cohen and Sinclair, 2001). To determine if BLM helicase activity is required for this rescue, a point mutation was introduced into the pYES-BLM vector that changed the conserved lysine within the Walker A box to glutamic acid. The Walker A box is required for ATP hydrolysis, and as a result of this mutation, the mutant BLM protein induced from the pYES-K695E BLM expression vector did not have helicase activity (data not shown). sgs1 est2 mutants expressing K695E BLM were created and analyzed in an identical manner as described above. All sgs1est2 survivors expressing K695E BLM were type I indicating that rescue of the type II pathway by BLM requires helicase activity.

IV. Discussion

Telomeres are thought to form loop structures termed T-loops, an intrachromosomal D-loop formed by insertion of the 3’-overhang into the preceding duplexed telomeric repeats. T-loops, along with telomere binding proteins, form a protective telomere “cap”. However, as telomeres shorten, the T-loop may be unable to form and telomere proteins may be unable to bind. Uncapped telomeres are highly recombinogenic and vulnerable to end-joining. TAs in cells with telomere lengthening defects, such as those found in Terc-/- MEFs, may have little or no detectable telomeric sequences at the junction (Blasco et al., 1997). In contrast, TAs in cells expressing a

64 Figure 14. BLM rescues the type II pathway in sgs1 est2 yeast mutants. SGS1 est2 and sgs1 est2 cultures (transformed with BLM expression vector, K695E BLM expression vector or pYES vector alone) were grown for greater than 300 generations to allow for generation of survivors. DNA from yeast was digested with XhoI and analyzed by Southern hybridization with a poly GT/CA probe. sgs1 est2 mutants transformed with BLM expression vector are type I (lanes 2-6) and type II survivors (lanes 7-9). As reported previously, sgs1 est2 mutants transformed with vector only produced survivors using the type I pathway (lanes 10-17). SGS1 est2 mutants transformed with BLM produce both type I (19-20, 22-26) and type II (21) survivors. By contrast, sgs1 est2 mutants transformed with an expression vector encoding the helicase-deficient K695E BLM mutant utilize only the type I pathway (lanes 27-34). TRF= terminal repeat fragment.

65 dominant-negative TRF2 have telomeric sequences at the junction, but are apparently fused due to erosion of the 3’-overhang and inability to form a T-loop structure (van

Steensel et al., 1998). Ironically, cells lacking HR and NHEJ proteins ATM, DNA-PKcs,

Ku70, Ku86, RAD54 and RAD50 (among others) display increased end fusions with and without telomere shortening (for review, see Ferreira et al., 2004). Thus, many of the proteins that induce TAs under dysfunctional circumstances, may also serve to lengthen and/or protect telomere structure.

A high frequency of TAs occur in metaphase spreads prepared from BS lymphoblasts suggesting a role for BLM in telomere maintenance as well. Telomere length heterogeneity in BS cells indicates that telomeres are hyper-recombinogenic, perhaps indicating telomere de-protection (Schawalder et al., 2003). In the present chapter, we have utilized telomere FISH to characterize TAs in a BS fibroblast cell line.

As described previously, the TAs in BS fibroblasts occur between homologous chromosome arms. Importantly, a clear telomeric signal is present at the junction of the

TAs indicating that TAs do not occur solely between critically shortened telomeres. The nature of these TAs suggest that telomeres in BS cells are uncapped at metaphases making homologous chromosome arms susceptible to end-joining and recombination.

Thus, BLM may be required to process telomeric ends to facilitate T-loop formation, a function which has also been ascribed to the RMN complex via association with TRF2

(Zhu et al., 2000). Alternatively, BLM may be required to resolve recombinant or entangled ends that occur when chromosome ends are exposed during replication. If the latter is true, BLM would be particularly important for maintaining telomere length

66 and structure under circumstances in which telomere recombination is upregulated, such as in cells using ALT.

Cells using ALT pathways may present unique circumstances for the telomere in which end-joining activity remains repressed but homologous recombination becomes activated to facilitate telomere lengthening. Variant PML bodies, termed APBs, are found in cells using ALT, containing telomeric DNA, TRF1 and TRF2, as well as proteins involved in recombination RAD51, RAD52, RPA, and the RMN complex (Wu et al.,

2003; Yeager et al., 1999). The incorporation of BrdU within APBs suggests they represent sites of recombination-mediated telomere lengthening (Wu et al., 2000). ALT may be temporally dissociated from normal replication, as APBs tend to be enriched in cells during G2/M (Grobelny et al., 2000). While BLM is associated with PML bodies, we demonstrate here, using co-localization with TRF2 and PML, that BLM is a component of APBs. Consistent with a function in telomere lengthening, the association of BLM with APBs increases during late S and G2/M. BLM and TRF2 also co-localize with foci of

DNA synthesis. Furthermore, a direct association of BLM with telomeric DNA and

TRF2 is shown by immunoprecipitation from cells using ALT. These data suggest that

BLM is a component of a telomere-associated complex which regulates recombination- mediated telomere lengthening.

The biochemical activities of BLM may promote or repress ALT. To test this, we utilized morpholino antisense technology to reduce BLM expression in cells in culture such that the function of BLM in ALT could be investigated. Antisense morpholinos were used to knock-down BLM expression levels. This strategy has not been previously

67 used to reduce significantly BLM expression in vivo. However, we were unable to demonstrate a significant change in telomere length in cells treated with antisense versus inverse antisense (negative control) morpholinos. A recent publication by Stravopolous et al. demonstrated that over-expression of BLM resulted in rapid extension of telomeres in cells using ALT suggesting that BLM promotes telomere synthesis (Stavropoulos et al., 2002). Our negative results may be attributable to the low sensitivity of our system.

First, telomeres in ALT cells are exceedingly long and hyper-variable making it difficult to discern any small gains or loss in telomere length using standard TRF analysis.

Stravopolous et al. used FLO-FISH, a flow cytometry-based telomere length assay which may be more sensitive for detecting small length changes in long telomeres. Secondly, the reduction of BLM for 3-7 population doublings may be an insufficient period of time to visualize any change or the small amount of BLM which remains after morpholino treatment may be sufficient for function in ALT. Finally, the related RecQ-like helicase

WRN may also participate in ALT via association with TRF2 (Opresko et al., 2002); therefore, it is possible that WRN has overlapping functions with BLM and may be used in its absence or reduction.

We next used S. cerevisiae as an alternative means for evaluating the function of

BLM in ALT. In S. cerevisiae, two recombination-mediated telomere lengthening pathways exist which elongate telomeres in the absence of telomerase. Both pathways require Rad52, but differentially require Rad51 or Rad50 (Le et al., 1999; Lundblad and

Blackburn, 1993). Yeast “survivors” utilizing the Rad51-dependent pathway (type I) have a short terminal repeat tract and amplified subtelomeric Y’-elements indicating

68 that these survivors utilize Y’-elements to extend critically shortened telomeres.

Whereas survivors utilizing the Rad50-dependent pathway (type II) have normal numbers of subtelomeric Y’-elements and long, heterogeneous terminal telomeric repeats. Telomeres in type II survivors more closely resemble the long and heterogeneous telomeres found in ALT cells (Bryan et al., 1995). Sgs1 is required for the type II pathway such that telomerase-negative sgs1 mutants can utilize only the type I pathway. In the current study, we find that BLM expression complements the type II pathway in sgs1est2 mutants. Furthermore, BLM helicase activity was required for this rescue as expression of the helicase-dead mutant K695E BLM does not result in the appearance of type II survivors. These data indicate that BLM and Sgs1 have conserved biochemical functions during ALT and most likely resolve similar types of structures.

Importantly, Sgs1 and BLM can resolve similar substrates in vitro, including synthetic

HJs (X-junctions), 3’-overhangs, D-loops and G-quadruplexes (Bennett et al., 1998; Sun et al., 1999). BIR is thought to mediate ALT in both yeast and humans (see Figure 5A). In fact, the division between the protein factors required for type I and type II pathways mimics the division between the Rad51-dependent and Rad50-dependent pathways for

BIR repair of DSBs that are characterized in S. cerevisiae (Ira and Haber, 2002; Kraus et al., 2001; Malkova et al., 1996; Signon et al., 2001).

Putative functions of BLM (and Sgs1) in a BIR-type mechanism for ALT are shown in Figure 15. First, BLM may be required to unwind the D-loop portion of the telomeric loop such that inter-chromosomal recombination can occur. Secondly, BLM may be involved in processing the 3’-overhang to facilitate strand invasion and pairing.

69 End processing may also be required for proper T-loop formation, and as discussed in the previous chapter, the TAs in BS cells may result from an inability to perform this function. Importantly, the RMN complex in eukaryotes has been implicated in end processing; the functional association with Rad50 in ALT may be conserved between

Sgs1 and BLM. BLM may also facilitate D-loop formation by stabilizing loops that are productive for replication, as has been suggested for BLM in HR repair (van Brabant et al., 2000). In human ALT, this may be accomplished by interaction of BLM with its protein partner, RAD51. Given its ability to facilitate branch migration (Karow et al.,

2000), BLM could additionally promote migration of the single HJ formed during BIR, thus promoting replication and proper resolution of recombinant telomeres. It is notable that Sgs1, while required for Rad50-dependent telomere lengthening, is not required for either BIR pathway for DSB repair in yeast (Kraus et al., 2001; Signon et al., 2001). Thus,

Sgs1, and perhaps BLM in human ALT, is specifically required to promote BIR through telomeric repeats. This requirement may reflect the G-rich nature of the telomeric repeats which may form secondary structures, such as G-quadruplexes, that would stall replication if left unresolved. Importantly, both Sgs1 and BLM are capable of unwinding G-quadruplexes in vitro (Sun et al., 1999; Sun et al., 1998).

Note in proof: some of the data presented in this Chapter are published (Lillard-

Wetherell et al., 2004).

70

A

B

C D

Figure 15. A model for the function of BLM in ALT. A BIR-like model for ALT is presented in which BLM may be required to (A) unwind the D-loop portion of the T- loop, (B) process the 3’overhang, perhaps in conjunction with the RMN complex, to facilitate strand invasion, (C) promote productive D-loop formation and facilitate resolution of the HJ by branch migration, and/or (D) resolve G-quadruplexes that may impede replication.

71 CHAPTER FOUR. Association and regulation of the BLM helicase by TRF1 and TRF2 in vitro.

I. Introduction

The exposed nature of the chromosome end makes it vulnerable to recognition as a DSB. A protective “cap” at the telomere is necessary to prevent chromosome end-to- end fusions and telomeric recombination. In mammalian cells, electron microscopy indicates that a looped structure exists at chromosome ends, presumably formed by insertion of the 3’-overhang into the preceding duplexed telomeric repeats (Griffith et al., 1999). Incubation of telomeric DNA with TRF2 results in formation of the “T-loop” in vitro (Stansel et al., 2001). TRF2 itself binds as a dimer to duplexed telomeric repeats via a MYB helix-turn-helix domain (Broccoli et al., 1997). During T-loop formation,

TRF2 is present at the end of the 3’-overhang and the junction of the D-loop, perhaps allowing it to both form and stabilize the T-loop (Stansel et al., 2001). Cells expressing a dominant-negative form of TRF2 have long telomeres, but an eroded G-rich 3’-overhang which results in chromosome end-to-end fusions (van Steensel et al., 1998). Similar to

TRF2, TRF1 dimerizes and binds to duplexed telomeric repeats via a MYB DNA-binding domain (Bianchi et al, 1999; Broccoli et al., 1997). TRF1 cannot promote T-loop formation; however, it can promote bending of DNA and pairing of duplexed telomeric repeats (Bianchi et al., 1997; Griffith et al., 1998; Griffith et al., 1999). Importantly, depletion of TRF1 results in an increased frequency of end-to-end fusions suggesting

72 that the biochemical properties of TRF1 may function to further package and/or protect chromosome ends (Iwano et al., 2004).

Over-expression of either TRF1 or TRF2 results in telomere shortening in telomerase-positive cells (Smogorzewska et al., 2000). Binding of TRF1 and TRF2 is proposed to induce a “closed” state at the telomere presumably by packaging the 3’- overhang making it unavailable for lengthening. Over time, telomeres shorten such that fewer TRF1 and TRF2 dimers are bound. This loss of TRF1 and TRF2 may allow the telomere to assume an open configuration that is accessible for telomere lengthening.

Ancelin et al. demonstrated cis repression of telomere lengthening by both TRF1 and

TRF2 by targeting of these proteins to a specific telomere using a lac operon system

(2002). TRF1 specifically represses telomerase, while TRF2 induces telomere shortening in both telomerase-positive and negative cells (Ancelin et al., 2002). However, a recent report indicates that TRF2 over-expression may delay senescence in telomerase-negative cells by protecting shortened telomeres from end fusion (Karlseder et al., 2002). Clearly,

TRF1 and TRF2 have critical roles in maintaining telomere structure and telomere length homeostasis.

Data presented in Chapter Three indicate an association of BLM and TRF2.

Other groups have also demonstrated co-localization of BLM and TRF1. In the present chapter, the association and regulation of BLM helicase activity by TRF1 and TRF2 are investigated by in vitro experiments. In vitro immunoprecipitations were performed to examine the interaction of BLM with TRF1 and TRF2. Substrates were designed and

73 created that resemble native telomeric conformations. The ability of purified BLM to unwind these substrates was examined in the presence and absence of TRF1 and TRF2.

Our results suggest a unique mechanism for regulation of BLM helicase activity at telomeres by TRF1 and TRF2.

II. Methods

Protein expression and purification. pYES-BLM expression vector (pJK1) was provided by I. Hickson (University of Oxford, Oxford, UK) (Karow et al., 1997). For expression and purification, pJK1 was transfected into the S. cerevisiae strain JEL1 provided by J.

