Functional characterization of MutL homologue mismatch ...

Functional characterization of MutL homologue mismatch

repair proteins and their variants

Mari K. Korhonen

Division of Genetics Department of Biological and Environmental Sciences

Faculty of Biosciences University of Helsinki

Academic dissertation

To be presented for public examination with the permission of the Faculty of Biosciences of the University of Helsinki in the auditorium 1041 of the Biocenter II,

Viikinkaari 5, Helsinki, on the 5th of June 2009 at 12 o´clock noon.

Supervisor Professor Minna Nyström Department of Biological and Environmental Sciences

Faculty of Biosciences University of Helsinki, Finland

Reviewers Docent Jukka Partanen

Research and Development Finnish Red Cross Blood Service Helsinki, Finland

Professor Katarina Pelin

Department of Biological and Environmental Sciences Faculty of Biosciences University of Helsinki, Finland

Opponent Docent Minna Pöyhönen

Department of Medical Genetics Haartman Institute University of Helsinki, Finland

and HUSLAB Department of Clinical Genetics Helsinki University Central Hospital Helsinki, Finland ISSN 1795-7079

ISBN 978-952-10-5494-5

ISBN 978-952-10-5495-2 (pdf)

Helsinki 2009, Helsinki University Printing House

The most beautiful adventures are not those we go to seek.

- Robert Louis Stevenson-

4

CONTENTS

LIST OF ORIGINAL PUBLICATIONS 6

ABBREVIATIONS 7

ABSTRACT 9

INTRODUCTION 11

REVIEW TO LITERATURE 13

DNA damage and repair 13

Different repair mechanisms 13

Mismatch repair mechanism 15

Overview of MMR 15

MMR in Escherichia coli 16

MMR in humans 17

Other functions of MMR proteins 20

Human MutL homologues 22

Different MutL homologues and their function

in MMR 22

Functional domains of MutL homologues 23

Nuclear localization signals in MutL homologues 25

Hereditary nonpolyposis colorectal cancer syndrome 27

Hereditary colorectal cancer 27

Diagnostic criteria for HNPCC 28

HNPCC and MMR 30

MMR gene mutations 30

Mutations in MutL homologue genes 31

Clarifying the pathogenicity of nontruncating

MMR mutations 33

AIMS OF THE STUDY 36

MATERIALS AND METHODS 37

MLH1, MLH3 and MSH2 mutations and associated families

(I, III, IV) 37

Protein expression (I-IV) 41

Mutagenesis and expression vectors (I-IV) 41

5

Baculoviral protein expression (I, III, IV) 42

Protein expression in human cells (I, II, IV) 43

Functional analyses (I-IV) 44

Nuclear localization (I, II, IV) 44

Western blot and coimmunoprecipitation analyses

(I, III, IV) 45

In vitro MMR repair assay (I, III, IV) 46

In silico comparative sequence analyses (I, IV) 48

RESULTS 49

Functional characterization of nontruncating MLH1,MLH3

and MSH2 mutations (I, III, IV) 49

Expression/stability of mutated proteins (I, III, IV) 49

Subcellular localization of mutated MLH1 proteins

(I, IV) 49

Interaction capability of mutated proteins (I, III) 50

Mismatch repair efficincy of mutated proteins

(I, III, IV) 50

In silico predictions of MLH1 and MSH2 mutations

(I, IV) 50

Subcellular localization of wild type MutL homologues MLH1,

PMS2 and MLH3 (II) 53

DISCUSSION 55

Pathogenicity of MLH1 mutations (I, IV) 55

In silico analyses of MLH1 mutations (I, IV) 58

Pathogenicity of MSH2 mutations (IV) 60

Pathogenicity of MLH3 mutations (III) 60

Nuclear localization of MLH1, PMS2 and MLH3 (II) 62

CONCLUSIONS 64

FUTURE PROSPECTS 65

ACKNOWLEDGEMENTS 66

REFERENCES 68

6

LIST OF ORGINAL PUBLICATIONS

I Raevaara TE, Korhonen MK, Hampel H, Lynch E, Lönnqvist KE, Holinski-Feder E,

Sutter C, McKinnon W, Duraisamy S, Gerdes A-M, Peltomäki P, Kohonen-Corish M,

Magold E, Macrae F, Greenblatt M, de la Chapelle A, and Nyström M. 2005. Functional

significance and clinical phenotype of nontruncating mismatch repair variants of MLH1.

Gastroenterology, 129: 537-549.

II Korhonen MK, Raevaara TE, Lohi H and Nyström M. 2007. Conditional nuclear

localization of hMLH3 suggests a minor activity in mismatch repair and supports its role

as a low risk gene in HNPCC. Oncology Reports, 17: 351-354.

III Korhonen MK, Vuorenmaa E and Nyström M. 2008. The first functional study of MLH3

mutations found in cancer patients. Genes Chromsomes & Cancer, 47:803-809.

IV Christensen LL*, Kariola R*, Korhonen MK*, Wikman FP, Sunde L, Gerdes A-M,

Okkels H, Brandt CA, Bernstein I, Hansen TV, Hageman-Madsen R, Andersen CL,

Nyström M and Ørntoft TF. Functional characterization of rare missense mutations in

MLH1 and MSH2 identified in Danish colorectal cancer patients. Submitted.

* Equal contribution

7

ABBREVIATIONS

AC I/II Amsterdam criteria I/II

ACVR2 Activin type II receptor

ADP Adenosine diphosphate

ATP Adenosine triphosphate

BER Base excision repair

cDNA Complementary DNA

CRC Colorectal cancer

DAPI 4´, 6´-diamidino-2-phenylindole

DSB Double strand break

dsDNA Double-stranded DNA

EC Endometrial cancer

EGFP Enhanced green fluorescent protein

ExoI Exonuclease I

FAP Familial adenomatous polyposis coli

GHKL Gyrase-HSP-Kinase-MutL ATPase superfamily

HDR Homology-directed repair

HMGB1 High-mobility group box 1

HNPCC Hereditary nonpolyposis colorectal cancer

IDL Insertion/deletion loop

IHC Immunohistochemistry

InSiGHT International Society for Gastrointestinal Hereditary Tumoros

LVOD Leiden Open Variation Database

MAPP Multivariate analysis of protein polymorphism

MGMT O6-methylguanine methyltransferase

MLH1/3 MutL homologue 1/3

MMR Mismatch repair

MMRUV Mismatch Repair Gene Unclassified Variants Database

MSI Microsatellite instability

MSI-H High microsatellite instability

MSH2/3/4/5/6 MutS homologue 2/3/4/5/6

MSS Microsatellite stable

8

MutHLS Mutator HLS

NE Nuclear extract

NER nucleotide excision repair

NHEJ nonhomologous end joining

NLS Nuclear localization signal

PCNA Proliferating cell nuclear antigen

PCR Polymerase chain reaction

PMS1/2 Post- meiotic segregation increased homolog 1/2

PolyPhen Polymorphism phenotyping

RFC Replication factor C

ROS reactive oxygen species

RPA Replication protein A

Sf9 Spodoptera frugiperda 9

SHM Somatic hypermutation

SIFT Sorting Intolerant From Tolerant program

Ssb single-strand DNA binding protein

ssDNA single-stranded DNA

TE Total extract

TGFB2 Transforming growth factor β receptor II

wt wild type

9

ABSTRACT

Mismatch repair (MMR) mechanisms repair DNA damage occurring during replication and

recombination. These errors, such as base/base mismatches and small insertion/deletion loops,

are in the newly synthesized strand and have escaped from the proofreading activity of DNA

polymerase. To date, five human MMR genes, the MutS homologues MSH2, MSH6 and

MSH3 and the MutL homologues MLH1 and PMS2, are known to be involved in the MMR

function. Human MMR proteins form 3 different heterodimers: MutSα (MSH2 and MSH6)

and MutSβ (MSH2 and MSH3), which are needed for mismatch recognition and binding, and

MutLα (MLH1 and PMS2), which is needed for mediating interactions between MutS

homologues and other MMR proteins. The other two MutL homologues, MLH3 and PMS1,

have been shown to heterodimerize with MLH1. However, the heterodimers MutLγ (MLH1

and MLH3) and MutLβ (MLH1 and PMS1) are able to correct mismatches only with low or

no efficiency, respectively.

A deficient MMR mechanism is associated with the hereditary colorectal cancer syndrome

called hereditary nonpolyposis colorectal cancer (HNPCC) or Lynch syndrome. HNPCC is

the most common hereditary colorectal cancer syndrome and accounts for 2-5% of all

colorectal cancer cases. Typical HNPCC tumors show high instability in microsatellites as a

consequence of MMR deficiency. HNPCC-associated mutations have been found in 5 MMR

genes: MLH1, MSH2, MSH6, PMS2 and MLH3. Most of the mutations have been found in

MLH1 and MSH2 (~90%) and are associated with typical HNPCC, while mutations in MSH6,

PMS2 and MLH3 are mainly linked to putative HNPCC families lacking the typical clinical

and molecular characteristics of the syndrome. Although MLH3 is suggested to be causative

in HNPCC, its relevance to MMR needs to be confirmed to reliably assess significance of the

inherited sequence variants in it.

Hundreds of different germline alterations have been found in HNPCC or putative HNPCC

families. A significant proportion consists of nontruncating changes associated with a

variability of clinical phenotypes, microsatellite instability and with the occasional presence

of residual protein in tumor tissue that suggests impaired functional activity but not total lack

of MMR. The difficulty in reliably assessing the pathogenicity of nontruncating mutations is a

major challenge in the diagnostics and counseling of HNPCC families.

10

In this study, 51 nontruncating mutations in the MLH1, MLH3 and MSH2 genes were

functionally characterized in order to address their pathogenic significance and mechanism of

pathogenicity. All the mutations were tested using in vitro MMR functional assays.

Protein/protein interaction, expression/stability and subcelluar localization analyses were

performed for most of the mutations, and the genetic and biochemical data were compared

with clinical data to find out the genotype-phenotype correlations.

Of the 36 MLH1 mutations, 22 were deficient in more than one assay, 2 mutations were

impaired only in one assay, and 12 mutations functioned like the wild type protein, whereas

all seven MLH3 mutants functioned like the wild type protein in the assays. To further clarify

the role and relevance of MLH3 in MMR, we analyzed the subcellular localization of the

native MutL homologue proteins. Our immunofluorescence analyses indicated that when all

the three MutL homologues are natively expressed in human cells, endogenous MLH1 and

PMS2 localize in the nucleus, whereas MLH3 stays in the cytoplasm. The coexpression of

MLH3 with MLH1 results in its partial nuclear localization. Only one MSH2 mutation was

pathogenic in the in vitro MMR assay.

Our study on MLH1 mutations could clearly distinguish nontruncating alterations with severe

functional defects from those not or only slightly impaired in protein function. However, our

study on MLH3 mutations suggest that MLH3 mutations per se are not sufficient to trigger

MMR deficiency and the continuous nuclear localization of MLH1 and PMS2 suggest that

MutLα has a major activity in MMR in vivo. Together with our functional assays, this

confirms that MutLγ is a less efficient MMR complex than MutLα.

11

INTRODUCTION

Multiple types of DNA damage occur in a cell, but these are usually mended by different

repair mechanism. The mismatch repair (MMR) mechanism repairs polymerase errors arising

during DNA replication and recombination. MMR-deficient cells have a mutator phenotype,

which means that mutations accumulate in the genome, especially in repetitive sequences.

The accumulation of mutations is associated with an increased risk of cancer. Based on

Knudson’s “two hit” model, both copies of a tumor suppressor gene need to be inactivated

before a cell can become malignant (Knudson 1971; 1996). An individual can either have

somatic mutations in both gene alleles or an inherited mutation in one and a somatic mutation

in the other allele. The inherited mutation is the first hit and thus individuals carrying this

mutation are more predisposed to cancer at a young age than indivuduals without the inherited

mutation. Hereditary nonpolyposis colorectal cancer (HNPCC) is a cancer syndrome

associated with dominantly inherited mutations in MMR genes, which are tumor supressor

genes (Peltomäki et al. 1993). Five MMR genes, MLH1, MHS2, MSH6 PMS2 and MLH3, are

associated with HNPCC (Peltomäki 2005).

Characteristic features of HNPCC are an early age of onset, numerous affected family

members and multiple primary tumors (Lynch & de la Chapelle 2003). In addition to

colorectal cancer (CRC), HNPCC families often display an excess of extracolonic tumors,

especially in the endometrium (Aarnio et al. 1999). HNPCC diagnosis has relied on the

international Amsterdam criteria (Vasen et al. 1991; Vasen et al. 1999). Hallmarks of

HNPCC tumors are microsatellite instability (MSI) and the loss of a functional MMR protein

(Lynch & de la Chapelle 2003). HNPCC patients with germline mutations in MLH1 or MSH2

have a very high penetrance and lifetime risk of developing colorectal or endometrial cancer.

The risk is lower in patients with mutations in MSH6 or PMS2 (Peltomäki 2005). Typical

HNPCC characteristics are found in MLH1 and MSH2 mutation carriers, whereas MHS6 and

PMS2 mutation carriers have more atypical features of HNPCC. The role of MLH3 in

HNPCC is controversial. Some MLH3 mutations are found in putative HNPCC families but

mainly with atypical features of HNPCC (Wu et al. 2001; Liu et al. 2003). Some other studies

could not find a correlation between MLH3 mutations and HNPCC (Loukola et al. 2000;

Akiyama et al. 2001; Lipkin et al. 2001; Hienonen et al. 2003). More data of MLH3

mutations are needed to assess the significance of its mutations in HNPCC.

12

Germline mutations in the 2 major HNPCC-associated genes, MLH1 and MSH2, are found in

two thirds of classical HNPCC that fulfill the stringent Amsterdam criteria and display MSI in

tumor tissue (Liu et al. 1996). A significant proportion of MMR gene mutations are

nontruncating (~30% of MLH1 and MSH6, ~20% of MSH2, ~50% of PMS2 and nearly 90%

of MLH3 mutations) and are associated with a variability of clinical phenotypes (Peltomäki &

Vasen 2004, Woods et al. 2007, http://www.insight-group.org). This type of mutation is a

major challenge for the diagnosis and counseling of HNPCC families. Thus, when

information of cosegregation with the disease is not available and the biochemical

significance of the alteration is uncertain, the nature of the mutation should be functionally

characterized. The results of functional analyses have shown how difficult it is to distinguish

pathogenic nontruncating mutations from harmless variants, as different mutations even in the

same codon can cause either complete elimination of MMR or have little or no effect on

protein function (Ellison et al. 2004).

In this PhD study, we functionally characterized the MutL homologue proteins MLH1, MLH3

and PMS2 by evaluating the pathogenicity of 36 nontruncating MLH1 and seven MLH3

mutations by in vitro MMR, protein expression/stability, protein interaction and subcellular

localization assays. In addition to that we also tested eight MSH2 mutations in the in vitro

MMR assay. Furthermore, we analyzed the subcellular localization of the native MutL

homologue proteins MLH1, MLH3 and PMS2 by immunofluorescence. These analyses

demonstrate that pathogenic MLH1 nontruncating mutations may interfere with different

biochemical mechanisms. The results also show that these MLH3 mutations alone are not

sufficient to trigger MMR deficiency and the results of the endogenous localization assays

support the idea that MutLα has a major role in MMR. Finally, the genetic and biochemical

data correlated well, showing clear differences between MLH1 mutations causing deficiency

of the protein, which are associated with characteristic HNPCC, and other types of mutation.

13

REVIEW OF THE LITERATURE

DNA damage and repair

Double-stranded DNA molecules contain all the information needed for the proper function of

a cell. Two functional sets of DNA exist in each cell and both copies are important. DNA

codes through RNA for the production of proteins. The final three-dimensional shape of a

protein depends on its amino acid sequence, and this precise native structure is important for

its function. Since DNA molecules are irreplaceable, damage to DNA needs to be repaired or

otherwise the function of the cell may be endangered. DNA strands are replicated during cell

division, forming two new copies of each chromosome which are then divided equally to the

two daughter cells thus mainting the full complement of chromosomes in each cell. Errors in

DNA replication lead to DNA mispairing. DNA can also be damaged by different chemically

reactive molecules, such as reactive oxygen species (ROS), which are a byproduct of

oxidative phosporylation, or by chemical modification, e.g. deamination. In addition to

endogenous damaging agents, DNA can be attacked by mutagens of foreign origin, such as

ionizing and ultraviolet radiation or chemical carcinogens. It is essential for cells to efficiently

respond to DNA damage. The accumulation of DNA lesions lead to the loss of genomic

integrity and promotion of tumorigenesis (reviewed in Thoms et al. 2007). Normally, DNA

damage leads to cell cycle checkpoint activation and the arrest of the cycle. This gives time to

repair the damage before DNA synthesis. DNA damage may also lead to cell senescence or

apoptosis, which are ways to prevent the replication of damaged DNA. Cells have several

hundred distinct proteins to ensure that damaged DNA is not transmitted to daughter cell.

Different repair mechanisms

Chemical adducts can be removed from DNA by enzyme-catalyzed direct reversal. This does

not involve multiple proteins or the excision of damaged bases (Sedgwick et al. 2007). The

best characterized DNA alkyltransferase is O6-methylguanine methyltransferase (MGMT),

which is able to remove alkylation adducts form the O6-position of guanine bases and thus

restore the normal base structure (Pegg & Byers 1992). The alkylation of MGMT leads to its

inactivation and one MGMT molecule can repair only a single alkyl adduct (Kaina et al.

14

2007). The unrepaired O6-methylated guanine mispairs with thymine, which leads to a G to A

transition mutation (Margison et al. 2002).

Double-strand breaks (DSB) are major lesions that can lead to cell death (Rich et al. 2000). In

DSBs, both strands are damaged (due to oxidative damage, ionizing radiation or the collapse

of the replication fork) and these breaks can lead to genome rearrangements such as

chromosome fusions, deletions and translocations (Jackson 2002; Povirk 2006). Homology-

directed repair (HDR) and nonhomologous end joining (NHEJ) are responsible for DSB

repair. HDR takes place during the late S and G2 phases of the cell cycle, using the

undamaged sister chromatid as a template. When the sister chromatid sequence is not present

(in the G0 and G1 phases of the cell cycle), DSBs are repaired by NHEJ, which results in the

fusion of the two double stranded DNA (dsDNA). This mechanism is very error-prone, but

NHEJ is important in the formation of functional antibodies and T-cell receptors (reviewed in

Soulas-Sprauel et al. 2007).

DNA damage can be in just one of the two strands. Base excision repair (BER), nucleotide

excision repair (NER) and mismatch repair (MMR) are mechanisms that work in single strand

damage repair. The repair mechanism removes the damaged nucleotide and replaces it with a

normal nucleotide complementary to the undamaged strand. Deficiencies in these repair

mechanisms lead to the accumulation of DNA damage in the genome and this in turn leads to

the development of tumors and cancer. BER repairs single nucleotide lesions and is mainly

involved in repairing endogenous damage in DNA caused by e.g. ROS or depurination.

Defects in BER are connected to cancer predisposition. NER repairs lesions affecting longer

stretches of bases (2-30), such as exogenous damage caused by chemical carcinogens (e.g.

polycyclic hydrocarbons, heterocyclic amines) and pyrimidine dimers formed by UV

radiation. Deficiency in NER predisposes to Xeroderma pigmentosa (XP). Patients suffering

from this disease are sensitive to UV radiation and have a highly increased risk of skin cancer

(Cleaver 1968). MMR corrects replication errors that have escaped the proofreading activity

of DNA polymerase. Especially important are mono- and dinucleotide repeat regions, which

are prone to errors in MMR deficiency. MMR deficiency leads to hereditary nonpolyposis

colorectal cancer (HNPCC). This will be discussed in detail in later chapters.

15

Mismatch repair mechanism

Overview of MMR

The MMR mechanism has a critical role in maintaining genetic stability and mutation

avoidance. The MMR mechanism repairs DNA errors that have not been corrected by the

proofreading function of DNA polymerase and lowers the mutation rate by a factor of 100-

1000 to one error per 1010 nucleotides synthesized (Modrich & Lahue 1996). The MMR

mechanism must specifically identify the incorrect base in the newly synthesized daughter

strand, because both bases at the error site are chemically normal components of DNA.

Different mispairings are recognized and repaired by the MMR mechanism with different

efficiencies, and e.g. G•T and A•C mispairings are very efficiently corrected in both

prokaryotic and eukaryotic cells, whereas C•C mismatches are not recognized by the MMR

proteins (Kramer et al. 1984; Dohet et al. 1985; Fang & Modrich 1993; Nakahara et al. 2000;

Brown et al. 2001). The G•T mismatch adopts a relatively stable and well-fitting

configuration, whereas the non-pairing C•C assumes poorly fitting arrangements (Johnson &

Beese 2004). Another class of mutations are insertion-deletion loops (IDLs), arising

especially during the replication of repetitive sequence motifs, so called microsatellites (e.g.

[A]n or [CA]n), which are present in our genome in large numbers (Kunkel 1993, Fishel et al.

1994). A loss of MMR leads to microsatellite instability (MSI) in tumor DNA. MSI can be

detected if many cells are affected by the same change (de la Chapelle 2003). The instability

is due to base substitution, or deletions or insertions of a few nucleotides. Microsatellites are

especially sensitive to mutations because they are extremely prone to polymerase slippage,

resulting in the accumulation of replication errors (Loeb 1994; Peltomäki 2001). This kind of

MSI has been found especially in many colon cancers but also in extracolonic cancers

(Boland et al. 1998; de la Chapelle 2003; Umar et al. 2004). Most microsatellites exist in non-

coding DNA regions, but they are also found in a number of human genes. One of the MSI

target genes is the type II receptor for transforming growth factor β (TGFβR2). MSI in this

gene may cause the change from adenoma to malignant tumor (Grady et al. 1998). TGFR2,

a regulator of growth and apopotosis, contains a microsatellite sequence within an exon and a

framshift mutation in this sequence truncates the protein (Markowitz et al. 1995). Another

TGF-β receptor superfamily member, Activin type II receptor (ACVR2), also contains a

microsatellite within an exon (Hempen et al. 2003; Jung et al. 2004), and both TGFβR2 and

16

ACVR2 are frequently mutated in colorectal cancers (Markowitz et al. 1995; Jung et al.

2004). Other genes that are altered in MSI tumors are e.g. PTEN (Shin et al. 2001), BAX

(Rampino et al. 1997), MSH6 and MSH3 (Ohmiya et al. 2001)

MMR in Escherichia coli

The key players of the Escherichia coli MMR system are MutS, MutL, MutH and DNA

helicase II (UvrD) (Lu et al. 1984). Other required factors are four exonucleases (ExoI,

ExoVII, ExoX and RecJ); single-strand DNA binding protein (Ssb) that protects the single-

strand DNA (ssDNA) gap resulting from endonuclease activity; β-clamb protein that is

required for processive DNA replication; γ-δ complex that loads the β-clamb onto DNA;

DNA polymerase III holoenzyme, and DNA ligase (Lahue et al. 1989; Burdett et al. 2001;

Kunkel & Erie 2005). MutS protein recognizes and binds to mismatches and IDLs. The MutL

protein is thought to be a molecular matchmaker, which mediates interactions between MutS

and MutH (Modrich 1991). Mismatch-bound homodimeric MutS recruits the MutL

homodimer in an ATP-dependent manner (Grilley et al. 1989). The MutS-MutL complex

activates monomeric MutH endonuclease, which incises a transiently unmethylated GATC

sequence at a site 5´or 3´ of the mismatch, located as far as 1 kb from the mismatch (Welsh et

al. 1987; Bruni et al. 1988; Grilley et al. 1993). The nick serves as the point of entry for the

MutL-activated helicase UvrD, which unwinds the DNA double helix from the nick to about

100 nucleotides past the mismatch (Welsh et al. 1987; Bruni et al. 1988; Dao & Modrich

1998). The Ssb protein stabilizes the single-stranded gap (Lahue et al. 1989). DNA is

degraded in the 5´→3´ direction by ExoVII or RecJ exonucleases, or in the 3´→5´ direction

by ExoI, ExoVII or ExoX exonucleases (Grilley et al. 1993; Burdett et al. 2001). Finally, the

single-stranded gap is filled by DNA polymerase III holoenzyme and the DNA ends are

sealed by DNA ligase (Modrich & Lahue 1996) (Figure 1).

