Comprehensive Summaries of Uppsala Dissertationsfrom the Faculty of Medicine 990

_____________________________ _____________________________

Genetic Studies of Two Inherited Human Phenotypes: Hearing Loss and Monoamine Oxidase Activity

BY

JORUNE BALCIUNIENE

ACTA UNIVERSITATIS UPSALIENSISUPPSALA 2001

Dissertation for the Degree of Doctor of Philosophy (Faculty of Medicine) in Medical Genetics at

Uppsala University in 2001

ABSTRACT

Balciuniene, J. 2001. Genetic Studies of Two Inherited Human Phenotypes: Hearing Loss and Monoamine Oxidase

Activity. Acta Universitatis Upsaliensis. Comprehensive Summaries of Uppsala Dissertations from the Faculty of

Medicine 990. 62pp. Uppsala. ISBN 91-554-4917-4.

This thesis focuses on the identification of genetic factors underlying two inherited human phenotypes: hearing loss

and monoamine oxidase activity.Non-syndromic hearing loss segregating in a Swedish family was tested for linkage to 13 previously reported

candidate loci for hearing disabilities. Linkage was found to two loci: DFNA12 (11q22-q24) and DFNA2 (1p32). A

detailed analysis of the phenotypes and haplotypes shared by the affected individuals supported the hypothesis ofdigenic inheritance of hearing disability in the Swedish family. Mutation screening of α-tectorin, a gene residing

within the DFNA12 region revealed a mutation of a conserved amino acid (Cys to Ser), that segregated with thedisease. The identification of the mutation added support to the involvement of α-tectorin in hearing disabilities. In

contrast, no mutations were identified in two candidate genes at the DFNA2 locus, that were reported to cause

hearing loss in other families. It is possible that the DFNA2 locus contains a third, not yet identified, hearing lossgene.

Monoamine oxidase A (MAOA) and B (MAOB) catalyze the degradation of certain neurotransmitters in the centralnervous system and are associated with specific behavioral and neuropsychiatric human traits. Activity levels of

both monoamine oxidases (MAO) are highly variable among humans and are determined by unknown geneticfactors. This study investigated the relationship of different MAO alleles with MAO mRNA levels and enzyme

activity in human brain. Several novel DNA polymorphisms were identified in a group of Swedish individuals.

Haplotypes containing several closely located MAOA polymorphisms were assessed in Asian, African, andCaucasian populations. The haplotype distribution and diversity pattern found among the three populations

supported the occurrence of a bottleneck during the dispersion of modern humans from Africa.Allelic association studies conducted on post-mortem human brain samples, revealed the association between a SNP

in the MAOB intron 13, and different levels of both MAO enzyme activities. This suggested that this SNP is inlinkage disequilibrium with at least one novel functional DNA polymorphism that controls MAO enzyme activities

in human brain. The identification of functional polymorphisms regulating the activity of these enzymes will help to

elucidate the involvement of MAO in human behavior and neuropsychiatric conditions.

Key words: linkage analysis, hearing loss, digenic inheritance, allelic association, monoamine oxidase, humangenetic diversity

Jorune Balciuniene, Department of Genetics and Pathology, Section of Medical Genetics, Rudbeck Laboratory, SE-

751 85 Uppsala, Sweden

© Jorune Balciuniene 2001

ISSN 0282-7476

ISBN 91-554-4917-4

Printed in Sweden by Tryck & Medier, Uppsala Universitet, Uppsala 2001

We dance around in a ring and suppose,

but the secret sits in the middle and knows.

Robert Frost

LIST OF PUBLICATIONS

This thesis is based on the following original publications, which will be referred to in the text bytheir Roman numerals:

I. Balciuniene J, Dahl N, Borg E, Samuelsson E, Koisti MJ, Pettersson U, Jazin EE (1998)Evidence for digenic inheritance of nonsyndromic hereditary hearing loss in a Swedishfamily. Am J Hum Genet 63: 786-93.

II. Balciuniene J, Dahl N, Jalonen P, Verhoeven K, Van Camp G, Borg E, Pettersson U,Jazin E (1999) Alpha tectorin involvement in hearing disabilities: One gene - twophenotypes. Human genetics 105: 211-16.

III. Balciuniene J, Syvänen A-C, McLeod HL, Pettersson U, Jazin E (2001) The geographicdistribution of monoamine oxidase haplotypes supports bottleneck during the dispersionof modern humans from Africa. Journal of Molecular Evolution, in press.

IV. Balciuniene J, Emilsson L, Oreland L, Pettersson U, Jazin EE (2001) A SNP in intron 13of Monoamine Oxidase B is associated with altered enzyme activity of both Monoamine

Oxidases in human brain. Manuscript.

Reprints were made with permissions from the publishers.

Table of contents

HUMAN DIVERSITY ........................................................................................................................................................1

GENE DISCOVERY...........................................................................................................................................................3

GENETIC MAPPING.............................................................................................................................................................3

Recombination .............................................................................................................................................................6

DNA markers................................................................................................................................................................6

Linkage analysis...........................................................................................................................................................8Parametric linkage analysis: LOD score method..................................................................................................................... 8

Non-parametric linkage analysis ............................................................................................................................................ 11

Linkage disequilibrium (allelic association) mapping .............................................................................................11Linkage disequilibrium phenomenon ...................................................................................................................................... 11

Allelic association studies: basic principles ........................................................................................................................... 12

POSITIONAL CANDIDATES ...............................................................................................................................................16Physical mapping.......................................................................................................................................................16

Identification of transcripts .......................................................................................................................................16

MUTATION IDENTIFICATION............................................................................................................................................17

FUNCTIONAL STUDIES .....................................................................................................................................................17IMPLICATIONS..................................................................................................................................................................18

HEARING AND HEARING IMPAIRMENT ...............................................................................................................19

HEARING MECHANISM.....................................................................................................................................................19HEARING LOSS.................................................................................................................................................................22

Genes causing hearing impairment...........................................................................................................................23

HUMAN MONOAMINE OXIDASES............................................................................................................................25

PROTEIN STUDIES ............................................................................................................................................................25

GENETIC STUDIES ............................................................................................................................................................26The MAO genes and identified polymorphisms ........................................................................................................26

Allelic association studies of the monoamine oxidase genes ...................................................................................29Association to enzyme activity and/or mRNA levels............................................................................................................... 29