Wang (Harvard University, Cambridge, MA) (Austin et al., 1995). Yeast were inoculated into –Ura DO + minimal base containing 2% glucose (Clontech, Cat. 8607-1, 8602-1) and grown to an OD600 of 0.6 in an orbital shaker at 30°C. Yeast were pelleted and resuspended in an equal volume of –Ura DO + minimal base containing 2% galactose

(Clontech, Cat. 8607-1, 8611-1) and grown for an additional 20-24 hours. Yeast were resuspended in cold Buffer A (50 mM KPO4 pH 7.0, 500 mM KCl, 10% glycerol and 1:500 mammalian protease inhibitors (Sigma, Cat. P-834) at a volume of 35 mL per liter of culture and lysed using a french press cell. Yeast lysates were bound to a charged

NiNTA resin (Novagen, Cat. 69670) in 1x binding buffer (Buffer A + 50mM imidazole) at

4°C for at least 2 hours. Resin was applied to the column and washed with 1x Binding

Buffer at least 5 times. BLM was eluted with 500 mM imidazole and dialyzed into Buffer

Z (60 mM Tris-HCl pH 7.5, 100 mM KCl, 1mM EDTA, 1mM DTT and 10% glycerol) for 4 hours at 4°C. Small aliquots were frozen and stored at –80°C. Purity was confirmed by

74 coomassie blue staining (Figure 16A). UvrD helicase was provided by S. Matson

(University of North Carolina, Chapel Hill, NC.

High titer baculoviral stocks containing constructs for His-tagged TRF1 and

TRF2 were provided by T. de Lange (The Rockefeller University). Insect cells (Sf9 strain) were infected with TRF2 baculovirus (multiplicity of infection of 5-10) and harvested after 48-72 hr. After lysis by sonication in buffer containing 20 mM Tris-HCl (pH 8.0),

500 mM NaCl, 5 µM imidazole, 0.5% Nonidet P-40 (NP-40), 10% glycerol, and 5 mM β- mercaptoethanol, lysates were clarified by centrifugation and the supernatants mixed in batch with NiNTA resin (Qiagen). The resin was washed extensively in the same buffer without NP-40. Finally, TRF2 eluted with buffer containing 300 mM imidazole, and was stored at –80°C. TRF1 was expressed and purified as previously described (Smucker and Turchi, 2001). Protein purity was confirmed by coomassie and silver stain (Figure

16B-C).

Expression constructs, in vitro transcription and translation (IVTT), and interaction site mapping by in vitro immunoprecipitation. IVTT was performed according to manufacturer instructions (Promega) with [35S]-methionine (Amersham). BLM-pET constructs used for IVTT (pET30A-BLM-C, pET24D-BLM-N, and pET24D-BLM-H) were generated and characterized as previously described (Langland et al., 2001). A template for IVTT synthesis of TRF1 was generated by PCR with a 5’ primer with a T7 promoter site. IVTT-generated proteins were verified for size by western analysis. In vitro immunoprecipitations with [35S]-methionine-labeled BLM segments (Figure 1E-F) were performed as previously described (Langland et al., 2001) using goat anti-TRF1 antibody

75 A. B. C.

Figure 16. Purified BLM (A), TRF2 (B), and TRF1 (C) analyzed by SDS-PAGE and staining of gel with Coomassie Brilliant Blue. Proteins were purified as described in methods. Mkr= Molecular weight marker.

76 (Santa Cruz, cat# sc-6165) and mouse anti-TRF2 antibody (Oncogene, Cat. OP129).

Normal goat IgG and normal mouse IgG were used as negative controls in the respective experiments.

In vitro immunoprecipitation of full length proteins. Purified BLM (500 ng) and/or

TRF2 (450 ng) were incubated in binding buffer (40 mM Tris (pH 8.0), 4 mM MgCl2, 0.1 mg/ml BSA, 5 mM DTT and 0.1% Nonidet-P40) for 1 h at 4°C, followed by an additional

1 h incubation with 2 µg of anti-TRF2 mouse monoclonal antibody (Upstate

Biotechnology, Lake Placid, NY). Equilibrated protein A-sepharose suspension was added to each sample, followed by incubation at 4°C for 1 h. Protein A-sepharose beads were collected by centrifugation and the supernatants removed for SDS-PAGE. After the pellet was washed three times with binding buffer, bound proteins were eluted by boiling the beads in binding buffer plus 2% SDS, 10% glycerol, 0.1% bromophenol blue

(0.1%), and 350 mM β-mercaptoethanol for 5 min. Proteins in the supernatant and pellet were separated by SDS-PAGE and detected by western blotting using rabbit anti-TRF2

(Santa Cruz Biotechnology, Santa Cruz, CA), rabbit anti-BLM (Novus Biologicals,

Littleton, CO) and donkey anti-rabbit-HRP conjugated antibodies followed by ECL detection (Amersham Pharmacia, Piscataway, NJ). In vitro immunoprecipitation with goat anti-TRF1 antibody (Santa Cruz, cat# sc-6165) was performed as above utilizing purified BLM (500 ng) and [35S]-labeled TRF1 made by IVTT (20 µL of a 50 µL IVTT reaction). BLM was detected by western blotting as described above. The portion of the

SDS-PAGE gel containing the [35S]-labeled TRF1 was dried and analyzed by autoradiography.

77 ELISA assay for detecting TRF1-BLM interaction. This assay was performed using standard techniques (Brosh et al., 2000; Opresko et al., 2002) with the following modifications. Briefly, purified recombinant BLM was diluted to 2.5 ng/µL in carbonate binding-buffer (0.016 M Na2CO3, 0.034 M NaHCO3 (pH 9.6)). 50 µL (125 ng) was added to individual wells of a 96-well microtiter plate and allowed to bind overnight at 4°C. In the same manner, BSA was bound to wells as a negative control, and TRF1 (0.25 ng/µL) was bound as a positive control. Plates were washed three times with PBST (PBS with

0.5% Tween 20) and blocked in blocking buffer (PBS with 0.5% Tween 20 and 3% BSA) for 2 h at 30°C. Plates were washed once more in blocking buffer prior to addition of

TRF1 (varying concentrations 0.25 to 1.5 ng/µL) diluted in binding buffer (50 mM Tris-

HCl (pH 7.4), 5 mM MgCl2, 5 mM ATP, 100 µg/ml BSA, 50 NaCl). TRF1 was incubated in plates for 1 h at 30°C. Where indicated, 10 units DNase I (Roche cat#766 785) were added to wells with TRF1. Plates were vigorously washed with wash buffer five times to remove unbound TRF1 from wells. TRF1 (bound fraction) was detected using anti-

TRF1 antibody (Santa Cruz, cat #sc-6165) which was diluted 1:1000 in blocking buffer and incubated at room temperature for 1 h. Unbound primary antibody was removed by three washes with wash buffer. Horseradish peroxidase-conjugated rabbit anti-goat antibody (Calbiochem, cat# 4015014) diluted 1:10000 was added to each well and incubated at room temperature for 1 h. Unbound secondary antibody was removed by three washes with wash buffer. Excess liquid was removed completely by tapping plates upside down before addition of 50 µL TACS-sapphire detection reagent

(Trevigen). After incubation for 10 min in the dark, reactions were stopped by addition

78 of an equal volume of 0.2 N HCl. Absorbance was read at 490 nm and all values were corrected based on background found in reactions containing no primary and secondary antibody.

DNA substrates. Telomeric D-loop substrate was prepared by a two-step method as described previously (Orren et al., 2002). Oligonucleotides were purchased from

Integrated DNA Technologies (Coralville, IA) and Operon (Alameda, CA).

Oligonucleotides (sequences in 5’ to 3’ orientation) used to generate the D-loop-like substrate were the partially complementary G80BUB21

(AGCTCCTAGGGTTACAAGCTTCACTAGGGTTGTCCAGTCACAGTCAGAGTCACA

GTCCTACACATGTAGGGTTGATCAGC) and C80

(GCTGATCAACCCTACATGTGTAGGTAACCCTAACCCTAACCCTAAGGACAACCC

TAGTGAAGCTTGTAACCCTAGGAGCT), as well as the A15G21 invading strand

(AAAAAAAAAAAAAAATTAGGGTTAGGGTTAGGGTTA). Either C80 or G80BUB21 was radiolabeled at its 5’ terminus with [γ-32P]ATP and T4 polynucleotide kinase prior to annealing. The telomeric 3’-overhang substrate was created by annealing two oligomers, a 38-mer

(T38:5’CCCTAACCCTAACCCTAACCCTAGAGGGAAAGGAAGAA3’) and a 68-mer

(T68:5’TTCTTCCTTTCCCTCTAGGGTTAGGGTTAGGGTTAGGGTTAGGGTTAGGGTT

AGGGTTAGGGTTAGGG3’). The non-telomeric 3’-overhang substrate was created by annealing a 38-mer

(NT38:5’ATGAGAAGCAGCCGTATCAGGAAGAGGGAAAGGAAGAA3’) to a 68-mer

(NT68:5’TTCTTCCTTTCCCTCTTCCTGATACGGCTGCTTCTCATCTACAACGTGATC

79 CGTCATTCGGAGTG3’). Before annealing, T38 and NT38 oligonucleotides were radiolabeled at its 5’ terminus with [γ-32P]ATP and T4 polynucleotide kinase using standard methods. Oligos were annealed by heating to 95°C for 3 min and slow cooling to 25°C. Annealed substrates were purified from a 4% agarose gel followed by ethanol precipitation.

DNase I footprinting. Telomeric D-loop substrate (2 fmol), labeled on the C80 oligomer, was incubated with BLM (7.5-168 nM) in BLM helicase buffer (10 µl) with

° ATPγS (2 mM) instead of ATP for 30 min at 4 C, followed by an additional 10 min

° incubation at 4 C with DNase I (0.7 unit). Reactions were stopped with an equal volume of formamide-containing loading buffer. DNA products were denatured by heat and separated by denaturing polyacrylamide (14%) gel electrophoresis, and visualized by phosphorimaging. Nucleotide size markers of 71, 60, 54, 23 and 18 nt were created by restriction enzyme digestion of the bubble substrate (C80 strand radiolabeled) as described previously (Machwe et al., 2002).

Helicase assays. DNA substrate (3.5 fmol) was incubated with BLM in helicase buffer

(final 1x buffer concentration of 20mM Tris-HCl pH 7.5, 2mM MgCl2, 0.1 mg/mL BSA

° and 1mM DTT) in 50 µL volume for 15 min at 37 C in the presence of ATP (2 mM). In reactions with both BLM and TRF1 or TRF2, substrate was first incubated either with

TRF1 or TRF2 for 5 min on ice followed by addition of BLM and an additional

° incubation at 37 C for 15 min in 50 µl of helicase buffer. Reactions were stopped with one-sixth volume of helicase stop dye (30% glycerol, 50 mM EDTA, 0.9% SDS, 0.25% bromophenol blue and 0.25% xylene cyanol). The DNA products were separated on

80 non-denaturing polyacrylamide (8%) gels and analyzed using a Storm 860

Phosphorimager and ImageQuant software. For telomeric D-loop substrate, unwinding was assessed by conversion of D-loop to bubble substrate caused by the displacement of the invading strand. For telomeric and non-telomeric 3’-overhang substrates, unwinding was assessed by conversion of duplexed substrate to single-stranded oligo.

For each reaction, percentage unwinding was determined by comparing the amount of bubble substrate or single-strand oligo produced to the total amount of substrate (D- loop plus bubble or duplexed substrate) in the reaction, after correcting for small amounts of (unannealed) substrate in untreated controls.

Gel shift assays. Gel shifts were performed with purified TRF1 and TRF2 as described by Li et al. (Li et al., 2000). Briefly, protein and DNA (3.5 fmol) were incubated in 1x gel shift buffer (4% Glycerol, 20 mM Glycine-KOH [pH 9.0] and 10 mM DTT per 50 µl) on ice for 15 min. 12 µl of each reaction was loaded onto a 0.6% agarose gel in 0.1x TBE and electrophoresed at 4°C. Gels were dried and analyzed using a Storm 860

Phosphorimager and ImageQuant software.

III. Results

TRF1 and TRF2 interact with BLM in vitro.

TRF1 and TRF2 are critical for telomere protection and telomere length control.

Each induces the formation of divergent higher order DNA structures from telomere sequences in vitro. TRF2 is postulated to protect telomere ends within T-loops, and can promote formation of looped telomeric structures in vitro (Griffith et al., 1999). While

81 TRF1 cannot mediate loop formation, it can promote bending of DNA and parallel pairing of telomere sequences (Bianchi et al., 1997; Griffith et al., 1998; Griffith et al.,

1999). In addition to co-localization with TRF2, BLM co-localizes in vivo with TRF1

(Stavropoulos et al., 2002; Yankiwski et al., 2000). Therefore, we tested whether there is a direct interaction between BLM and TRF1 in vitro using ELISA. (Figure 17A). BLM protein was bound to wells of microtiter plates, and after blocking, purified recombinant

TRF1 added in increasing concentrations. Bound TRF1 was detected using an anti-TRF1 antibody, HRP-conjugated secondary and colorimetric substrate. TRF1 bound to wells coated with BLM in a dose-dependent manner; DNase treatment did not interfere with this interaction. TRF1 did not bind significantly to wells coated with BSA, indicating that TRF1 binding is mediated by specific interaction with BLM. To confirm these data, in vitro immunoprecipitations were performed with purified recombinant BLM protein and [35S]-labeled TRF1 made by in vitro transcription and translation (IVTT) (Figure 17B).