17

Figure 1. Mismatch repair in E. coli. The newly synthesized DNA strand contains a mispair. The MutS-MutL-ATP complex activates MutH endonuclease, which incises the nearest unmethylated GATC sequence (either 5´ or 3´from the mispair) in the newly synthesized strand. Exonucleases degrade the nascent strand from the nick to ~100 nucleotides past the mispair and the resulting single-stranded gap is bound by Ssb. DNA polymerase III fills the gap and the remaining gap is sealed by DNA ligase. Finally the new strand is methylated by Dam methylase (modified from Jiricny 2006). MMR in humans

The eukaryotic MMR mechanism is evolutionarily well conserved and it has many features in

common with prokaryotic MMR. However, unlike in E. coli, in eukaryotes MutS and MutL

homologues function as heterodimers. There are two different complexes of MutS

homologues, a heterodimer of MSH2 and MSH6 (MutSα) and a heterodimer of MSH2 and

MSH3 (MutSβ) (Acharya et al. 1996; Palombo et al. 1996), and three different MutL

complexes: a heterodimer of MLH1 and PMS2 (MutLα), a heterodimer of MLH1 and MLH3

Mismatch or IDL

MutS, MutL, ATP

CH3

MutH

CH3MutS, MutL, ATPExo, UvrD, Ssb

CH3

CH3

DNA pol III

CH3

CH3

DNA ligase, Dam

18

(MutLγ) and a heterodimer of MLH1 and PMS1 (MutLβ) (Li & Modrich 1995; Räschle et al.

1999; Lipkin et al. 2000). Yeast Pms1 is the closest homologue to human PMS2, and yeast

Mlh2 is most similar to human PMS1 (Wang et al. 1999). E. coli and human MMR proteins

and their functions are listed in Table 1.

Table 1. Functions of MMR proteins in E.coli and human E.coli Function Human

homologue

Function

MutS mismatch and small IDL

recognition and binding

MutSα

(MSH2+MSH6)

mismatch and small IDL

recognition and binding

MutSβ

(MSH2+MSH3)

IDL binding

MutL molecular matchmaker;

activates MutH endolytic

activity in MutS-, ATP- and

mismatch-dependent manner

MutLα

(MLH1+PMS2)

Moleculars matchmaker

Endonuclease

Meiotic recombination

MutLβ

(MLH1+PMS1)

Function in human MMR not

known

MutLγ

(MLH1+MLH3)

Meiotic recombination

Possible backup for MutLα in

MMR

MutH endonuclease, incises the

unmethylated strand at

hemimethylated GATC sites

none -

DNA helicase

II

Unwinds DNA to allow

excision of ssDNA

none -

γ-δ complex loads β-clamb onto DNA RFC Loads PCNA onto DNA

β-clamp Enhances processivity of

DNA pol III

Interacts with MutS

PCNA interacts with MSH2, MSH6,

MSH3 and MLH1

initiation of DNA resynthesis

Ssb Single-stranded gap protection RPA

HMGB1

ssDNA gap protection

ExoI, ExoX,

RecJ, Exo VII

Exonuclease EXOI Exonuclease

Interacts with MSH2 and MLH1

DNA pol III DNA resynthesis DNA pol δ DNA resynthesis

DNA ligase nick ligation after completion

of DNA resynthesis

DNA ligase I nick ligation after completion of

DNA resynthesis

IDL, insertion deletion loop; PCNA, proliferating cell nuclear antigen; RFC, replication factor C; RPA, replication protein A; HMGB, high-mobility group box 1

19

In vitro studies have suggested that also in eukaryotes repair is directed by a strand

discontinuity (either 3´ or 5´ of the mismatch) (Thomas et al. 1991; Kunkel & Erie 2005).

How strand discrimination takes place is not clear, since MutH is found only in gram-negative

bacteria. In humans neither MutH nor DNA helicase homologues have been identified, and

there is no evidence that CpG methylation would have a similar effect on the rate of mismatch

correction as adenine methylation in E. coli (Drummond & Bellacosa 2001). Strand breaks,

such as the 5´ or 3´ termini of Okazaki fragments in the lagging strand or the 3´ terminus of

the leading strand, could direct mismatch processing (Modrich & Lahue 1996). Strand

discontinuity might also be directed by PCNA, which interacts with MutS and MutL

homologues (Umar et al. 1996). Recently, MutLα was found to possess a latent endonuclease

activity, which is activated by MutSα, PCNA and RFC in an ATP-dependent manner

(Kadyrov et al. 2006).

In humans, mispairs are recognized by MutSα or MutSβ, depending on the type of mutation.

MutSα primarily recognizes single base-base mismatches and small (< 2 bp) IDLs, whereas

MutSβ recognizes only IDLs (Acharya et al. 1996; Palombo et al. 1996; Marra et al. 1998;

Umar et al. 1998). The ADP-bound form of MutS forms a stable complex with mismatched

DNA (Lamers et al. 2000; Junop et al. 2001) and an ATP/ADP switch controls the DNA

binding activity of MutSα (Gradia et al. 1997; Iaccarino et al. 2000). After mismatch

recognition, MutSα and MutLα form an ATP-dependent ternary complex and this complex

may travel along the DNA (Li & Modrich 1995; Räschle et al. 2002; Plotz et al. 2002). MutS

and MutL homologues are the main proteins in the MMR mechanism, but additional factors

are also needed. PCNA interacts with RFC and is loaded onto the 3´ terminus of an Okazaki

fragment or onto the 3´ end of the leading strand. PCNA interacts with MutS homologues and

it may help localize MutSα or MutSβ to mispairs in newly replicated DNA (Umar et al. 1996;

Clark et al. 2000; Kleczkowska et al. 2001; Lau & Kolodner 2003). Exonucleases of both

3´→ 5´ and 5´→3´ polarity are needed for strand degradation. Exonuclease-1 (ExoI) catalyzes

the 5´→3´ degradation of the discontinuous strand and is the only exonuclease that has been

directly implicated in MMR in both in vivo and in vitro (Genschel et al. 2002; Wei et al.

2003). MutSα is needed for ExoI activation (Genschel & Modrich 2003). RPA binds and

protects the single-stranded DNA region. Reconstitution of the bidirectional MMR systems

have proven that MutSα, MutLα, ExoI, RPA, PCNA, RFC and DNA polymerase δ are needed

for in vitro repair (Constantin et al. 2005; Zhang et al. 2005). Human PMS1 has not been

shown to be involved in DNA mismatch repair (Räschle et al. 1999).

20

Figure 2. A model of the MMR mechanism in eukaryotes. MMR requires the DNA to be nicked either 5´or 3 ́of the mismatch. MutSα heterodimer binds the mismatch. MutSα together with PCNA and RFC activates MutLα and the MutSα-MutLα complex undergoes an ATP-dependent conformational switch. This allows the complex to slide away from the mismatch and to search the discontinuous strand. This generates entry points for MutSα-activated ExoI, which degrades the nicked strand. RPA protects the single-stranded gaps. DNA Polδ fills the gap with the help of its cofactors PCNA and RFC, and finally ligase I seals the remaining nick (modified from Kadyrov et al. 2006; Kunz et al. 2009).

Other functions of MMR proteins

A number of different interacting partners of MLH1, PMS2 and PMS1 have been found,

indicating that the MutL homologue proteins play more roles in humans than are currently

known (Cannavo et al. 2007). MMR proteins are thought to function in the efficiency and

5´

Mismatch/ IDL

5´

ATP

ADP + Pi

5´

5´

ATP

ADP + Pi

Resyntehesisand ligation

5´

5´

MutSα MutLα PCNA RFC RPAExoI

21

fidelity of both meiotic and mitotic recombination, DNA-damage signaling, apoptosis and

cell-type specific processes, e.g. class-switch recombination and somatic hypermutation

(reviewed in Jiricny 2006).

MMR-deficient cells are much more resistant than MMR-proficient cells to cell death induced

by methylating agents such as N-methyl-N´-nitro-N-nitroguanidine (MNNG) and N-methyl-

N-nitrosourea (MNU) (Kat et al. 1993). These methylating agents cause DNA damage by

forming O6-methylguanine, which pairs with C or T during replication. MMR proteins

recognize the damage and mediate G2/M cell cycle checkpoint activation and apoptosis. In

MMR deficiency, DNA damage accumulates in cell, but does not trigger cell death (Kat et al.

1993; Hawn et al. 1995; D´Atri et al. 1998; Cejka et al. 2003).

Somatic hypermutation (SHM) and class-switch recombination (CSR) occur during antigen

stimulated B-cell differentiation (Cascalho et al. 1998; Li et al. 2004a). The SHM mechanism

creates extra variability in antibody genes by introducing point mutations or small deletions

into the variable (V) regions, and in CSR, immune cells can change the constant region of the

produced antibodies. In SHM, the activation-induced deaminase (AID) generates G•U

mismatches. The dU bases can be removed by uracil-N-glycosylase (UNG), creating abasic

sites that can be bypassed by error-prone DNA polymerases, which can generate a transition

or transversion mutation to a G•C base pair (Rada et al. 1998). The G•U mismatches are also

recognized by MMR proteins, resulting in A•T base mutations (Cascalho et al. 1998; Rada et

al. 1998; Li et al. 2006). In mice, MSH2 or MSH6 deficiency shows a decreased frequency of

SHM, whereas in MLH1 or PMS2-deficient mice the mutation frequency is less dramatic

(Rada et al. 1998; Martin & Scharff 2002). This suggests that MutSα promotes mutagenesis

in SHM. Differing from other MMR proteins, mouse studies suggest that MLH3 normally

plays an inhibitory role in SHM (Li et al. 2006). MMR proteins are also needed for efficient

CSR. Mouse studies have shown that MSH2 deficiency causes a 2- to 10-fold reduction in

isotype switching (Ehrenstein & Neuberger 1999; Schrader et al. 1999). Especially the

ATPase activity of MSH2 is thought to be important for CSR (Martin et al. 2003), and MSH6,

PMS2 and MLH1-deficient mice also show a reduction in switching (Schrader et al. 1999;

Ehrenstein et al. 2001; Li et al. 2004b).

Two genes homologous to MutS, MSH4 and MSH5, which have not been shown to have roles

in DNA repair, function in the early steps of meiotic recombination and may have some role

22

in crossover formation in mammals (Bocker et al. 1999; Kneitz et al. 2000; Snowden et al.

2004). Excluding PMS1-/- mice, knockout mice lacking MutL homologue genes are usually

infertile (Prolla et at 1998). PMS2 -/- male mice and both MLH1-/- and MLH3-/- mice of both

sexes are infertile (Prolla et al. 1998; Lipkin et al. 2002), suggesting a role in meiosis for

these proteins. Yeast Mlh3 and Mlh1 are needed for meiotic crossover products (Hunter &

Borts 1997, Wang et al. 1999, Argueso et al. 2004), and based on the localization studies and

phenotypes of MLH1 and MLH3 mutant mice, most probably also mammalian MLH1 and

MLH3 are required for crossover formation (Baker et al. 1996; Kolas & Cohen 2004).

Meiotic recombination is mediated at least in part by the MMR proteins MSH4-MSH5 and

MLH1-MLH3 (Lipkin et al. 2002, Santucci-Darmanin et al. 2002). MSH4-MSH5 might

stabilize Holliday junction intermediates (Snowden et al. 2004), and MLH1-MLH3 activates

and directs an unknown downstream factor (Hoffmann & Borts 2004).

Human MutL homologues Different MutL homologues and their function in MMR

MutL-like proteins in humans are MutL homologue 1 and 3 (MLH1 and MLH3) and post-

meiotic segregation proteins 1 and 2 (PMS2 and PMS1). These four proteins form three

different heterodimer complexes: MutLα (MLH1 and PMS2), MutLβ (MLH1 and PMS1) and

MutLγ (MLH1 and MLH3). In human MMR, MutLα is the most important MutL complex in

MMR (Li & Modrich 1995; Harfe & Jinks 2000; Kondo et al. 2001). It has been suggested

that the MutLγ complex may act as a backup for MutLα (Flores-Rozas & Kolodner 1998;

Chen et al. 2005; Cannavo et al. 2005). The MutLβ complex does not participate in MMR in

vitro (Räschle et al. 1999).

The functions of MutL homologues have been difficult to clarify. So far, MutL homologues

and especially MutLα are believed to play an essential role in MMR by functioning as

“molecular matchmaker” (Jiricny & Nyström-Lahti 2000). In support of, the lack of MLH1 or

PMS2 causes MMR deficiency both in vitro and in vivo (Lindblom et al. 1993; Li & Modrich

1995; Nicolaides 1998). The reconstitution of the 5´-directed MMR mechanism in vitro

occurs without MutLα (Genschel & Modrich 2003; Constantin et al. 2005, Zhang et al. 2005),

whereas 3´-directed reactions require MutLα (Constantin et al. 2005). This should mean that

cell extracts lacking MutLα are proficient in MMR when using 5´-nicked substrates, but this

23

is not the case (Nyström-Lahti et al. 2002; Tomer et al. 2002; Raevaara et al. 2002; 2003;

2004). A recent study has shown that MutLα harbors a latent endonuclease, which is activated

in a mismatch-dependent manner requiring MutSα, RFC, PCNA and ATP and provides the 5´

break for the excision initiation by ExoI (Kadyrov et al. 2006). This could partially explain

the activity seen in the reconstituted 5’-directed MMR, but not why cells lacking MutLα are

MMR deficient.

MLH1 is needed in all three MutL complexes and is the most important MutL homologue in

MMR. PMS2 is thought to be the major partner of MLH1 in the MMR mechanism (Lipkin et

al. 2001; Cannavo et al. 2005). The role of MLH3 in human cells is not so well known.

Mammalian MLH3 was first described by Lipkin et al. (2000). Mouse models suggest that

MLH3 deficiency causes microsatellite instability (Lipkin et al. 2000; Chen et al. 2005). The

amount of MLH3 in human cell lines is much lower than that of other MMR proteins,

approximately 60 times less than PMS2, whereas PMS1 is only ~ 10 times lower than PMS2

(Cannavo et al. 2005). Some studies indicate that MLH3 has some role in tumorigenesis.

Human MutLγ has been shown to repair G•T mismatches, but not IDLs, in vitro (Cannavo et

al. 2005). However, in yeast, the MLH1-MLH3 complex repairs only IDLs (Flores-Rozas &

Kolonder 1998), and also in mice, MLH3 deficiency increases the IDL mutation frequency

(Chen et al. 2008).

MLH1-/- knockout mice are MMR deficient and the rate of spontaneous mutations and tumor

development are increased (Prolla et al. 1996). Double-knockout mice that lack both MLH1

and PMS2 do not show a difference in mutator phenotype compared to MLH1-/- mice (Yao et

al. 1999). The mutation frequency of MLH1-/- mice is two to three-fold higher than that of

PMS2-/- mice. Furthermore, PMS2-/- mice expressed MLH1 normally, whereas MLH1-/- mice

did not express PMS2 (Yao 1999). The double knockout MLH3-/- ; PMS2-/- mice have more

severe phenotypes than either MLH3-/- or PMS2 -/- mice, and thus it has been suggested that

PMS2 and MLH3 might have overlapping functions (Chen et al. 2005).

Functional domains of MutL homologues

A crystal structure of full-length MutL is not available. There is a highly conserved ATPase

domain at the N-terminus and a less conserved C-terminal region, which contains the

dimerization domain (Ban & Yang 1998, Ban et al. 1999; Kondo et al. 2001; Wu et al. 2003;

24

Guarne et al. 2004). Ban et al. have determined the crystal structure of the 40-kDa N-terminal

part of E. coli MutL (LN40) and demonstrated that MutL binds and hydrolyzes ATP to ADP

(Ban & Yang 1998, Ban et al. 1999). MutL is a member of the GHKL (Gyrase, Hsp90,

histidine kinase and MutL) ATPase superfamily, which contains four short conserved

sequence motifs. These motifs might have an important role in ATP binding and hydrolysis

(Dutta &Inouye 2000). The ATPase activity of MutL is stimulated by DNA and probably acts

as a switch to coordinate DNA mismatch repair (Ban et al. 1999; Plotz et al. 2002).

Eukaryotic MutLα interacts with MutSα in an ATP-dependent manner and the interaction is

dependent on DNA (Blackwell et al. 2001; Plotz et al. 2002; Räschle et al. 2002).

It has been found that yeast MutLα can adopt four different conformations (extended, one-

armed, semicondensed and condensed) under varying conditions (different ADP or ATP

concentrations) and that human MutLα displays similar conformational changes (Sacho et al.

2008). Sacho et al. propose a model for the ATPase cycle of MutLα, where the binding of the

first ATP to MutLα, likely to MLH1, drives the formation of the one-armed state. The binding

of a second ATP condenses PMS2 forming the condensed state (or semicondensed state,

which is related to the condensed state by a conformational rearrangement of the protein). The

hydrolysis and release of both bound adenine nucleotides returns the protein to the extended

conformation. It has been suggested that the binding of ATP to one subunit of MutLα induces

the formation of a secondary structure element and promotes the interaction between a linker

arm and the N- and C-termini (Sacho et al. 2008). The interactions between MutSα and

MutLα are mediated by the N-teminal domain of MLH1 and are not linked to the ATPase

activity of MutLα (Räschle et al. 2002; Plotz et al. 2003).

The interaction domain of MLH1 is in the C-terminal part of the protein between residues 492

and 742 (Kondo et al. 2001). MLH1 can heterodimerize with PMS2 and MLH3 through this

particular interaction domain. The domain of PMS2, which interacts with MLH1 is located

either between residues 675-850 (Guerrette et al. 1999) or 612-674 (Kondo et al. 2001). An

MLH3 domain between residues 1399 and 1453 interacts strongly with MLH1 and is most

probably the interaction domain of MLH3. This region partially overlaps with domain 1390-

1424, which is homologous to one of the interaction regions of PMS2 (612-674). In MLH3,

two other regions (1-224 and 860-895) show weak interactions with MLH1 (Kondo et al.

2001).

25

The PMS2 protein contains the highly conserved metal-binding motif (DQHA(X)2E(X)4E)

implicated in the endonuclease activity of MutLα (Kadyrov et al. 2006). This region is also

found in MLH3 but not in MLH1 (Kadyrov et al. 2006; Nishant et al. 2008). Mutations in this

region in PMS2 (D699N and E705K) cause problems in Fe2+ binding and these mutants are

defective in ATP and Mn2+-dependent endonuclease activity (Kadyrov et al. 2006). Two other

conserved metal-binding motifs, ACR and C(P/N)HGRP, are found in the carboxy terminal

part of PMS2, and there is one further motif, FxR, in MLH1 (Kosinski et al. 2008). Kosinski

et al. suggested that DQHA(X) 2E(X) 4E, ACR, C(P/N)HGRP and FxR motifs together form

one functional site in MutLα. They also showed that PMS2 mutations in these regions

(D699N, H701Q, E705Q, C812S, C843S, H845Q and R847E) cause MMR deficiency.

Nuclear localization signals in MutL homologues

While all proteins are synthesized on ribosomes in the cytoplasm, nuclear proteins such as the

MMR proteins need to be imported into the nucleus in order to function. Selective and active

transport to the nucleus is mediated by a signal sequence called the nuclear localization signal

(NLS) (Dingwall & Laskey 1998). Nuclear import is mediated by a number of proteins that

cycle between the cytosol and the nucleus, e.g. importin α and β, which are soluble receptors

for nuclear proteins (Figure 3) (Görlich et al. 1995).

26

Figure 3. Nuclear transport. A protein containing an NLS is bound by a complex of importin α and β, which then binds to a nuclear pore. Translocation is mediated by the Ran GTPase. Inside the nucleus, importin β dissociates first from importin α, followed by importin α dissociation from the nuclear protein.

The first NLS was identified in the SV40 large T antigen (Kalderon 1984). NLS can be mono-

or bipartite, for example the sequence PKKKRKVE is monopartite. In bipartite NLSs, the

motifs are separated by a 10-12 amino acid stretch, as in KRPAAIKKAGQAKKKKLD. The

first bipartite NLS was identified in nuceloplasmin (Dingwall & Laskey 1998; Mattaj &

Englmeier 1998). The NLS usually consists of one or more positively charged amino acid

clusters and is located either in the C- or N- terminal region of a protein (Lange et al. 2007).

Human MutLα seems to contain at least two NLS sequences, one in MLH1 and one in PMS2

(Wu et al. 2003, Briger et al. 2005; Leong et al. 2009). Different results concerning the NLS

sequences of MLH1 have been observed. Briger et al. found a monopartite NLS sequence in

both MLH1 and PMS2 (Briger et al. 2005), whereas Leong et al. found a bipartite NLS in

MLH1 and a monopartite NLS in PMS2 (Leong et al. 2009). In addition, the localization of

MutL homologues is controversial. Early studies showed that MLH1 and PMS2 dimerize in

Nuclear protein

NLS Cytoplasm

Nucleus

Importin β

Importin αNuclear envelope

GTP GDP + Pi

RanNuclear pore complex

27

the cytoplasm and that the nuclear localization of MutLα is dependent on this dimerization

(Wu et al. 2003). A recent study, however, suggests that both PMS2 and MLH1 are capable

of translocating to the nucleus alone (Leong et al. 2009). The presence of NLS regions in

MLH3 is not known.

Hereditary nonpolyposis colorectal cancer syndrome

Hereditary colorectal cancer

Cancer is one of the most common causes of death in the world. In Finland, every third or

fourth person will be affected by cancer in their lifetime and the risk of cancer increases with

age. In Finland, there were approximately 27,000 new cancer cases in the year 2006

(www.cancerregistry.fi), and in the US the estimated number in the year 2008 was nearly

1,500,000 (Jemal et al. 2008). The five-year survival rate from cancer diagnosis is about 66%

(Jemal et al. 2008, www.cancer.org, www.cancerregistry.fi). The three most common cancer

types in Finnish men are prostate, lung and colorectal, and in women these are breast,

endometrial and colorectal cancers (www.cancerregistry.fi). The same cancer types are

relevant in all Western countries.

Normally cancer development takes time and requires a large number of changes in the

genome. Changes during tumorigenesis involve both the activation of oncogenes and

inactivation of tumor suppressor genes. Tumor suppressor genes are a large group of genes

which slow down cell division, repair DNA mistakes and/or induce apoptosis. The loss of

tumor suppressor genes can occur either through genetic mutation or the epigenetic silencing

of genes and leads to uncontrolled growth. Knudsons “two hit” theory is based on the idea

that both copies of a given tumor suppressor gene have to be inactivated before tumorigenesis

begins (Knudson 1971; Knudson 1996). In hereditary cancer syndromes, one mutant allele

(the “first hit”) is inherited, which means that the person is born with one mutated and one

wild type copy of the cancer susceptibility gene. Tumor formation begins after the

inactivation of the wild type allele (the “second hit”).