Association to behavioral and neuropsychiatric human phenotypes .................................................................................... 29

AIMS....................................................................................................................................................................................33

PART I. HEARING DISABILITIES.......................................................................................................................................33

PART II. MONOAMINE OXIDASE ACTIVITY .....................................................................................................................33

PRESENT INVESTIGATIONS.......................................................................................................................................34

PART I. HEARING DISABILITIES.......................................................................................................................................34

Family I. Results ........................................................................................................................................................34DFNA12 locus .......................................................................................................................................................................... 35DFNA2 locus ............................................................................................................................................................................ 36

Family II. Results .......................................................................................................................................................38

Discussion: three puzzles...........................................................................................................................................38I. Phenotypic variation............................................................................................................................................................. 38

II. α-Tectorin............................................................................................................................................................................ 39

III. DFNA2................................................................................................................................................................................ 39

Future directions........................................................................................................................................................40

PART II. MONOAMINE OXIDASE ACTIVITY .....................................................................................................................40

Results ........................................................................................................................................................................41

Discussion: MAO genetic variation ..........................................................................................................................42I. Implications for human evolution and linkage disequilibrium studies .............................................................................. 42

II. Implications for functional variation.................................................................................................................................. 44

Future directions........................................................................................................................................................45

ACKNOWLEDGEMENTS..............................................................................................................................................46

REFERENCES...................................................................................................................................................................48

Abbreviations11q22 chromosome 11, long arm, region 2, band 21p32 chromosome 1, short arm, region 3, band 2A/C/G/T adenine/cytosine/guanine/thymineBLAST Basic Local Alignment Search Toolbp base paircDNA complementary DNAcM centiMorganCNS central nervous systemCSF cerebrospinal fluidCys or C cysteinedB decibel

DFNA12 autosomal dominant deafness locus 12DFNA2 autosomal dominant deafness locus 2DNA deoxyribonucleic acidEST expressed sequence tagGJB gap junction protein, betaHUGO Human Genome OrganizationHz (kHz) hertz (kilohertz)IBD identical by descentkb kilobaseKCNQ4 potassium channel, voltage-gated, subfamily q, member 4LOD logarithm of the odds ratio

MAO monoamine oxidaseMAOA monoamine oxidase AMAOB monoamine oxidase BMb megabasemRNA messenger ribonucleic acidPCR polymerase chain reactionRFLP restriction fragment length polymorphismRNA ribonucleic acidSer or S serineSNP (SNPs) single nucleotide polymorphism (more than one SNP)TECTA α-Tectorin

VNTR variable number of tandem repeatsZ LOD scoreΘ theta, recombination fraction

Human diversity

Human genetic material consists of two sets of 3x109 nucleotides. It has been suggested that 1 in

250 to 1 in 1000 nucleotides are different among two DNA sequences taken at

random120,22,112,207. This gives a minimum of 3x106 differences between two randomly taken

human haploid genomes. Assuming that the differences are due to two equally frequent alleles of

the same 3x106 nucleotides in all human beings, it would lead to 23x106 different combinations of

haploid genomes. In comparison, there are about 234 haploid human genomes currently living on

the Earth. Apparently, none of the living or deceased humans who emerged from differentzygotes had the same sequence of nucleotides. Our appearance and other physical or mentalcharacteristics are products of our genes and the environment. Some of these characteristics arepurely genetic, like eye color, some of them are purely environmental, like teeth brushing habit,but most of them are the result of the combined action between genetics and environment, likebody weight or height. Human genetics studies human characteristics that are at least partlycaused by genetic determinants.

The predicted number of genes in the human genome ranges from 35 000 to 120 000 genes43,92.

Careful analysis of the complete human genome after the completion of the HUGO project willprovide the final number of genes. Meanwhile, assuming that the average length of a geneincluding introns is 10 kb, this would suggest that genes occupy 10%- 40% of the humangenome, while coding sequences (assuming an average protein size of 500 amino acids) make uponly 2-6% of the genome. DNA variation in coding sequences as well as in regulatory elementsmay alter the gene expression and, therefore, give rise to variability at the protein level. As weknow, proteins are the main determinants in the development and maintenance of our organism.Consequently, variation in proteins leads to changes in our physical and mental characteristics(phenotype or trait). These phenotypes may be neutral for individual fitness, like size of ears orthickness of hair, or be disadvantageous like various kinds of disorders, e.g. haemophilia ormental retardation. Further classification of human inherited phenotypes is based on the number

of genes or the presence of environmental factors that are required to produce the phenotype.Also important are the genomic locations of the genes: mitochondria, autosomes or sexchromosomes, and the inheritance pattern of the phenotype: dominant, semi-dominant orrecessive (Figure 1).Most of the phenotypes, like body height, rheumatoid arthritis or schizophrenia, are products of acombined action of several genetic loci and environmental factors, i.e. multifactorial or complextraits. Most often such phenotypes display a wide spectrum of overlapping characteristics amonghumans. Moreover, some of them, like body weight or blood pressure, are present in all humansand have a continuous pattern of values (quantitative phenotype).

Figure 1. Classification of inherited human phenotypes.

Genetic variation also provides information about the evolutionary history of our species128.

DNA is constantly being subjected to mutagenic factors, such as DNA replication errors or

mutagenic chemical attacks. The probability of a mutation is around 2.5x10-8 mutations per

nucleotide117. Usually, a specific mutation occurs in a single person. Its chances to survive (andspread) in future generations depends on the probability that it is transmitted to successivegenerations. For neutral mutations (without effect on individual fitness) this chance corresponds

to the probabilities for mendelian segregation. In many cases, the more frequent and broadlydistributed a mutation is, the earlier in history it has occurred (i.e. the older it is). The geneticcomposition of modern humans living today contains numerous DNA differences of varyingfrequencies. Having determined DNA sequences for a number of individuals from differenthuman populations, one can trace back the genealogy of the sequences and find the most

probable ancestral sequence70. The time passed since the appearance of the ancestral sequence(ancestor) will be reflected in the number of differences accumulated, while the geographical

Autosomal X-linked gene is on gene is on autosome X chromosome Mitochondrial

gene is in mitochondrial genome

Y-linkedgene is on

Y chromosome

MonogenicCaused by one gene

Multifactorial(Complex)