Immunoprecipitations with anti-TRF1 antibody were analyzed by western blotting with anti-BLM antibody and autoradiography to detect [35S]-labeled TRF1. In the absence of

TRF1, no BLM is immunoprecipitated with the anti-TRF1 antibody. By contrast, a significant amount of BLM is immunoprecipitated when TRF1 is present in the reaction.

These results demonstrate that BLM and TRF1 interact directly.

Our in vivo experiments presented in Chapter Three indicate that TRF2 and BLM associate in cells using ALT. In order to test whether BLM and TRF2 interact directly, immunoprecipitations with purified recombinant BLM and TRF2 were performed

(Figure 17C). Immunoprecipitations with anti-TRF2 antibody were analyzed by western

82 blotting with anti-BLM and anti-TRF2 antibodies. Both pellet and supernatant from each reaction were analyzed, representing bound and unbound protein, respectively. In the absence of TRF2, the majority of BLM protein remains in the supernatant. By contrast, a significant amount of BLM is immunoprecipitated when TRF2 is present in the reaction. These results demonstrate that BLM and TRF2 interact directly and confirm previously published ELISA results (Opresko et al., 2002).

To identify the regions of BLM responsible for interaction with TRF1 and TRF2, three separate [35S]-labeled segments of BLM were synthesized by IVTT, representing the amino-terminus, the helicase domain and the carboxy-terminus (Figure 18A). [35S]- labeled TRF1 and BLM segments were incubated together, immunoprecipitated with anti-TRF1 antibody and analyzed by autoradiography (Figure 18B). The BLM helicase segment co-immunoprecipitates with TRF1. No protein was non-specifically immunoprecipitated with IgG or with anti-TRF1 when TRF1 was not added (data not shown). Likewise, BLM protein segments from IVTT and purified TRF2 were incubated together, immunoprecipitated with anti-TRF2 antibodies and analyzed by autoradiography (Figure 18C). The BLM C-terminal segment and the BLM helicase segment co-immunoprecipitate with TRF2. The interaction of TRF2 with the BLM C- terminal segment may be more significant than its interaction with the helicase segment, as judged by the efficiency of immunoprecipitation. No BLM segments were non-

83 A.

B.

C.

Figure 17. BLM interacts with TRF1 and TRF2 in vitro. A.) Interaction between BLM and TRF1 detected by ELISA. BLM (2.5 ng/mL) or BSA (2.5 ng/mL) was pre-bound to individual wells of a microtiter plate. TRF1 was incubated in wells and binding to BLM or BSA was detected using anti-TRF1 antibody and a colorimetric substrate (read at OD 450nm). Black bars represent average OD 450nm for wells with BLM and white bars represent values for wells with BSA at indicated concentrations. As a positive control for the antibody, TRF1 was coated directly onto the plate (gray bar). B.) BLM immunoprecipitates with TRF1 in vitro. Samples containing BLM and/or [35S]-labeled TRF1 were incubated, followed by addition of anti-TRF1 antibody and protein A sepharose beads. Immunoprecipitations were separated by SDS-PAGE and analyzed by western with anti-BLM antibody and autoradiography to visualize [35S]-labeled TRF1. The input lane represents approximately 25% of each protein added to the respective reactions. C.) Co- immunoprecipitation of BLM with TRF2 in vitro. Samples containing BLM and/or TRF2 were incubated, followed by addition of anti-TRF2 antibody and protein A sepharose beads. Supernatant and pellet represent unbound and bound proteins, respectively. Purified BLM and TRF2 were loaded on the gels as markers. Parallel gels were subjected to western blotting using anti-BLM antibody (top) and anti- TRF2 antibody (bottom).

84 A.

B.

C.

Figure 18. Mapping of BLM domains that mediate interaction with TRF1 and TRF2. A.) Graphical representation of the BLM segments made by in vitro transcription and translation. BLM N spans amino acids 1-587, BLM Hel spans amino acids 588-1000, and BLM C spans amino acids 996-1417. The helicase domain is highlighted in light gray, the RecQ C-terminal domain (RecQ) in dark gray, and RNaseD C-terminal (HRDC) domain and the nuclear localization signal (NLS) are labeled. B.) Mapping of the interaction between TRF1 and BLM in vitro. [35S]-labeled BLM segments (shown in 1D) and [35S]-labeled TRF1 were incubated together and immunoprecipitated using goat IgG (lane 5) or anti-TRF1 Ab (lane 6-8). Immunoprecipitations were separated by SDS- PAGE and analyzed by autoradiography. Individual proteins are as markers (lanes 1-4). C.) Mapping of the interaction between TRF2 and BLM in vitro. [35S]-labeled BLM segments and purified TRF2 were incubated together and immunoprecipitated using mouse IgG (lane 4) or anti-TRF2 Ab (lane 5-7). Immunoprecipitations were separated by SDS-PAGE and analyzed by autoradiography. Individual proteins are loaded as mar- kers (lanes 1-3).

85 specifically immunoprecipitated with IgG or with anti-TRF2 when TRF2 was not added

(data not shown). These results demonstrate direct interactions between BLM and TRF1 mediated by the BLM helicase domain, and between BLM and TRF2 mediated by the

BLM C-terminus and helicase domains.

BLM unwinds substrates that resemble native telomere conformations.

Telomeric DNA isolated from cells demonstrates the existence of a 3’-overhang of the G-rich strand. To investigate a role for BLM in maintaining such a structure, we constructed a DNA duplex that resembles a native linearized telomere with a 3’-single- stranded region composed of five 5’-TTAGGG-3’ repeats and a 38-bp duplexed region containing 23-bp of telomeric repeats for TRF1 and TRF2 binding (Bianchi et al., 1999).

(Figure 19A). BLM unwinds this substrate efficiently, reaching 60% unwinding of total substrate at the highest concentrations of BLM tested (Figure 19B, 19E).

In mammalian cells, telomeric 3’-overhangs may be folded back and inserted into the duplex telomeric sequences to form “T-loops” in order to protect chromosome ends from degradation and end-joining (Griffith et al., 1999; Stansel et al., 2001). At the site of insertion, the single-stranded overhang displaces the G-rich strand of the duplex to form a D-loop structure that is stabilized by TRF2 binding. Additionally, telomere D- loops may be present as intermediates in recombination during ALT (Lundblad, 2002).

T-loops must be resolved in order for chromosome ends to be replicated, thus helicases to unwind these structures are likely essential for proper DNA replication. To investigate a role for BLM in telomere loop metabolism, a model substrate was created

86 in vitro that approximates a D-loop (Figure 19C). Our substrate was constructed by annealing partially complementary oligomers to form a duplex with bubble, then annealing a third oligomer complementary to the bubble. This substrate contains a 21- bp stretch of perfect telomeric repeats within the D-loop and imperfect telomeric repeat sequences on both duplex arms for TRF1 and TRF2 binding (Bianchi et al., 1999). DNase

I footprinting analysis demonstrates that BLM protects a small area on the C-rich strand within the D-loop region near the point where the invading strand exits the duplex; this suggests that BLM binds specifically to the D-loop structure (Figure 20). Helicase assays demonstrate that BLM efficiently displaces the invading strand of the telomeric D-loop substrate and converts it to a “bubble-containing” duplex (Figure 19D, 19E). This displacement occurs in a dose-dependent manner and is nearly complete at a BLM concentration of 10 nM.

TRF1 and TRF2 oppositely regulate BLM activity on telomeric substrates.

Since both TRF1 and TRF2 interact directly with BLM, we tested whether these proteins affect BLM unwinding activity on telomeric substrates. Substrates were pre- incubated with varying amounts of TRF1 or TRF2 before starting the helicase reaction with a fixed concentration of BLM. In a dose-dependent manner, TRF2 stimulated BLM unwinding of the telomeric 3’-overhang at concentrations between 4-20 nM (Figure 21A,

21C) and the telomere D-loop at concentrations between 2.5-10 nM (Figure 21D, 21F).

Conversely, TRF1 inhibited BLM unwinding of the telomeric 3’-overhang at concentrations between 6-60 nM (Figure 21B, 21C) and the telomere D-loop at

87 A. B.

C. D.

E.

Figure 19. BLM helicase unwinds telomere-like substrates. A.) Sequence and structure of the telomeric 3’-overhang substrate. B.) BLM unwinds the telomeric 3’-overhang structure. BLM (0, 1, 2, 4, 6, 8, 10 nM; lanes 1-7) was incubated with the telomeric 3’- overhang substrate and the DNA products resolved by non-denaturing polyacrylamide gel electrophoresis. Unwinding is evidenced by the conversion of duplexed substrate to faster migrating single-stranded oligo. Reaction with BLM, but no ATP is shown in lane 8. The migration of substrates is shown on the left by drawings. Quantitation is shown in 19E. C.) Structure and sequence of the telomeric D-loop substrate. D.) BLM displaces the invading strand of the telomeric D-loop substrate. Reactions were performed with BLM (0, 0.12, 0.25, 0.62, 1.25, 2.5, 5, 10 nM; lanes 1-8) as described in 4B. Displacement of the invading strand (unwinding) is evidenced by the conversion of the D-loop to a faster migrating bubble substrate. Bubble substrate without the inserted strand was loaded as a marker (lane 9). The migration of substrates are shown on the left by drawings. Quantitation is shown in 19E. E.) Quantitation of BLM unwinding of telomere substrates. Unwinding of 3’-overhang substrate (dotted line) was calculated by comparing the amount of single-stranded substrate produced to the total amount of the substrate in the reaction, with correction for any unannealed substrate in untreated controls. Unwinding of D-loop substrate (solid line) was calculated by comparing the amount of bubble substrate produced to the total amount of the substrate in the reaction, with correction for small amounts of unannealed bubble substrate in untreated controls.

88

Figure 20. BLM binding to the telomeric D-loop substrate demonstrated by DNase I footprinting. A) Telomeric D-loop substrate labeled on the C-rich strand was incubated with BLM (0, 7.5, 21, 42, 84, or 168 nM) followed by DNase I. DNA products were resolved by denaturing gel electrophoresis. The position of the D-loop region and the BLM footprint are defined by curved and right-angle brackets, respectively. Nucleotide size markers are shown at left. B) Diagram of telomeric-loops substrate indicating region of BLM binding determined in A.

89

Figure 21. Effect of TRF1 and TRF2 on BLM unwinding of telomere substrates. A.) TRF2 timulates unwinding of telomeric 3’-overhang substrate by BLM. TRF2 (0, 4, 8, 20, 40 nM; lanes 1-5) was pre-incubated with DNA substrate followed by addition of BLM (4 nM). Control reaction with TRF2 alone is shown in lane 6. Quantitation is shown in 20C. B.) TRF1 inhibits unwinding of telomeric 3’-overhang substrate by BLM. As with TRF2, TRF1 (0, 6, 12, 24, 60 nM; lanes 1-5) was pre-incubated with DNA substrate followed by addition of BLM (6 nM). Control reaction with TRF1 alone is shown in lane 6. Quantitation is shown in 20C. C.) BLM unwinding in the presence of TRF1 (dotted line) and TRF2 (solid line) derived from data shown in A and B, quantitated as described in 19E. D.) TRF2 stimulates disruption of telomeric D-loop by BLM. Reactions with TRF2 (0, 1.25, 2.5, 5, 10, 15 nM; lanes 2-7) and BLM (1nM) were performed in the same manner as telomeric 3’-overhang. Control reaction with TRF2 alone is shown in lane 8. Quantitation is shown in 20F. E.) TRF1 inhibits disruption of telomeric D-loop by BLM. Reactions with TRF1 (0, 3, 6, 15 nM; lanes 2-5) and BLM (1 nM) were performed in the same manner as telomeric 3’-overhang substrate. Quantitation is shown in 20F. F.) BLM unwinding in the presence of TRF1 (dotted line) and TRF2 (solid line) were derived from data shown in 5D and 5E, and quantitated as described in 19E.

90

Figure 22. TRF1 and TRF2 do not regulate UvrD helicase activity. UvrD (6 nM) was incubated with telomeric 3’-overhang substrate in the presence of TRF2 (0, 6, 12, 24, 48 nM; lanes 1-5) or TRF1 (0, 6, 12, 24, 48 nM; lanes 7-11). Reactions were analyzed as described in Methods. Quantitation of unwinding was determined as described in 19E.

91 concentrations between 3-15 nM (Figure 21E, 21F). When these experiments were performed utilizing the 3’-5’ helicase UvrD, TRF1 and TRF2 had no significant effect on unwinding at concentrations between 6-48 nM (Figure 22). Thus, TRF1 and TRF2 effects on unwinding are not universally applicable to DNA helicases.

TRF2 but not TRF1 affects BLM unwinding of a non-telomeric 3’-overhang.

We next determined whether the regulation of BLM unwinding by TRF1 and

TRF2 was related to the ability of TRF1 or TRF2 to bind stably to DNA-containing telomeric repeats. A non-telomeric 3’-overhang substrate with a 38-bp duplexed region was generated which is the same length and structure as the telomeric 3’-overhang substrate (see Figure 19 and 21), but without telomeric repeats. BLM unwinds the 3’- overhang substrate in a dose-dependent manner at concentrations between 1 and 10 nM

(Figure 23A). The non-telomeric and telomeric 3’-overhang substrates are unwound with similar efficiency, reaching near 60% unwinding of total substrate at the highest concentrations of BLM tested (Figure 23B). Gel shift assays confirmed that TRF1 and

TRF2 bind stably to the telomeric substrate, but not to non-telomeric substrate (Figure

24). We next tested the ability of BLM to unwind the non-telomeric 3’-overhang substrate in the presence of TRF1 and TRF2. When TRF2 is added to helicase reactions with non-telomeric substrate, TRF2 (at concentrations between 4-40 nM) stimulates BLM unwinding of this substrate (Figure 25A, 25C). In contrast, TRF1 (at concentrations between 6-60 nM) has no effect on BLM unwinding of this substrate (Figure 25B, 25C).