Colorectal cancer (CRC) is globally the third most common type of cancer and one of the

most common causes of cancer-related deaths in Western countries (www.cancerregistry.fi,

28

www.cancer.org). Cancer itself is not hereditary, but in hereditary cancer syndromes the risk

of developing cancer is higher than normal. When one mutation is inherited, cancer can

develop much faster than in sporadic cases and it is the reason why the age of cancer onset is

usually much lower in hereditary cancer syndromes. Other typical features of inherited cancer

syndromes are numerous affected family members and multiple primary tumors. However,

the frequencies of hereditary cancers are low compared to sporadic cancers. Approximately

10-15% of all colorectal cancer patients have a family history of this type of cancer and 5% of

patients have early-onset colorectal cancer (Vasen 2007). Environmental factors, for example

a high-fat diet or long-term smoking, are known to be associated with an increased risk for

colorectal cancer and only in a small fraction of cases do genetic factors play a dominant role

(Potter 1999; Weitz et al. 2005; Vasen 2007).

Hereditary nonpolyposis colorectal cancer (HNPCC or Lynch syndrome) and Familial

adenomatous polyposis (FAP) are the main inherited colorectal cancers. HNPCC accounts for

2-5% of all CRC cases (Mecklin 1987; Aaltonen et al. 1998; Lynch & de la Chappelle 2003;

Hampel et al. 2005) and is associated with germline mutations in MMR genes (Aaltonen et al.

1998; Lynch & de la Chapelle 2003). FAP accounts for ~ 1% of all CRCs and is associated

with germline mutations in the APC gene (Groden et al. 1991; Bisgaard et al. 1994). Other

hereditary colorectal cancers (e.g. MYH-associated polyposis, Juvenile polyposis, Peutz-

Jeghers and Cowden syndrome) together account for less than 1% of all CRC cases (Lynch et

al. 2008)

Diagnostic criteria for HNPCC

The most common hereditary colon cancer is HNPCC and it was also the first hereditary

cancer syndrome to be discovered, being described already in the mid-1960s (Lynch et al.

1993). HNPCC is caused by a germline mutation in one of the following MMR genes: MLH1,

MSH2, MSH6, PMS2 or MLH3 (Peltomäki & Vasen 2004; Peltomäki 2005). A deficiency of

the MMR mechanism leads to a mutator phenotype. The progression from adenoma to

carcinoma in HNPCC patients may take less than 3 years, whereas in sporadic cases it takes

10-15 years (Vasen et al. 1996). While the most common tumors in HNPCC are colorectal

and endometrial tumors, the risk for a broad spectrum of extra-colonic cancers like stomach,

ovarian, small bowel, hepatobiliary tract, pancreatic, ureter and renal pelvis cancer is also

increased (Watson & Lynch 1993; Aarnio et al. 1999). HNPCC is inherited in an autosomal

29

dominant manner and mutation carriers have a 50-80% lifetime risk of developing CRC, a 50-

60% risk of endometrial cancer and less than a 15% risk of other tumors (Watson & Lynch

1993; Aarnio et al. 1999). The average age at CRC onset in HNPCC is ~45 years, which is

~20 years earlier than in sporadic cases (Watson & Lynch 1993; Lynch & de la Chapelle

2003). The characteristic for typical HNPCC is high microsatellite instability (MSI) in tumors

due to a deficient MMR mechanism (Aaltonen et al. 1993).

Table 2: The international criteria for the diagnosis of HNPCC

Amsterdam criteria (Vasen et al. 1991; 1999) Revised Bethesda guidelines (Umar et al. 2004)

At least three relatives with HNPCC-

associated cancer (in colorectum,

endometrium, small bowel, ureter and

renal pelvis)

One should be a first-degree relative of

the other two

At least two successive affected

generations

At least one HNPCC syndrome-

associated cancer diagnosed under 50

years of age

FAP should be excluded

Tumors should be verified by

pathological examination

Colorectal cancer diagnosed in a patient

before 50 years of age

Presence of synchronous, metachronous

colorectal, or other HNPCC associated

tumors, regardless of age

Colorectal cancer with the MSI-H

histology diagnosed in a patient before

60 years of age

Colorectal cancer diagnosed in one or

more first-degree relatives with an

HNPCC-associated tumor, with one of

the cancers diagnosed before age 50

Colorectal cancer diagnosed in two or

more first- or second-degree relatives

with HNPCC-associated tumors,

regardless of age

The first international criteria for diagnosis of HNPCC, the Amsterdam criteria I (AC I), were

developed in 1991 (Vasen et al. 1991). Later those criteria were modified to the Amsterdam

criteria II (AC II) (Vasen et al. 1999) (Table 2). AC II contains a wider spectrum of HNPCC-

related cancers than AC I. The Bethesda guidelines are other diagnostic criteria for HNPCC

(Rodriguez-Bigas et al. 1997) (Table 2), and these guidelines have later been modified (Umar

et al. 2004). The Bethesda guidelines are used for indentification of CRC patients who should

be tested for MSI. The Amsterdam criteria and Bethesda guidelines together form good

diagnostic rules for finding HNPCC families. The typical HNPCC phenotype is usually

associated with MLH1 and MSH2 mutations. Of the families that fulfill AC I, 63% have a

30

susceptibility mutation in MLH1, 50% in MSH2 and less than 20% in MHS6 or PMS2

(Peltomäki & Vasen 2004).

HNPCC and MMR

MMR gene mutations

HNPCC is linked to MMR malfunction. The germline mutations found in HNPCC patients

are in the MMR genes MLH1, MSH2, MSH6, PMS2 or MLH3. Overall, MLH1 and MSH2

cover the major part (~90%) of HNPCC mutations (Peltomäki 2005; Woods et al. 2007).

Over 1500 different variants associated with HNPCC have been identified. By February 2007,

659 variants in MLH1 (44% of all identified MMR gene), 595 in MSH2 (39%), 216 in MSH6

(14%), 45 in PMS2 (3%) and 34 in MLH3 had been published (Woods et al. 2007;

http://www.insight-group.org).

The majority of MMR gene mutations are specific to only one HNPCC family (Peltomäki

&Vasen 2004). Most of these mutations are nonsense or frame-shift mutations, causing

truncation and loss-of-function of the polypeptide. However, a significant proportion of

HNPCC-related mutations are nontruncating; 32% of MLH1, 18% of MSH2 and 38% of

MSH6, 49% of PMS2 and 87% of MLH3 mutations are of this kind (Peltomäki & Vasen 2004,

Woods et al. 2007), and the proportion of these mutations is growing all the time. The

interpretation of nontruncating alterations in MMR genes is a major challenge to cancer

diagnostics and genetic counseling, which rely on the assessment of sequence variants found

in cancer patients. The effect of a nontruncating mutation can vary from none to the complete

loss of protein function.

Due to the large amount of different genetic variations in MMR genes, different databases to

store this information have been created. The first HNPCC mutation database was created by

the International Society for Gastrointestinal Hereditary Tumors (InSiGHT)

(http://www.insight-group.org), and it relies on entries of original data from investigators.

Another database was created by Woods et al., who have assembled all published MMR

mutations in one database (Woods et al. 2007; http://med.mun.ca/MMRvariants). Information

on the functional assays of different missense variants are collected in the Mismatch Repair

31

Gene Unclassified Variants Database (MMRUV) (Ou et al. 2008: http://www.mmruv.info).

The different databases contain partially overlapping information, and lately an effort to

merge the information in these three databases with the Leiden Open Variation Database

(LVOD) has been started. However, the conjunction of the data in these three databases is not

finished yet and the LVOD contains some of the information on some mutations twice. These

duplications complicate the estimation of different MMR variants.

Mutations in MutL homologue genes

MLH1 is the most common susceptibility gene in HNPCC. In the year 2007, 659 different

MLH1 variants located throughout the MLH1 polypeptide were published (Woods et al.

2007). Missense mutations cover 24%, insertion/deletion mutations 22%, and large genomic

deletions or duplications covers 25% of all found MLH1 mutations

(http://www.med.mun.ca/MMRvariants). The MLH1 mutations are mostly associated with

typical HNPCC phenotypes (Liu et al. 1996). Founder mutations, similar mutations affecting

several families in a typical geographical area, are rare. However, in Finland the majority of

found HNPCC mutations are founder mutations, such as a splice-site mutation in MLH1 exon

6 and a big genomic deletion eliminating exon 16 (Nyström-Lahti et al. 1995; Moisio et al.

1996; Aaltonen et al. 1998). Nevertheless, the growing amount of found missense mutations

from all over the world is a challenge for genetic counseling and HNPCC diagnosis.

In sporadic CRC with high MSI (MSI-H), MLH1 is often silenced via promoter

hypermethylation, and it correlates well with the loss of the MLH1 protein in these sporadic

tumors (Kane et al. 1997; Cunningham et al. 1998). Epigenetic changes in MLH1 have also

been reported, and most of those individuals were suspected of being HNPCC patients

(Gazzoli et al. 2002; Miyakura et al. 2004; Suter et al. 2004; Hitchins et al. 2007; Morak et

al. 2008; Gylling et al. 2009). Epigenetic changes are caused by promoter hypermethylation

and differ from germline mutations in that epigenetic changes can revert to the normal state or

cause mosaicism, which is a common feature in epigenetics. Some studies provide evidence

for the heritabilitiy or transgenerational inheritance of epigenetic changes (Chan et al. 2006;

Hitchins et al. 2007; Morak et al. 2008). There are also studies where no evidence of

32

heritability has been found (Miayakura et al. 2004; Suter etal 2004; Hitchins et al. 2005;

Valle et al. 2007; Joensuu et al. 2008; Morak et al. 2008)

Eighteen putative pathogenic MLH3 mutations, found in cancer patients, have been published

(Wu et al. 2001; Liu et al. 2003: 2006; Taylor et al. 2006; Kim et al. 2007) and 34 different

mutations are reported in database (http://www.insight-grop.org). The majority of found

MLH3 mutations are of the missense type (~90%), and the rest are frame-shift mutations

(Peltomäki & Vasen 2004; http://www.insight-group.org). The majority of found MLH3

mutations are located in exon 2 (http://www.insight-group.org). The pathogenicity of these

MLH3 mutations is not clear. Some studies suggest that MLH3 is a susceptibility gene for

HNPCC. In these studies, some of the MLH3 mutations were found in patients lacking the

typical molecular characteristics of the HNPCC syndrome and/or without a family history

suggestive of inherited cancer susceptibility, whereas others were associated with the HNPCC

phenotype and MSI-H in the tumors (Wu et al. 2001; Liu et al. 2003; 2006; Taylor et al.

2006; Kim et al. 2007). The presence of a simple repeat sequence in the MLH3 coding region,

as in the MSH6 and MSH3 genes, might indicated that MLH3 itself is a MSI target gene, and

that MLH3 mutations may be secondary to other MMR gene mutation (Akiyama et al. 2001).

It has recently been shown that PMS2 mutations might have a bigger role in cancer

susceptibility than was earlier thought (Clendenning et al. 2006). Due to there being many

pseudogenes in the same chromosome region, PMS2 mutations have been difficult to screen

(Nicolaides et al. 1995a; 1995b; Nakagawa et al. 2004). Now, when new methods have made

it possible to find PMS2 mutations (Clendenning et al. 2006; Etzler et al. 2008; Senter et al.

2008), the amount of found mutations has increased. In the year 2007, 45 PMS2 mutations

had been published (Woods et al. 2007), and new variants have been found since (Jeske et al.

2008; Senter et al. 2008). The majority of PMS2 mutations are of the missense type (49%),

and large genomic deletions or duplications cover 19% of all PMS2 mutations

(http://www.med.mun.ca/MMRvariants). The majority of the found PMS2 mutations are

located in exon 11 (http://www.insight-group.org). However, the penetrance in PMS2

mutation carriers seems to be lower than in MLH1 and MSH2 mutation carriers (Truninger et

al. 2005; Auclair et al. 2007; Senter et al. 2008).

Also PMS1 was earlier belived to be an HNPCC susceptibility gene. However, only one

PMS1 mutation has been published (Nicolaides et al. 1994). A later study found that the same

33

family contained a large MSH2 deletion, which most probably is the predisposing mutation in

that family (Liu et al. 2001). This suggests that PMS1 is not an HNPCC susceptibility gene.

Clarifying the pathogenicity of nontruncating MMR mutations

MSI and IHC in HNPCC diagnostic

A low age of cancer onset, a family history of cancer and multiple primary tumors in the

colon or endometrium indicate HNPCC. Identification of CRC patients carrying an MMR

gene defect is important for HNPCC diagnosis, design of appropriate treatment, counseling

and follow-up. Characteristic for MMR defects in tumors are MSI-H and loss of protein.

However, MSI and loss of MLH1 are also found in sporadic CRC cases due to the

methylation of the MLH1 promoter (Kane et al. 1997). However, sporadic methylated MLH1

tumors probably have a different etiology from HNPCC tumors with an MLH1 germline

mutation (Hofstra et al. 2008).

Generally, the Bethesda panel, containing five microsatellite markers, is used in MSI analysis.

If two or more markers are unstable, the tumor is classified as MSI-H (Boland et al. 1998).

Immunohistochemistry (IHC) is another method used for detecting MMR deficiency in a

tumor/tumors. Analysis of the expression of four genes (MLH1, MHS2, MHS6 and PMS2)

give a good indication of the mutated MMR gene (Hampel et al. 2005). However, MMR

proteins function as heterodimers and MSH6 and PMS2 degrade without their counterpart

(Chang et al. 2000). For example, the loss of MSH2 leads to loss of MSH6 and similarly, the

loss of MLH1 is accompanied by loss of PMS2 (Jass 2008). However, the expression of a

protein in a tumor does not always mean that the protein is functional, and the mutant protein

may cause an MMR defect even if it is expressed in the tumor tissue (Mangold et al. 2005).

In silico comparative analyses

Different computational analyses have been created to predict the pathogenicity of missense

mutations. In silico prediction is based on comparative sequence or protein structure analysis.

Most of the current algorithms rely on protein multiple sequence alignments (PMSAs) of the

gene of interest across multiple species. For use with alignment-based tools, the PMSAs must

be sufficient in size and carefully constructed and curated so that the sequence data can be

used properly and is biologically sound (Tavtigian et al. 2008). Many different in silico

34

programs are available, e.g. sorting intolerant from tolerant, SIFT (Ng & Henikoff 2001);

polymorphism phenotyping, PolyPhen (Ramensky et al. 2002); PMut (Ferrer-Costa et al.

2005); and multivariate analysis of protein polymorphism, MAPP (Stone & Sidow 2005). In

silico analyses are a valuable tool for classification of mutations when in vitro or genetic data

on the pathogenicity of a mutation is not available. However, the validation of computer-

based methods requires the simultaneous use of in silico and functional analyses of mutations.

This validation helps to quantify the degree to which in silico analyses can be used as a

clinical tool (Chan et al. 2007).

Functional tests

Different biochemical assays have been developed to investigate whether the nontruncating

variants found in HNPPC or CRC patients cause functional defects. Assays can be divided

into two groups: 1) assays that analyze the MMR repair capacity of the mutated protein as a

complete process in vivo or in vitro, and 2) assays that test specific biochemical functions of

MMR proteins (Ou et al. 2007).

Yeast-based in vivo functional assays are based on the fact that the MMR system is

evolutionarily conserved. The found MMR gene mutation can be introduced into the yeast

MMR ortholog and the consequent mutation rate can be determined (Drotschmann et al.

1999, Shcherbakova & Kunkel 1999, Gammie et al. 2007). The yeast test can also be based

on the fact that functional human MMR proteins interact with the yeast MMR system and

interfere with its function and cause a mutator phenotype (Clark et al. 1999, Shimodaira et al.

1998, Takahashi et al. 2007). Yeast-based in vivo tests are limited to variants in conserved

regions of the MMR homologue protein (Takahashi et al. 2007).

The in vitro assays measure the DNA repair capacity of human mutant MMR proteins and are

not limited to conserved amino acids. In these assays, an artificial substrate containing a

mismatch is incubated with human MMR-deficient cell extract, which is complemented with

the mutated MMR protein. The successful repair of the substrate can be determined by the use

of appropriate restriction enzymes that uniquely cleave the repaired molecule (Lahue et al.

1989; Holmes et al. 1990; Nyström-Lahti et al. 2002). Another in vitro assay uses a

bacteriophage heteroduplex containing a mismatch in the lacZ α-complementation domain as

the substrate. In this in vitro assay, the repaired substrate changes the reading frame of the

lacZ gene (Thomas et al. 1991; Marra et al. 1998; Trojan et al. 2002).

35

Several methods are available for detecting a specific biochemical function of mutated MMR

proteins. The expression level of a mutated protein might be lower than the wild type protein

(Brieger et al. 2002; Nyström-Lahti et al. 2002; Raevaara et al. 2002; Trojan et al. 2002;

Ollila et al. 2006). Interaction assays are important, since MutSα, MutSβ, MutLα and MutLγ

function as heterodimers. The interactions can be detected by several methods: the yeast two-

hybrid (Kondo et al. 2003), GST pull-down (Guerrette et al. 1998; 1999) and

immunoprecipitation assays (Nyström-Lahti et al. 2002; Raevaara et al. 2002; Kariola et al.

2004; Ollila et al. 2006). The capacity of the MutS heterodimer to bind or release mismatched

DNA can be measured by a bandshift assay or electrophoretic mobility shift assay (EMSA)

(Clark et al. 1999; Drotchmann et al. 1999; Heinen et al.2002, Ollila et al. 2008a; 2008b).

Protein localization experiments show in vivo whether the protein has a problem in nuclear

localization (Brieger et al. 2005; Gammie et al. 2007).

36

AIMS OF THE STUDY

The main aim of this PhD study was to functionally characterize MutL homologue proteins

and their variants. The specific aims were:

1. To analyze the subcellular localization of the MutL homologue proteins MLH1,

MLH3 and PMS2 to estimate their roles in MMR (II)

2. To determine the functional significance and clinical phenotypes of MLH1 germline

missense mutations (I, IV)

3. To address the question of whether and how MLH3 germline missense mutations

cause susceptibility to HNPCC (III)

37

MATERIALS AND METHODS

MLH1, MLH3 and MSH2 mutations and associated families (I, III and IV)

The mutations included in this study consisted of thirty MLH1 and seven MLH3 missense

mutations and six small in-frame deletions in MLH1. Thirty-three MLH1 alterations were

found in putative HNPCC families and collected for functional studies through an

international collaboration, and three alterations were selected from the InSight database

(http://www.insight-group.org) (I, IV). The genetic and clinical data on the seven MLH3

mutations were collected from two published studies (Wu et al. 2001; Liu et al. 2003). These

mutations were found in colorectal or endometrial cancer patients and suggested to be

pathogenic (III). In addition to these MLH1 and MLH3 mutations, eight MSH2 mutations

found in colorectal cancer patients were also included in this study (IV). The genetic and

clinical data on all 51 mutations are collected in Table 3. All studies were approved by the

institutional review boards of the collaborating universities or local ethical committees.

MLH1 mutations were found in 58 putative HNPCC families, of which half did not fullfill the

Amsterdam criteria I or II. The mean age of index patients at cancer onset varied between 19

to 76 years. Fifty index patients had colorectal carcinoma wheras only four index patients had

endometrial carcinoma. The MSI phenotypes were high in 39, low in 4, and stable in 2

studied tumors (Table 3). The MSI status of the tumor was not available for 10 index patients.

IHC staining showed loss of MLH1 protein in 24 and decreased expression in 4 tumors (Table

3). In 15 cases the IHC staining of MLH1 was not available. We also included in the study

three MLH1 variations which were reported in the international HNPCC mutation database

(www.insight-group.org): 1) I219V, which was shown to be nonpathogenic in previous

studies (Shimodaira et al. 1998; Ellison et al. 2001; Trojan et al. 2002; Kondo et al. 2003)

and was used here as a functional control 2) K681T, the pathogenicity of which remained

partly unsettled (Guerrette et al. 1998; Shimodaira et al. 1998; Trojan et al. 2002; Kondo et

al. 2003) and 3) del633-663, comprising exon 17, which was previously found to be

pathogenic (Nyström-Lahti et al. 2002) and was used here as a nonfunctional control.

38

Table 3. Genetic and clinical data of MMR gene alterations under study.

MLH1 mutaion cDNA change Amino acid

change

AC I/II1

index patient

age/cancer

MSI status2

Immunohistochemistry Other MMR gene mutation found in the

family MLH1 MSH2 MSH6

P28L c.83C>T Pro→Leu + 30/CRC High NA NA NA P28L c.83C>T Pro→Leu - 27/CRC High No loss No loss NA A29S c.85G>T Ala→Ser + 37/CRC High Loss NA NA MLH1,

g -27C>A TSI45-47CF c.133-

141delACAAGTATT ins TGTTTT

Thr, Ser, Ile→

Cys, Phe

+ 32/CRC NA NA No loss NA

D63E c.189C>A Asp→Glu - 44/CRC High Decr. No loss NA G67R c.199G>A Gly→Arg - 36/CRC High NA NA NA del 71 c.211-213delGAA del Glu - 24/CRC High Loss No loss NA del 71 c.211-213delGAA del Glu + 19/CRC High NA NA NA C77R c.229T>C Cys→Arg + 22/CRC High NA NA NA F80V c.238T>G Phe→Val + 51/CRC High No loss No loss NA K84E c.250A>G Lys→Glu - 32/CRC High NA NA NA S93G c.277A>G Ser→Gly + 70/CRC NA NA NA NA I107R c.320T>G Ile→Arg + 46/EC High Loss NA NA I107R c.320T>G Ile→Arg + 53/CRC High Loss NA NA I107R c.320T>G Ile→Arg + 46/CA NA NA NA NA I107R c.320T>G Ile→Arg + 30/CRC High Loss NA NA L155R c.467T>G Leu→Arg + 33/CRC High Loss No loss NA V187G c.554T>G Val→Gly + 43/CRC High Decr. No loss No loss V213M c.637G>A Val→Met - 64/CRC Low No loss No loss No loss V213M c.637G>A Val→Met - 44/EC High No loss Loss Loss V213M c.637G>A Val→Met - 67/CRC High Loss No loss No loss V213M c.637G>A Val→Met - 46/CRC High No loss No loss No loss I219V3 c.655A>G Ile→Val S247P c.739T>C Ser→Pro - 43/CRC NA NA NA NA S247P c.739T>C Ser→Pro + 42/CRC High Decr. No loss NA H329P c.986A>C His→Pro + 32/CRC High Loss No loss NA Idel330 c.988-990delATC del Ile + 29/CRC High Loss No loss No loss K443Q c.1327A>C Lys→Gln - 57/CRC High Loss No loss No loss E460A c.1379A>C Glu→Ala + 34/CRC NA No loss Loss Loss MSH2,

exon 8 del E460A c.1379A>C Glu→Ala + - NA NA NA NA MSH2, M663fs L550P c.1649T>C Leu→Pro - 45/CRC High Loss No loss NA A589D c.1766C>A Ala→Asp - 34/CRC High Loss No loss NA

V612del c.1834-1836delTTG del Val + 47/CRC NA NA NA NA K616del c.1846-1848delAAG del Lys + 44/CRC High Loss No loss No loss K618A c.1852-1853AA>GC Lys→Ala - 43/CRC High No loss No loss NA K618A c.1852-1853AA>GC Lys→Ala - 33/CRC NA NA NA NA K618A c.1852-1853AA>GC Lys→Ala - 76/CRC Low Loss No loss No loss K618A c.1852-1853AA>GC Lys→Ala - 72/CRC High Loss No loss No loss K618A c.1852-1853AA>GC Lys→Ala - 69/CRC Low No loss No loss No loss K618A c.1852-1853AA>GC Lys→Ala + 44/CRC MSS No loss No loss NA K618A c.1852-1853AA>GC Lys→Ala + 32/CRC High Loss No loss No loss MLH1, R659Q K618T3 c.1853A>C Lys→Thr

del633-6633 c.1897-1989del(exon17)