Caused by several genes and environmental factors

OligogenicCaused by several genes

Dominantmanifest in

heterozygotes

Recessivemanifest in homozygotes

Semi-Dominantintermediate between

hetero - and homozygotes

Neutral or Pathological

Human InheritedPhenotypes

location will point to the place of the ancestral origin. On the other hand, the number andfrequency of mutations will also depend on selection acting on the mutation or loci linked to it,as well as demographic events like population bottlenecks, expansion or migration that occurred

during human history59. With help of mathematical modeling, one can predict an expected

pattern of human genetic diversity under certain evolutionary and demographic conditions, andcompare it with the observed pattern. The information obtained can help to uncover the modernhuman evolution. We do not yet have the technical possibilities to examine the complete pictureof the DNA variation among humans. Nonetheless, a lot of helpful data can be extracted by

studying variation of different genes scattered throughout the genome128. So far, most of thestudies investigating DNA variation in different regions of the human genome point to theAfrican origin of modern humans followed by a subsequent migration and settlement to the rest

of the world171.All in all, inherited human variation is the object for two disciplines: human molecular evolutionand human molecular genetics. The first discipline uses DNA variation to determine our

evolutionary history, while the second discipline aims to dissect genetic components beyond ourphenotypes that are virtual determinants of our life quality.

Gene discovery

Phenotypic variation is the first aspect that we notice in humans. It troubles us if it imposes ahandicap and awakens our curiosity if it is harmless but extreme. This variation is the ultimaterequirement for genetic studies that attempt to find genes lying behind it. If there are no solidindications about proteins that are involved in the phenotype under study, and if the inspection ofchromosomes in the affected individuals does not result in the detection of any grosschromosomal aberrations, the outline of the gene discovery process in human beings can besummarized as it shown in Figure 2.

Genetic mapping

The aim of genetic mapping is to position unidentified genetic loci relative to known positionswithin the genome. Two main statistical approaches are linkage analysis and linkage

disequilibrium (allelic association) analysis. The first one analyzes recombination events infamilies to determine whether two genetic loci are close to each other, while the second approachattempts to identify the non-random association of specific alleles at different loci in apopulation.

Figure 2. Main steps of a gene discovery process in humans. Continues on the next page

1

2

3

Critical region

Clonecontigs

Transcripts

Constructionof haplotypes

Genotyping

Genetic mapping

Physical mapping

123

Pol

ymor

ph

ic m

arke

rs

or

Explore Human Genome databases:Draft Human Genome Browser http://genome.ucsc.edu/goldenPath/hgTracks.htmlHuman Genome sequencing http://www.ncbi.nlm.nih.gov/genome/seq/etc

Figure 2. Main steps of a gene discovery process in humans.

Mutation identification

Functional studies

Sequencing

Genomic organization and sequence

Monogenic traits were the first to be studied by human geneticist and statistical methodology for

such investigation was proposed as early as in 1922 (for a historical review see ref172). Since theearly seventies, the rapid development in DNA screening techniques accelerated gene discoveryprocess of monogenic traits. In contrast, studies of complex human traits had to wait for more

sophisticated statistical methodologies and DNA techniques, and many discoveries are yet to bemade.

Recombination

As a rule, each of us humans has 46 chromosomes: two homologous sets of 22 autosomal

chromosomes and a pair of sex chromosomes. To each of our children, we transfer a set ofautosomes and a sex chromosome. However, the exact DNA content will differ in eachtransferred set. This is because of two events that occur during gametogenesis: independentassortment of parental chromosomes (when one chromosome of each homologous pair israndomly distributed to the gametes) and recombination (meiotic crossover). During therecombination, two homologous chromosomes, each consisting of two chromatids, exchangeDNA segments in a random manner (Figure 3). As a result, new “mosaic” chromosomes thatcontain DNA segments from both of the parental chromosomes are formed. The recombination ismore likely to occur between loci that are further apart. Therefore, the recombination frequency(Θ) between two loci can be used to estimate the distance between the loci. This distance is

called genetic distance and it is measured in centimorgans (cM). One percent of recombination(Θ=0.01) is defined as 1 cM. Roughly, it is approximated that 1 cM = 1 Mb, but it differs

substantially between different regions in the genome41. Two loci located on differentchromosomes segregate independently (=are not linked) and the recombination frequencybetween them is 50% (Θ=0.5). This is because there is a 50% chance for two independent

chromosomes to end up in the same gamete. For linked loci the recombination frequency will belower than 0.5 (Θ<0.5).

DNA markers

Recombination events (and recombinant individuals) in human pedigrees can be spotted trackingthe segregation of chromosomal fragments through generations. This can be easily achieved bymonitoring familial segregation of a specific genetic locus (DNA marker) that has more than oneallelic variant among humans, i.e. polymorphic marker. One can determine which marker allelesare present in the parents and which alleles have been transmitted to the offspring (Figure 4).This is possible only if the parents are heterozygous for the studied polymorphism. Therefore,

the more alleles a certain polymorphism has, the more information it can provide to scorerecombinants. This marker characteristic is called informativeness, and it is usually reflected bytwo measures called marker heterozygosity (H) and polymorphism information content (PIC) thatcan be estimated using the following formulas:

pi and pj denote allele frequencies of i and j alleles, respectively.The human genome contains a variety of polymorphic markers: microsatellites (di-, tri-, ortetranucleotide repeats), variable number of tandem repeats (VNTR or minisatellites) and singlenucleotide polymorphisms (SNPs).

Figure 3. Meiotic crossover. Each homologous chromosome consists of two sister chromatids. Three different

genetic loci are indicated by different letters. Every locus consists of two alleles that are denoted by capital andsmall case letters. A recombination occurred between loci B and C and generated two recombinant sister

chromatids.