92 A.

B.

Figure 23. BLM unwinds the non-telomeric 3’-overhang substrate. A.) Reactions were performed with BLM (0, 1, 2, 4, 6, 8, 10 nM; lanes 1-7). Reaction with BLM, but no ATP is shown in lane 8. The migration of substrates is shown on the left by drawings. B.) BLM unwinding of non-telomeric 3’-overhang substrate, calculated as described in Figure 19E. Unwinding of non-telomeric 3’-overhang is represented by a solid line. Unwinding of the telomeric 3’-overhang is shown for comparison (data derived from Figure 19B), represented by a dashed line.

93 Figure 24. Gel shift analysis demonstrating binding of TRF1 and TRF2 to telomeric 3’- overhang substrate (upper panel) and lack of binding to non-telomeric 3’-overhang substrate (lower panel). Positive control for gel shift is Hela nuclear extract (NE) shown in first lane of each gel.

94 A.

B.

C.

Figure 25. Effects of TRF1 and TRF2 on unwinding of non-telomeric 3’-overhang by BLM. A) TRF2 stimulates unwinding of non-telomeric substrate by BLM. TRF2 (0, 4, 8, 20, 40 nM; lanes 1-5) was pre-incubated with DNA substrate followed by addition of BLM (4 nM). Reactions with TRF2 only and no protein are shown in lanes 6 and 7, respectively. B.) TRF1 has no effect on unwinding of non-telomeric substrate by BLM. TRF1 (0, 6, 12, 30 and 60 nM; lanes 1-5) was pre-incubated with DNA substrate followed by addition of BLM (6 nM). Reactions with TRF1 only and no protein are shown in lanes 6 and 7, respectively. C.) BLM unwinding in the presence of TRF1 (dotted line) and TRF2 (solid line) derived from data shown in A and B, quantitated as described in 19E.

95 These results indicate that TRF1 inhibition of BLM unwinding on telomeric substrate is due to specific binding to telomere repeat sequences. On the other hand, TRF2 stimulation of BLM unwinding is not specific for telomeric substrates, suggesting that stable binding to DNA is not required for this stimulatory effect.

Effects of TRF1 and TRF2 in combination on BLM helicase activity vary with relative concentrations.

Our results indicate that TRF1 and TRF2 act in opposition to one another to regulate BLM helicase activity on telomere substrates. To address the mechanism coordinating BLM activity, we performed helicase assays using the telomeric 3’- overhang substrate with both TRF1 and TRF2 in varying molar ratios (Figure 26). When equimolar concentrations of TRF1 and TRF2 (4 nM each) are pre-incubated with substrate, TRF2 stimulates BLM unwinding. TRF2 stimulates BLM unwinding when it is present in excess of TRF1. However, when TRF1 is present in excess of TRF2, the stimulatory effect of TRF2 on BLM unwinding is inhibited. These data argue that regulation of BLM unwinding by TRF1 and TRF2 is dependent on the relative molar ratio of these proteins at the telomere.

IV. Discussion

The co-localization of BLM with TRF1 and TRF2, reported here and elsewhere

(Yankiwski et al., 2000; Stavropoulos et al., 2002), prompted us to examine whether these proteins interact directly and the effects of such interactions on BLM activity. We find that BLM interacts directly with TRF2 in vitro in agreement with a previous report

96

Figure 26. Effect of combined TRF1 and TRF2 on BLM unwinding of telomeric substrate. Helicase assays were performed using the telomeric 3’-overhang substrate. TRF1 and/or TRF2 was pre-incubated with DNA substrate followed by addition of BLM (4 nM fixed). Reaction are as follows: Lane 1: BLM only, Lane 2: BLM and TRF2 (4nM), Lane 3: BLM and TRF1 (4 nM), Lane 4: BLM, TRF1 (4nM) and TRF2 (4 nM), Lane 5: BLM, TRF1 (4nM) and TRF2 (8nM), Lane 6: BLM, TRF1 (4nM) and TRF2 (12nM), Lane 7: BLM, TRF2 (4nM) and TRF1 (8nM), Lane 8: BLM, TRF2 (4nM) and TRF1 (12 nM), Lane 9: no protein control. Percent unwinding is calculated as described in 19E; graph bars represent average percent unwinding and error bars represent standard deviation for three sepa- rate experiments.

97 (Opresko et al., 2002; Stavropoulos et al., 2002). Our experiments have mapped this interaction primarily to the BLM C-terminus and secondarily to the helicase domain.

Interestingly, TRF2 interacts with another RecQ helicase, WRN, through a portion of the

WRN helicase and RecQ C-terminal domains, protein regions conserved between BLM and WRN (Opresko et al., 2002).

In this chapter, we have characterized the binding and unwinding of telomeric substrates in vitro. Two telomeric substrates were created, a 3’-overhang and a telomeric

D-loop. Helicase assays demonstrated efficient unwinding of both telomeric substrates by BLM. Additionally, TRF2 stimulates BLM unwinding of both telomeric substrates.

TRF2 also stimulates BLM unwinding of a non-telomeric substrate, suggesting that physical interaction between BLM and TRF2 primarily mediates this stimulatory effect.

Importantly, TRF2 binds stably to both telomere substrates indicating that stable DNA binding, although not required, does not hinder the ability of TRF2 to stimulate BLM activity. This result is significant since TRF2 is bound to telomeric DNA in vivo.

Additionally, we find that TRF2 stimulates BLM unwinding of substrates with a short duplex (21-bp; telomeric D-loop), as well as a long duplex (38-bp; 3’-overhang substrates), suggesting that TRF2 stimulates BLM enzymatic activity and processivity.

A dual mechanism of stimulation by TRF2 may be required to promote BLM unwinding of long telomeric DNA tracts, such as those found in cells using ALT. These findings contrast with those of Opresko et al. (2002) who reports that TRF2 stimulates BLM unwinding of short duplexes (22-bp) that do not bind TRF2, but does not stimulate BLM unwinding of long duplexes (34-bp) that bind TRF2. Our substrates were designed to

98 more approximate more closely native telomere conformations and are distinct from the forked substrates used by Opresko et al. Therefore, substrate composition and structure may be a critical factor for analyzing TRF2 stimulation of BLM activity.

We demonstrate an additional direct interaction between BLM and TRF1 mediated specifically by the BLM helicase domain. In contrast to TRF2, we find that

TRF1 inhibits BLM unwinding of two types of telomeric substrates but does not affect

BLM unwinding of a non-telomeric substrate. These data suggest that TRF1 binding to telomeric DNA is necessary for its inhibition of BLM helicase activity. TRF1 binding to telomeric substrate may make the DNA more difficult to unwind; however, TRF1 does not inhibit unwinding of telomeric substrates by the UvrD helicase. Taken together, these results suggest that TRF1 binding to both BLM and telomeric DNA might sterically hinder the ability of BLM to unwind telomeric substrates.

Finally, our data demonstrate that TRF2 stimulates BLM unwinding in the presence of an equimolar concentration TRF1 and suggest that TRF2 is the dominant regulator of BLM activity. However, the stimulatory effect of TRF2 on BLM unwinding is blocked when TRF1 is present in excess. Without TRF2, BLM may be unable to traverse efficiently and unwind long, GC-rich telomeric tracts. Therefore, regulation of

BLM activity at the telomere may be dependent on the relative concentrations of TRF1 and TRF2. It remains to be determined how TRF1 and TRF2 might coordinate BLM activity at telomeres in vivo. One mechanism is suggested by the recent finding that poly

(ADP)-ribosylation of TRF1 by tankyrase directs ubiquitin-dependent degradation of

TRF1 and telomere lengthening (Smith et al., 2000; Chang et al., 2003). Degradation of

99 TRF1 would then relieve TRF1-mediated inhibition and allow stimulation of BLM by

TRF2 during telomere lengthening when BLM activity might be required.

Note in proof: data presented in this Chapter are published (Lillard-Wetherell et al.,

2004).

100 Chapter Five. Thesis Summary

Somatic cells from persons with the inherited chromosome breakage syndrome

Bloom syndrome (BS) feature excessive chromosome breakage and intra-and inter- chromosomal homologous exchanges (German, 1993). The gene mutated in BS, BLM, encodes a RecQ-like ATP-dependent 3’-to-5’ helicase that presumably functions in some types of DNA transactions (Ellis et al., 1995). As the absence of BLM is associated with excessive recombination (German et al., 1977), BLM is thought to suppress recombination and/or resolve recombination intermediates. In support of the latter,

BLM promotes branch migration of Holliday junctions (HJs), resolves D-loops and unwinds G-quadruplex DNA in vitro (Karow et al., 2000; Mohaghegh et al., 2001; Sun et al., 1998; van Brabant et al., 2000).

An increase in telomeric associations (TAs) between homologous chromosome arms and anaphase/telophase bridges first suggested that BLM may be required to maintain telomere protection (German and Schonberg, unpublished data; German and

Crippa, 1966). Later studies revealed hyper-variability of telomere length in clonal BS cell lines which may be a cause or consequence of TAs (Schawalder et al., 2003). In

Chapter Three, FISH analysis confirmed the presence of TAs between homologous chromosomes in BS fibroblasts and demonstrated the presence of telomeric sequences at the junction of TAs. These cytogenetic data confirm that TAs are elevated significantly in primary and transformed BS cells. The presence of telomeric sequences at TAs suggested a function for BLM in maintaining telomere protection, perhaps by facilitating

3’-end processing and formation of the protective T-loop. The latter function has also

101 been ascribed to the RMN complex via its association with TRF2 (Zhu et al., 2000).

Alternatively, BLM may be required to resolve recombinant or entangled ends that occur when chromosome ends are exposed, such as during replication. If the latter is true, we hypothesize that BLM would be particular important for maintaining telomere length and structure under circumstances in which telomere recombination is upregulated, such as in cells using alternative lengthening of telomeres (ALT).

ALT was first discovered in human cells when a small percentage of transformed, immortalized cell lines were found without telomerase activity (Bryan et al., 1997; Bryan et al., 1995; Dunham et al., 2000; Lundblad and Blackburn, 1993).

Telomeres in these cells were hyper-variable in length and were frequently greater than

50 kb (Bryan et al., 1995). Homologous recombination was increased at telomeres in cells using ALT and these cells displayed the unique ability to transfer tagged telomeric sequences between chromosomes (Bailey et al., 2004; Bechter et al., 2004; Bechter et al.,

2003; Dunham et al., 2000; Londono-Vallejo et al., 2004) ALT may occur within variant

PML bodies termed ALT-associated PML bodies (APBs) in which telomeric DNA, telomeric proteins (TRF1 and TRF2), and homologous recombination proteins (RMN complex, RAD51, RAD52 and RPA) reside (Yeager et al., 1999). APBs are enriched and undergo BrdU incorporation during G2/M suggesting that telomeric synthesis occurs within APBs after replication of the majority of the genome is complete (Grobelny et al.,

2000; Wu et al., 2000).

Immunofluorescence was used to determine whether BLM associated with telomeres in cells using ALT, telomerase, and in primary cells with neither pathway

102 activated. By immunofluorescence, BLM associated with telomeres and TRF2 in cells using ALT, but not in telomerase-positive or primary cell lines. BLM and TRF2 also co- localized with PML, confirming that these proteins resided within APBs. Additionally,

BLM associates with telomeres and TRF2 in cells using ALT. These data agree with the hypothesis that BLM is required to resolve telomere associations under circumstances in which recombination is upregulated. Importantly, BLM and TRF2 also co-localize with foci of DNA synthesis during late S and G2/M, establishing a spatial and temporal association of BLM and TRF2 with telomere synthesis in cells using ALT. We were unable to discern changes in telomere length after knock-down of BLM with antisense morpholinos. Over-expression of BLM in cells using ALT has been demonstrated by others to facilitate rapid extension of telomeres, their work suggested that BLM promotes telomere synthesis by ALT (Stavropoulos et al., 2002). Our failure to demonstrate a change in telomere length may be attributed to the limited sensitivity of our telomere length assay. Additionally, the small amounts of BLM remaining after antisense treatment may be sufficient to support ALT. Lastly, other helicases, particularly WRN may have overlapping functions with BLM and may promote ALT when BLM is low or absent. Importantly, WRN interacts with TRF2 and localizes to

APBs indicating a function in ALT (Opresko et al., 2002).

To bypass these limitations, we used S. cerevisiae mutants to investigate BLM function in ALT. Sgs1, the sole RecQ-like helicase in S. cerevisiae, is required for type II,

Rad50-dependent telomere lengthening pathway in telomerase-negative yeast; sgs1 mutants utilize only the type I, Rad51-dependent telomere lengthening pathway (Cohen

103 and Sinclair, 2001; Huang et al., 2001; Johnson et al., 2001). Our aim was to determine whether the functions of BLM and Sgs1 are conserved with respect to recombination- mediated telomere lengthening. sgs1est2 mutant yeast were created by mating and tetrad dissection. EST2 encodes an essential catalytic subunit of telomerase; therefore, est2 mutants cannot utilize telomerase to extend telomeres. BLM expression in pre-crisis sgs1est2 mutants resulted in the appearance of type II survivors indicating that BLM rescues the type II pathway in the absence of Sgs1. Conversely, expression of K695E

BLM, encoding a “helicase-dead” BLM protein, could not rescue type II recombination.

These data indicate that BLM and Sgs1 have similar biochemical functions in ALT and perhaps recognize and resolve similar DNA structures in yeast and humans.