Del of aa 633-663

Y646C c.1937A>G Tyr→Cys - 36/CRC High No loss No loss No loss P648L c.1943C>T Pro→Leu - 43/CRC High No loss No loss NA P648S c.1942C>T Pro→Ser + 54/CRC High Loss No loss No loss P654L c.1961C>T Pro→Leu - 35/CRC NA NA No loss NA P654L c.1961C>T Pro→Leu - 31/CRC NA Loss NA NA P654L c.1961C>T Pro→Leu - 38/CRC High Loss No loss NA P654L c.1961C>T Pro→Leu + 41/CRC High Loss No loss NA R659P c.1976G>C Arg→Pro + 35/CRC High Loss NA NA R659Q c.1976G>A Arg→Gln + 32/CRC High Loss No loss No loss A681T c.2041G>A Ala→Thr - 38/CRC High NA No loss NA R687W c.2059C>T Arg→Trp - 48/EC High Loss/Red. No loss Red. V716M c.2146G>A Val→Met - 65/CRC Low No loss No loss No loss V716M c.2146G>A Val→Met - 39/EC High Loss No loss No loss V716M c.2146G>A Val→Met + 43/CRC High Loss No loss NA V716M c.2146G>A Val→Met + 52/CRC MSS NA No loss NA

39

MLH3 mutation cDNA change Amino acid

change AC I/II

index patient age/cancer MSI

Immunohistochemistry Other

MMR gene mutation

found in the family

MLH1 MSH2 MSH6

Q24E c.70C>G Gln-→Glu + 50/CRC, 50/BL

NA No loss No loss No loss

R647C c.1939C>T Arg→Cys + 49/EC, 49/OV NA No loss No loss NA R647C c.1939C>T Arg→Cys - 54/ EC High NA NA NA S817G c.2449A>G Ser→Gly + 46/EC, 46/OV NA No loss No loss Loss MSH6,

IVS9+43ins10bp

G933C c.2797G>T Gly→Cys - 61/CRC MSS No loss No loss No loss W1276R c.3826T>C Trp→Arg + 29/CRC MSS NA NA NA MSH2,

E198G W1276R c.3826T>C Trp→Arg - CRC NA NA NA NA A1394T c.4180G>A Ala→Thr - 44/CRC NA No loss No loss No loss E1451K c.4351G>A Glu→Lys + 45/CRC NA No loss No loss No loss MSH6,

V878A E1451K c.4351G>A Glu→Lys + 41/CRC NA No loss No loss Loss MSH6,

650insT E1451K c.4351G>A Glu→Lys - 76/CRC MSS No loss No loss No loss MSH2

mutation

T44M c.131C>T Thr→Met - NA/colon

(adenomas) NA NA NA NA

A45V c.134C>T Ala→Val - 45/CRC NA NA NA NA L187R c.560T>G Leu→Arg + 31/CRC High No loss Loss Loss F519L c.1555T>C Phe→Leu + 40/CRC MSS No loss No loss No loss M688V c.2062A>G Met→Val - 54/CRC NA No loss No loss No loss

M688V c.2062A>G Met→Val + 46/CRC High Loss No loss NA

MLH1, T117M MSH6,

A1889V V722I c.2164G>A Val→Ile + 30/CRC MSS NA No loss NA A848S c.2542G>T Ala→Ser - NA/CRC MSS NA NA NA

E886G c.2657A>G Glu→Gly + 34/CRC NA No loss No loss No Loss

/Red MSH6 (NA)

E886G c.2657A>G Glu→Gly + 41/CRC NA No loss Loss/Red. Loss MSH6 (NA)

References are in the original articles I, III and IV.1AC, Amsterdam criteria I/ II; 2MSI status of the tumor from an index patient if available. MSI analysis was carried out using the Bethesda panel (Markers BAT25, BAT26, D2S123, D5S346 and D17S250 and in some cases D18S69). Tumors with two or more unstable markers were considered as MSI-H. * MSI was examined with markers D3S283, D9S171, D17S250, D3S2456 and D6S1279; 3Mutations from international database (http://www.insight-group.org); Red, reduced; CRC, colorectal cancer; EC, endometrial cancer; BL, bladder cancer; OV, ovarian cancer; MSS, microsatellite stable; NA, not available The MLH1 mutations are scattered throughout the MLH1 polypeptide (Figure 4). Mutation

clusters can be seen in two regions, in the ATP-binding and hydrolysis region in the amino

terminus (Ban et al. 1999, Tran & Liskay 2000, Räschle et al. 2002) and in the interaction

region the carboxy terminus (Guerrette et al. 1999, Kondo et al. 2001). Although the MMR

functionality of the MLH1 protein mutations del71, C77R, S93G, I107R, del 616, del633-663,

P648S and R659P had already been tested in previous studies (Nyström-Lahti et al. 2002;

Raevaara et al. 2002; 2003 and 2004), they were included in the present study for further

functional characterization.

40

Figure 4. Schematic representation of the studied MLH1 mutations and known functional domains along the MLH1 polypeptide (Ban et al. 1999; Tran et al. 2000; Räschle et al. 2002; Guerrette et al. 1999; Kondo et al. 2001).

Of the seven MLH3 mutations, three are located in known or putative functional domains

(Figure 5). MLH3 Q24E is in the putative ATP-binding and hydrolysis region in the amino

terminus (Ban et al. 1999) and the mutations A1394T and E1451K in the MLH3/MLH1

interaction domain in the carboxy-terminal region (Kondo et al. 2001). The MLH3 mutations

were found in 11 separate cancer families (Table 3), of which half fulfilled the Amsterdam

criteria I or II. Mutations in MLH1, MSH2 and MSH6 were also studied in these MLH3

mutation carriers and four of them (S817G, W1276R and two patients which E1451K from

unrelated families) also showed a mutation in another MMR gene, either MSH2 or MSH6.

The MSI status was stable in carriers with the mutations G933C, W1276R or E1451K (Liu et

al. 2003), whereas R647C was associated with the MSI-H phenotype (Taylor et al. 2006). In

the Wu et al. (2001) study, 77 % of tumors (including tumors associated with the mutations

Q24E, R647C, S817G, A1394T and E1451K) showed the MSI-H phenotype with the marker

panel D3S1283, D9S171, D17S250, D3S2456, and D6S1279, whereas using the Bethesda

panel, only 39% of tumors showed the MSI-H phenotype. The MSI status of the index

patients was not specified. The expression of MLH3 or PMS2 was not studied in the tumors.

In the two patients with compound heterozygosity of MLH3 and MSH6 mutations, MSH6

expression was also lost in the tumors. Loss of heterozygosity of MLH3 was found in two

tumors (Q24E and A1394T) (Wu et al. 2001).

41

Figure 5. Schematic representation of the studied MLH3 mutations along the MLH3 polypeptide and known or putative functional domains (Ban et al. 1999; Kondo et al. 2001)

Protein expression (I-IV)

Mutagenesis and expression vectors (I-IV)

All studied mutations were generated in human MLH1 or MLH3 complemetary DNA

(cDNA), cloned between the BamHI and NotI sites of the pFastBacI plasmid (Invitrogen) or

human MSH2 cDNA cloned between the BamHI and XhoI sites of the pFastBacI plasmid,

using PCR-based site-directed mutagenesis as detailed in the original articles. The mutated

sites in MLH1 or MSH2 wild type (wt) cDNA were created using two primer pairs, forward-A

with reverse-A and forward-B with reverse-B. The correct fragment sizes were verified by

agarose gel electrophoresis. In the second PCR, the A and B fragments were used as

templates, with the primers forward-A and reverse-B. The second PCR product, which

contained the mutations, was cloned into the original plasmid between the appropriate

restriction sites. The MLH3 mutations were created using the PhusionTM site-directed

mutagenesis kit according to the manufacturer´s instructions (Finnzymes). All constructs were

verified by sequencing (ABIPrism 3100 Genetic Analyser, Applied Biosystems).

For protein expression in human cells and for immunofluorescence, MLH1 (wt or mutated),

PMS2 (wt) and MLH3 (wt) cDNAs were cloned from the pFastBac1-plasmid into the vector

42

pEGFP-N1 (BD Biosciences) between the BamHI and NotI sites, so that the enhanced green

fluorescent protein (EGFP) gene was replaced. These constructs were named as MLH1-N1,

PMS2-N1 and MLH3-N1, recpectively. For direct fluorescent protein production, MLH1 (wt

or mutated) and PMS2 (wt) cDNAs were fused to the EGFP gene. For cloning, stop codons

were removed and the appropriate restriction sites were generated in the MLH1 and PMS2

cDNAs by site-directed mutagenesis using pFastBacI-MLH1 and pFastBac1-PMS2 as

templates. An NheI restriction site was generated upstream of the MLH1 start codon and a

SacI site was introduced to replace the MLH1 stop codon. The products were cloned into the

pFastBasI-MLH1 vector between either the BamHI and PvuII or NsiI and NotI sites,

respectively. In PMS2, the stop codon was replaced by an AgeI restriction site and the product

was cloned into the pFastBacI-PMS2 vector between its SpeI and NotI sites. The MLH1 and

PMS2 cDNAs without stop codons were cloned into the pEGFP-N1 vector (BD Bioscience)

between the NheI and SacI and BamHI and AgeI sites. The resulting constructs were named

MLH1-EGFP and PMS2-EGFP. The cloning sites of MLH1 mutations are detailed in the

original articles.

Baculoviral protein expression (I, III, IV)

The recombinant baculoviruses for protein production in Spodoptera frugiperda 9 (Sf9) insect

cells were generated using the Bac-to-Bac system (Invitrogen). Sf9 cell were grown at 27°C in

Grace´s Insect Medium (Invitrogen/GIBCO) in the presence of 10% fetal bovine serum, 2

mM L-glutamine and 50 units/ml of penicillin-streptomycin. The cDNAs of MLH1 (wt or

mutated), MLH3 (wt or mutated), MSH2 (wt or mutated), PMS2 (wt) and MSH6 (wt) were

transformed to DH10Bac E. coli cells (Invitrogen), which contain bacmid DNA with a mini-

attTn7 target site and a helper plasmid. The isolated bacmid DNA was used to transfect Sf9

cells. The baculoviruses were collected after 3 days and amplified in Sf9 cells for 5 days. For

protein production, Sf9 were co-infected with two recombinant baculoviruses, either wt or

mutated MLH1 together with wt PMS2 or wt or mutated MLH3 together wt MLH1 or wt or

mutated MSH2 together with wt MSH6, in order to produce functional complexes. As

controls, MLH1 wt, PMS2 wt, and MLH3 wt were produced alone. After 50 hours of culture,

the total protein extracts (TEs), including the overexpressed heterodimeric MutLα or MutLγ,

were produced by incubating the cells in ice-cold lysis buffer (25 mM Hepes, 2 mM β-

mercaptoethanol, 0.5 mM spermidine, 0.15 mM spermine, 0.5 mM phenylmethylsulfonyl

fluoride (PMSF), 2× Complete protease inhibitor cocktail (Roche)) for 30 min. After

43

centrifugation at 12,000 g for 50 min at +4° C, the supernatant was collected and NaCl (to

100 mM) and glycerol (to 10%) were added. Protein fractions were stored at -80° C.

Protein expression in human cells (I, II, IV)

The cell lines that were used in this study were HeLa (cervical carcinoma), HCT116

(colorectal carcinoma) and 293T (human embryonic kidney). The HeLa cell line was used as

a positive control, since in this cell line all three MutL genes are natively expressed, whereas

HCT116 cells lack MLH1 and PMS2 and 293T cells lack MLH1, PMS2 and MLH3 (Cannavo

et al. 2005). HeLa cells were cultured in MEM + Earle’s medium (Invitrogen/GIBCO),

HCT116 cells in McCoy´s 5a medium (Invitrogen/GIBCO) with 10 % fetal bovine serum

(Invitrogen/GIBCO), and 293T cells in Dulbecco´s Modified Eagle Medium/F12 medium

(Invtrogen/GIBCO) with 5 % fetal bovine serum, 2 mM L-glutamine and penicillin-

streptomycin (50 units/ml) at 37° C in a 5% CO2- humidified atmosphere.

The human cells were used to study the subcellular localization of MutL wt homologues

MLH1, PMS2 and MLH3 by immunofluorescence and mutated MutLα variants by direct

fluorescence. The stability of mutated MutLα variants was also studied in 293T cells.

Transfections were done either with Fugene6 (Roche) or TurboFectTM in vitro transfection

reagent (Fermentas). For transfections, 1 x 105 (or 3 x 105 for the stability studies) cells were

seeded into a 35-mm well and after 4 hours of culture, the cells were transfected with 1 µg of

the appropriate vectors. After transfection the cells were cultured for 24 hours for localization

studies and 48 hours for stability studies.

For TE preparation from 293T, the trypsinated cells were washed twice with ice cold PBS and

lysed in 60 µl of ice-cold extraction buffer (50mK Tris-Cl pH 8.0; 350 mM NaCl, 0.5%

Nonidet-P40, 1x Complete protease inhibitor cocktail (Roche), 1 mM PMSF, 2 µg/ml

aprotinin and 0.7 µg/ml pepstatin) for 25 minutes. Cells were then centrifuged at 16,000 g for

3 minutes and the supernatant was preserved.

44

Functional analyses (I-IV)

Nuclear localization (I, II, IV)

Direct fluorescence (I, IV)

Direct fluorescence was used to study the localization of MutLα mutations in 293T and

HCT116 cells. The experiments with each MLH1 mutant were performed in three different

combinations: 1) each MLH1-EGFP alone, 2) with PMS2-N1 and 3) each MLH1-N1 with

pPMS2-EGFP.

For the detection of fluorescent MLH1-EGFP and PMS2-EGFP proteins, 24 hours after

transfection the 293T cells were washed twice with PBS and fixed with 4 %

paraformaldehyde in PBS for 20 minutes at room temperature. After fixation, the cells were

washed with PBS, and the nuclei were stained by incubating cells in PBS with 300 nM 4´, 6´-

diamindino-2-phenylindole (DAPI) (Sigma Aldrich) for three minutes. Slides were mounted

with Fluorescence Mounting Medium (DAKO).

Immunofluorescence (II)

Immunofluorescence was used to detect the native MutL homologues MLH1, PMS2 and

MLH3 proteins in HeLa cells and transfected wt MLH1, PMS2 and MLH3 proteins in HeLa

and HCT116 cells. In HeLa cells, MLH3-N1 was transfected separately and together with

MLH1-N1. In HCT116 cells, each of the three expression vectors were transfected separately,

and MLH3-N1 together with MLH1-N1 or PMS2-N1.

Twenty-four hours after transfection or seeding onto coverslips, cells were washed with PBS,

fixed with ice-cold methanol for 15 minutes at -20°C, and washed again with PBS. The cells

were blocked with 1 % bovine serum albumin (Sigma Aldrich) in PBS at +37 °C for at least 1

hour (h). Incubations with primary and secondary antibodies were performed at + 37 °C for 1

h each. Antibodies were diluted in PBS including 1% bovine serum albumin as follows: anti-

MLH1 1 µg/ml (G168-728 BD Biosciences Pharmingen), anti-PMS2 2 µg/ml (clone A16-4

BD Biosciences Pharmingen,), anti-MLH3 10 µg/ml (sc-25313 Santa Cruz Biotechnologies),

and fluorescein-conjugated AffinityPure donkey anti-mouse immunoglobulin G (IgG) 4 µg/ml

(the secondary antibody used, from Jackson ImmunoResearch Labs Inc.). The nuclei were

45

stained by incubating cells in PBS with 300 nM DAPI (Sigma Aldrich) for 3 minutes. Slides

were mountained with Fluorescence Mounting Medium (Dako).

Fluorescence detection (I, II, IV)

Intracellular localization of proteins was analyzed by direct fluorescence using an Axioplan 2

microscope (Carl Zeiss) with a 63 objective. Each experiment was repeated at least three

times, and at least 100 cells from each individual experiment were analyzed. Representative

images were taken with Isis5 software (Metasystems).

Western blot and coimmunoprecipitation analyses (I, III, IV)

The expression levels and correct sizes of the recombinant protein were detected by sodium

dodecylsulfate polyacrylamide gel electrophpresis (SDS-PAGE) and Western blot analysis.

The protein extracts were run in an 8 % SDS-PAGE gel, blotted onto a Hybond C membrane

(Amersham Pharmacia Biotech) and detected with the appropriate antibody (anti-MLH1 0.3

µg/ml clone 168-15, BD Biosciences/PharMingen, and anti-PMS2 0.2 µg/ml, Ab-

1,Calbiochem/Oncogene Research, anti-MLH3 1.5 µg/ml, H-2, Santa Cruz Biotechnologies,

anti-MSH2 0,4 µg/ml Ab-2, Na27 Calbiochem Oncogene Research, anti-MSH6 0,17µg/ml

clone 44, BD Transduction Laboratories). Horseradish peroxidase-linked anti-mouse

immunoglobulin was used as a secondary antibody (Amersham Pharmacia Biotech) and the

ECL Western blotting analysis system was used for visualization.

For coimmunoprecipitation, Sf9 protein extracts, which were estimated to contain equal

amounts of mutated or wt protein, were incubated for one hour at +4C on a rotating wheel

with 0.5 µg of anti-MLH1 antibody (clone G168-728, BD Biosciences, Palo Alto, CA) or

anti-PMS2 antibody (clone A16-4, BD Biosciences, Palo Alto, CA) in a total volume of 1 ml

in RIPA lysis buffer (150 mM NaCl, 1% Nonidet P-40, 0.5% sodium deoxycholate, 0.1%

SDS, 50 mM Tris-Cl pH 8.0) or NP40 lysis buffer (50 mmol/l Tris-HCl pH 8,0, 125 mmol/l

NaCl, 1 % NP40, 2 mmol/l EDTA, 1 mmol/l PMSF, 1 x complete protease inhibitor cocktail

(Roche)). A total of 20 µl of protein A/G agarose suspension (Santa Cruz Biotechnologies)

was added and incubation was continued for a further 1.5 hours. The precipitates were

centrifuged for 5 minutes at 2,500 g, washed three times with cold RIPA or NP40 lysis buffer,

and run on an 8% SDS polyacrylamide gel. The interactions between MLH1 and PMS2 or

MLH1 and MLH3 were detected by Western blot analysis.

46

In vitro MMR repair assay (I, III, IV)

Nuclear protein extraction Nuclear protein extracts (NEs) contains all the proteins needed for the MMR reaction. NEs

from HeLa and TK6 cells were used as functional controls and HCT116 (lacks MutLα), 293T

(lacks MutLα and MutLγ) and LoVo (lacks MSH2) were used as negative controls.

Appropriate amounts of cells (~5-6× 108) were collected, centrifuged at 500 g for 10 min at +

4° C and washed with PBS. The pellets were resuspended in 3.5 ml of ice-cold isotonic buffer

(20mM Hepes pH 7.9, 5 mM KCl, 1.5 mM MgCl, 250 mM sucrose, 0.2 mM PMSF, 1×

Complete EDTA-free protease inhibitor mixture (Roche), 0.25 µg/ml aprotinin, 0.7 µg/ml

pepstatin, 0.5 µg/ml leupeptin, 1 mM DTT) and centrifuged as above. The pellets were

suspended in ice-cold hypotonic buffer (isotonic buffer without sucrose). After centrifugation

cells were resuspended in ice-cold isotonic buffer and incubated on ice for 5 min. The cell

membranes were disrupted using syringe with a narrow gauge needle (No. 27). The nuclei

were collected by centrifugation at 3000 g for 10 min at +4°C. The pellets were suspended in

ice-cold extraction buffer (25mM Hepes pH 7.5, 10% sucrose, 1 mM PMSF, 0.5 mM DTT, 1

µg/ml leupeptine) and NaCl was added up to 155 mM. The suspension was rotated for 1 h at

+ 4°C and then centrifuged for 20 min at 14,500 g at +4°C. The supernatant was dialyzed for

2 hours against ice-cold dialysis buffer (25mM Hepes pH 7.5, 50 mM KCl, 0.1 mM EDTA

pH 8, 10% sucrose, 1 mM PMSF, 2 mM DTT, 1 µg/ml leupeptin). The dialyzed extract was

centrifuged at 16,000 g for 15 min at + 4°C and the supernatant containing the nuclear

proteins was preserved.

Heteroduplex preparation

The circular DNA heteroduplex containing a G•T mispair and single-strand nick (Figure 6)

were prepared as reported previously (Lu et al. 1983, Lahue et al. 1989, Holmes et al. 1990;

Baerenfaller et al. 2006).

47

Figure 6. The 5´ G•T substrate used in the in vitro MMR assays. The uncut strand contains a complete BglII restriction site, whereas the nicked strand has G in the place of A, resulting in a G•T mismatch. Correction of the mismatch generates a complete BglII site. The 5 ́nick is 370 bp from the BglII site. BsaI cuts both repaired and unrepaired plasmids 1,360 bp from the BglII site.

The substrates were constructed from pGEM-TA plasmid with an intact BglII restriction site

and pGEM-CG plasmid carries a G instead of T in its BglII site. Single-stranded (ss) DNA

was produced in the E. coli strain XL1 Blue with the helper phage M13K07 (New England

Biolabs) using the pGEM-TA plasmid as a template. Double-stranded (ds) pGEM-CG was

linearized with the BanII restriction enzyme to achieve dsDNA which was linearized 370 bp

from the BglII site. The ssDNA (from pGEM-TA) and linear dsDNA (from pGEM-GC) were

annealed, generating circular heteroduplex DNA with a 5´ single-stranded nick and a G•T

mispair.

MMR assay

The ability of MutLα, MutLγ and MutSα variants to repair the G•T mismatch were tested in

the MMR assay. The repair reaction included 100 ng of circular heteroduplex DNA with the

G•T mismatch and 75 µg of HeLa, TK6, 293T, HCT116 or LoVo nuclear extract (NE). The

functionality of the mutated MLH1, MLH3, and MSH2 proteins was studied by

complementing the appropriate NE with the Sf9 TE including overexpressed MutLα, MutLγ

or MutSα complex. The TEs used for the MMR assay were adjusted to contain equal

quantities of the recombinant MMR complex. The repair reaction was performed for 30 min

at 37 °C, after which the DNA was extracted, purified, and digested with BsaI and BglII

restriction enzymes. The BsaI enzyme was used to linearize the DNA. A successful repair

reaction converts the G•T heteroduplex, which is not susceptible to cleavage by the

AG TCTTC AGA

G

T5´

BglII

BsaI

48

endonuclease BglII, to an A•T homoduplex, which is cleaved by BglII. Thus, the efficiency of

repair can be measured by the cleavage efficiency. The digested DNA was run in 1% agarose

gels and the repair efficiencies were quantified by comparing the intensities of the repaired

and unrepaired fragments using an Image-Pro 4.0 system (Media Cybernetics, Silver Spring,

MD). The repair percentages were calculated as an average of at least 3 independent

experiments and the repair percentages of the proficient controls were used as a reference

level.

In silico comparative sequence analyses (I, IV)

In silico comparative sequence analyses were performed by our collaborators. In study I, the

sequence similarity of the aligments was used to predict the tolerability of all potential MLH1

amino acid substitutions using the program SIFT (http://blocks.fhcrc.org/sift/SIFT.htm) (Ng

& Henikoff 2001). In study IV, the predictions were performed using the programs SIFT,

Polyphen (http://coot.embl.de/PolyPhen) (Ramensky et al. 2002), and PMut

(http://mmb2.pcb.ub.es.8080/PMut/) (Ferrer-Costa et al. 2005).