A

B

C

a

b

c

a

b

c

A

B

C

A

B

C

a

b

c

a

b

c

A

B

C

Two homologues chromosomes

Crossing-over Two recombinant chromatids

A

B

c

a

B

c

a

C

b

A

B

C

a

b

c

H p PIC p p pii

n

ii

n

ij i

n

i

n

j= − = − −= = = +=∑ ∑ ∑∑1 1 22

1

2

1

2

11

2 and ,

For genetic mapping purposes, the genomic position of a DNA marker should be determined. Aset of ordered DNA markers scattered throughout the genome constitutes a marker map. Markermaps may differ in the method used to order and estimate the distance between markers. Ingenetic maps, distance and order of DNA markers are based on recombination analysis. DNAmarker maps are used to locate a gene causing a certain phenotype in the human genome (this is

called genome scan or genome-wide mapping). Usually, for initial mapping purposes, a set ofmarkers spaced at 10 cM distance is sufficient. The HUGO project has generated up to 10 000

highly polymorphic genetic markers and placed them on a map25,26. In parallel, The SNPconsortiuma has released a map that contains around 300 000 SNPs. Microsatellite markers are agood choice for genome-wide mapping as they are highly polymorphic and are easy to genotypeby fluorescent PCR-based methods. On the other hand, SNPs, though being bi-alellic, are veryabundant in the human genome and can be genotyped by high-throughput DNA chip technology.Moreover, because of their abundance and low mutability, SNPs can be the best choice for finemapping and/or linkage disequilibrium studies.

Linkage analysis

Parametric linkage analysis: LOD score method

The inheritance pattern of monogenic phenotypes can provide information on certaincharacteristics of a gene that causes this phenotype. For instance, it can give clues on whether thegene is autosomal or sex-linked and whether one phenotype-causing allele (if dominantphenotype) or two phenotype-causing alleles (if recessive phenotype) are present in theindividuals manifesting the phenotype. In other words, we can assume a certain genetic modelfor the studied trait. Further analysis will depend on the assumed model and it is calledparametric linkage analysis.Let’s assume that we have a family with a dominant condition (Figure 4). In this case, weassume that affected individuals harbor one trait-causing allele (for recessive trait both allelesshould be trait-causing) in an unknown gene. Members of the family are genotyped for a set ofpolymorphic markers spread throughout the genome, and the proportion of recombinant

individuals between marker alleles and the envisioned trait allele is estimated. As we see, person4 has inherited a disease from his mother (person 1) together with alleles “3” at the polymorphicloci A, B and C. Then, affected person 6 is non-recombinant as it has inherited the “3” alleles atall the three polymorphisms. Similarly, individuals 5, 7 and 9 are also non-recombinant as theyhave inherited alleles “4” for A, B and C loci, that come from the healthy grandfather. In

a http://snp.cshl.org/

contrast, affected persons 8 and 10 are recombinants, as they have received alleles “4” at the lociC and A, respectively, that are coming from the healthy grandfather. In general, a DNA marker islikely to be linked to a disease gene if the recombination fraction between them is less than 50%(Θ< 0.5). In the drawn case, the disease gene is in complete linkage with marker B, as they have

not recombined. Moreover, one can determine boundaries of a chromosomal fragment that islikely to contain a trait gene (gray bars), i.e. the fragment that consists of markers with Θ=0 and

is bordered in each side by DNA markers that have recombined with the trait locus. We see thatin this pedigree, the disease allele has not recombined with alleles at marker B, but recombinedwith alleles at loci A and C. Therefore, we conclude that the disease gene lies between the DNAmarkers A and C.

Figure 4. Recombination analysis in a hypothetical family. Letters “A”, “B”, and “C” denote three polymorphic

DNA markers, with several alleles indicated in different numbers. Individuals manifesting a certain phenotype (inthis case, a disease) are shown in black symbols. The disease is inherited in an autosomal dominant manner. “D”

denotes an allele that causes the disease, while “+” is a normal allele. Alleles on the same chromosome constitute ahaplotype that is represented as a vertical bar. Gray-filled bars represent the disease-associated haplotype.

Identification number for each individual is given below a symbol.

1 4

+ + D +

+ + + D + + + D + + + D

1 2 3 4

1 4 2 3 1 4 2 4 2 4 1 3

1 4 2 3 1 4 2 3 2 4 1 41 4 2 3 1 4 2 3 2 4 1 3

1 2 3 41 2 3 4

AB

CGene

AB

CGene

3 4

5 6 7 8 9 10

3 2 1 4

3 2 1 43 2

AB

CGene

1 2

+ +D +

However, a recombination rate of Θ< 0.5 can also be observed for unlinked loci just by chance.

Therefore, it is important to perform statistical calculations to evaluate the significance of theobserved results. Usually, the statistics LOD score (Logarithm of the ODds) is calculated forthese purposes. LOD score (Z) is a logarithm of likelihood ratio that compares the probability ofobserved recombinants and non-recombinants in a family given that a marker and a trait loci arelinked (at certain recombination fraction Θ) versus the probability that the marker and the trait

gene are not linked. In an ideal situation, when recombinants (R) and non-recombinants (N) inthe family can be determined unambiguously, LOD score can be estimated by the followingformula:

In the example above, LOD score for marker B at Θ=0 will be:

Z≥3 is considered a significant evidence for linkage between a tested marker and a trait locus.Z≥3 means that the probability of the observed family genotypes assuming linkage is 1000 morelikely compared to the probability under no linkage and this Z value corresponds to asignificance level of p=0.05. LOD scores are being computed for a range of Θ and the maximum

Z is selected. Z<-2 is considered to be a significant evidence for no linkage (exclusion criteria).For genome-wide mapping, Z≥3.3 is required to achieve a corresponding significance level of

p=0.0584.In our hypothetical case, marker B gave Z=1.8 and this is less than the accepted significance

criteria (Z≥3), though it is the maximum that can be achieved with this family. This means thatthe sample material should be increased: either more members belonging to the same familyshould be identified, or additional families with the same phenotype should be ascertained toobtain statistically significant results.Scoring of the recombinant individuals in human pedigrees is not a trivial task. The DNAmarkers used are not always informative enough to unambiguously determine the recombinants.Also, very often some family members are missing and their genotypes have to be inferred.Nowadays, the scoring of recombinants is being done by a variety of sophisticated computerprograms that can be chosen depending on the collected family material and the phenotype

studied105,126,172. The LINKAGE package of programs86 is the most often used especially forsimpler, for instance monogenic, traits. There are also other programs available like Genetic

ZR N

R N=−

≤ <+lg

( )( . )

, .Θ Θ

Θ1

0 50 0 5 where

ZΘ= +=−

0

0 6

0 6

0 1 00 5

lg( )

( . )= lg64 = 1.8

Analysis Software (GASa), MENDEL85, and VITESSE124 that can perform similar LOD scorecalculations.