Both yeast and human ALT may occur by BIR. Therefore, we propose putative function for BLM (and Sgs1) in ALT using a BIR model for telomere lengthening in

Chapter Three. Based on biochemical preferences, BLM may 1) unwind the D-loop portion of the T-loop, 2) process telomeric ends to facilitate strand invasion (possibly in conjunction with the RMN complex), 3) facilitate formation of a D-loop that is accessible for replication, 4) facilitate replication and resolution of the single Holliday Junction by branch migration, and/or 5) resolve G-quadruplexes that may disrupt replication. Sgs1 is required for efficient replication of the G-rich rDNA repeats (Versini et al., 2003), and both Sgs1 and BLM are important for maintaining rDNA integrity (Heo et al., 1999;

Schawalder et al., 2003; Sinclair and Guarente, 1997). Therefore, the requirement for these helicases in ALT may reflect the difficulties associated with replication of the G- rich, telomeric repeats.

104 In Chapter Three, the association of BLM with TRF2 in cells using ALT was established by co-localization and immunoprecipitation. Other groups have also reported co-localization of BLM with TRF1 in cells using ALT. The aim of Chapter Four was to determine if these proteins associate with BLM directly and/or regulate BLM helicase activity. Co-immunoprecipitation of BLM with both TRF1 and TRF2 in vitro indicated a physical association of BLM with each of these proteins. Next, we demonstrated the ability of BLM to unwind two telomeric substrates designed to resemble native telomere structures, a 3’-overhang and a telomeric D-loop. The effects of TRF1 and TRF2 on BLM unwinding were tested using telomeric substrates. While

TRF2 stimulated BLM activity, TRF1 inhibited unwinding of both telomeric substrates.

Stimulation by TRF2 may occur by physical interaction, as it is also able to stimulate

BLM unwinding of non-telomeric substrates. Importantly, TRF2 stimulates unwinding of both long and short duplexes indicating TRF2 promotes both BLM helicase activity and processivity. We propose that this dual stimulation may be required to allow BLM to unwind long, GC-rich telomeric sequences.

In contrast to TRF2, TRF1 does not regulate BLM unwinding of non-telomeric substrates. TRF1 binding may directly inhibit BLM or induce the formation of a secondary structure that blocks BLM activity. TRF1 does not inhibit activity of another

3’-5’ helicase, UvrD. This suggests that TRF1 inhibits BLM by dually interacting with

BLM and telomeric DNA, perhaps sterically blocking access of BLM to the substrate.

Importantly, when both TRF1 and TRF2 are present, TRF1 can block stimulation of BLM

105 by TRF2. The regulation of BLM at telomeres may then be coordinated by the relative quantity of TRF1 and TRF2.

Using data presented in Chapter Three and Four, we propose a model for the regulation of BLM by TRF1 and TRF2 at telomeres in cells using ALT (Figure 27). In telomerase-positive cells, it is well established that TRF1 inhibits telomere lengthening, perhaps by blocking telomerase access to the telomere (van Steensel and de Lange,

1997). TRF1 may additionally block access by replicative polymerases during C-strand synthesis (Smucker and Turchi, 2001). Recent reports find that poly (ADP)-ribosylation of TRF1 by tankyrase directs ubiquitin-dependent degradation of TRF1 and telomere lengthening (Chang et al., 2003; Smith and de Lange, 2000). We suggest that BLM activity is inhibited at the telomere by TRF1 when telomere replication is not occurring.

Degradation of TRF1 would relieve this inhibition and allow stimulation of BLM by

TRF2 specifically during telomere synthesis. DNase footprinting indicates that BLM binds to the junction of our telomeric D-loop substrate, near the point where the invading strand enters the “bubble.” Intriguingly, TRF2 binds to the junction of the T- loop as well (Griffith et al., 1999), placing it in a convenient orientation to interact with

BLM and facilitate unwinding of the telomeric loop during ALT. TRF2 additionally binds to a telomeric duplexes, near the point where the 3’-overhang begins. Thus, TRF2 may function with BLM (and perhaps the RMN complex), to process the 3’-overhang to facilitate D-loop formation. TRF2 promotes D-loop formation between telomere repeats on separate DNA strands in vitro (Stansel et al., 2001); therefore, BLM and TRF2 may cooperate to promote D-loop formation, replication and branch migration. After

106 telomere synthesis is complete, the telomere is remodeled into a T-loop and TRF1 repression is re-established.

In conclusion, this work has established a functional link between BLM and telomere synthesis in cells using ALT, demonstrated a conserved function for BLM and

Sgs1 in telomere lengthening, and identified two telomeric proteins that interact with and regulate BLM activity in vitro. Based on these data, two models have been presented outlining putative function for BLM in ALT and the coordinated regulation of

BLM at telomeres during ALT by TRF1 and TRF2. Future studies will be designed to test the function and regulation of BLM in ALT and to better elucidate the mechanism by which ALT occurs. First, we will investigate the regulation of ALT by TRF1 and

TRF2. Secondly, we will investigate whether interaction of BLM with TRF1 or TRF2 is necessary for BLMs function in ALT and/or the localization or stability of BLM at telomeres. Thirdly, we will determine what proteins associate with BLM and TRF2 in cells using ALT and determine whether these proteins interact with and/or regulate

BLM activity in vitro. Finally, while several models have been proposed, the precise mechanism by which human ALT occurs remains unclear. Of particular interest, is the accumulation of extrachromosomal telomeric circles in cells using ALT and the possibility that these circular DNAs could participate in rolling circle replication of new telomeric repeats. This model is particularly intriguing as it may explain the rapid and extensive elongations which occur in cells using ALT. Additionally, Rad51, Rad52 and

Rad50 function in telomere lengthening in yeast; however, little is known regarding the function or requirement of these proteins in human ALT beyond their association with

107 APBs. Investigations into the requirement of these protein factors in human ALT may help to elucidate mechanistic pathways by which ALT occurs. Future studies will be directed to address these basic questions such that the function(s) for BLM in ALT can be more precisely defined.

108

TRF1 inhibition TRF1 degradation TRF2 stimulation X X X X (A) (B)

(C) & END PROCESSING

(D)

TRF1

TRF2

BLM

Figure 27. Model for the coordinated regulation of BLM by TRF1 and TRF2 at telomeres in cells using ALT. This model is based on the proposed functions for BLM in ALT, as presented in Figure 15. Presence of TRF1 at telomeres may prevent BLM unwinding activity of the T-loop, thus blocking telomere replication (A). Degradation of TRF1 by poly (ADP)-ribosylation and ubiquitin-dependent degradation relieves repression of BLM (B). In the absence of TRF1, TRF2 stimulates BLM to unwind the T-loop, promote D-loop formation, replication and branch migration during ALT (C) After telomere synthesis, the telomere is remodeled and TRF1 repression is re-established.

109 CHAPTER SIX. Bibliography

Ababou, M., Dumaire, V., Lecluse, Y., and Amor-Gueret, M. (2002). Bloom's syndrome protein response to ultraviolet-C radiation and hydroxyurea-mediated DNA synthesis inhibition. Oncogene 21, 2079-88.

Ababou, M., Dutertre, S., Lecluse, Y., Onclercq, R., Chatton, B., and Amor-Gueret, M. (2000). ATM-dependent phosphorylation and accumulation of endogenous BLM protein in response to ionizing radiation. Oncogene 19, 5955-63.

Adams, M. D., McVey, M., and Sekelsky, J. J. (2003). Drosophila BLM in double-strand break repair by synthesis-dependent strand annealing. Science 299, 265-7.

Ancelin, K., Brunori, M., Bauwens, S., Koering, C. E., Brun, C., Ricoul, M., Pommier, J. P., Sabatier, L., and Gilson, E. (2002). Targeting assay to study the cis functions of human telomeric proteins: evidence for inhibition of telomerase by TRF1 and for activation of telomere degradation by TRF2. Mol. Cell Biol. 22, 3474-87.

Aurias, A., Antoine, J. L., Assathiany, R., Odievre, M., and Dutrillaux, B. (1985). Radiation sensitivity of Bloom's syndrome lymphocytes during S and G2 phases. Cancer. Genet. Cytogenet. 16, 131-6.

Austin, C. A., Marsh, K. L., Wasserman, R. A., Willmore, E., Sayer, P. J., Wang, J. C., and Fisher, L. M. (1995). Expression, domain structure, and enzymatic properties of an active recombinant human DNA topoisomerase II beta. J. Biol. Chem. 270, 15739-46.

Bailey, S. M., Brenneman, M. A., and Goodwin, E. H. (2004). Frequent recombination in telomeric DNA may extend the proliferative life of telomerase-negative cells. Nucleic Acids Res. 32, 3743-51.

Bachrati, C.Z. and Hickson, I.D. (2003). RecQ helicases: suppressors of tumorigenesis and premature ageing. Biochem. J. 374, 577-606.

Baro, P. R., Bastart, F. M., Bartrina, J. R., Mateo, J. M., and Vidal, M. T. (1989). Case report 529: Osteosarcoma of calcaneus with Rothmund-Thompson syndrome (RTS). Skeletal Radiol. 18, 136-9.

Beadle, G. F., Mackay, I. R., Whittingham, S., Taggart, G., Harris, A. W., and Harrison, L. C. (1978). Werner's syndrome: a model of premature aging? J. Med. 9, 377-404.

Beamish, H., Kedar, P., Kaneko, H., Chen, P., Fukao, T., Peng, C., Beresten, S., Gueven, N., Purdie, D., Lees-Miller, S., Ellis, N., Kondo, N., and Lavin, M. F. (2002). Functional

110 link between BLM defective in Bloom's syndrome and the ataxia-telangiectasia mutated protein, ATM. J. Biol. Chem. 28, 28.

Bechter, O. E., Shay, J. W., and Wright, W. E. (2004). The frequency of homologous recombination in human ALT cells. Cell Cycle 3, 547-9.

Bechter, O. E., Zou, Y., Shay, J. W., and Wright, W. E. (2003). Homologous recombination in human telomerase-positive and ALT cells occurs with the same frequency. EMBO Rep. 4, 1138-43.

Bennett, R. J., Noirot-Gros, M. F., and Wang, J. C. (2000). Interaction between yeast sgs1 helicase and DNA topoisomerase III. J. Biol. Chem. 275, 26898-905.

Bennett, R. J., Sharp, J. A., and Wang, J. C. (1998). Purification and characterization of the Sgs1 DNA helicase activity of Saccharomyces cerevisiae. J. Biol.Chem. 273, 9644-50.

Bennett, R. J., and Wang, J. C. (2001). Association of yeast DNA topoisomerase III and Sgs1 DNA helicase: studies of fusion proteins. Proc. Natl. Acad. Sci. U S A 98, 11108-13.

Bianchi, A., Smith, S., Chong, L., Elias, P., and de Lange, T. (1997). TRF1 is a dimer and bends telomeric DNA. EMBO J. 16, 1785-94.

Bianchi, A., Stansel, R. M., Fairall, L., Griffith, J. D., Rhodes, D., and de Lange, T. (1999). TRF1 binds a bipartite telomeric site with extreme spatial flexibility. EMBO J. 18, 5735- 44.

Bischof, O., Galande, S., Farzaneh, F., Kohwi-Shigematsu, T., and Campisi, J. (2001). Selective cleavage of BLM, the bloom syndrome protein, during apoptotic cell death. J. Biol. Chem. 276, 12068-75.

Bischof, O., Kim, S. H., Irving, J., Beresten, S., Ellis, N. A., and Campisi, J. (2001). Regulation and localization of the Bloom syndrome protein in response to DNA damage. J. Cell Biol. 153, 367-80.

Blasco, M. A., Lee, H. W., M.P., H., Samper, E., Lansdorp, P. M., DePinho, R. A., and Greider, C. W. (1997). Telomere shortening and tumor formation by mouse cells lacking telomerase RNA. Cell 91, 25-34.

Bloom, D. (1954). Congenital telangiectatic erythema resembling lupus erythematosus in dwarfs. Am. J. Dis. Child. 88, 754.

111 Braybrooke, J. P., Li, J. L., Wu, L., Caple, F., Benson, F. E., and Hickson, I. D. (2003). Functional interaction between the Bloom's syndrome helicase and the RAD51 paralog, RAD51L3 (RAD51D). J. Biol. Chem. 278, 48357-66.

Broccoli, D., Smogorzewska, A., Chong, L., and de Lange, T. (1997). Human telomeres contain two distinct Myb-related proteins, TRF1 and TRF2. Nat. Genet. 17, 231-5.

Brosh, R. M., Jr., Li, J. L., Kenny, M. K., Karow, J. K., Cooper, M. P., Kureekattil, R. P., Hickson, I. D., and Bohr, V. A. (2000). Replication protein A physically interacts with the Bloom's syndrome protein and stimulates its helicase activity. J. Biol. Chem. 275, 23500- 8.

Bryan, T. M., Englezou, A., Dalla-Pozza, L., Dunham, M. A., and Reddel, R. R. (1997). Evidence for an alternative mechanism for maintaining telomere length in human tumors and tumor-derived cell lines. Nat. Med. 3, 1271-4.

Bryan, T. M., Englezou, A., Gupta, J., Bacchetti, S., and Reddel, R. R. (1995). Telomere elongation in immortal human cells without detectable telomerase activity. EMBO J. 14, 4240-8.

Chaganti, R. S., Schonberg, S., and German, J. (1974). A manyfold increase in sister chromatid exchanges in Bloom's syndrome lymphocytes. Proc. Natl. Acad. Sci. U S A 71, 4508-12.

Chang, C. C., Kuo, I. C., Ling, I. F., Chen, C. T., Chen, H. C., Lou, P. J., Lin, J. J., and Chang, T. C. (2004). Detection of quadruplex DNA structures in human telomeres by a fluorescent carbazole derivative. Anal. Chem. 76, 4490-4.