49

RESULTS

Functional characterization of nontruncating MLH1, MLH3 and MSH2

mutations (I, III, IV)

Expression/ stability of mutated proteins (I, III, IV)

To determine whether the MLH1, MLH3 or MSH2 mutations affected the expression levels or

stability of the transcripts or the proteins, we coexpressed the mutations with their counterpart

in order to get functional heterodimer complexes. In study I, the MLH1 mutations, excluding

MLH1-E460A and MLH1-R687W, were expressed in both Sf9 and 293T cells. In studies III

and IV, where the MLH3 mutations and the two MLH1 mutations excluded from study I,

respectively, were expressed in Sf9 insect cells only. Among the 36 studied MLH1 mutations,

20 (P28L, TSI45-47CF, D63E, G67R, del71, C77R, I107R, L155R, V185G, S247P, H329P,

del330, L550P, A589D, del612, del616, del633-663, P648L, P648S and P654L) affected

either the expression or stability of the protein so that the quantities of the respective

mutations were lower than the quantity of wt MutLα (I). All the seven MLH3 mutants

together with their MLH1 counterpart were expressed similarly to wt MutLγ (III). Two

studied MSH2 mutations (L187R and M688V) showed reduced expression in Sf9 cells (IV).

Subcellular localization of mutated MLH1 proteins (I, IV)

To study whether the MLH1 mutations affect the subcellular localization of MutLα,

fluorescent proteins were expressed in 293T and HCT116 cells. Three different combinations

of vectors were used: the MLH1-EGFP mutation alone, the MLH1-EGFP mutation together

with wt PMS2-N1, and wt PMS2-EGFP with the MLH1-N1 mutation. In this way we were

able to analyze the location of MLH1 alone, the location of MLH1 when PMS2 was present,

and the location of PMS2 when expressed with MLH1 mutations.

We analyzed the subcellular localization of 35 MLH1 mutations in the 293T cell line. The

nuclear localization was decreased in MLH1 mutations P28L, TSI45-47CF, D63E, G67R,

del71, C77R, K84E, I107R, L155R, V185G, S247P, H329P, del330, L550P, A589D, del612,

del616, del633-663, P648L, P648S, and R659P (Table 4, I). The nuclear localization of

50

MutLα mutations was shown to be abnormal in two different ways: either the mutated MLH1

protein was found in the cytoplasm with and without PMS2, and coexpressed PMS2 was also

mainly in the cytoplasm; or MLH1 expressed alone was mainly in the nucleus, and the

nuclear proportion of MLH1 was not increased when coexpressed with PMS2, which

remained mainly in the cytoplasm.

Interaction capability of mutated proteins (I, III)

Combined coimmunoprecipitation and Western blot assays were perfomed to study the effects

of mutated MLH1 and MLH3 proteins on the MLH1/PMS2 and MLH1/MLH3 interactions,

respectively. The assays were done using total extracts of Sf9 cells co-expressing MLH1

mutations with wt PMS2 or MLH3 mutations with wt MLH1. Wt MLH1, wt PMS2 and wt

MLH3 expressed alone were used as controls. Of 34 MLH1 and 7 MLH3 mutations, only

MLH1 mutations del633-663 and R659P interfered with the MLH1/PMS2 interaction (Table

4, I, III).

Mismatch repair efficiency of mutated proteins (I, III, IV)

The functionality of the MutLα and MutLγ complexes were examined in the in vitro MMR

assay. An appropriate amount of NE (HCT116,293T or LoVo) was complemented with Sf9

TE containing the overexpressed protein complex. TEs were adjusted to contain similar

quantities of the studied proteins. Of the 36 MLH1 mutations, 15 disrupted the MMR function

of the MLH1 protein. These MMR-deficient mutations were P28L, TSI45-47CF, D63E,

G67R, del71, C77R, F80V, K84E, I107R, L155R, V185G, S247P, del330, del633-663 and

P659P. The repair efficiencies of MLH1 H329P and del612 were slightly reduced (Table 4,

I). The seven MLH3 mutations did not interfere with the repair efficiency of MutLγ.

Significantly, the overall repair efficiency of wt MutLγ was much lower than the efficiency of

wt MutLα (Table 4, III). Only one MSH2 mutant, L187R, was deficient in the MMR assay

(Table 4, IV).

In silico predictions of MLH1 and MSH2 mutations (I, IV)

The in silico predictions were compared with the results of the functional assays. The in silico

analyses correctly predicted the functional results for 24 out of 30 MLH1 mutations, for an

51

overall predictive value (OPV) of 80%. The positive predictive value (PPV) was 88% (16 of

18 predicted deleterious mutations had reduced function) and the negative predictive value

(NPV) was 72% (8 of 11 predicted tolerated variations functioned as wt). Predictions for the

MLH1 mutations H329P, A589D, K681T, Y646C, A681T and R687W were not confirmed by

the functional assays (Table 4, I, IV). The mutations H329P, A589D and K681T were

predicted to be tolerated, while in the functional assays these mutations had problems in

localization and H329P and A589D also had decreased expression levels. However, the

mutations Y646C, A681T and R687W, which were predicted as deleterious, functioned as the

wild type. The OPV of eight MSH2 mutations was 62%. The predicitons for the MSH2

mutations A45V, L187R, F519L, V722I and E886G were confirmed by the in vitro MMR

assay. The mutation M688V predicted to be deleterious, displayed problems in expression/

stability, but was functional in the in vitro MMR assay (IV). The results of all the functional

analyses and in silico predictions are summarized in table 4.

52

Table 4.The results of functional analyses

MLH1 mutation

(I, IV)

Expression Interaction MMR

capability

Nuclear

localization

In silico

prediciton

P28L decreased1 normal2 decreased decreased3 Del

A29S normal normal normal normal Tol

TSI45-47CF decreased normal decreased decreased4 NA

D63E decreased normal decreased decreased4 Del

G67R decreased normal decreased decreased4 Del

del71 decreased normal decreased decreased3 NA

C77R decreased normal decreased decreased4 Del

F80V normal normal decreased normal Del

K84E normal normal decreased decreased5 Del

S93G normal normal normal normal Tol

I107R decreased normal decreased decreased4 Del

L155R decreased normal decreased decreased4 Del

V187G decreased normal decreased decreased4 Del

V213M normal normal normal normal Tol

I219V normal normal normal normal Tol

S247P decreased normal decreased decreased4 Del

H329P decreased normal normal decreased4 Tol

del330 decreased normal decreased decreased4 NA

K443Q normal normal normal normal Tol

E460A normal NA normal NA Tol

L550P decreased normal normal decreased6 Del

A589D decreased normal normal decreased6 Tol

612del decreased normal normal decreased6 NA

616del decreased normal normal decreased6 NA

K618A normal normal normal normal Tol

K618T normal normal normal decreased6 Tol

del633-663 decreased decreased decreased decreased6 NA

Y646C normal normal normal normal Del

P648L decreased normal normal decreased6 Del

P648S decreased normal normal decreased6 Del

P654L decreased normal normal decreased6 Del

R659P normal decreased decreased decreased6 Del

R659Q normal normal normal normal Tol

A681T normal normal normal normal Del

R687W normal NA normal normal Del

V716M normal normal normal normal Tol

53

MLH3 mutation

(III)

Expression Interaction MMR

capability

Nuclear

localization

In silico

prediciton

Q24E normal normal normal NA NA

R647C normal normal normal NA NA

S817G normal normal normal NA NA

G933C normal normal normal NA NA

W1276R normal normal normal NA NA

A1394T normal normal normal NA NA

E1451K normal normal normal NA NA

MSH2 mutation

(IV)

T44M normal NA normal NA Del A45V normal NA normal NA Tol L187R decreased NA decreased NA Del F519L normal NA normal NA Tol M688V decreased NA normal NA Del V722I normal NA normal NA Tol A848S normal NA normal NA Del E886G normal NA normal NA Tol 1 mutated protein functioned abnormally; 2mutated protein functioned like wild type; 3 only localization data of PMS2-EGFP available; 4the proportions of nucleus-localized MLH1 and PMS2 were decreased; 5 the proportion of nucleus-localized MLH1 expressed alone was slightly decreased; 6coexpressed PMS2-EGFP was mainly cytoplasmic and nuclear localization of coexpressed MLH1-EGFP was slightly decreased, but MLH1-EGFP expressed alone was mainly nuclear; NA not available; Del, deleterious; Tol, tolerated. Subcellular localization of wild type MutL homologues MLH1, PMS2 and

MLH3 (II)

To find out the subcellular localization of the native and transfected wt MutL homologues

MLH1, PMS2 and MLH3, we used appropriate cell lines and immunofluorescence. The

endogenous localizations of MutL homologues were studied in HeLa, which express all three

protein, and in HCT116 cells, which express only MLH3. In HeLa cells, MLH1 and PMS2

clearly localized in the nucleus, whereas MLH3 was mostly cytoplasmic. The endogenous

localization of MLH3 in HCT116 cells was also mostly cytoplasmic.

The endogenous amount of MLH3 in cells is much lower than the amounts of MLH1 and

PMS2. We wanted to study what happens if more protein were expressed in the cell, hence we

transfected MLH3 into HeLa and HCT116 cells. After transfection, MLH3 started to go into

the nucleus in HeLa cells, but not in HCT116 cells. When cotransfected with MLH1, MLH3

54

was partially nuclear in both cell lines and MLH1 was always found in the nucleus. MLH1

was able to go to the nucleus alone, but both MLH3 and PMS2 seemed to require

heterodimerization with MLH1 for nuclear localization. When PMS2 and MLH3 were

transfected into HCT116 cells, both proteins seemed to be in the cytoplasm.

55

DISCUSSION

A significant proportion of DNA mutations (32% of MLH1, 18% of MSH2, 38 % of MSH6

and 87% of MLH3 mutations) found in suspected HNPPC patients are missense mutations or

small in-frame deletions (Peltomäki & Vasen 2004, http://www.insight-group.org). These

nontruncating mutations are a major challenge for the diagnostic and genetic counseling of

HNPCC patients. Different nontruncating alterations appear to be associated with a wide

variety of clinical phenotypes, from no to highly increased cancer risk. Different functional

analyses have demonstrated the difficulty of distinguishing nontruncating pathogenic

mutations from harmless variants. Even different base substitutions in the same codon can

cause different functional consequences, from complete elimination of the MMR activity to

little or no effect on protein function (Ellison et al. 2004). Functional characterization of

mutated proteins gives significant information about nontruncating mutations and by

comparing functional results with the clinical data helps to interpret the pathogenicity of the

mutation.

MLH1 is the main susceptibility gene in HNPCC and about 650 different MLH1 variations

associated with HNPCC patients have been reported in the international HNPCC database

(Woods et al. 2007; http://www.insight-group.org). In studies I and IV, we determined the

functional significance of different nontruncating MLH1 germline mutations found in putative

HNPCC patients identified through international collaboration. So far only 34 different MLH3

variations have been reported in the HNPCC database (http://www.insight-group). Since the

role of the MLH3 protein in MMR is not quite clear, the pathogenicity of the found MLH3

missense mutations has been difficult to interpret. In study III we evaluated the pathogenicity

of seven inherited alterations in MLH3, which were reported in two publications (Wu et al.

2001; Liu et al. 2003). All mutations were found in putative HNPCC families.

Pathogenicity of MLH1 mutations (I, IV)

The MLH1 mutations were located throughout the whole gene. Interestingly, our results

showed that the most amino-terminal MLH1 mutations (P28L, TSI45-47CF, D63E, G667R,

del71, C77R, I107R, L155R, V185G, S247P and del330) caused a deficiency in MMR and

instability of the protein when expressed in human cells. The mutated variations F80V and

56

K84E were defective in MMR, but stable. Our results are in concordance with the knowledge

that the four ATP-binding sites are located in the amino-terminal region of MLH1 (Bergerat

et al. 1997, Mushegian et al. 1997, Dutta& Inouye 2000) and that ATP binding and

hydrolysis is critical for MutLα stability and function (Ban et al. 1999; Tran & Liskay 2000;

Räschle et al. 2002). In view of this, these amino-terminal missense mutations most likely

disrupt the binding and/or hydrolysis of ATP molecules.

The MLH1 interaction domain is located in the carboxy-terminal part of the protein, and

mutations in this region are associated with a defective interaction between MLH1 and PMS2

(Guerrette et al. 1999; Nyström-Lahti et al. 2002; Kondo et al. 2003). However, not all amino

acid substitutions in this region have been shown to interfere with the MLH1/PMS2

interaction (Ellison et al. 2001; Kondo et al. 2003). In our study, only two (R659P and

del633-663) out of 15 mutations located in the interaction region interfered with the

MLH1/PMS2 interaction in the coimmunoprecipitation assay. However, eight carboxy-

terminal MLH1 mutations (L550P, A589D, del612, del616, K618T, P648L, P648S and

P654L) that were proficient in the in vitro MMR assay interacted with PMS2 but had

abnormal expression/stability and nuclear localization. The heterodimerization of MLH1 and

PMS2 has been shown to be important for PMS2 localization (Table 4) (Wu et al. 2003; Luo

et al. 2004). These eight mutations especially affected the transfer of PMS2 into the nucleus,

suggesting incorrect MLH1/PMS2 dimerization. Published functional results of the mutation

K618T are controversial, suggesting either pathogenicity or nonpathogenicity (Guerrette et al.

1999; Shimodaira et al. 1998; Kondo et al. 2003; Brieger et al. 2002; Trojan et al. 2002). In

our assays, otherwise normally functioning MLH1 K618T only slightly reduced the nuclear

localization of PMS2. This kind of mild deficiency might explain the previous controversial

conclusions about its pathogenicity.

Overall, the 22 mutations which were shown to be pathogenic in more than one assay were all

associated with typical HNPCC characteristics such as an early age at onset and MSI-H in

tumors of mutation carriers. The mutated protein F80V was nonfunctional in only one assay,

and it seemed to cause MSI-H but not loss of MLH1 protein in tumors.

In our functional tests, 12 variants (A29S, S93G, V213M, K443Q, E460A K618A, Y646C,

R659Q, A681T, R687W, V716M and the functional control I219V) acted like the wt MLH1

protein. The functional control I219V and the mutations S93G and V213M were also shown

57

to be nonpathogenic in previous functional assays (Shimodaira et al. 1998, Ellison et al. 2001,

Nyström-Lahti et al. 2002; Troijan et al. 2002, Kondo et al. 2003).

The mutations which did not interfere with any tested functions but were associated with

typical HNPCC characteristics indicate how important it is to interpret the biochemical results

together with the clinical and genetic data on the families under study. In the typical HNPCC

family carrying the mutation A29S, another MLH1 mutation (g. -27C>A) in the promoter was

later found and shown to segregate with the cancer. The pathogenicity of this mutation should

still be characterized by an RNA-based method. Also the family carrying the mutation K618A

carries another mutation, R659Q, in MLH1. However, neither of these mutations showed

pathogenicity in our assays. The pathogenicity of R659Q might be associated with aberrant

pre-mRNA splicing and skipping of the whole exon 17 (del633-663), which is caused by

three other alterations of the same codon R659X, R659P and R659L (Kohonen-Corish et al.

1996; Nyström-Lahti et al. 1999). If this is the case, the aberrant splicing could not be

detected in our functional assays which are cDNA-based. However, the resulting protein

variant without exon 17 (del633-663) has been shown to be MMR-deficient and unstable

(Nyström-Lahti et al. 2002).

The proficient mutation E460A was identified in individuals that carried deleterious

mutations in one of the other MMR genes which most probably were the cause of their

cancer. However, the mutation E460A was not found in the normal population (Christensen et

al. 2008). The mutation R687W has been reported in several CRC families (Furukawa et al.

2002; Jakubowska et al. 2001; Caldes et al. 2002; Lagersted et al. 2007). This mutation was

found as homozygous in 3 siblings with gastrointestinal cancers and features of

neurofibromatosis type I at a very low age but the parents, who were heterozygous, were

healthy (Gallinger et al. 2004). Moreover, in two studied families the mutation did not

segregate with the disease (Lagersted et al. 2007). The lack of MLH1 protein in their tumor

tissue might have been caused by promoter methylation.

The mutations S93G, V213M, K443Q, K618A, Y646C and A681T were associated with

atypical or mild features of HNPCC. The variation S93G was found in a family in which the

mean age of cancer onset was higher than normal (65 years when normally in HNPCC it is

<50 years). The variation V213M was reported in four families who did not fulfill the

Amsterdam criteria I or II (Vasen et al. 1991; Vasen et al. 1999). K443Q was found in a

58

family who did not fulfill the Amsterdam criteria, and in whom the mean age of cancer onset

was also high (62 years). K618A was found in seven families whose mean age at onset varied

between 38 and 77 years, whose MSI phenotypes in tumors were high, low or microsatellite

stable (MSS), and in whom IHC analysis revealed loss or no loss of MLH1 in tumors. The

mutations Y646C and A681T were both found only in one cancer patient. The mutation

V716M was found in four families, where the mean age of cancer onset varied between 45

and 67 years and the MSI status and IHC results varied between tumors.

In contrast to yeast assays, the present study performed in the human homologous system

sorts out not only pathogenic and nonpathogenic mutations but also the mechanism of the

pathogenicity and its effect on the clinical phenotype in mutation carriers. Some of the

alterations had been previously studied by different yeast-based functional assays and

interpreted as pathogenic (P28L, G67R, C77R, K84E, I107R, V185G, del616, K618T, P654L

and R659P) or nonpathogenic (A29S, V213M; I219V and K618A) (Shimodaira et al. 1998;

Guerrette et al. 1999; Scherbakova & Kunkel 1999; Ellison et al. 2001; Kondo et al. 2003;

Takahashi et al. 2007). These results are in accordance with our results. Different functional

results have been published for the mutations A681T and R687W. Yeast-based assays have

shown these mutations to be either pathogenic or the results have been inconclusive

(Shimodaira et al. 1998; Guerrette et al. 1999; Kondo et al. 2003; Takahashi et al. 2007).

In summary, families carrying MLH1 variations which were shown to be nonpathogenic in

our functional analyses usually have mild or atypical features of HNPCC. We conclude that

these mutations are either harmless or may in vivo cause a mild deficiency, which is not

detected by the methods used in the present study. Some families carried nonpathogenic

variations, although these families seemed to be typical HNPCC families with an early age of

onset and high MSI in tumors. The possibility of other MMR gene mutations in these families

should be excluded.

In silico analyses of MLH1 mutations (I, IV)

Statistically valid conclusions based on sequence homology in disease-related genes are

usually limited by the insufficient sample size of tested variants, insufficient size of sequence

databases, insufficient epidemiological data or lack of reliable functional assays (Cooper et al.

2003; Greenblatt et al. 2003; Flemming et al. 2003; Goldgar et al. 2004; Chan et al. 2007;

59

Tavtigian et al. 2008). In study I these limitations were not a problem and hence, multiple

sequence alignment analyses of human MLH1 with 20 other eukaryotic species were

constructed using the ClustalW software (Thompson & French 1994). The sequence similarity

of the aligments was used to predict the tolerability of all potential MLH1 amino acid

substitutions using the SIFT program (Ng and Henikoff 2001). SIFT correctly predicted our

functional results for 23 out of 28 MLH1 variants.

In study IV, in silico predictions were made using the SIFT, PolyPhen and Pmut programs. A

mutation was classified as deleterious when two thirds of the prediction programs predicted a

pathogenic status. The amino acid substitution of glutamic acid to alanine in codon 460

(E460A) was shown to be tolerated, which correlates with our functional results, whereas the

in silico prediction of the substitution of arginine to tryptophan in codon 687 (R687W) did

not.

Some of the mutations studied here were also included in the study by Chao et al. (2008),

where their pathogenicity was predicted by the MAPP-MMR (Multivariate Analysis of

Protein Polymorphisms-MisMatch Repair) program. The MAPP-MMR program predicted

eight (S93G, I219V, K443Q, Y646C, R659Q, A681T, R687W, and V716M) mutations as

nonpathogenic. These predictions are in accordance with our functional results, but not with

our in silico predictions of 3 mutations (Y646C, A681T, and R687W). MAPP-MMR

predicted 18 (D63E, G67R, C77R, F80V, K48E, I107R, L155R, V185G, S247P, H329P,

L550P, A589D, K618A, K618T, P648L, P648S, P645L, R659P) mutations as pathogenic

(Chao et al. 2008). Again, these predictions correlate quite well with our functional results,

since only K618A, which was predicted to be deficient, was functional in our assays.

However, four predictions (H329P, A589D, K618A and K618T) were different from our

SIFT predicitons.

Overall, the in silico predictions’ PPV (88%) and NPV (72%) values correlated well with the

functional results, suggesting that especially when functional studies are not possible, in silico

analyses might give a good indication for the pathogenicity of a mutation. The amount of

different missense-type variations is growing all the time and quick, cheap and reliable

methods for clarifying their pathogenicity are needed. Different in silico methods have

improved considerably and carefully validated algorithms can be a good tool for the

classification of missense mutations (Tavtigian et al. 2008)

60

Pathogenicity of MSH2 mutations (IV)

In addition to MutL homologue analyses we also tested eight MSH2 missense mutations in the

in vitro MMR assay. Only mutation L187R was deficient in the functional assay as well its

expression was reduced and this mutation can be classified as pathogenic. Mutation M688V

showed reduced expression, but was functional in the in vitro MMR assay. The other six

mutations (T44M, A45V, F519L, V722I, A848S, and E886G) were like the wild type. The

mutations M688V and E886G were indentified in individuals that also carried a deleterious

mutation in one of the other MMR genes and these secondary deleterious mutations might

explain the HNPCC phenotypes of these families.

The in silico analysis of MHS2 mutation predicted four mutations (T44M, L187R, M688V,

and A848S) to be deleterious and four other mutation (A45V, F419L, V722I, and E886G) as

tolerated using the SIFT, PolyPhen and Pmut programs. Our functional results correlated well

with the predictions, only two mutations (T44M and A848S), which were predicted to be

pathogenic, functioned like the wild type. The MAPP-MMR program predicted mutation

T44M to be tolerated (Chao et al. 2008), which is in concordance with our functional results.

Pathogenicity of MLH3 mutations (III)

The role of human MLH3 in MMR is not clear, although it has been shown to correct

base/base mismatches in the in vitro assays. Furthermore, in the human system, MLH3 is not

able to repair IDLs (Cannavo et al. 2005), even thought both yeast (Flores-Rozas & Kolodner

1998) and mouse studies suggest that MLH3 has a role in IDL repair (Chen et al. 2008).

However, MLH3 is a MutL homologue gene and mutations in it are thought to interfere with

the MMR function and thus predispose to cancer. Some putative pathogenic MLH3 mutations

have been found (Wu et al. 2001; Liu et al. 2003, www.insight-group.org). In many cases, the

mutations are marked in the database as both pathogenic and nonpathogenic. Clinical

phenotypes vary, but generally MLH3 mutations are associated with atypical HNPCC. There

are also studies in which no association between an MLH3 mutation and CRC was found

(Loukola et al. 2000, Akiyama et al. 2001; Lipkin et al. 2001).

61

Here, we studied seven MLH3 variations (Q24E, R647C, S817G, G933C, W1276R, A1394T,

E1451K), which were reported to be found in putative HNPCC patients (Wu et al. 2001 and

Liu et al. 2003). The variations were suggested to be pathogenic, although some of them were

reported to cause only a low risk for CRC (Wu et al. 2001, Liu et al. 2003, 2006; Taylor et al.

2006; www.insight-group.org). The mutations were found in 11 separate families but mostly

without a family history suggestive of inherited cancer susceptibility. The mutations were not

found in control individuals or in sporadic CRC cases (Wu et al. 2001; Liu et al. 2003, 2006;

Taylor et al. 2006), except for the variation E1451K, which was later found in both familial

CRC patients and healthy controls in the Korean population (Kim et al. 2007). Mutations

Q24E, R647C, S817G, G933C, A1394T, and E1451K were associated with MSI-H (Wu et al.

2001, Taylor et al. 2006). In summary, neither the genetic nor clinical data of the mutation

carriers and their families was conclusive concerning the pathogenicity of these mutations.