Non-parametric linkage analysis

For the above described LOD score method, one should a priori assume a model for a trait-causing gene. However, this is not trivial for complex phenotypes, as they are products of manygenes and environmental factors. Therefore, segregation analysis of a complex phenotype in afamily may not yield information on the underlying genotype. Also, it may be difficult to judgewhich individuals are clearly unaffected, since some individuals can carry the same traitgenotype as affected but do not manifest trait characteristics (non-penetrant). In this case, onecan use model-free or non-parametric linkage analysis methods that do not require an explicitlyspecified genetic model for a trait. These analyses use only affected relative pairs (usually sib-

pairs) and investigate what is the frequency of marker alleles, that are coming from the same

ancestor (identical by descent (IBD)), among the affected pairs46. If affected individuals have amarker for which certain IBD alleles are shared more often among affected individuals that itwould be expected by chance, the marker may be linked to the trait gene(s). If a quantitativephenotype is studied, IBD allele sharing can be investigated in individuals with extreme

phenotypic values144. Again, several computer programs can be implemented for these purposes:

ASPEX154, MAPMAKER/SIBS80, and GENEHUNTER79.

Linkage disequilibrium (allelic association) mapping

Linkage disequilibrium phenomenon

Linkage disequilibrium is defined as a non-random association between specific alleles at two or

more loci. In other words, specific alleles at two or more loci are in linkage disequilibrium ifthey appear together more frequently than what is predicted from individual frequencies of thealleles by Hardy-Weinberg distribution. Most often, linkage disequilibrium is generated by amutation that, due to genetic drift or natural selection, expands in a population. Alleles of otherloci, that are on the same chromosome as the mutation, will be transmitted together as ahaplotype, and thus will be in linkage disequilibrium. Alleles that are farther apart from themutation will be quickly separated by meiotic recombination as they pass through generations,while closest loci will stay in linkage disequilibrium with the mutation longer. When the time

aAvailable at ftp://ftp.ebi.ac.uk/pub/software/linkage_and_mapping

goes on, there will be enough recombination events to shear all the loci initially linked to themutation, and eventually linkage equilibrium will be reached. Therefore, linkage disequilibriumis strongest for the alleles that have occurred recently. In a population, linkage disequilibrium isa dynamic process, as new mutations are being generated constantly followed by their expansion,and eventually it decays due to recombination. Another cause for linkage disequilibrium in a

population can be a sizable migration from the outside, that can bring in new haplotypes usuallywith new alleles or alleles that are different in frequency compared to the existing alleles. Also,some other demographic events, such as population stratification, may mimic the presence oflinkage disequilibrium in a population.

Allelic association studies: basic principles

Linkage disequilibrium analysis (or allelic association studies) can be performed on a genomewide scale to map an unknown phenotype-causing gene, or can be used to test the relationship

between alleles at a specific candidate gene and a certain phenotype. In contrast to linkageanalysis, allelic association mapping can be performed on unrelated affected individuals comingfrom the same population. The basic assumption is that a certain phenotype in tested individuals

is due to the same mutation, i.e. IBD allele of a phenotype-causing gene174. The simplest form ofallelic association analysis is case-control studies, where the distribution of alleles at a variableDNA locus is compared between a sample of unrelated cases (individuals with a certainphenotype) and a sample of well-matched controls.Let’s consider a study that investigates the relationship between a certain phenotype and acandidate gene. To achieve this, polymorphic loci in the gene are identified and they are testedfor association with the phenotype. As a rule, association between a tested polymorphism and a

certain phenotype can be detected, if the tested allele has a direct functional effect on a testedphenotype (functional variant), or if it lies in a close vicinity to an unknown functional mutation,i.e. is in linkage disequilibrium (Figure 5). If the tested DNA variant causes the phenotype, it willbe present in a group of cases, but it will not be present among the control individuals, and theinterpretation of the findings will be straightforward. However, in the case that the tested DNAvariant does not have an effect on the phenotype but rather serves as a marker for an unknownfunctional variant, i.e. is located in the vicinity, the results of the association studies may not beas clear-cut. Figure 5 shows a DNA marker with two alleles indicated as solid and open circles.The functional DNA variant (solid triangle) occurred on one of the chromosomes with the“solid” marker allele. Consequently, three types of non-recombinant chromosomes (haplotypes)will be present in the population: the functional variant together with the “solid” allele, and two

types of chromosomes that differ in the alleles (solid and open circles) at the marker locus but donot contain the functional variant. In a case-control study, the cases will have under-representation of the “open” allele (if any), while controls will contain both marker alleles.

Figure 5. Allelic association. Each bar represents a single chromosome. The solid triangle indicates a functional

variant that causes a certain phenotype. The circles represent a neutral SNP: the solid circle indicates an allele that is

on the same chromosome as the functional variant, while the open circle denotes alternative alleles at the SNP locus.

In general, the success of such allelic association studies will rely on the degree of linkagedisequilibrium between a functional variant and the linked neutral marker. In turn, the degree oflinkage disequilbrium among the two variants will depend on the mutation rate of the variants,the rate of recombination between them, their frequency and their age.Most of the association studies use microsatellites and/or minisatellites because they are the mostinformative DNA polymorphisms. In some cases, these markers themselves may have a direct

effect on the expression of a gene product28. Expanded tri-nucleotide repeats are excellent

examples of polymorphisms shown to cause a variety of neurological disorders such as

Huntington’s disease, and fragile-X syndrome139. However, if repeat polymorphisms do notrepresent functionally significant DNA variants, but serve as markers for hidden causativepolymorphisms, their use in allelic association studies of complex disorders may be quite limited(Figure 6A). The mutation rate of mini/microsatellites on average is 7x10-3 per gamete/per

generation68,12. This is a high mutation rate compared to 2.5x10-8 per nucleotide for non-

repeated sequences117. Over generations, the higher mutation rate for mini/microsatellites wouldintroduce new alleles linked to the same functional variant obscuring association studies (Figure6A). Usage of other types of DNA polymorphisms, such as small deletions or SNP, may be

better. SNP have also the advantage of frequent occurrence throughout the genome, and they are

amenable to automation83,193. Similar to mini/microsatellites, they can represent functionalvariants as such, or can serve as markers for a presence of unknown phenotype-related alleles.Another limitation to association studies may emerge if recombination has occurred between thefunctional variant and an initially linked marker allele (Figure 6B). In this case, the initiallinkage disequilibrium between the marker and the functional variant may be destroyed. Thus,the distribution of the marker alleles may even out between cases and controls, and noassociation would be detected.