Chang, W., Dynek, J. N., and Smith, S. (2003). TRF1 is degraded by ubiquitin-mediated proteolysis after release from telomeres. Genes Dev. 17, 1328-33.

Chen, Q., Ijpma, A., and Greider, C. W. (2001). Two survivor pathways that allow growth in the absence of telomerase are generated by distinct telomere recombination events. Mol. Cell Biol. 21, 1819-27.

Chester, N., Kuo, F., Kozak, C., O'Hara, C. D., and Leder, P. (1998). Stage-specific apoptosis, developmental delay, and embryonic lethality in mice homozygous for a targeted disruption in the murine Bloom's syndrome gene. Genes Dev. 12, 3382-93.

Chisholm, C. A., Bray, M. J., and Karns, L. B. (2001). Successful pregnancy in a woman with Bloom syndrome. Am. J. Med. Genet. 102, 136-8.

112 Cohen, H., and Sinclair, D. A. (2001). Recombination-mediated lengthening of terminal telomeric repeats requires the Sgs1 DNA helicase. Proc. Natl. Acad. Sci. U S A 98, 3174-9.

Cox, M. M. (2001). Historical overview: searching for replication help in all of the rec places. Proc. Natl. Acad. Sci. U S A 98, 8173-80.

Cumin, I., Cohen, J. Y., David, A., Mechinaud, F., Avet-Loiseau, H., and Harousseau, J. L. (1996). Rothmund-Thomson syndrome and osteosarcoma. Med. Pediatr. Oncol. 26, 414-6.

Der Kaloustian, V. M., McGill, J. J., Vekemans, M., and Kopelman, H. R. (1990). Clonal lines of aneuploid cells in Rothmund-Thomson syndrome. Am. J. Med. Genet. 37, 336-9.

Dunham, M. A., Neumann, A. A., Fasching, C. L., and Reddel, R. R. (2000). Telomere maintenance by recombination in human cells. Nat. Genet. 26, 447-50.

Dutertre, S., Ababou, M., Onclercq, R., Delic, J., Chatton, B., Jaulin, C., and Amor-Gueret, M. (2000). Cell cycle regulation of the endogenous wild type Bloom's syndrome DNA helicase. Oncogene 19, 2731-8. el-Khoury, J. M., Haddad, S. N., and Atallah, N. G. (1997). Osteosarcomatosis with Rothmund-Thomson syndrome. Br. J. Radiol. 70, 215-8.

Ellis, N. A., Groden, J., Ye, T. Z., Straughen, J., Lennon, D. J., Ciocci, S., Proytcheva, M., and German, J. (1995). The Bloom's syndrome gene product is homologous to RecQ helicases. Cell 83, 655-66.

Ellis, N. A., Groden, J., Ye, T. Z., Straughen, J., Lennon, D. J., Ciocci, S., Proytcheva, M., and German, J. (1995). The Bloom's syndrome gene product is homologous to RecQ helicases. Cell 83, 655-66.

Ellis, N. A., Lennon, D. J., Proytcheva, M., Alhadeff, B., Henderson, E. E., and German, J. (1995). Somatic intragenic recombination within the mutated locus BLM can correct the high sister-chromatid exchange phenotype of Bloom syndrome cells. Am. J. Hum. Genet. 57, 1019-27.

Ferreira, M. G., Miller, K. M., and Cooper, J. P. (2004). Indecent exposure: when telomeres become uncapped. Mol. Cell 13, 7-18.

Franchitto, A., and Pichierri, P. (2002). Bloom's syndrome protein is required for correct relocalization of RAD50/MRE11/NBS1 complex after replication fork arrest. J. Cell Biol. 157, 19-30.

113 Frei, C., and Gasser, S. M. (2000). The yeast Sgs1p helicase acts upstream of Rad53p in the DNA replication checkpoint and colocalizes with Rad53p in S-phase-specific foci. Genes Dev. 14, 81-96.

Freire, R., d'Adda Di Fagagna, F., Wu, L., Pedrazzi, G., Stagljar, I., Hickson, I. D., and Jackson, S. P. (2001). Cleavage of the Bloom's syndrome gene product during apoptosis by caspase-3 results in an impaired interaction with topoisomerase IIIalpha. Nucleic Acids Res. 29, 3172-80.

Gangloff, S., McDonald, J. P., Bendixen, C., Arthur, L., and Rothstein, R. (1994). The yeast type I topoisomerase Top3 interacts with Sgs1, a DNA helicase homolog: a potential eukaryotic reverse gyrase. Mol. Cell Biol. 14, 8391-8.

Garkavtsev, I. V., Kley, N., Grigorian, I. A., and Gudkov, A. V. (2001). The Bloom syndrome protein interacts and cooperates with p53 in regulation of transcription and cell growth control. Oncogene 20, 8276-80.

Gaymes, T. J., North, P. S., Brady, N., Hickson, I. D., Mufti, G. J., and Rassool, F. V. (2002). Increased error-prone non homologous DNA end-joining--a proposed mechanism of chromosomal instability in Bloom's syndrome. Oncogene 21, 2525-33.

German, J. (1993). Bloom syndrome: a mendelian prototype of somatic mutational disease. Medicine (Baltimore) 72, 393-406.

German, J. (1995). Bloom's syndrome. Dermatol. Clin. 13, 7-18.

German, J. (1997). Bloom's syndrome. XX. The first 100 cancers. Cancer. Genet. Cytogenet. 93, 100-6.

German, J., and Crippa, L. P. (1966). Chromosomal breakage in diploid cell lines from Bloom's syndrome and Fanconi's anemia. Annales de Genetique 9, 143-154.

German, J., and Ellis, N. A. (2001). Bloom Syndrome. In The Genetic Basis of Human Cancer, B. Volgelstein and K. W. Kinzler, eds. (New York: McGraw-Hil).

German, J., and Passarge, E. (1989). Bloom's syndrome. XII. Report from the Registry for 1987. Clin. Genet. 35, 57-69.

German, J., Schonberg, S., Louie, E., and Chaganti, R. S. (1977). Bloom's syndrome. IV. Sister-chromatid exchanges in lymphocytes. Am.J. Hum. Genet. 29, 248-55.

Giannelli, F., Pawsey, S. A., and Botcherby, P. K. (1981). Tendency to high levels of UVR- induced unscheduled DNA synthesis in Bloom syndrome. Mutat. Res. 81, 229-41.

114

Goss, K. H., Risinger, M. A., Kordich, J. J., Sanz, M. M., Straughen, J. E., Slovek, L. E., Capobianco, A. J., German, J., Boivin, G. P., and Groden, J. (2002). Enhanced tumor formation in mice heterozygous for Blm mutation. Science 297, 2051-3.

Goto, M. (2000). Werner's syndrome: from clinics to genetics. Clin. Exp. Rheumatol. 18, 760-6.

Goto, M., Miller, R. W., Ishikawa, Y., and Sugano, H. (1996). Excess of rare cancers in Werner syndrome (adult progeria). Cancer Epidemiol. Biomarkers Prev. 5, 239-46.

Griffith, J., Bianchi, A., and de Lange, T. (1998). TRF1 promotes parallel pairing of telomeric tracts in vitro. J. Mol. Biol. 278, 79-88.

Griffith, J. D., Comeau, L., Rosenfield, S., Stansel, R. M., Bianchi, A., Moss, H., and de Lange, T. (1999). Mammalian telomeres end in a large duplex loop. Cell 97, 503-14.

Grobelny, J. V., Godwin, A. K., and Broccoli, D. (2000). ALT-associated PML bodies are present in viable cells and are enriched in cells in the G(2)/M phase of the cell cycle. J. Cell Sci. 113 Pt 24, 4577-85.

Groden, J., and German, J. (1992). Bloom's syndrome. XVIII. Hypermutability at a tandem-repeat locus. Hum Genet 90, 360-367.

Groden, J., Nakamura, Y., and German, J. (1990). Molecular evidence that homologous recombination occurs in proliferating human somatic cells. Proc Natl Acad Sci U S A 87, 4315-4319.

Gruber, S. B., Ellis, N. A., Scott, K. K., Almog, R., Kolachana, P., Bonner, J. D., Kirchhoff, T., Tomsho, L. P., Nafa, K., Pierce, H., Low, M., Satagopan, J., Rennert, H., Huang, H., Greenson, J. K., Groden, J., Rapaport, B., Shia, J., Johnson, S., Gregersen, P. K., Harris, C. C., Boyd, J., and Offit, K. (2002). BLM heterozygosity and the risk of colorectal cancer. Science 297, 2013.

Hand, R., and German, J. (1975). A retarded rate of DNA chain growth in Bloom's syndrome. Proc. Natl. Acad. Sci. U S A 72, 758-62.

Harley, C. B., Futcher, A. B., and Greider, C. W. (1990). Telomeres shorten during ageing of human fibroblasts. Nature 345, 458-60.

Hayakawa, S., Kaneko, H., Fukao, T., Kasahara, K., Matsumoto, T., Furuichi, Y., and Kondo, N. (2000). Characterization of the nuclear localization signal in the DNA helicase responsible for Bloom syndrome. Int. J. Mo.l Med. 5, 477-84.

115

Helleday, T. (2003). Pathways for mitotic homologous recombination in mammalian cells. Mutat. Res. 532, 103-15.

Henson, J. D., Neumann, A. A., Yeager, T. R., and Reddel, R. R. (2002). Alternative lengthening of telomeres in mammalian cells. Oncogene 21, 598-610.

Heo, S. J., Tatebayashi, K., Ohsugi, I., Shimamoto, A., Furuichi, Y., and Ikeda, H. (1999). Bloom's syndrome gene suppresses premature ageing caused by Sgs1 deficiency in yeast. Genes Cells 4, 619-25.

Hoehn, H., Bryant, E. M., Au, K., Norwood, T. H., Boman, H., and Martin, G. M. (1975). Variegated translocation mosaicism in human skin fibroblast cultures. Cytogenet. Cell. Genet. 15, 282-98.

Honma, M., Tadokoro, S., Sakamoto, H., Tanabe, H., Sugimoto, M., Furuichi, Y., Satoh, T., Sofuni, T., Goto, M., and Hayashi, M. (2002). Chromosomal instability in B- lymphoblasotoid cell lines from Werner and Bloom syndrome patients. Mutat. Res. 520, 15-24.

Hook, G. J., Kwok, E., and Heddle, J. A. (1984). Sensitivity of Bloom syndrome fibroblasts to mitomycin C. Mutat. Res. 131, 223-30.

Hsu, H. L., Gilley, D., Blackburn, E. H., and Chen, D. J. (1999). Ku is associated with the telomere in mammals. Proc. Natl. Acad. Sci. U S A 96, 12454-8.

Hu, P., Beresten, S. F., van Brabant, A. J., Ye, T. Z., Pandolfi, P. P., Johnson, F. B., Guarente, L., and Ellis, N. A. (2001). Evidence for BLM and Topoisomerase IIIalpha interaction in genomic stability. Hum. Mol. Genet. 10, 1287-98.

Huang, P., Pryde, F. E., Lester, D., Maddison, R. L., Borts, R. H., Hickson, I. D., and Louis, E. J. (2001). SGS1 is required for telomere elongation in the absence of telomerase. Curr. Biol. 11, 125-9.

Huang, S., Li, B., Gray, M. D., Oshima, J., Mian, I. S., and Campisi, J. (1998). The premature ageing syndrome protein, WRN, is a 3'-->5' exonuclease. Nat. Genet. 20, 114-6.

Huber, M. D., Lee, D. C., and Maizels, N. (2002). G4 DNA unwinding by BLM and Sgs1p: substrate specificity and substrate-specific inhibition. Nucleic Acids Res. 30, 3954- 61.

116 Hutteroth, T. H., Litwin, S. D., and German, J. (1975). Abnormal immune responses of Bloom's syndrome lymphocytes in vitro. J. Clin. Invest. 56, 1-7.

Ira, G., and Haber, J. E. (2002). Characterization of RAD51-independent break-induced replication that acts preferentially with short homologous sequences. Mol. Cell Biol. 22, 6384-92.

Ira, G., Malkova, A., Liberi, G., Foiani, M., and Haber, J. E. (2003). Srs2 and Sgs1-Top3 suppress crossovers during double-strand break repair in yeast. Cell 115, 401-11.

Iwano, T., Tachibana, M., Reth, M., and Shinkai, Y. (2004). Importance of TRF1 for functional telomere structure. J. Biol. Chem. 279, 1442-1448.

Johnson, F. B., Marciniak, R. A., McVey, M., Stewart, S. A., Hahn, W. C., and Guarente, L. (2001). The Saccharomyces cerevisiae WRN homolog Sgs1p participates in telomere maintenance in cells lacking telomerase. EMBO J. 20, 905-13.

Karlseder, J., Smogorzewska, A., and de Lange, T. (2002). Senescence induced by altered telomere state, not telomere loss. Science 295, 2446-9.

Karow, J. K., Chakraverty, R. K., and Hickson, I. D. (1997). The Bloom's syndrome gene product is a 3'-5' DNA helicase. J. Biol. Chem. 272, 30611-4.

Karow, J. K., Constantinou, A., Li, J. L., West, S. C., and Hickson, I. D. (2000). The Bloom's syndrome gene product promotes branch migration of Holliday junctions. Proc. Natl. Acad.Sci. U S A 97, 6504-8.

Karow, J. K., Newman, R. H., Freemont, P. S., and Hickson, I. D. (1999). Oligomeric ring structure of the Bloom's syndrome helicase. Curr. Biol. 9, 597-600.

Kauli, R., Prager-lewin, R., Kaufman, H., and Laron, Z. (1977). Gonadal function in Bloom's syndrome. Clin. Endocrinol. (Oxf) 6, 285-9.