The seven mutations included in our study were spread throughout the MLH3 gene. The

mutation Q24E was in the putative ATP-binding and hydrolysis domain (Ban et al. 1999), and

A1394T and E1451K were in the MLH1/MLH3 interaction domain (Kondo et al. 2001).

None of the mutated MLH3 proteins lost their ability to interact with MLH1 and they were

able to repair G•T mismatches in the in vitro MMR assay. Thus, the location of these MLH3

mutations did not appear to make any difference in functionality, as was the case for the

MLH1 mutations.

Our results suggest that the MLH3 missense mutations alone do not interfere with MMR. A

recently reported study supports our results (Ou et al. 2009). In four families, the index

patient was a heterozygote with two different MMR gene germline mutations (Table 3).The

patient with the mutation S817G and the two unrelated families carrying the mutation

E1451K showed another mutation in MSH6. Moreover, the MSH6 protein was lost in both the

tumors with the MLH3-E1451K mutation, indicating that the mutation in MSH6 would be the

cause of cancer in these families (Wu et al. 2001). However, the possible compound effect of

mutations in two low- penetrance MMR genes, such as MLH3 and MSH6, cannot be ruled

out. The mutation MLH3 W1276R was found together with another mutation in MSH2

(E198G) (Liu et al. 2003), which was later shown to be pathogenic (Gammie et al. 2007;

Chao et al. 2008).

62

Our results support previous data on the importance of MutLα in the MMR mechanism (Li &

Modich 1995; Harfe & Jinks-Robertson 2000 Lipkin et al. 2001; Cannavo et al. 2005). We

showed that the in vitro mismatch repair efficiency of wild type MutLα was much higher than

of wild type MutLγ. Previous studies have described a reconstitution of the human MMR

system needing seven purified activities including MutS (MSH2/MSH6) or MutSβ

(MSH2/MSH3) and MutL (MLH1/PMS2) (Constantin et al. 2005, Zhang et al. 2005). The

MutSβ complex was shown to play a limited role in repair of base-base mismatches, but it

processed insertion/deletion mispairs much more efficiently than MutS, which competently

corrected both types of mismatches (Acharaya et al. 1996, Palomobo et al. 1996). This is at

least one explanation why mutations in MSH2 and MLH1 account for most HNPCC families,

whereas mutations in MSH6, MSH3, PMS2 and MLH3 are associated with no or low cancer

susceptibility. The fact that MutLγ is able to correct mismatches leads to the idea of MutLγ

functioning as a “backup” in the repair system (Cannavo et al. 2005). Moreover, human

MLH1 interacts with 36 homologous amino acid residues within PMS2, MLH3 and PMS1

(Kondo et al. 2001), supporting the idea of a functional overlap of MLH3 between PMS2.

Moreover, mouse studies have shown that the heterodimer of MLH1 and MLH3 contributes

to the tumor suppression mechanism (Chen et al. 2005; 2008). Altough it seems that while

PMS2 is the major partner of MLH1 in MMR, MLH3-MLH1 complex may have a bigger role

in meiosis (Prolla et al. 1998; Lipkin et al. 2001; 2002). However, MLH3 may also function

in the MMR mechanism and PMS2 might have some role at least in male meiosis (Prolla et

al. 1998; Cannavo et al. 2005).

In conclusion, MLH3 mutations alone seem not to be sufficient to trigger MMR deficiency,

but may take part in vivo in the compound pathogenicity of two MMR gene mutations, but

this needs to be confirmed. All in all, MLH3 is a no or a low risk gene for HNPCC.

Nuclear localization of MLH1, PMS2 and MLH3 (II)

MMR occurs in the nucleus, and thus all required proteins need to be in the nucleus during

repair. We wanted to characterize subcellular localization of the MutL homologues MLH1,

PMS2 and MLH3 to find out their contribution in MMR. Consistently with previously

published data that the endogenous MLH3 is ~60 times less abundant than PMS2 (Cannavo et

al. 2005), the MLH3 protein was difficult to detect by immunofluorescence. The higher levels

63

of MLH1 and PMS2 already suggest that MutLα is a more important complex than MutLγ in

the human MMR system. Our results showed that native MLH1 and PMS2 localize in the

nucleus, while MLH3 is more cytoplasmic. Moreover, if MLH3 is expressed alone (HCT116)

it stays in the cytoplasm, whereas coexpression with MLH1 results in its partial nuclear

localization. The results suggest that the endogenous amount of MLH3 is too low for its

nuclear localization. It also seems that the nuclear localization of MLH3 is dependent on the

interaction with MLH1. PMS2 and MLH3 might also compete for binding to MLH1. The

reported results on the nuclear localization of MutL homologues are controversial. Although

our results showed that PMS2 and MLH3 are not localized in the nucleus without MLH1,

other studies suggest that both PMS2 and MLH3 are capable of entering the nucleus without

MLH1 (Leong et al. 2009; Ou et al. 2009), even though the MutLα heterodimer might be

imported more efficiently than MLH1 or PMS2 monomers (Leong et al. 2009). In accordance

with our results, other studies have suggested that the nuclear localization of MutLα is

dependent on heterodimerization and the localization of PMS2 is dependent on MLH1 (Wu et

al. 2003; Luo et al. 2004). Our results indicate that MLH1 and heterodimerization is

important in the nuclear localization of both PMS2 and MLH3. The low amount of

endogenous MLH3 in cells could explain the different results observed for its localization

with and without transfection. Our results show that after transfection and overexpression,

MLH3 was partially located in the nucleus, but only when MLH1 was present.

In summary, the nuclear localization of PMS2 and MLH3 is dependent on MLH1. The

continuous localization of MLH1 and PMS2 (MutLα) in the nucleus suggest a major activity

in MMR. The observation that overexpressed MLH3 localizes partially in the nucleus

suggests its role as a no or low risk gene in HNPCC. However, overlapping functionalities for

PMS2 and MLH3 cannot be ruled out.

64

CONCLUSIONS

This PhD study was undertaken to evaluate functionality of MutL homologue MMR proteins.

The following conclusions can be made based on the results of this work:

The functional studies of different MLH1 missense mutations showed that

Mutations in the amino-terminal part of MLH1 frequently caused repair deficiency,

most probably by affecting the ATP binding/hydrolysis capability of MLH1.

The carboxy-terminal MLH1 mutations often affect the stability and subcellular

localization of the protein.

The functional data on the MLH1 variants correlate quite well with clinical data.

The nuclear localization of both PMS2 and MLH3 seems to be MLH1-dependent.

MLH3 mutations alone do not interfere with MMR and most probably do not predispose to

HNPCC. MLH3 mutations may still have some compound affect with mutations in other

MMR genes in predisposing to cancer.

MutLγ might have some conditional activity but MutLα seems to have a major role in the

MMR mechanism.

65

FUTURE PROSPECTS

To further characterize the MLH1 mutations whose pathogenicity remained unclear in

the present study (e.g. RNA level studies, MLH1 methylation)

To study the compound effects of two different MMR mutations

To study whether MLH1 mutations affect MutSα-MutLα ternary complex formation

To study how different substrates (e.g. mismatches and IDLs nicked either on the 5´or

3´ side of the error) are repaired in the in vitro MMR assay

To study the functional significance of PMS2 mutations

66

ACKNOWLEDGEMENTS

This work has been carried out at the Division of Genetics, Department of Biological and

Environmental Sciences, University of Helsinki. The work was financially supported by the

Helsinki Graduate School in Biotechnology, the Sigrid Juselius Foundation, the Academy of

Finland and Finnish cancer organizations.

My warmest thanks go to my supervisor Professor Minna Nyström, whom I thank for giving

me the opportunity to work in an interesting research project and for supervision. Her support

and encouragement during these years have been very important to me.

Docent Jukka Partanen and Professor Katarina Pelin are acknowledged for reviewing and

commenting on this thesis. I would like to thank Docent Nina Horelli-Kuitunen and Professor

Katarina Pelin for participating in my graduate school thesis committee, and for their interest

in my work. Jack Leo is acknowledged for skillful revision of the English language of this

thesis.

I thank all present and former members of our research group. Thanks go to Tiina Raevaara

for all her help and guidance especially at the very beginning of my laboratory work and of

course for her contributions to my work. I am thankful to Reetta Kariola and Elina

Vuorenmaa, who contributed to my work. Laura Sarantaus, Saara Ollila, Jukka Kantelinen

and Minttu Kansikas are wonderful people and colleagues. Thank you all for the help,

kindness and patience with me. It has been great working with you. I am also very grateful to

our collaborators around the world.

The Division of Genetics, led by Professor Tapio Palva, has provided excellent facilities for

research. I thank the people at the Division of Genetics for the warm atmosphere there. Arja

Ikävalko and Arja Välimäki are acknowledged for their help in many general issues.

I thank my dear friends for showing me that there is life outside the laboratory and for their

endless friendship and support. You’re all invaluable! Special thanks to Jenni Küblbeck, who

has shared so many scientific and non-scientific moments with me - thanks also for those

numerous e-mails.

67

My parents Martti and Eeva, big brothers Jussi and Antti and their families deserve lots of

thanks for their constant love, support and encouragement. Special thanks go to Antti who

always has time for me and who has helped me uncountable times. I thank my niece Maija

and nephews Olli and Ilkka for happy moments and for their great sense of humor. The whole

family Korhonen is acknowledged for their kindness and help.

Finally, I thank my husband Sami, whose love, understanding and support are the most

important things to me. You are the sunshine of my life.

Espoo, March 2009

Mari Korhonen

68

REFERENCES

Aaltonen, LA, Peltomäki, P, Leach, FS, Sistonen, P, Pylkkanen, L, Mecklin, JP, Järvinen, H, Powell, SM, Jen, J, Hamilton, SR. 1993. Clues to the pathogenesis of familial colorectal cancer. Science 260: 812-816.

Aaltonen, LA. 1998. Molecular epidemiology of hereditary nonpolyposis colorectal cancer in Finland. Recent Results Cancer Res. 154: 306-311.

Aarnio, M, Sankila, R, Pukkala, E, Salovaara, R, Aaltonen, LA, de la Chapelle, A, Peltomäki, P, Mecklin, JP, Järvinen, HJ. 1999. Cancer risk in mutation carriers of DNA-mismatch-repair genes. Int.J.Cancer 81: 214-218.

Acharya, S, Wilson, T, Gradia, S, Kane, MF, Guerrette, S, Marsischky, GT, Kolodner, R, Fishel, R. 1996. hMSH2 forms specific mispair-binding complexes with hMSH3 and hMSH6. Proc.Natl.Acad.Sci.U.S.A. 93: 13629-13634.

Akiyama, Y, Nagasaki, H, Nakajima, T, Sakai, H, Nomizu, T, Yuasa, Y. 2001. Infrequent frameshift mutations in the simple repeat sequences of hMLH3 in hereditary nonpolyposis colorectal cancers. Jpn.J.Clin.Oncol. 31: 61-64.

Argueso, JL, Wanat, J, Gemici, Z, Alani, E. 2004. Competing crossover pathways act during meiosis in Saccharomyces cerevisiae. Genetics 168: 1805-1816.

Auclair, J, Leroux, D, Desseigne, F, Lasset, C, Saurin, JC, Joly, MO, Pinson, S, Xu, XL, Montmain, G, Ruano, E, Navarro, C, Puisieux, A, Wang, Q. 2007. Novel biallelic mutations in MSH6 and PMS2 genes: gene conversion as a likely cause of PMS2 gene inactivation. Hum.Mutat. 28: 1084-1090.

Baerenfaller, K, Fischer, F, Jiricny, J. 2006. Characterization of the "mismatch repairosome" and its role in the processing of modified nucleosides in vitro. Methods Enzymol. 408: 285-303.

Baker, SM, Plug, AW, Prolla, TA, Bronner, CE, Harris, AC, Yao, X, Christie, DM, Monell, C, Arnheim, N, Bradley, A, Ashley, T, Liskay, RM. 1996. Involvement of mouse Mlh1 in DNA mismatch repair and meiotic crossing over. Nat.Genet. 13: 336-342.

Ban, C & Yang, W. 1998. Crystal structure and ATPase activity of MutL: implications for DNA repair and mutagenesis. Cell 95: 541-552.

Ban, C, Junop, M, Yang, W. 1999. Transformation of MutL by ATP binding and hydrolysis: a switch in DNA mismatch repair. Cell 97: 85-97.

Bergerat, A, de Massy, B, Gadelle, D, Varoutas, PC, Nicolas, A, Forterre, P. 1997. An atypical topoisomerase II from Archaea with implications for meiotic recombination. Nature 386: 414-417.

Bisgaard, ML, Fenger, K, Bulow, S, Niebuhr, E, Mohr, J. 1994. Familial adenomatous polyposis (FAP): frequency, penetrance, and mutation rate. Hum.Mutat. 3: 121-125.

Blackwell, LJ, Wang, S, Modrich, P. 2001. DNA chain length dependence of formation and dynamics of hMutSalpha.hMutLalpha.heteroduplex complexes. J.Biol.Chem. 276: 33233-33240.

Bocker, T, Barusevicius, A, Snowden, T, Rasio, D, Guerrette, S, Robbins, D, Schmidt, C, Burczak, J, Croce, CM, Copeland, T, Kovatich, AJ, Fishel, R. 1999. hMSH5: a human MutS homologue that forms a novel heterodimer with hMSH4 and is expressed during spermatogenesis. Cancer Res. 59: 816-822.

Boland, CR, Thibodeau, SN, Hamilton, SR, Sidransky, D, Eshleman, JR, Burt, RW, Meltzer, SJ, Rodriguez-Bigas, MA, Fodde, R, Ranzani, GN, Srivastava, S. 1998. A National Cancer Institute Workshop on

69

Microsatellite Instability for cancer detection and familial predisposition: development of international criteria for the determination of microsatellite instability in colorectal cancer. Cancer Res. 58: 5248-5257.

Brieger, A, Trojan, J, Raedle, J, Plotz, G, Zeuzem, S. 2002. Transient mismatch repair gene transfection for functional analysis of genetic hMLH1 and hMSH2 variants. Gut 51: 677-684.

Brieger, A, Plotz, G, Raedle, J, Weber, N, Baum, W, Caspary, WF, Zeuzem, S, Trojan, J. 2005. Characterization of the nuclear import of human MutLalpha. Mol.Carcinog. 43: 51-58.

Brown, J, Brown, T, Fox, KR. 2001. Affinity of mismatch-binding protein MutS for heteroduplexes containing different mismatches. Biochem.J. 354: 627-633.

Bruni, R, Martin, D, Jiricny, J. 1988. d(GATC) sequences influence Escherichia coli mismatch repair in a distance-dependent manner from positions both upstream and downstream of the mismatch. Nucleic Acids Res. 16: 4875-4890.

Burdett, V, Baitinger, C, Viswanathan, M, Lovett, ST, Modrich, P. 2001. In vivo requirement for RecJ, ExoVII, ExoI, and ExoX in methyl-directed mismatch repair. Proc.Natl.Acad.Sci.U.S.A. 98: 6765-6770.

Caldes, T, Godino, J, de la Hoya, M, Garcia Carbonero, I, Perez Segura, P, Eng, C, Benito, M, Diaz-Rubio, E. 2002. Prevalence of germline mutations of MLH1 and MSH2 in hereditary nonpolyposis colorectal cancer families from Spain. Int.J.Cancer 98: 774-779.

Cannavo, E, Marra, G, Sabates-Bellver, J, Menigatti, M, Lipkin, SM, Fischer, F, Cejka, P, Jiricny, J. 2005. Expression of the MutL homologue hMLH3 in human cells and its role in DNA mismatch repair. Cancer Res. 65: 10759-10766.

Cannavo, E, Gerrits, B, Marra, G, Schlapbach, R, Jiricny, J. 2007. Characterization of the interactome of the human MutL homologues MLH1, PMS1, and PMS2. J.Biol.Chem. 282: 2976-2986.

Cascalho, M, Wong, J, Steinberg, C, Wabl, M. 1998. Mismatch repair co-opted by hypermutation. Science 279: 1207-1210.

Cejka, P, Stojic, L, Mojas, N, Russell, AM, Heinimann, K, Cannavo, E, di Pietro, M, Marra, G, Jiricny, J. 2003. Methylation-induced G(2)/M arrest requires a full complement of the mismatch repair protein hMLH1. EMBO J. 22: 2245-2254.

Chan, PA, Duraisamy, S, Miller, PJ, Newell, JA, McBride, C, Bond, JP, Raevaara, T, Ollila, S, Nyström, M, Grimm, AJ, Christodoulou, J, Oetting, WS, Greenblatt, MS. 2007. Interpreting missense variants: comparing computational methods in human disease genes CDKN2A, MLH1, MSH2, MECP2, and tyrosinase (TYR). Hum.Mutat. 28: 683-693.

Chan, TL, Yuen, ST, Kong, CK, Chan, YW, Chan, AS, Ng, WF, Tsui, WY, Lo, MW, Tam, WY, Li, VS, Leung, SY. 2006. Heritable germline epimutation of MSH2 in a family with hereditary nonpolyposis colorectal cancer. Nat.Genet. 38: 1178-1183.

Chang, DK, Ricciardiello, L, Goel, A, Chang, CL, Boland, CR. 2000. Steady-state regulation of the human DNA mismatch repair system. J.Biol.Chem. 275: 18424-18431.

Chao, EC, Velasquez, JL, Witherspoon, MS, Rozek, LS, Peel, D, Ng, P, Gruber, SB, Watson, P, Rennert, G, Anton-Culver, H, Lynch, H, Lipkin, SM. 2008. Accurate classification of MLH1/MSH2 missense variants with multivariate analysis of protein polymorphisms-mismatch repair (MAPP-MMR). Hum.Mutat. 29: 852-860.

Chen, PC, Dudley, S, Hagen, W, Dizon, D, Paxton, L, Reichow, D, Yoon, SR, Yang, K, Arnheim, N, Liskay, RM, Lipkin, SM. 2005. Contributions by MutL homologues Mlh3 and Pms2 to DNA mismatch repair and tumor suppression in the mouse. Cancer Res. 65: 8662-8670.

70

Chen, PC, Kuraguchi, M, Velasquez, J, Wang, Y, Yang, K, Edwards, R, Gillen, D, Edelmann, W, Kucherlapati, R, Lipkin, SM. 2008. Novel roles for MLH3 deficiency and TLE6-like amplification in DNA mismatch repair-deficient gastrointestinal tumorigenesis and progression. PLoS Genet. 4: e1000092.

Christensen, LL, Madsen, BE, Wikman, FP, Wiuf, C, Koed, K, Tjonneland, A, Olsen, A, Syvanen, AC, Andersen, CL, Orntoft, TF. 2008. The association between genetic variants in hMLH1 and hMSH2 and the development of sporadic colorectal cancer in the Danish population. BMC Med.Genet. 9: 52.

Clark, AB, Cook, ME, Tran, HT, Gordenin, DA, Resnick, MA, Kunkel, TA. 1999. Functional analysis of human MutSalpha and MutSbeta complexes in yeast. Nucleic Acids Res. 27: 736-742.

Clark, AB, Valle, F, Drotschmann, K, Gary, RK, Kunkel, TA. 2000. Functional interaction of proliferating cell nuclear antigen with MSH2-MSH6 and MSH2-MSH3 complexes. J.Biol.Chem. 275: 36498-36501.

Cleaver, JE. 1968. Defective repair replication of DNA in xeroderma pigmentosum. Nature 218: 652-656.

Clendenning, M, Hampel, H, LaJeunesse, J, Lindblom, A, Lockman, J, Nilbert, M, Senter, L, Sotamaa, K, de la Chapelle, A. 2006. Long-range PCR facilitates the identification of PMS2-specific mutations. Hum.Mutat. 27: 490-495.

Constantin, N, Dzantiev, L, Kadyrov, FA, Modrich, P. 2005. Human mismatch repair: reconstitution of a nick-directed bidirectional reaction. J.Biol.Chem. 280: 39752-39761.

Cooper, GM, Brudno, M, NISC Comparative Sequencing Program, Green, ED, Batzoglou, S, Sidow, A. 2003. Quantitative estimates of sequence divergence for comparative analyses of mammalian genomes. Genome Res. 13: 813-820.

Cunningham, JM, Christensen, ER, Tester, DJ, Kim, CY, Roche, PC, Burgart, LJ, Thibodeau, SN. 1998. Hypermethylation of the hMLH1 promoter in colon cancer with microsatellite instability. Cancer Res. 58: 3455-3460.

Dao, V&Modrich, P. 1998. Mismatch-, MutS-, MutL-, and helicase II-dependent unwinding from the single-strand break of an incised heteroduplex. J.Biol.Chem. 273: 9202-9207.

D'Atri, S, Tentori, L, Lacal, PM, Graziani, G, Pagani, E, Benincasa, E, Zambruno, G, Bonmassar, E, Jiricny, J. 1998. Involvement of the mismatch repair system in temozolomide-induced apoptosis. Mol.Pharmacol. 54: 334-341.

de la Chapelle, A. 2003. Microsatellite instability. N.Engl.J.Med. 349: 209-210.

Dingwall, C & Laskey, RA. 1998. Nuclear import: a tale of two sites. Curr.Biol. 8:922-924.

Dohet, C, Wagner, R, Radman, M. 1985. Repair of defined single base-pair mismatches in Escherichia coli. Proc.Natl.Acad.Sci.U.S.A. 82: 503-505.

Drotschmann, K, Clark, AB, Tran, HT, Resnick, MA, Gordenin, DA, Kunkel, TA. 1999. Mutator phenotypes of yeast strains heterozygous for mutations in the MSH2 gene. Proc.Natl.Acad.Sci.U.S.A. 96: 2970-2975.

Drummond, JT & Bellacosa, A. 2001. Human DNA mismatch repair in vitro operates independently of methylation status at CpG sites. Nucleic Acids Res. 29: 2234-2243.

Dutta, R & Inouye, M. 2000. GHKL, an emergent ATPase/kinase superfamily. Trends Biochem.Sci. 25: 24-28.

Ehrenstein, MR & Neuberger, MS. 1999. Deficiency in Msh2 affects the efficiency and local sequence specificity of immunoglobulin class-switch recombination: parallels with somatic hypermutation. EMBO J. 18: 3484-3490.

71

Ehrenstein, MR, Rada, C, Jones, AM, Milstein, C, Neuberger, MS. 2001. Switch junction sequences in PMS2-deficient mice reveal a microhomology-mediated mechanism of Ig class switch recombination. Proc.Natl.Acad.Sci.U.S.A. 98: 14553-14558.

Ellison, AR, Lofing, J, Bitter, GA. 2001. Functional analysis of human MLH1 and MSH2 missense variants and hybrid human-yeast MLH1 proteins in Saccharomyces cerevisiae. Hum.Mol.Genet. 10: 1889-1900.

Etzler, J, Peyrl, A, Zatkova, A, Schildhaus, HU, Ficek, A, Merkelbach-Bruse, S, Kratz, CP, Attarbaschi, A, Hainfellner, JA, Yao, S, Messiaen, L, Slavc, I, Wimmer, K. 2008. RNA-based mutation analysis identifies an unusual MSH6 splicing defect and circumvents PMS2 pseudogene interference. Hum.Mutat. 29: 299-305.

Fang, WH & Modrich, P. 1993. Human strand-specific mismatch repair occurs by a bidirectional mechanism similar to that of the bacterial reaction. J.Biol.Chem. 268: 11838-11844.

Ferrer-Costa, C, Gelpi, JL, Zamakola, L, Parraga, I, de la Cruz, X, Orozco, M. 2005. PMUT: a web-based tool for the annotation of pathological mutations on proteins. Bioinformatics 21: 3176-3178.