Cases

Controls

Figure 6. Two genetic events that disturb the allelic association between two linked variable loci. A) The

tandem open triangles represent repeats in different alleles of a microsatelite marker. The functional variant (solidtriangle) is linked to the microsatelite alleles. During evolution, an allele with three repeats may mutate (to two or

four repeats). Consequently, all three marker alleles will segregate on the same chromosome as the functional

variant, and therefore, all three will be present among cases. B) The circles indicate alleles of a SNP. Recombinationmay occur between the functional variant and an initially linked SNP allele. Consequently, both SNP alleles will

segregate with the functional variant.

The frequency of the analyzed DNA marker, in many cases proportional to its evolutionary age,will play a significant role in the detection of association. For example, if a DNA polymorphismhas occurred on the same chromosome as the functional variant but much later during evolution,it will be very rare among cases. Consequently, a much larger sample of cases would be requiredto reach statistically significant association.Simultaneous analysis of several linked DNA markers (e.g. SNPs) covering the candidate region

should be used to decrease the problems of association studies mentioned above (Figure 7). As itis seen in the figure, the independent analysis of two DNA markers, symbolized as squares andcircles, would not have revealed the existing association with a functional variant (symbolized assolid triangles). Analysis of a third marker, illustrated with diamonds, would expose theassociation only if a very large sample of cases was analyzed. Detection of the functionalpolymorphism can be more efficient if, instead of testing individual DNA markers, combinedhaplotypes generated with these three markers are analyzed simultaneously for association with a

A) DNA marker mutation B) Recombination

phenotype. In Figure 7, two specific haplotypes would be detected in a group of cases but not incontrols, indicating the presence of a functional polymorphism that is in linkage disequilibriumwith these two particular haplotypes. The problem may be the correct determination of thehaplotypes, especially when the genotypes from the parents are missing. A collection of

computer programs, such as GENEHUNTER79, SIMCROSS, SIMWALK195, and

TRANSMIT24, may be used to estimate haplotypes given genotype information.Another problem of such studies is to obtain well-matched controls. The solution can beprovided by family-based linkage disequilibrium approaches, such as transmission

disequilibrium (TDT)162 or haplotype relative risk (HRR)173 tests. Both approaches, in a slightlydifferent manner, use parental alleles that were not transmitted to the affected offspring as acontrol population.In general, an allelic association approach is more widely used to dissect the genetic basis ofcomplex phenotypes. Many different genetic loci together with diverse environmental factorsmay interplay to produce a given complex phenotype. This makes the allelic association mapping

even more complicated. Many additional parameters will influence the power of such allelicassociation studies. For example, the density of tested SNPs at a given candidate locus, theproportion of variance of a given phenotype attributable to a causative polymorphism within thetested locus, the size of a tested sample, the sampling scheme, and the selection of a statisticaldesign are some of the factors that determine the success of an association

mapping174,78,101,153,172,197.

Figure 7. Use of haplotypes for association studies. Each bar symbolizes a single haplotype. The circles, the

squares, and the diamonds, represent three neutral polymorphic loci that are being tested for association with a

certain phenotype caused by an unknown functional variant indicated by the solid triangle. Cases would contain onlyhaplotypes 1 and 2, that carry the functional variant.

Cases

Controls

1

2

3

4

Positional candidates

Once the genetic location of a trait-causing gene has been mapped, one should identify all thegenes residing in this region and screen them for mutations in affected individuals. Upon theaccomplishment of the Human Genome Project, identification of genes located in the candidateregions will be a simple task, as all human DNA sequences will be available in genomedatabases. However, just a year ago (1999), the situation was not so smooth and one had to spenda lot of efforts to identify all genes in the candidate region.

Physical mapping

After having mapped a trait gene to a chromosomal region as narrow as possible (optimally 1 cMdistance), one should locate the gene physically in the genome, i.e. to find a DNA clone(s) thatspans over the mapped region. A variety of ordered human DNA clones (referred as physicalmap) corresponding to a specific chromosomal location are available. The most handy ones areclones in yeast artificial chromosomes (YAC) that contain up to 2 Mb DNA insert, while clones

in bacterial artificial chromosomes (BAC) hold up to 300 kb DNA inserts. After a relevant YACor BAC clones have been obtained, the next step is to identify all genes that are present in theseDNA fragments.

Identification of transcripts

Methods used to identify genes in a given DNA sequence are based on generalized features thatdistinguish coding elements from non-coding. These features are the ability to produce EST(expressed sequence tag), a high degree of conservation among species, and the presence of CpGislands.Potential genes can be identified by bioinformatic approaches. Special computer programs such

as GRAIL2180 and GENSCAN16 can predict genes and CpG islands in a DNA sequence. A lotof information can be obtained by searching for homology to expressed sequences, such as EST,full-length mRNA and/or proteins, deposited to different databases (BLAST searcha). Also,

comparative sequence analysis by using programs such as ALFRESCO67 or PIPMAKER155 can

help to discover genes known in other species, or to identify unknown evolutionary conservedsequences that are likely to contain functionally important DNA elements.

a http://www.ncbi.nlm.nih.gov:80/blast/

Having identified all the possible transcripts in the chromosomal region of interest, the next stepis to select candidate gene(s) relevant for the studied trait. Common sense and knowledge ofother biological aspects of the studied trait or similar traits can give some clues about function orexpression pattern of the “hunted” gene or suggest potential molecular pathways in which thecandidate gene may be involved. Also, knowledge on genes involved in similar traits in model

organisms can be extremely helpful in pointing to a candidate human gene.