Keller, C., Keller, K. R., Shew, S. B., and Plon, S. E. (1999). Growth deficiency and malnutrition in Bloom syndrome. J. Pediatr. 134, 472-9.

Kitao, S., Shimamoto, A., Goto, M., Miller, R. W., Smithson, W. A., Lindor, N. M., and Furuichi, Y. (1999). Mutations in RECQL4 cause a subset of cases of Rothmund-Thomson syndrome. Nat. Genet. 22, 82-4.

Kondo, J., Adachi, W., Umeda, S., Sunami, T., and Takenaka, A. Crystal structures of a DNA octaplex with I-motif of G-quartets and its splitting into two quadruplexes suggest a folding mechanism of eight tandem repeats. Nucleic Acids Res. 8, 2541-9

117

Kondo, N., Asano, J., Kimura, S., Asano, T., and Orii, T. (1991). Insulin-dependent diabetes developed in a young man with Bloom's syndrome. Clin. Genet. 40, 251-2.

Kondo, N., Motoyoshi, F., Mori, S., Kuwabara, N., Orii, T., and German, J. (1992). Long- term study of the immunodeficiency of Bloom's syndrome. Acta Paediatr. 81, 86-90.

Kondo, N., Ozawa, T., Kato, Y., Motoyoshi, F., Kasahara, K., Kameyama, T., and Orii, T. (1992). Reduced secreted mu mRNA synthesis in selective IgM deficiency of Bloom's syndrome. Clin. Exp. Immunol. 88, 35-40.

Kraus, E., Leung, W. Y., and Haber, J. E. (2001). Break-induced replication: a review and an example in budding yeast. Proc. Natl. Acad. Sci. U S A 98, 8255-62.

Kurihara, T., Tatsumi, K., Takahashi, H., and Inoue, M. (1987). Sister-chromatid exchanges induced by ultraviolet light in Bloom's syndrome fibroblasts. Mutat. Res. 183, 197-202.

Kusunoki, Y., Hayashi, T., Hirai, Y., Kushiro, J., Tatsumi, K., Kurihara, T., Zghal, M., Kamoun, M. R., Takebe, H., Jeffreys, A., and et al. (1994). Increased rate of spontaneous mitotic recombination in T lymphocytes from a Bloom's syndrome patient using a flow- cytometric assay at HLA-A locus. Jpn. J. Cancer. Res. 85, 610-8.

Kyoizumi, S., Kusunoki, Y., Seyama, T., Hatamochi, A., and Goto, M. (1998). In vivo somatic mutations in Werner's syndrome. Hum. Genet. 103, 405-10.

Kyoizumi, S., Nakamura, N., Takebe, H., Tatsumi, K., German, J., and Akiyama, M. (1989). Frequency of variant erythrocytes at the glycophorin-A locus in two Bloom's syndrome patients. Mutat. Res. 214, 215-22.

Langland, G., Elliott, J., Li, Y., Creaney, J., Dixon, K., and Groden, J. (2002). The BLM helicase is necessary for normal DNA double-strand break repair. Cancer Res. 62, 2766- 70.

Langland, G., Kordich, J., Creaney, J., Goss, K. H., Lillard-Wetherell, K., Bebenek, K., Kunkel, T. A., and Groden, J. (2001). The Bloom's syndrome protein (BLM) interacts with MLH1 but is not required for DNA mismatch repair. J. Biol. Chem. 276, 30031-5.

Langlois, R. G., Bigbee, W. L., Jensen, R. H., and German, J. (1989). Evidence for increased in vivo mutation and somatic recombination in Bloom's syndrome. Proc. Natl. Acad. Sci. U S A 86, 670-4.

118 Le, S., Moore, J. K., Haber, J. E., and Greider, C. W. (1999). RAD50 and RAD51 define two pathways that collaborate to maintain telomeres in the absence of telomerase. Genetics 152, 143-52.

Lee, S. K., Johnson, R. E., Yu, S. L., Prakash, L., and Prakash, S. (1999). Requirement of yeast SGS1 and SRS2 genes for replication and transcription. Science 286, 2339-42.

Li, B., Oestreich, S., and de Lange, T. (2000). Identification of human Rap1: implications for telomere evolution. Cell 101, 471-83.

Li, J. L., Harrison, R. J., Reszka, A. P., Brosh, R. M., Jr., Bohr, V. A., Neidle, S., and Hickson, I. D. (2001). Inhibition of the Bloom's and Werner's syndrome helicases by G- quadruplex interacting ligands. Biochemistry 40, 15194-202.

Lillard-Wetherell, K., Machwe, A., Langland, G.T., Combs, K.A., Behbehani, G.K., Schonberg, S.A., German, J., Turchi, J. J., Orren, D.K. and Groden, J. (2004). Association and regulation of the BLM helicase by the telomere proteins TRF1 and TRF2. Hum. Mol. Genet. 17, 1919-1932.

Lindor, N. M., Furuichi, Y., Kitao, S., Shimamoto, A., Arndt, C., and Jalal, S. (2000). Rothmund-Thomson syndrome due to RECQ4 helicase mutations: report and clinical and molecular comparisons with Bloom syndrome and Werner syndrome. Am. J. Med. Genet. 90, 223-8.

Liu, Z., Macias, M. J., Bottomley, M. J., Stier, G., Linge, J. P., Nilges, M., Bork, P., and Sattler, M. (1999). The three-dimensional structure of the HRDC domain and implications for the Werner and Bloom syndrome proteins. Structure Fold Des. 7, 1557- 66.

Londono-Vallejo, J. A., Der-Sarkissian, H., Cazes, L., Bacchetti, S., and Reddel, R. R. (2004). Alternative lengthening of telomeres is characterized by high rates of telomeric exchange. Cancer Res. 64, 2324-7.

Lundblad, V. (2002). Telomere maintenance without telomerase. Oncogene 21, 522-31.

Lundblad, V., and Blackburn, E. H. (1993). An alternative pathway for yeast telomere maintenance rescues est1- senescence. Cell 73, 347-60.

Lundblad, V., and Szostak, J. W. (1989). A mutant with a defect in telomere elongation leads to senescence in yeast. Cell 57, 633-43.

119 Luo, G., Santoro, I. M., McDaniel, L. D., Nishijima, I., Mills, M., Youssoufian, H., Vogel, H., Schultz, R. A., and Bradley, A. (2000). Cancer predisposition caused by elevated mitotic recombination in Bloom mice. Nat. Genet. 26, 424-9.

MacEachern, E., and Iyer, B. (2001). Short telomeres in yeast are highly recombinogenic. Mol. Cell 7, 695-704.

Malkova, A., Ivanov, E. L., and Haber, J. E. (1996). Double-strand break repair in the absence of RAD51 in yeast: a possible role for break-induced DNA replication. Proc. Natl. Acad. Sci. U S A 93, 7131-6.

Mamada, A., Kondo, S., and Satoh, Y. (1989). Different sensitivities to ultraviolet light- induced cytotoxicity and sister chromatid exchanges in xeroderma pigmentosum and Bloom's syndrome fibroblasts. Photodermatol. 6, 124-30.

McDaniel, L. D., Chester, N., Watson, M., Borowsky, A. D., Leder, P., and Schultz, R. A. (2003). Chromosome instability and tumor predisposition inversely correlate with BLM protein levels. DNA Repair (Amst) 2, 1387-404.

McKenna, D. J., Gallus, M., McKeown, S. R., Downes, C. S., and McKelvey-Martin, V. J. (2003). Modification of the alkaline Comet assay to allow simultaneous evaluation of mitomycin C-induced DNA cross-link damage and repair of specific DNA sequences in RT4 cells. DNA Repair (Amst) 2, 879-90.

Meetei, A. R., Sechi, S., Wallisch, M., Yang, D., Young, M. K., Joenje, H., Hoatlin, M. E., and Wang, W. (2003). A multiprotein nuclear complex connects Fanconi anemia and Bloom syndrome. Mol. Cell Biol. 23, 3417-26.

Miozzo, M., Castorina, P., Riva, P., Dalpra, L., Fuhrman Conti, A. M., Volpi, L., Hoe, T. S., Khoo, A., Wiegant, J., Rosenberg, C., and Larizza, L. (1998). Chromosomal instability in fibroblasts and mesenchymal tumors from 2 sibs with Rothmund-Thomson syndrome. Int. J. Cancer 77, 504-10.

Mohaghegh, P., Karow, J. K., Brosh Jr, R. M., Jr., Bohr, V. A., and Hickson, I. D. (2001). The Bloom's and Werner's syndrome proteins are DNA structure-specific helicases. Nucleic Acids Res. 29, 2843-9.

Mori, S., Kondo, N., Motoyoshi, F., Yamaguchi, S., Kaneko, H., and Orii, T. (1990). Diabetes mellitus in a young man with Bloom's syndrome. Clin. Genet. 38, 387-90.

Myung, K., Datta, A., Chen, C., and Kolodner, R. D. (2001). SGS1, the Saccharomyces cerevisiae homologue of BLM and WRN, suppresses genome instability and homeologous recombination. Nat. Genet. 27, 113-6.

120

Myung, K., and Kolodner, R. D. (2002). Suppression of genome instability by redundant S-phase checkpoint pathways in Saccharomyces cerevisiae. Proc. Natl. Acad. Sci. U S A 99, 4500-7.

Nakayama, H., Nakayama, K., Nakayama, R., Irino, N., Nakayama, Y., and Hanawalt, P. C. (1984). Isolation and genetic characterization of a thymineless death-resistant mutant of Escherichia coli K12: identification of a new mutation (recQ1) that blocks the RecF recombination pathway. Mol. Gen. Genet. 195, 474-80.

Natarajan, S., and McEachern, M. J. (2002). Recombinational telomere elongation promoted by DNA circles. Mol. Cell Biol. 22, 4512-21.

Niedernhofer, L. J., Odijk, H., Budzowska, M., van Drunen, E., Maas, A., Theil, A. F., de Wit, J., Jaspers, N. G., Beverloo, H. B., Hoeijmakers, J. H., and Kanaar, R. (2004). The structure-specific endonuclease Ercc1-Xpf is required to resolve DNA interstrand cross- link-induced double-strand breaks. Mol. Cell Biol. 24, 5776-87.

Nishijo, K., Nakayama, T., Aoyama, T., Okamoto, T., Ishibe, T., Yasura, K., Shima, Y., Shibata, K. R., Tsuboyama, T., Nakamura, T., and Toguchida, J. (2004). Mutation analysis of the RECQL4 gene in sporadic osteosarcomas. Int. J. Cancer 111, 367-72.

Ohhata, T., Araki, R., Fukumura, R., Kuroiwa, A., Matsuda, Y., Tatsumi, K., and Abe, M. (2000). Cloning, genomic structure and chromosomal localization of the gene encoding mouse DNA helicase RecQ helicase protein-like 4. Gene 261, 251-8.

Olovnikov, A. M. (1973). A theory of marginotomy. The incomplete copying of template margin in enzymic synthesis of polynucleotides and biological significance of the phenomenon. J. Theor. Biol. 41, 181-90.

Onclercq-Delic, R., Calsou, P., Delteil, C., Salles, B., Papadopoulo, D., and Amor-Gueret, M. (2003). Possible anti-recombinogenic role of Bloom's syndrome helicase in double- strand break processing. Nucleic Acids Res. 31, 6272-82.

Onoda, F., Seki, M., Miyajima, A., and Enomoto, T. (2000). Elevation of sister chromatid exchange in Saccharomyces cerevisiae sgs1 disruptants and the relevance of the disruptants as a system to evaluate mutations in Bloom's syndrome gene. Mutat. Res. 459, 203-9.

Opresko, P. L., von Kobbe, C., Laine, J. P., Harrigan, J., Hickson, I. D., and Bohr, V. A. (2002). Telomere-binding protein TRF2 binds to and stimulates the Werner and Bloom syndrome helicases. J. Biol. Chem. 277, 41110-9.

121 Orren, D. K., Theodore, S., and Machwe, A. (2002). The Werner syndrome helicase/exonuclease (WRN) disrupts and degrades D-loops in vitro. Biochemistry 41, 13483-8.

Parshad, R., Sanford, K. K., and Jones, G. M. (1983). Chromatid damage after G2 phase x- irradiation of cells from cancer-prone individuals implicates deficiency in DNA repair. Proc. Natl. Acad. Sci. U S A 80, 5612-6.

Pedrazzi, G., Perrera, C., Blaser, H., Kuster, P., Marra, G., Davies, S. L., Ryu, G. H., Freire, R., Hickson, I. D., Jiricny, J., and Stagljar, I. (2001). Direct association of Bloom's syndrome gene product with the human mismatch repair protein MLH1. Nucleic Acids Res. 29, 4378-86.

Piquero-Casals, J., Okubo, A. Y., and Nico, M. M. (2002). Rothmund-thomson syndrome in three siblings and development of cutaneous squamous cell carcinoma. Pediatr. Dermatol. 19, 312-6.

Prache-de-Carrere, B., Teillac-Hamel, D., Capesius, C., Castelneau, J. P., Robert, C., Fraitag, S., and de Prost, Y. (1996). [Rothmund-Thomson syndrome with reduced DNA repair capacity]. Ann. Dermatol. Venereol. 123, 395-7.

Reddel, R. R., Bryan, T. M., Colgin, L. M., Perrem, K. T., and Yeager, T. R. (2001). Alternative lengthening of telomeres in human cells. Radiat. Res. 155, 194-200.

Richardson, C., and Jasin, M. (2000). Coupled homologous and nonhomologous repair of a double-strand break preserves genomic integrity in mammalian cells. Mol Cell Biol 20, 9068-75.