Fishel, R, Ewel, A, Lee, S, Lescoe, MK, Griffith, J. 1994. Binding of mismatched microsatellite DNA sequences by the human MSH2 protein. Science 266: 1403-1405.

Fleming, MA, Potter, JD, Ramirez, CJ, Ostrander, GK, Ostrander, EA. 2003. Understanding missense mutations in the BRCA1 gene: an evolutionary approach. Proc.Natl.Acad.Sci.U.S.A. 100: 1151-1156.

Flores-Rozas, H & Kolodner, RD. 1998. The Saccharomyces cerevisiae MLH3 gene functions in MSH3-dependent suppression of frameshift mutations. Proc.Natl.Acad.Sci.U.S.A. 95: 12404-12409.

Furukawa, T, Konishi, F, Shitoh, K, Kojima, M, Nagai, H, Tsukamoto, T. 2002. Evaluation of screening strategy for detecting hereditary nonpolyposis colorectal carcinoma. Cancer 94: 911-920.

Gallinger, S, Aronson, M, Shayan, K, Ratcliffe, EM, Gerstle, JT, Parkin, PC, Rothenmund, H, Croitoru, M, Baumann, E, Durie, PR, Weksberg, R, Pollett, A, Riddell, RH, Ngan, BY, Cutz, E, Lagarde, AE, Chan, HS. 2004. Gastrointestinal cancers and neurofibromatosis type 1 features in children with a germline homozygous MLH1 mutation. Gastroenterology 126: 576-585.

Gammie, AE, Erdeniz, N, Beaver, J, Devlin, B, Nanji, A, Rose, MD. 2007. Functional characterization of pathogenic human MSH2 missense mutations in Saccharomyces cerevisiae. Genetics 177: 707-721.

Gazzoli, I, Loda, M, Garber, J, Syngal, S, Kolodner, RD. 2002. A hereditary nonpolyposis colorectal carcinoma case associated with hypermethylation of the MLH1 gene in normal tissue and loss of heterozygosity of the unmethylated allele in the resulting microsatellite instability-high tumor. Cancer Res. 62: 3925-3928.

Genschel, J, Bazemore, LR, Modrich, P. 2002. Human exonuclease I is required for 5' and 3' mismatch repair. J.Biol.Chem. 277: 13302-13311.

Genschel, J & Modrich, P. 2003. Mechanism of 5'-directed excision in human mismatch repair. Mol.Cell 12: 1077-1086.

Goldgar, DE, Easton, DF, Deffenbaugh, AM, Monteiro, AN, Tavtigian, SV, Couch, FJ, Breast Cancer Information Core (BIC) Steering Committee. 2004. Integrated evaluation of DNA sequence variants of unknown clinical significance: application to BRCA1 and BRCA2. Am.J.Hum.Genet. 75: 535-544.

Gradia, S, Acharya, S, Fishel, R. 1997. The human mismatch recognition complex hMSH2-hMSH6 functions as a novel molecular switch. Cell 91: 995-1005.

72

Grady, WM, Rajput, A, Myeroff, L, Liu, DF, Kwon, K, Willis, J, Markowitz, S. 1998. Mutation of the type II transforming growth factor-beta receptor is coincident with the transformation of human colon adenomas to malignant carcinomas. Cancer Res. 58: 3101-3104.

Greenblatt, MS, Beaudet, JG, Gump, JR, Godin, KS, Trombley, L, Koh, J, Bond, JP. 2003. Detailed computational study of p53 and p16: using evolutionary sequence analysis and disease-associated mutations to predict the functional consequences of allelic variants. Oncogene 22: 1150-1163.

Grilley, M, Welsh, KM, Su, SS, Modrich, P. 1989. Isolation and characterization of the Escherichia coli mutL gene product. J.Biol.Chem. 264: 1000-1004.

Grilley, M, Griffith, J, Modrich, P. 1993. Bidirectional excision in methyl-directed mismatch repair. J.Biol.Chem. 268: 11830-11837.

Groden, J, Thliveris, A, Samowitz, W, Carlson, M, Gelbert, L, Albertsen, H, Joslyn, G, Stevens, J, Spirio, L, Robertson, M. 1991. Identification and characterization of the familial adenomatous polyposis coli gene. Cell 66: 589-600.

Guarne, A, Ramon-Maiques, S, Wolff, EM, Ghirlando, R, Hu, X, Miller, JH, Yang, W. 2004. Structure of the MutL C-terminal domain: a model of intact MutL and its roles in mismatch repair. EMBO J. 23: 4134-4145.

Guerrette, S, Wilson, T, Gradia, S, Fishel, R. 1998. Interactions of human hMSH2 with hMSH3 and hMSH2 with hMSH6: examination of mutations found in hereditary nonpolyposis colorectal cancer. Mol.Cell.Biol. 18: 6616-6623.

Guerrette, S, Acharya, S, Fishel, R. 1999. The interaction of the human MutL homologues in hereditary nonpolyposis colon cancer. J.Biol.Chem. 274: 6336-6341.

Gylling, A, Ridanpaa, M, Vierimaa, O, Aittomaki, K, Avela, K, Kaariainen, H, Laivuori, H, Poyhonen, M, Sallinen, SL, Wallgren-Pettersson, C, Järvinen, HJ, Mecklin, JP, Peltomäki, P. 2009. Large genomic rearrangements and germline epimutations in Lynch syndrome. Int.J.Cancer 124: 2333-2340.

Görlich, D, Vogel, F, Mills, AD, Hartmann, E, Laskey, RA. 1995. Distinct functions for the two importin subunits in nuclear protein import. Nature 377: 246-248.

Hampel, H, Frankel, WL, Martin, E, Arnold, M, Khanduja, K, Kuebler, P, Nakagawa, H, Sotamaa, K, Prior, TW, Westman, J, Panescu, J, Fix, D, Lockman, J, Comeras, I, de la Chapelle, A. 2005. Screening for the Lynch syndrome (hereditary nonpolyposis colorectal cancer). N.Engl.J.Med. 352: 1851-1860.

Harfe BD & Jinks-Robertson S. 2000. DNA mismatch repair and genetic instability. Annu Rev Genet. 34: 359-399.

Hawn, MT, Umar, A, Carethers, JM, Marra, G, Kunkel, TA, Boland, CR, Koi, M. 1995. Evidence for a connection between the mismatch repair system and the G2 cell cycle checkpoint. Cancer Res. 55: 3721-3725.

Heinen, CD, Wilson, T, Mazurek, A, Berardini, M, Butz, C, Fishel, R. 2002. HNPCC mutations in hMSH2 result in reduced hMSH2-hMSH6 molecular switch functions. Cancer.Cell. 1: 469-478.

Hempen, PM, Zhang, L, Bansal, RK, Iacobuzio-Donahue, CA, Murphy, KM, Maitra, A, Vogelstein, B, Whitehead, RH, Markowitz, SD, Willson, JK, Yeo, CJ, Hruban, RH, Kern, SE. 2003. Evidence of selection for clones having genetic inactivation of the activin A type II receptor (ACVR2) gene in gastrointestinal cancers. Cancer Res. 63: 994-999.

73

Hienonen, T, Laiho, P, Salovaara, R, Mecklin, JP, Järvinen, H, Sistonen, P, Peltomäki, P, Lehtonen, R, Nupponen, NN, Launonen, V, Karhu, A, Aaltonen, LA. 2003. Little evidence for involvement of MLH3 in colorectal cancer predisposition. Int.J.Cancer 106: 292-296.

Hitchins, M, Williams, R, Cheong, K, Halani, N, Lin, VA, Packham, D, Ku, S, Buckle, A, Hawkins, N, Burn, J, Gallinger, S, Goldblatt, J, Kirk, J, Tomlinson, I, Scott, R, Spigelman, A, Suter, C, Martin, D, Suthers, G, Ward, R. 2005. MLH1 germline epimutations as a factor in hereditary nonpolyposis colorectal cancer. Gastroenterology 129: 1392-1399.

Hitchins, MP, Wong, JJ, Suthers, G, Suter, CM, Martin, DI, Hawkins, NJ, Ward, RL. 2007. Inheritance of a cancer-associated MLH1 germ-line epimutation. N.Engl.J.Med. 356: 697-705.

Hoffmann, ER & Borts, RH. 2004. Meiotic recombination intermediates and mismatch repair proteins. Cytogenet.Genome Res. 107: 232-248.

Hofstra, RM, Spurdle, AB, Eccles, D, Foulkes, WD, de Wind, N, Hoogerbrugge, N, Hogervorst, FB, IARC Unclassified Genetic Variants Working Group. 2008. Tumor characteristics as an analytic tool for classifying genetic variants of uncertain clinical significance. Hum.Mutat. 29: 1292-1303.

Holmes, J,Jr, Clark, S, Modrich, P. 1990. Strand-specific mismatch correction in nuclear extracts of human and Drosophila melanogaster cell lines. Proc.Natl.Acad.Sci.U.S.A. 87: 5837-5841.

Hunter, N&Borts, RH. 1997. Mlh1 is unique among mismatch repair proteins in its ability to promote crossing-over during meiosis. Genes Dev. 11: 1573-1582.

Iaccarino, I, Marra, G, Dufner, P, Jiricny, J. 2000. Mutation in the magnesium binding site of hMSH6 disables the hMutSalpha sliding clamp from translocating along DNA. J.Biol.Chem. 275: 2080-2086.

Jackson, SP. 2002. Sensing and repairing DNA double-strand breaks. Carcinogenesis 23: 687-696.

Jakubowska, A, Gorski, B, Kurzawski, G, Debniak, T, Hadaczek, P, Cybulski, C, Kladny, J, Oszurek, O, Scott, RJ, Lubinski, J. 2001. Optimization of experimental conditions for RNA-based sequencing of MLH1 and MSH2 genes. Hum.Mutat. 17: 52-60.

Jass, JR. 2008. Colorectal polyposes: from phenotype to diagnosis. Pathol.Res.Pract. 204: 431-447.

Jemal, A, Siegel, R, Ward, E, Hao, Y, Xu, J, Murray, T, Thun, MJ. 2008. Cancer statistics, 2008. CA Cancer.J.Clin. 58: 71-96.

Jeske, YW, So, A, Kelemen, L, Sukor, N, Willys, C, Bulmer, B, Gordon, RD, Duffy, D, Stowasser, M. 2008. Examination of chromosome 7p22 candidate genes RBaK, PMS2 and GNA12 in familial hyperaldosteronism type II. Clin.Exp.Pharmacol.Physiol. 35: 380-385.

Jiricny, J&Nyström-Lahti, M. 2000. Mismatch repair defects in cancer. Curr.Opin.Genet.Dev. 10: 157-161.

Jiricny, J. 2006. The multifaceted mismatch-repair system. Nat.Rev.Mol.Cell Biol. 7: 335-346.

Joensuu, EI, Abdel-Rahman, WM, Ollikainen, M, Ruosaari, S, Knuutila, S, Peltomäki, P. 2008. Epigenetic signatures of familial cancer are characteristic of tumor type and family category. Cancer Res. 68: 4597-4605.

Johnson, SJ & Beese, LS. 2004. Structures of mismatch replication errors observed in a DNA polymerase. Cell 116: 803-816.

Jung, B, Doctolero, RT, Tajima, A, Nguyen, AK, Keku, T, Sandler, RS, Carethers, JM. 2004. Loss of activin receptor type 2 protein expression in microsatellite unstable colon cancers. Gastroenterology 126: 654-659.

74

Junop, MS, Obmolova, G, Rausch, K, Hsieh, P, Yang, W. 2001. Composite active site of an ABC ATPase: MutS uses ATP to verify mismatch recognition and authorize DNA repair. Mol.Cell 7: 1-12.

Kadyrov, FA, Dzantiev, L, Constantin, N, Modrich, P. 2006. Endonucleolytic function of MutLalpha in human mismatch repair. Cell 126: 297-308.

Kaina, B, Christmann, M, Naumann, S, Roos, WP. 2007. MGMT: key node in the battle against genotoxicity, carcinogenicity and apoptosis induced by alkylating agents. DNA Repair (Amst) 6: 1079-1099.

Kalderon, D, Richardson, WD, Markham, AF, Smith, AE. 1984. Sequence requirements for nuclear location of simian virus 40 large-T antigen. Nature 311: 33-38.

Kane, MF, Loda, M, Gaida, GM, Lipman, J, Mishra, R, Goldman, H, Jessup, JM, Kolodner, R. 1997. Methylation of the hMLH1 promoter correlates with lack of expression of hMLH1 in sporadic colon tumors and mismatch repair-defective human tumor cell lines. Cancer Res. 57: 808-811.

Kariola, R, Hampel, H, Frankel, WL, Raevaara, TE, de la Chapelle, A, Nyström-Lahti, M. 2004. MSH6 missense mutations are often associated with no or low cancer susceptibility. Br.J.Cancer 91: 1287-1292.

Kat, A, Thilly, WG, Fang, WH, Longley, MJ, Li, GM, Modrich, P. 1993. An alkylation-tolerant, mutator human cell line is deficient in strand-specific mismatch repair. Proc.Natl.Acad.Sci.U.S.A. 90: 6424-6428.

Kim, JC, Roh, SA, Yoon, YS, Kim, HC, Park, IJ. 2007. MLH3 and EXO1 alterations in familial colorectal cancer patients not fulfilling Amsterdam criteria. Cancer Genet.Cytogenet. 176: 172-174.

Kleczkowska, HE, Marra, G, Lettieri, T, Jiricny, J. 2001. hMSH3 and hMSH6 interact with PCNA and colocalize with it to replication foci. Genes Dev. 15: 724-736.

Kneitz, B, Cohen, PE, Avdievich, E, Zhu, L, Kane, MF, Hou, H,Jr, Kolodner, RD, Kucherlapati, R, Pollard, JW, Edelmann, W. 2000. MutS homolog 4 localization to meiotic chromosomes is required for chromosome pairing during meiosis in male and female mice. Genes Dev. 14: 1085-1097.

Knudson, AG. 1971. Mutation and cancer: statistical study of retinoblastoma. Proc.Natl.Acad.Sci.U.S.A. 68: 820-823.

Knudson, AG. 1996. Hereditary cancer: two hits revisited. J.Cancer Res.Clin.Oncol. 122: 135-140.

Kohonen-Corish, M, Ross, VL, Doe, WF, Kool, DA, Edkins, E, Faragher, I, Wijnen, J, Khan, PM, Macrae, F, St John, DJ. 1996. RNA-based mutation screening in hereditary nonpolyposis colorectal cancer. Am.J.Hum.Genet. 59: 818-824.

Kolas, NK & Cohen, PE. 2004. Novel and diverse functions of the DNA mismatch repair family in mammalian meiosis and recombination. Cytogenet.Genome Res. 107: 216-231.

Kondo, E, Horii, A, Fukushige, S. 2001. The interacting domains of three MutL heterodimers in man: hMLH1 interacts with 36 homologous amino acid residues within hMLH3, hPMS1 and hPMS2. Nucleic Acids Res. 29: 1695-1702.

Kondo, E, Suzuki, H, Horii, A, Fukushige, S. 2003. A yeast two-hybrid assay provides a simple way to evaluate the vast majority of hMLH1 germ-line mutations. Cancer Res. 63: 3302-3308.

Kosinski, J, Plotz, G, Guarne, A, Bujnicki, JM, Friedhoff, P. 2008. The PMS2 subunit of human MutLalpha contains a metal ion binding domain of the iron-dependent repressor protein family. J.Mol.Biol. 382: 610-627.

Kramer, B, Kramer, W, Fritz, HJ. 1984. Different base/base mismatches are corrected with different efficiencies by the methyl-directed DNA mismatch-repair system of E. coli. Cell 38: 879-887.

75

Kunkel, TA. 1993. Nucleotide repeats. Slippery DNA and diseases. Nature 365: 207-208.

Kunkel, TA & Erie, DA. 2005. DNA mismatch repair. Annu.Rev.Biochem. 74: 681-710.

Kunz, C, Saito, Y, Schar, P. 2009. DNA repair in mammalian cells: Mismatched repair: variations on a theme. Cell Mol.Life Sci. 66: 1021-1038.

Lagerstedt Robinson, K, Liu, T, Vandrovcova, J, Halvarsson, B, Clendenning, M, Frebourg, T, Papadopoulos, N, Kinzler, KW, Vogelstein, B, Peltomäki, P, Kolodner, RD, Nilbert, M, Lindblom, A. 2007. Lynch syndrome (hereditary nonpolyposis colorectal cancer) diagnostics. J.Natl.Cancer Inst. 99: 291-299.

Lahue, RS, Au, KG, Modrich, P. 1989. DNA mismatch correction in a defined system. Science 245: 160-164.

Lamers, MH, Perrakis, A, Enzlin, JH, Winterwerp, HH, de Wind, N, Sixma, TK. 2000. The crystal structure of DNA mismatch repair protein MutS binding to a G x T mismatch. Nature 407: 711-717.

Lange, A, Mills, RE, Lange, CJ, Stewart, M, Devine, SE, Corbett, AH. 2007. Classical nuclear localization signals: definition, function, and interaction with importin alpha. J.Biol.Chem. 282: 5101-5105.

Lau, PJ & Kolodner, RD. 2003. Transfer of the MSH2.MSH6 complex from proliferating cell nuclear antigen to mispaired bases in DNA. J.Biol.Chem. 278: 14-17.

Leong, V, Lorenowicz, J, Kozij, N, Guarne, A. 2009. Nuclear import of human MLH1, PMS2, and MutLalpha: Redundancy is the key. Mol.Carcinog. in press.

Li, GM & Modrich, P. 1995. Restoration of mismatch repair to nuclear extracts of H6 colorectal tumor cells by a heterodimer of human MutL homologs. Proc.Natl.Acad.Sci.U.S.A. 92: 1950-1954.

Li, Z, Scherer, SJ, Ronai, D, Iglesias-Ussel, MD, Peled, JU, Bardwell, PD, Zhuang, M, Lee, K, Martin, A, Edelmann, W, Scharff, MD. 2004b. Examination of Msh6- and Msh3-deficient mice in class switching reveals overlapping and distinct roles of MutS homologues in antibody diversification. J.Exp.Med. 200: 47-59.

Li, Z, Woo, CJ, Iglesias-Ussel, MD, Ronai, D, Scharff, MD. 2004a. The generation of antibody diversity through somatic hypermutation and class switch recombination. Genes Dev. 18: 1-11.

Li, Z, Peled, JU, Zhao, C, Svetlanov, A, Ronai, D, Cohen, PE, Scharff, MD. 2006. A role for Mlh3 in somatic hypermutation. DNA Repair (Amst) 5: 675-682.

Lindblom, A, Tannergard, P, Werelius, B, Nordenskjold, M. 1993. Genetic mapping of a second locus predisposing to hereditary non-polyposis colon cancer. Nat.Genet. 5: 279-282.

Lipkin, SM, Wang, V, Jacoby, R, Banerjee-Basu, S, Baxevanis, AD, Lynch, HT, Elliott, RM, Collins, FS. 2000. MLH3: a DNA mismatch repair gene associated with mammalian microsatellite instability. Nat.Genet. 24: 27-35.

Lipkin, SM, Wang, V, Stoler, DL, Anderson, GR, Kirsch, I, Hadley, D, Lynch, HT, Collins, FS. 2001. Germline and somatic mutation analyses in the DNA mismatch repair gene MLH3: Evidence for somatic mutation in colorectal cancers. Hum.Mutat. 17: 389-396.

Lipkin, SM, Moens, PB, Wang, V, Lenzi, M, Shanmugarajah, D, Gilgeous, A, Thomas, J, Cheng, J, Touchman, JW, Green, ED, Schwartzberg, P, Collins, FS, Cohen, PE. 2002. Meiotic arrest and aneuploidy in MLH3-deficient mice. Nat.Genet. 31: 385-390.

Liu, B, Parsons, R, Papadopoulos, N, Nicolaides, NC, Lynch, HT, Watson, P, Jass, JR, Dunlop, M, Wyllie, A, Peltomäki, P, de la Chapelle, A, Hamilton, SR, Vogelstein, B, Kinzler, KW. 1996. Analysis of mismatch repair genes in hereditary non-polyposis colorectal cancer patients. Nat.Med. 2: 169-174.

76

Liu, HX, Zhou, XL, Liu, T, Werelius, B, Lindmark, G, Dahl, N, Lindblom, A. 2003. The role of hMLH3 in familial colorectal cancer. Cancer Res. 63: 1894-1899.

Liu, T, Yan, H, Kuismanen, S, Percesepe, A, Bisgaard, ML, Pedroni, M, Benatti, P, Kinzler, KW, Vogelstein, B, Ponz de Leon, M, Peltomäki, P, Lindblom, A. 2001. The role of hPMS1 and hPMS2 in predisposing to colorectal cancer. Cancer Res. 61: 7798-7802.

Loeb, LA. 1994. Microsatellite instability: marker of a mutator phenotype in cancer. Cancer Res. 54: 5059-5063.

Loukola, A, Vilkki, S, Singh, J, Launonen, V, Aaltonen, LA. 2000. Germline and somatic mutation analysis of MLH3 in MSI-positive colorectal cancer. Am.J.Pathol. 157: 347-352.

Lu, AL, Clark, S, Modrich, P. 1983. Methyl-directed repair of DNA base-pair mismatches in vitro. Proc.Natl.Acad.Sci.U.S.A. 80: 4639-4643.

Lu, AL, Welsh, K, Clark, S, Su, SS, Modrich, P. 1984. Repair of DNA base-pair mismatches in extracts of Escherichia coli. Cold Spring Harb.Symp.Quant.Biol. 49: 589-596.

Luo, Y, Lin, FT, Lin, WC. 2004. ATM-mediated stabilization of hMutL DNA mismatch repair proteins augments p53 activation during DNA damage. Mol.Cell.Biol. 24: 6430-6444.

Lynch, HT, Smyrk, TC, Watson, P, Lanspa, SJ, Lynch, JF, Lynch, PM, Cavalieri, RJ, Boland, CR. 1993. Genetics, natural history, tumor spectrum, and pathology of hereditary nonpolyposis colorectal cancer: an updated review. Gastroenterology 104: 1535-1549.

Lynch, HT & de la Chapelle, A. 1999. Genetic susceptibility to non-polyposis colorectal cancer. J.Med.Genet. 36: 801-818.

Lynch, HT, Lynch, JF, Lynch, PM, Attard, T. 2008. Hereditary colorectal cancer syndromes: molecular genetics, genetic counseling, diagnosis and management. Fam.Cancer. 7: 27-39.

Mangold, E, Pagenstecher, C, Friedl, W, Mathiak, M, Buettner, R, Engel, C, Loeffler, M, Holinski-Feder, E, Muller-Koch, Y, Keller, G, Schackert, HK, Kruger, S, Goecke, T, Moeslein, G, Kloor, M, Gebert, J, Kunstmann, E, Schulmann, K, Ruschoff, J, Propping, P. 2005. Spectrum and frequencies of mutations in MSH2 and MLH1 identified in 1,721 German families suspected of hereditary nonpolyposis colorectal cancer. Int.J.Cancer 116: 692-702.

Margison, GP, Santibanez Koref, MF, Povey, AC. 2002. Mechanisms of carcinogenicity/chemotherapy by O6-methylguanine. Mutagenesis 17: 483-487.

Markowitz, S, Wang, J, Myeroff, L, Parsons, R, Sun, L, Lutterbaugh, J, Fan, RS, Zborowska, E, Kinzler, KW, Vogelstein, B. 1995. Inactivation of the type II TGF-beta receptor in colon cancer cells with microsatellite instability. Science 268: 1336-1338.

Marra, G, Iaccarino, I, Lettieri, T, Roscilli, G, Delmastro, P, Jiricny, J. 1998. Mismatch repair deficiency associated with overexpression of the MSH3 gene. Proc.Natl.Acad.Sci.U.S.A. 95: 8568-8573.