Mutation identification

The selected candidate gene(s) can be screened for DNA variants (mutations) that may cause the

phenotype. One can sequence regions that are important for the gene function (such as promoter,coding regions, splice sites). Any type of mutations, e.g. deletions, insertions, missensemutations, nonsensse mutations, splice site mutations, frameshift mutations, and dynamicmutations, can produce the phenotype. For monogenic pathological conditions one would expectthe identified mutation to be present in affected individuals but not in controls. However, forcomplex traits, the same DNA variant can be frequent in the control group. Therefore, theidentification of a trait-associated mutation makes the selected gene a stronger candidate, butadditional proof is required. For example, a strong proof can be provided if human mutationsintroduced by genetic manipulations into a model animal would result in a phenotype similar tothe human trait under investigation.

Functional studies

Validation of a candidate gene leads to the next stage of research that attempts to unravel thefunction of the gene. Useful hints on gene function can be gained by searching in electronicdatabases, for example by BLASTa, for sequences with known function that are homologous tothe DNA sequence or protein sequence of the candidate gene. Also, some computer programs

(some are available at Pedro's BioMolecular Research Toolsb) can scan the gene for the presenceof DNA or protein motifs that are good predictors for a specific function. Studies of mRNAexpression patterns in human derived material would also provide information on the function.This information can be related to what tissues or cells express the gene, in which developmentalor differentiation stage the gene is expressed, what external stimuli does the gene respond to andso on. Methods that can be used for transcript expression studies are Northern blot analysis, in

situ hybridization203, DNA microarray hybridization151, mRNA differential display93, and 5’-

a http://www.ncbi.nlm.nih.gov:80/blast/bhttp://www.public.iastate.edu/~pedro/research_tools.html

nuclease assay19. In addition, human derived samples can be used to investigate expression ofthe encoded protein that can be visualized using specific antibodies in immunocytochemical,immunofluorescent or Western blot methods. Moreover, various methods, such as DNase I

footprinting50 and gel retardation assay156 can identify DNA sequences that bind proteins. Yeast

two-hybrid system44 is a very powerful approach to find other proteins that physically interactwith the protein encoded by the gene. The mentioned approaches can discover gene function at acellular level, though would not tell anything about the function at the whole organism level.The function of a gene at the organism level can be studied in model organisms, such asnematode, fruitfly, zebrafish, frog, or mouse. An animal with altered function of a studied genecan be constructed. For example, it is possible to shut off the gene completely (knockout) or to

shut it off only in specific tissues and/or specific time points (conditional gene knockout100).Also, it is possible to over-express the gene (to insert extra copies of the gene), to express itectopically, or to insert an altered form of a gene to a model animal. Consequent phenotypes of

such gene alterations can provide additional information on the function of a gene. The drawbackof such studies is that some genes may not be present in animals. If present, they may befunctionally unimportant or may have a divergent role. Also, the lack of some genes may notproduce a phenotype at all or may result in embryonic lethality. Moreover, phenotypes such asbehavior may be very difficult to interpret, and the information obtained may not always beapplicable to humans.

Implications

The identification of a gene involved in a certain phenotype and the discovery of its functionmay contribute substantially to the progress in medicine. New molecular procedures that wouldimprove clinical and prenatal diagnosis as well as genetic counseling can be invented. Moreover,a disease-causing gene is a direct target for new therapeutic agents or gene therapies. This is theultimate goal when the genetic cause of a disease is identified.Lastly, knowledge on genes involved in any phenotype broadens our horizons on basic molecularmechanisms that rule our bodies, minds and consequently destinies.

Hearing and hearing impairment

Hearing mechanism

Ear structures first appeared in fish over 400 million years ago131. It is speculated that hearinghas evolved for a better survival in an environment where one is constantly threatened of beingeaten: hearing enables to “detect and localize” moving objects without seeing or physicallycontacting them. For us humans, hearing was also the prerequisite in developing and maintainingour speech.

Figure 8. Sound and its physical characteristics.

Sound (Figure 8) is propagated as a wave in a physical medium (gas, liquid or solid) and is theresult of a variation in atmospheric pressure. Normally, humans can hear sounds ranging from 20to 20 000 Hz, with the highest sensitivity in the range between 1000-4000 Hz.

Ear is an organ that has evolved to catch these mechanical waves and to transform them into asignal (electrical signal) that can be understood by the brain. The ear consists of three parts:outer, middle and inner ear (Figure 9).Sound waves are captured by the outer ear (auricle) and are passed through the external acousticduct to the tympanic membrane. Vibration of the tympanic membrane induced by a sound istransmitted through the three ossicles: malleus, incus, and stapes, to the oval window of the innerear. The inner ear consists of the bony and the membraneous labyrinths. The bony labyrinth isfilled with fluid called perilymph and it is comprised of three major parts: cochlear, vestibule andsemicircular canals.

Loudness=Amplitude=

Sound pressure level (dB)= 20lg(P/20µPa)

Pitch =Frequency (Hz)

= 1/T

Sound

Period ofcycle (T)

Time

Pre

ssu

re (

P)

Figure 9. Structure of the human ear. Adapted from Petit (1996)133.

The auditory machinery is located in the cochlear, while the vestibular system resides insemicircular canals, utricle and saccule. The membranous labyrinth resides inside the bonylabyrinth and is filled with endolymph.The cochlear duct, the membranous part of the cochlear, contains the organ of Corti (Figure 10)

that harbors the main sensory components of the auditory system. The mechanical energygenerated by a sound wave is converted into electrical impulse by hair cells that are located onthe basilar membrane of the organ of Corti. The hair cells are of two types: inner and outer. Eachhair cell contains a bundle of actin filled stiff microvilli, called stereocilia, that are connectedwith each other by actin links (Figure 11). Tips of the stereocilia are covered by acellulargelatinous membrane referred to as tectorial membrane. Vibration of the basilar membranecaused by an oscillation of the perilymph induces sheering of the tectorial membrane (Figure 11,depolarization). This causes a deflection of the stereocilia that results in mechanical opening ofunidentified potassium channels (Figure 11). Potassium ions flux into the hair cell through theopened channels and switch the influx of calcium ions. This triggers the release of

Vestibularapparatus

Tympanicmembrane

Cochlearduct

External ear Middle ear

Inner ear

Ovalwindow

Cochlear

Perilymph

Endolymph

Utricle

Saccule

Semicircularcanals

External acoustic duct

Osicles

neurotransmitters from the hair cell that activate the acoustic nerve conducting the actionpotential to the auditory cortex.

Figure 10. Crossection of the cochlea duct. Adapted from Willems, PJ (2000)201.