Rong, S. B., Valiaho, J., and Vihinen, M. (2000). Structural basis of Bloom syndrome (BS) causing mutations in the BLM helicase domain. Mol. Med. 6, 155-64.

Schawalder, J., Paric, E., and Neff, N., F. (2003). Telomere and ribosomal DNA repeats are chromosomal targets of the bloom syndrome DNA helicase. BMC Cell Biol. 4, 15.

Schoen, E. J., and Shearn, M. A. (1967). Immunoglobulin deficiency in Bloom's syndrome. Am. J. Dis. Child 113, 594-6.

Sengupta, S., Linke, S. P., Pedeux, R., Yang, Q., Farnsworth, J., Garfield, S. H., Valerie, K., Shay, J. W., Ellis, N. A., Wasylyk, B., and Harris, C. C. (2003). BLM helicase- dependent transport of p53 to sites of stalled DNA replication forks modulates homologous recombination. EMBO J. 22, 1210-22.

122 Sharma, S., Sommers, J. A., Wu, L., Bohr, V. A., Hickson, I. D., and Brosh, R. M., Jr. (2004). Stimulation of flap endonuclease-1 by the Bloom's syndrome protein. J. Biol. Chem. 279, 9847-56.

Shay, J. W., and Wright, W. E. (1996). Telomerase activity in human cancer. Curr. Opin. Oncol. 8, 66-71.

Shen, J. C., Gray, M. D., Oshima, J., and Loeb, L. A. (1998). Characterization of Werner syndrome protein DNA helicase activity: directionality, substrate dependence and stimulation by replication protein A. Nucleic Acids Res. 26, 2879-85.

Shiloh, Y., and Kastan, M. B. (2001). ATM: genome stability, neuronal development, and cancer cross paths. Adv. Cancer Res. 83, 209-54.

Shiraishi, Y., and Sandberg, A. A. (1978). Effects of mitomycin C on sister chromatid exchange in normal and Bloom's syndrome cells. Mutat. Res. 49, 233-8.

Shuttleworth, D., and Marks, R. (1987). Epidermal dysplasia and skeletal deformity in congenital poikiloderma (Rothmund-Thomson syndrome). Br. J. Dermatol. 117, 377-84.

Signon, L., Malkova, A., Naylor, M. L., Klein, H., and Haber, J. E. (2001). Genetic requirements for RAD51- and RAD54-independent break-induced replication repair of a chromosomal double-strand break. Mol. Cell Biol. 21, 2048-56.

Signon, L., Malkova, A., Naylor, M. L., Klein, H., and Haber, J. E. (2001). Genetic requirements for RAD51- and RAD54-independent break-induced replication repair of a chromosomal double-strand break. Mol. Cell Biol. 21, 2048-56.

Sinclair, D. A., and Guarente, L. (1997). Extrachromosomal rDNA circles--a cause of aging in yeast. Cell 91, 1033-42.

Sinclair, D. A., Mills, K., and Guarente, L. (1997). Accelerated aging and nucleolar fragmentation in yeast sgs1 mutants. Science 277, 1313-6.

Smith, P. J., and Paterson, M. C. (1982). Enhanced radiosensitivity and defective DNA repair in cultured fibroblasts derived from Rothmund Thomson syndrome patients. Mutat. Res. 94, 213-28.

Smith, S., and de Lange, T. (2000). Tankyrase promotes telomere elongation in human cells. Curr. Biol. 10, 1299-302.

123 Smogorzewska, A., van Steensel, B., Bianchi, A., Oelmann, S., Schaefer, M. R., Schnapp, G., and de Lange, T. (2000). Control of human telomere length by TRF1 and TRF2. Mol. Cell Biol. 20, 1659-68.

Smucker, E. J., and Turchi, J. J. (2001). TRF1 inhibits telomere C-strand DNA synthesis in vitro. Biochemistry 40, 2426-32.

Stansel, R. M., de Lange, T., and Griffith, J. D. (2001). T-loop assembly in vitro involves binding of TRF2 near the 3' telomeric overhang. EMBO J. 20, 5532-40.

Stavropoulos, D. J., Bradshaw, P. S., Li, X., Pasic, I., Truong, K., Ikura, M., Ungrin, M., and Meyn, M. S. (2002). The Bloom syndrome helicase BLM interacts with TRF2 in ALT cells and promotes telomeric DNA synthesis. Hum. Mol. Genet. 11, 3135-44.

Sun, H., Bennett, R. J., and Maizels, N. (1999). The Saccharomyces cerevisiae Sgs1 helicase efficiently unwinds G-G paired DNAs. Nucleic Acids Res. 27, 1978-84.

Sun, H., Karow, J. K., Hickson, I. D., and Maizels, N. (1998). The Bloom's syndrome helicase unwinds G4 DNA. J.Biol. Chem. 273, 27587-92.

Tachibana, A., Tatsumi, K., Masui, T., and Kato, T. (1996). Large deletions at the HPRT locus associated with the mutator phenotype in a Bloom's syndrome lymphoblastoid cell line. Mol. Carcinog. 17, 41-7.

Teng, S. C., and Zakian, V. A. (1999). Telomere-telomere recombination is an efficient bypass pathway for telomere maintenance in Saccharomyces cerevisiae. Mol. Cell Biol. 19, 8083-93.

Tokutake, Y., Matsumoto, T., Watanabe, T., Maeda, S., Tahara, H., Sakamoto, S., Niida, H., Sugimoto, M., Ide, T., and Furuichi, Y. (1998). Extra-chromosomal telomere repeat DNA in telomerase-negative immortalized cell lines. Biochem. Biophys. Res. Commun. 247, 765-72.

Ui, A., Satoh, Y., Onoda, F., Miyajima, A., Seki, M., and Enomoto, T. (2001). The N- terminal region of Sgs1, which interacts with Top3, is required for complementation of MMS sensitivity and suppression of hyper-recombination in sgs1 disruptants. Mol. Genet. Genomics 265, 837-50. van Brabant, A. J., Ye, T., Sanz, M., German, I. J., Ellis, N. A., and Holloman, W. K. (2000). Binding and melting of D-loops by the Bloom syndrome helicase. Biochemistry 39, 14617-25.

124 van Steensel, B., and de Lange, T. (1997). Control of telomere length by the human telomeric protein TRF1. Nature 385, 740-3. van Steensel, B., Smogorzewska, A., and de Lange, T. (1998). TRF2 protects human telomeres from end-to-end fusions. Cell 92, 401-13.

Vanderschueren-Lodeweyckx, M., Fryns, J. P., Van den Berghe, H., Eggermont, E., and Eeckels, R. (1984). Bloom's syndrome. Possible pitfalls in clinical diagnosis. Am. J. Dis. Child 138, 812-6.

Varughese, M., Leavey, P., Smith, P., Sneath, R., Breatnach, F., and O'Meara, A. (1992). Osteogenic sarcoma and Rothmund Thomson syndrome. J. Cancer Res. Clin. Oncol. 118, 389-90.

Versini, G., Comet, I., Wu, M., Hoopes, L., Schwob, E., and Pasero, P. (2003). The yeast Sgs1 helicase is differentially required for genomic and ribosomal DNA replication. EMBO J. 22, 1939-49. von Kobbe, C., Karmakar, P., Dawut, L., Opresko, P., Zeng, X., Brosh, R. M., Jr., Hickson, I. D., and Bohr, V. A. (2002). Colocalization, physical, and functional interaction between Werner and Bloom syndrome proteins. J. Biol. Chem. 277, 22035-44.

Wang, L. L., Gannavarapu, A., Kozinetz, C. A., Levy, M. L., Lewis, R. A., Chintagumpala, M. M., Ruiz-Maldanado, R., Contreras-Ruiz, J., Cunniff, C., Erickson, R. P., Lev, D., Rogers, M., Zackai, E. H., and Plon, S. E. (2003). Association between osteosarcoma and deleterious mutations in the RECQL4 gene in Rothmund-Thomson syndrome. J. Natl. Cancer Inst. 95, 669-74.

Wang, L. L., Levy, M. L., Lewis, R. A., Chintagumpala, M. M., Lev, D., Rogers, M., and Plon, S. E. (2001). Clinical manifestations in a cohort of 41 Rothmund-Thomson syndrome patients. Am. J. Med. Genet. 102, 11-7.

Wang, X. W., Tseng, A., Ellis, N. A., Spillare, E. A., Linke, S. P., Robles, A. I., Seker, H., Yang, Q., Hu, P., Beresten, S., Bemmels, N. A., Garfield, S., and Harris, C. C. (2001). Functional interaction of p53 and BLM DNA helicase in apoptosis. J. Biol. Chem. 276, 32948-55.

Wang, Y., Cortez, D., Yazdi, P., Neff, N., Elledge, S. J., and Qin, J. (2000). BASC, a super complex of BRCA1-associated proteins involved in the recognition and repair of aberrant DNA structures. Genes Dev. 14, 927-39.

125 Waters, R., Regan, J. D., and German, J. (1978). Increased amounts of hybrid (heavy/heavy) DNA in Bloom's syndrome fibroblasts. Biochem. Biophys. Res. Commun. 83, 536-41.

Watt, P. M., Hickson, I. D., Borts, R. H., and Louis, E. J. (1996). SGS1, a homologue of the Bloom's and Werner's syndrome genes, is required for maintenance of genome stability in Saccharomyces cerevisiae. Genetics 144, 935-45.

Weemaes, C. M., Bakkeren, J. A., Haraldsson, A., and Smeets, D. F. (1991). Immunological studies in Bloom's syndrome. A follow-up report. Ann. Genet. 34, 201-5.

Wright, W. E., Piatyszek, M. A., Rainey, W. E., Byrd, W., and Shay, J. W. (1996). Telomerase activity in human germline and embryonic tissues and cells. Dev. Genet. 18, 173-9.

Wu, G., Jiang, X., Lee, W. H., and Chen, P. L. (2003). Assembly of functional ALT- associated promyelocytic leukemia bodies requires Nijmegen Breakage Syndrome 1. Cancer Res. 63, 2589-95.

Wu, G., Lee, W. H., and Chen, P. L. (2000). NBS1 and TRF1 colocalize at promyelocytic leukemia bodies during late S/G2 phases in immortalized telomerase-negative cells. Implication of NBS1 in alternative lengthening of telomeres. J. Biol. Chem. 275, 30618-22.

Wu, L., Davies, S. L., Levitt, N. C., and Hickson, I. D. (2001). Potential role for the BLM helicase in recombinational repair via a conserved interaction with RAD51. J. Biol. Chem. 276, 19375-81.

Wu, L., and Hickson, I. D. (2002). The Bloom's syndrome helicase stimulates the activity of human topoisomerase IIIalpha. Nucleic Acids Res. 30, 4823-9.

Wu, L., and Hickson, I. D. (2003). The Bloom's syndrome helicase suppresses crossing over during homologous recombination. Nature 426, 870-4.

Yang, Q., Zhang, R., Wang, X. W., Linke, S. P., Sengupta, S., Hickson, I. D., Pedrazzi, G., Perrera, C., Stagljar, I., Littman, S. J., Modrich, P., and Harris, C. C. (2004). The mismatch DNA repair heterodimer, hMSH2/6, regulates BLM helicase. Oncogene 23, 3749-56.

Yang, Q., Zhang, R., Wang, X. W., Spillare, E. A., Linke, S. P., Subramanian, D., Griffith, J. D., Li, J. L., Hickson, I. D., Shen, J. C., Loeb, L. A., Mazur, S. J., Appella, E., Brosh, R. M., Jr., Karmakar, P., Bohr, V. A., and Harris, C. C. (2002). The processing of Holliday junctions by BLM and WRN helicases is regulated by p53. J. Biol. Chem. 21, 21.

126 Yankiwski, V., Marciniak, R. A., Guarente, L., and Neff, N. F. (2000). Nuclear structure in normal and Bloom syndrome cells. Proc. Natl. Acad. Sci. U S A 97, 5214-9.

Yankiwski, V., Noonan, J. P., and Neff, N. F. (2001). The C-terminal domain of the Bloom syndrome DNA helicase is essential for genomic stability. BMC Cell Biol. 2, 11.

Yeager, T. R., Neumann, A. A., Englezou, A., Huschtscha, L. I., Noble, J. R., and Reddel, R. R. (1999). Telomerase-negative immortalized human cells contain a novel type of promyelocytic leukemia (PML) body. Cancer Res. 59, 4175-9.

Yeager, T. R., Neumann, A. A., Englezou, A., Huschtscha, L. I., Noble, J. R., and Reddel, R. R. (1999). Telomerase-negative immortalized human cells contain a novel type of promyelocytic leukemia (PML) body. Cancer Res. 59, 4175-9.

Ying, K. L., Oizumi, J., and Curry, C. J. (1990). Rothmund-Thomson syndrome associated with trisomy 8 mosaicism. J. Med. Genet. 27, 258-60.

Yu, C. E., Oshima, J., Fu, Y. H., Wijsman, E. M., Hisama, F., Alisch, R., Matthews, S., Nakura, J., Miki, T., Ouais, S., Martin, G. M., Mulligan, J., and Schellenberg, G. D. (1996). Positional cloning of the Werner's syndrome gene. Science 272, 258-62.

Zhong, S., Hu, P., Ye, T. Z., Stan, R., Ellis, N. A., and Pandolfi, P. P. (1999). A role for PML and the nuclear body in genomic stability. Oncogene 18, 7941-7.

Zhu, X. D., Kuster, B., Mann, M., Petrini, J. H., and Lange, T. (2000). Cell-cycle-regulated association of RAD50/MRE11/NBS1 with TRF2 and human telomeres. Nat. Genet. 25, 347-52. .

127