Martin, A & Scharff, MD. 2002. Somatic hypermutation of the AID transgene in B and non-B cells. Proc.Natl.Acad.Sci.U.S.A. 99: 12304-12308.

Martin, A, Li, Z, Lin, DP, Bardwell, PD, Iglesias-Ussel, MD, Edelmann, W, Scharff, MD. 2003. Msh2 ATPase activity is essential for somatic hypermutation at a-T basepairs and for efficient class switch recombination. J.Exp.Med. 198: 1171-1178.

Mattaj, IW & Englmeier, L. 1998. Nucleocytoplasmic transport: the soluble phase. Annu.Rev.Biochem. 67: 265-306.

77

Mecklin, JP. 1987. Frequency of hereditary colorectal carcinoma. Gastroenterology 93: 1021-1025.

Miyakura, Y, Sugano, K, Akasu, T, Yoshida, T, Maekawa, M, Saitoh, S, Sasaki, H, Nomizu, T, Konishi, F, Fujita, S, Moriya, Y, Nagai, H. 2004. Extensive but hemiallelic methylation of the hMLH1 promoter region in early-onset sporadic colon cancers with microsatellite instability. Clin.Gastroenterol.Hepatol. 2: 147-156.

Modrich, P. 1991. Mechanisms and biological effects of mismatch repair. Annu.Rev.Genet. 25: 229-253.

Modrich, P & Lahue, R. 1996. Mismatch repair in replication fidelity, genetic recombination, and cancer biology. Annu.Rev.Biochem. 65: 101-133.

Moisio, AL, Sistonen, P, Weissenbach, J, de la Chapelle, A, Peltomäki, P. 1996. Age and origin of two common MLH1 mutations predisposing to hereditary colon cancer. Am.J.Hum.Genet. 59: 1243-1251.

Morak, M, Schackert, HK, Rahner, N, Betz, B, Ebert, M, Walldorf, C, Royer-Pokora, B, Schulmann, K, von Knebel-Doeberitz, M, Dietmaier, W, Keller, G, Kerker, B, Leitner, G, Holinski-Feder, E. 2008. Further evidence for heritability of an epimutation in one of 12 cases with MLH1 promoter methylation in blood cells clinically displaying HNPCC. Eur.J.Hum.Genet. 16: 804-811.

Mushegian, AR, Bassett, DE,Jr, Boguski, MS, Bork, P, Koonin, EV. 1997. Positionally cloned human disease genes: patterns of evolutionary conservation and functional motifs. Proc.Natl.Acad.Sci.U.S.A. 94: 5831-5836.

Nakagawa, H, Lockman, JC, Frankel, WL, Hampel, H, Steenblock, K, Burgart, LJ, Thibodeau, SN, de la Chapelle, A. 2004. Mismatch repair gene PMS2: disease-causing germline mutations are frequent in patients whose tumors stain negative for PMS2 protein, but paralogous genes obscure mutation detection and interpretation. Cancer Res. 64: 4721-4727.

Nakahara, T, Zhang, QM, Hashiguchi, K, Yonei, S. 2000. Identification of proteins of Escherichia coli and Saccharomyces cerevisiae that specifically bind to C/C mismatches in DNA. Nucleic Acids Res. 28: 2551-2556.

Ng, PC & Henikoff, S. 2001. Predicting deleterious amino acid substitutions. Genome Res. 11: 863-874.

Nicolaides, NC, Papadopoulos, N, Liu, B, Wei, YF, Carter, KC, Ruben, SM, Rosen, CA, Haseltine, WA, Fleischmann, RD, Fraser, CM. 1994. Mutations of two PMS homologues in hereditary nonpolyposis colon cancer. Nature 371: 75-80.

Nicolaides, NC, Carter, KC, Shell, BK, Papadopoulos, N, Vogelstein, B, Kinzler, KW. 1995a. Genomic organization of the human PMS2 gene family. Genomics 30: 195-206.

Nicolaides, NC, Kinzler, KW, Vogelstein, B. 1995b. Analysis of the 5' region of PMS2 reveals heterogeneous transcripts and a novel overlapping gene. Genomics 29: 329-334.

Nicolaides, NC, Littman, SJ, Modrich, P, Kinzler, KW, Vogelstein, B. 1998. A naturally occurring hPMS2 mutation can confer a dominant negative mutator phenotype. Mol.Cell.Biol. 18: 1635-1641.

Nishant, KT, Plys, AJ, Alani, E. 2008. A mutation in the putative MLH3 endonuclease domain confers a defect in both mismatch repair and meiosis in Saccharomyces cerevisiae. Genetics 179: 747-755.

Nyström-Lahti, M, Kristo, P, Nicolaides, NC, Chang, SY, Aaltonen, LA, Moisio, AL, Järvinen, HJ, Mecklin, JP, Kinzler, KW, Vogelstein, B. 1995. Founding mutations and Alu-mediated recombination in hereditary colon cancer. Nat.Med. 1: 1203-1206.

78

Nyström-Lahti, M, Holmberg, M, Fidalgo, P, Salovaara, R, de la Chapelle, A, Jiricny, J, Peltomäki, P. 1999. Missense and nonsense mutations in codon 659 of MLH1 cause aberrant splicing of messenger RNA in HNPCC kindreds. Genes Chromosomes Cancer 26: 372-375.

Nyström-Lahti, M, Perrera, C, Räschle, M, Panyushkina-Seiler, E, Marra, G, Curci, A, Quaresima, B, Costanzo, F, D'Urso, M, Venuta, S, Jiricny, J. 2002. Functional analysis of MLH1 mutations linked to hereditary nonpolyposis colon cancer. Genes Chromosomes Cancer 33: 160-167.

Ohmiya, N, Matsumoto, S, Yamamoto, H, Baranovskaya, S, Malkhosyan, SR, Perucho, M. 2001. Germline and somatic mutations in hMSH6 and hMSH3 in gastrointestinal cancers of the microsatellite mutator phenotype. Gene 272: 301-313.

Ollila, S, Sarantaus, L, Kariola, R, Chan, P, Hampel, H, Holinski-Feder, E, Macrae, F, Kohonen-Corish, M, Gerdes, AM, Peltomäki, P, Mangold, E, de la Chapelle, A, Greenblatt, M, Nyström, M. 2006. Pathogenicity of MSH2 missense mutations is typically associated with impaired repair capability of the mutated protein. Gastroenterology 131: 1408-1417.

Ollila, S, Dermadi Bebek, D, Greenblatt, M, Nyström, M. 2008a. Uncertain pathogenicity of MSH2 variants N127S and G322D challenges their classification. Int.J.Cancer 123: 720-724.

Ollila, S, Dermadi Bebek, D, Jiricny, J, Nyström, M. 2008b. Mechanisms of pathogenicity in human MSH2 missense mutants. Hum.Mutat. 29: 1355-1363.

Ou, J, Niessen, RC, Lutzen, A, Sijmons, RH, Kleibeuker, JH, de Wind, N, Rasmussen, LJ, Hofstra, RM. 2007. Functional analysis helps to clarify the clinical importance of unclassified variants in DNA mismatch repair genes. Hum.Mutat. 28: 1047-1054.

Ou, J, Niessen, RC, Vonk, J, Westers, H, Hofstra, RM, Sijmons, RH. 2008. A database to support the interpretation of human mismatch repair gene variants. Hum.Mutat. 29: 1337-1341.

Ou, J, Rasmussen, M, Westers, H, Andersen, SD, Jager, PO, Kooi, KA, Niessen, RC, Eggen, BJ, Nielsen, FC, Kleibeuker, JH, Sijmons, RH, Rasmussen, LJ, Hofstra, RM. 2009. Biochemical characterization of MLH3 missense mutations does not reveal an apparent role of MLH3 in Lynch syndrome. Genes Chromosomes Cancer 48: 340-350.

Palombo, F, Iaccarino, I, Nakajima, E, Ikejima, M, Shimada, T, Jiricny, J. 1996. hMutSbeta, a heterodimer of hMSH2 and hMSH3, binds to insertion/deletion loops in DNA. Curr.Biol. 6: 1181-1184.

Pegg, AE & Byers, TL. 1992. Repair of DNA containing O6-alkylguanine. FASEB J. 6: 2302-2310.

Peltomäki, P, Aaltonen, LA, Sistonen, P, Pylkkanen, L, Mecklin, JP, Järvinen, H, Green, JS, Jass, JR, Weber, JL, Leach, FS. 1993. Genetic mapping of a locus predisposing to human colorectal cancer. Science 260: 810-812.

Peltomäki, P. 2001. Deficient DNA mismatch repair: a common etiologic factor for colon cancer. Hum.Mol.Genet. 10: 735-740.

Peltomäki, P & Vasen, H. 2004. Mutations associated with HNPCC predisposition -- Update of ICG-HNPCC/INSiGHT mutation database. Dis.Markers 20: 269-276.

Peltomäki, P. 2005. Lynch syndrome genes. Fam.Cancer. 4: 227-232.

Plotz, G, Raedle, J, Brieger, A, Trojan, J, Zeuzem, S. 2002. hMutSalpha forms an ATP-dependent complex with hMutLalpha and hMutLbeta on DNA. Nucleic Acids Res. 30: 711-718.

Plotz, G, Raedle, J, Brieger, A, Trojan, J, Zeuzem, S. 2003. N-terminus of hMLH1 confers interaction of hMutLalpha and hMutLbeta with hMutSalpha. Nucleic Acids Res. 31: 3217-3226.

79

Potter, JD. 1999. Colorectal cancer: molecules and populations. J.Natl.Cancer Inst. 91: 916-932.

Povirk, LF. 2006. Biochemical mechanisms of chromosomal translocations resulting from DNA double-strand breaks. DNA Repair (Amst) 5: 1199-1212.

Prolla, TA, Abuin, A, Bradley, A. 1996. DNA mismatch repair deficient mice in cancer research. Semin.Cancer Biol. 7: 241-247.

Prolla, TA, Baker, SM, Harris, AC, Tsao, JL, Yao, X, Bronner, CE, Zheng, B, Gordon, M, Reneker, J, Arnheim, N, Shibata, D, Bradley, A, Liskay, RM. 1998. Tumour susceptibility and spontaneous mutation in mice deficient in Mlh1, Pms1 and Pms2 DNA mismatch repair. Nat.Genet. 18: 276-279.

Rada, C, Ehrenstein, MR, Neuberger, MS, Milstein, C. 1998. Hot spot focusing of somatic hypermutation in MSH2-deficient mice suggests two stages of mutational targeting. Immunity 9: 135-141.

Raevaara, TE, Timoharju, T, Lonnqvist, KE, Kariola, R, Steinhoff, M, Hofstra, RM, Mangold, E, Vos, YJ, Nyström-Lahti, M. 2002. Description and functional analysis of a novel in frame mutation linked to hereditary non-polyposis colorectal cancer. J.Med.Genet. 39: 747-750.

Raevaara, TE, Vaccaro, C, Abdel-Rahman, WM, Mocetti, E, Bala, S, Lonnqvist, KE, Kariola, R, Lynch, HT, Peltomäki, P, Nyström-Lahti, M. 2003. Pathogenicity of the hereditary colorectal cancer mutation hMLH1 del616 linked to shortage of the functional protein. Gastroenterology 125: 501-509.

Raevaara, TE, Gerdes, AM, Lonnqvist, KE, Tybjaerg-Hansen, A, Abdel-Rahman, WM, Kariola, R, Peltomäki, P, Nyström-Lahti, M. 2004. HNPCC mutation MLH1 P648S makes the functional protein unstable, and homozygosity predisposes to mild neurofibromatosis type 1. Genes Chromosomes Cancer 40: 261-265.

Ramensky, V, Bork, P, Sunyaev, S. 2002. Human non-synonymous SNPs: server and survey. Nucleic Acids Res. 30: 3894-3900.

Rampino, N, Yamamoto, H, Ionov, Y, Li, Y, Sawai, H, Reed, JC, Perucho, M. 1997. Somatic frameshift mutations in the BAX gene in colon cancers of the microsatellite mutator phenotype. Science 275: 967-969.

Räschle, M, Marra, G, Nyström-Lahti, M, Schar, P, Jiricny, J. 1999. Identification of hMutLbeta, a heterodimer of hMLH1 and hPMS1. J.Biol.Chem. 274: 32368-32375.

Räschle, M, Dufner, P, Marra, G, Jiricny, J. 2002. Mutations within the hMLH1 and hPMS2 subunits of the human MutLalpha mismatch repair factor affect its ATPase activity, but not its ability to interact with hMutSalpha. J.Biol.Chem. 277: 21810-21820.

Rich, T, Allen, RL, Wyllie, AH. 2000. Defying death after DNA damage. Nature 407: 777-783.

Rodriguez-Bigas, MA, Boland, CR, Hamilton, SR, Henson, DE, Jass, JR, Khan, PM, Lynch, H, Perucho, M, Smyrk, T, Sobin, L, Srivastava, S. 1997. A National Cancer Institute Workshop on Hereditary Nonpolyposis Colorectal Cancer Syndrome: meeting highlights and Bethesda guidelines. J.Natl.Cancer Inst. 89: 1758-1762.

Sacho, EJ, Kadyrov, FA, Modrich, P, Kunkel, TA, Erie, DA. 2008. Direct visualization of asymmetric adenine-nucleotide-induced conformational changes in MutL alpha. Mol.Cell 29: 112-121.

Santucci-Darmanin, S, Neyton, S, Lespinasse, F, Saunieres, A, Gaudray, P, Paquis-Flucklinger, V. 2002. The DNA mismatch-repair MLH3 protein interacts with MSH4 in meiotic cells, supporting a role for this MutL homolog in mammalian meiotic recombination. Hum.Mol.Genet. 11: 1697-1706.

Schrader, CE, Edelmann, W, Kucherlapati, R, Stavnezer, J. 1999. Reduced isotype switching in splenic B cells from mice deficient in mismatch repair enzymes. J.Exp.Med. 190: 323-330.

80

Sedgwick, B, Bates, PA, Paik, J, Jacobs, SC, Lindahl, T. 2007. Repair of alkylated DNA: recent advances. DNA Repair (Amst) 6: 429-442.

Senter, L, Clendenning, M, Sotamaa, K, Hampel, H, Green, J, Potter, JD, Lindblom, A, Lagerstedt, K, Thibodeau, SN, Lindor, NM, Young, J, Winship, I, Dowty, JG, White, DM, Hopper, JL, Baglietto, L, Jenkins, MA, de la Chapelle, A. 2008. The clinical phenotype of Lynch syndrome due to germ-line PMS2 mutations. Gastroenterology 135: 419-428.

Shcherbakova, PV&Kunkel, TA. 1999. Mutator phenotypes conferred by MLH1 overexpression and by heterozygosity for mlh1 mutations. Mol.Cell.Biol. 19: 3177-3183.

Shimodaira, H, Filosi, N, Shibata, H, Suzuki, T, Radice, P, Kanamaru, R, Friend, SH, Kolodner, RD, Ishioka, C. 1998. Functional analysis of human MLH1 mutations in Saccharomyces cerevisiae. Nat.Genet. 19: 384-389.

Shin, KH, Park, YJ, Park, JG. 2001. PTEN gene mutations in colorectal cancers displaying microsatellite instability. Cancer Lett. 174: 189-194.

Snowden, T, Acharya, S, Butz, C, Berardini, M, Fishel, R. 2004. hMSH4-hMSH5 recognizes Holliday Junctions and forms a meiosis-specific sliding clamp that embraces homologous chromosomes. Mol.Cell 15: 437-451.

Soulas-Sprauel, P, Rivera-Munoz, P, Malivert, L, Le Guyader, G, Abramowski, V, Revy, P, de Villartay, JP. 2007. V(D)J and immunoglobulin class switch recombinations: a paradigm to study the regulation of DNA end-joining. Oncogene 26: 7780-7791.

Stone, EA & Sidow, A. 2005. Physicochemical constraint violation by missense substitutions mediates impairment of protein function and disease severity. Genome Res. 15: 978-986.

Suter, CM, Martin, DI, Ward, RL. 2004. Germline epimutation of MLH1 in individuals with multiple cancers. Nat.Genet. 36: 497-501.

Takahashi, M, Shimodaira, H, Andreutti-Zaugg, C, Iggo, R, Kolodner, RD, Ishioka, C. 2007. Functional analysis of human MLH1 variants using yeast and in vitro mismatch repair assays. Cancer Res. 67: 4595-4604.

Tavtigian, SV, Greenblatt, MS, Lesueur, F, Byrnes, GB, IARC Unclassified Genetic Variants Working Group. 2008. In silico analysis of missense substitutions using sequence-alignment based methods. Hum.Mutat. 29: 1327-1336.

Taylor, NP, Powell, MA, Gibb, RK, Rader, JS, Huettner, PC, Thibodeau, SN, Mutch, DG, Goodfellow, PJ. 2006. MLH3 mutation in endometrial cancer. Cancer Res. 66: 7502-7508.

Thomas, DC, Roberts, JD, Kunkel, TA. 1991. Heteroduplex repair in extracts of human HeLa cells. J.Biol.Chem. 266: 3744-3751.

Thompson, DI & French, RM. 1994. The use of computerized risk adjustment tools in clinical process improvement. Healthc.Inf.Manage. 8: 45-52.

Thoms, KM, Kuschal, C, Emmert, S. 2007. Lessons learned from DNA repair defective syndromes. Exp.Dermatol. 16: 532-544.

Tomer, G, Buermeyer, AB, Nguyen, MM, Liskay, RM. 2002. Contribution of human mlh1 and pms2 ATPase activities to DNA mismatch repair. J.Biol.Chem. 277: 21801-21809.

Tran, PT & Liskay, RM. 2000. Functional studies on the candidate ATPase domains of Saccharomyces cerevisiae MutLalpha. Mol.Cell.Biol. 20: 6390-6398.

81

Trojan, J, Zeuzem, S, Randolph, A, Hemmerle, C, Brieger, A, Raedle, J, Plotz, G, Jiricny, J, Marra, G. 2002. Functional analysis of hMLH1 variants and HNPCC-related mutations using a human expression system. Gastroenterology 122: 211-219.

Truninger, K, Menigatti, M, Luz, J, Russell, A, Haider, R, Gebbers, JO, Bannwart, F, Yurtsever, H, Neuweiler, J, Riehle, HM, Cattaruzza, MS, Heinimann, K, Schar, P, Jiricny, J, Marra, G. 2005. Immunohistochemical analysis reveals high frequency of PMS2 defects in colorectal cancer. Gastroenterology 128: 1160-1171.

Umar, A, Buermeyer, AB, Simon, JA, Thomas, DC, Clark, AB, Liskay, RM, Kunkel, TA. 1996. Requirement for PCNA in DNA mismatch repair at a step preceding DNA resynthesis. Cell 87: 65-73.

Umar, A, Risinger, JI, Glaab, WE, Tindall, KR, Barrett, JC, Kunkel, TA. 1998. Functional overlap in mismatch repair by human MSH3 and MSH6. Genetics 148: 1637-1646.

Umar, A, Boland, CR, Terdiman, JP, Syngal, S, de la Chapelle, A, Ruschoff, J, Fishel, R, Lindor, NM, Burgart, LJ, Hamelin, R, Hamilton, SR, Hiatt, RA, Jass, J, Lindblom, A, Lynch, HT, Peltomäki, P, Ramsey, SD, Rodriguez-Bigas, MA, Vasen, HF, Hawk, ET, Barrett, JC, Freedman, AN, Srivastava, S. 2004. Revised Bethesda Guidelines for hereditary nonpolyposis colorectal cancer (Lynch syndrome) and microsatellite instability. J.Natl.Cancer Inst. 96: 261-268.

Valle, L, Carbonell, P, Fernandez, V, Dotor, AM, Sanz, M, Benitez, J, Urioste, M. 2007. MLH1 germline epimutations in selected patients with early-onset non-polyposis colorectal cancer. Clin.Genet. 71: 232-237.

Vasen, HF, Mecklin, JP, Khan, PM, Lynch, HT. 1991. The International Collaborative Group on Hereditary Non-Polyposis Colorectal Cancer (ICG-HNPCC). Dis.Colon Rectum 34: 424-425.

Vasen, HF, Wijnen, JT, Menko, FH, Kleibeuker, JH, Taal, BG, Griffioen, G, Nagengast, FM, Meijers-Heijboer, EH, Bertario, L, Varesco, L, Bisgaard, ML, Mohr, J, Fodde, R, Khan, PM. 1996. Cancer risk in families with hereditary nonpolyposis colorectal cancer diagnosed by mutation analysis. Gastroenterology 110: 1020-1027.

Vasen, HF, Watson, P, Mecklin, JP, Lynch, HT. 1999. New clinical criteria for hereditary nonpolyposis colorectal cancer (HNPCC, Lynch syndrome) proposed by the International Collaborative group on HNPCC. Gastroenterology 116: 1453-1456.

Vasen, HF. 2007. Review article: The Lynch syndrome (hereditary nonpolyposis colorectal cancer). Aliment.Pharmacol.Ther. 26 Suppl 2: 113-126.

Wang, TF, Kleckner, N, Hunter, N. 1999. Functional specificity of MutL homologs in yeast: evidence for three Mlh1-based heterocomplexes with distinct roles during meiosis in recombination and mismatch correction. Proc.Natl.Acad.Sci.U.S.A. 96: 13914-13919.

Watson, P & Lynch, HT. 1993. Extracolonic cancer in hereditary nonpolyposis colorectal cancer. Cancer 71: 677-685.

Wei, K, Clark, AB, Wong, E, Kane, MF, Mazur, DJ, Parris, T, Kolas, NK, Russell, R, Hou, H,Jr, Kneitz, B, Yang, G, Kunkel, TA, Kolodner, RD, Cohen, PE, Edelmann, W. 2003. Inactivation of Exonuclease 1 in mice results in DNA mismatch repair defects, increased cancer susceptibility, and male and female sterility. Genes Dev. 17: 603-614.

Weitz, J, Koch, M, Debus, J, Hohler, T, Galle, PR, Buchler, MW. 2005. Colorectal cancer. Lancet 365: 153-165.

Welsh, KM, Lu, AL, Clark, S, Modrich, P. 1987. Isolation and characterization of the Escherichia coli mutH gene product. J.Biol.Chem. 262: 15624-15629.

82

Woods, MO, Williams, P, Careen, A, Edwards, L, Bartlett, S, McLaughlin, JR, Younghusband, HB. 2007. A new variant database for mismatch repair genes associated with Lynch syndrome. Hum.Mutat. 28: 669-673.

Wu, X, Platt, JL, Cascalho, M. 2003. Dimerization of MLH1 and PMS2 limits nuclear localization of MutLalpha. Mol.Cell.Biol. 23: 3320-3328.

Wu, Y, Berends, MJ, Sijmons, RH, Mensink, RG, Verlind, E, Kooi, KA, van der Sluis, T, Kempinga, C, van dDer Zee, AG, Hollema, H, Buys, CH, Kleibeuker, JH, Hofstra, RM. 2001. A role for MLH3 in hereditary nonpolyposis colorectal cancer. Nat.Genet. 29: 137-138.

Yao, X, Buermeyer, AB, Narayanan, L, Tran, D, Baker, SM, Prolla, TA, Glazer, PM, Liskay, RM, Arnheim, N. 1999. Different mutator phenotypes in Mlh1- versus Pms2-deficient mice. Proc.Natl.Acad.Sci.U.S.A. 96: 6850-6855.

Zhang, Y, Yuan, F, Presnell, SR, Tian, K, Gao, Y, Tomkinson, AE, Gu, L, Li, GM. 2005. Reconstitution of 5'-directed human mismatch repair in a purified system. Cell 122: 693-705.