Figure 11. Transduction process of a in a hair cell. Three stages of a hair cell activity are represented: resting,

depolarized and repolarized phase. See the text for more detail explanation.

Scala vestibuli

Scala tympani

Scala media

Basilarmembrane

Inner haircell

Outer haircells

Cochlearnerve

Tectorialmembrane

Supportingcells

Striavascularis

Organ ofcorti

K+

K+

K+

K+

K+

K+

K+

K+

K+

K+

K+

Ca+ K+

K+

K+

K+

K+

K+

K+

K+

K+

K+

K+

Endolymph

Stereocilia

Nerve

Synapticvesicles

Connexinchannels

Basilarmembrane

Actinlinks

K+

K+

Perilymph

Potassiumchannel

Supportingcells

Tectorialmembrane

Depolarization RepolarizationResting

Ca+

Ca+Ca+

Ca+

A pattern of action potential encodes certain characteristics of the sound, such as intensity,frequency, and time course. The cells repolarize when potassium ions exit the cells through gapjunctions to the supporting cells and subsequently to the endolymph (Figure 11, repolarization).The inner hair cells act only as receptor cells that transmit electrical signals to the cochlear nerveleading to the auditory cortex. The outer hair cells, in addition to receptor function, amplify the

sound induced motion of the basilar membrane, therefore increasing hearing sensitivity. Eachhair cell is specific to certain sound frequencies. Therefore, damage of different hair cells resultsin inability to detect different sounds. For a detailed description of the hearing mechanism,

see65,122.

Hearing loss

Hearing loss is one of the most frequent sensory disorders in humans133. Its negative influenceon individual’s life quality highly depends on the severity of the illness and the age of onset. Ifthe disease develops early in childhood, it will prevent acquisition of speech and this, in turn,will effect cognition and psychosocial development.Classification of hearing loss is based on several different criteria (Table 1). The incidence of

pre-lingual deafness is 1 in 1000 newborn. Of them, 60% have a genetic cause104. Of the genetic

cases, 30% constitute syndromic hearing loss111. The prevalence of post-lingual hearing lossincreases with age: it affects 1% of young adults, 10% of persons at the age of 60, and 50% of

persons at the age of 75 and older160. It has been estimated that 85% of pre-lingual hearing loss

is due to recessive mutations133. In general, autosomal dominant forms of hearing loss showdifferent disease course compared to recessive forms. Recessive forms tend to manifest as

congenital (pre-lingual) non-progressive severe to profound or profound hearing loss168, whilepost-lingual progressive mild to severe hearing impairment prevails among dominant

cases182,185.

Several tests are used to assess the nature and severity of hearing impairment. Audiometry is abasic test that evaluates hearing loss severity, frequencies affected, and which ear structures areinvolved. During this test, a person is presented with certain pure tones of several differentfrequencies. The softest level of sound that a person can hear (hearing threshold in decibels (dB))is recorded, and the data is plotted in a graph called audiogram. Other tests include impendanceaudiometry that evaluates the integrity and function of the ear structures, electrocochleographyand evoked-response audiometry that directly measures electrical response of the ear or brain to

a sound stimulus148.

Table 1. Classification of hearing impairment. As presented in Willems, PJ (2000)201.

Criteria Classification Description

Nongenetic Environmental factors, e.g. infectiousdiseases, acoustic trauma, ototoxic drugs

Cause

Genetic Monogenic (autosomal:dominant andrecessive; X-linked, mitochondrial),oligogenic, multifactorial

Syndromic Present together with other phenotypes, e.g.

blindness, pigmentary defects, goiterManifestation

Non-syndromic Solely hearing loss

Prelingual Develops before speech acquisitionOnset

Postlingual Develops after speech acquisition

Conductive Defect in outer or middle ear

Sensorineural Defect in inner ear or neurotransduction

Type of ear defect

Mixed Conductive and sensorineural

Mild Loss of 21-40 dB

Moderate Loss of 41-60 dB

Severe moderate Loss of 61-80 dB

Severe Loss of 81-100 dB

Severity

Profound Loss of >100 dB

Low <500 Hz

Middle 501-2000 Hz

Frequencies affected

High >2000 Hz

Genes causing hearing impairment

Genetic hearing impairment is highly diverse among humans. Various types of mendeliangenetic transmission have been observed in families with non-syndromic hearing loss. It hasbeen postulated that several dozens of genes may be responsible for hearing impairment in

different families111. A comprehensive and constantly updated summary of achievements inhearing loss mapping and gene identification can be viewed at the Hereditary Hearing

Homepagea181. Several reviews have also been published72,201. Briefly, up to 30 different lociresponsible for autosomal dominant hearing loss, 28 genes responsible for autosomal recessive

a http://dnalab-www.uia.ac.be/dnalab/hhh/

forms of the loss, and 5 genes linked to X-chromosome have been mapped. Among them, 15genes have been identified (Table 2). For many of these genes, different mutations have beenfound in a number of different families. Mutations in Connexin 26 (GJB2) are one of the most

common causes for congenital deafness explaining almost 50% of the cases137. In addition,

GJB2 is mutated in 10% of the sporadic congenital cases89.

Table 2. Genes responsible for non-syndromic hearing impairment.

Gene Function

Connexin 26 (GJB2)Connexin 31 (GJB3)Connexin 30 (GJB6)

Gap junction proteins that enablesexchange of small molecules between cells

Myosin 7AMyosin 15

Involved in intracellular transport, cellularmovement, muscular contraction and etc

KCNQ4 Potassium channel

α-tectorin; COCH gene Extracellular matrix proteins

POU3F4; POU4F3 Transcription factors

Pendrin Iodide and chloride ions transport

Diaphanous Involved in maintenance of cytoskeleton

Otoferlin Probably transport of membrane vesicles

Collagen COL11A2 Probably one of the components of thetectorial membrane

ICERE-1 Unknown

Another interesting observation is that different mutations in the same gene can result in differentforms of hearing loss. For example, different allelic variants of the Connexin 26 can cause either

recessive74 or dominant hearing loss38. Similarly, different mutation in Myosin7A can result in

recessive or dominant form of hearing impairment96,97 or can cause Usher syndrome that is

characterized by hearing impairment and blindness196. More strinkingly, different alleles of

GJB3 can underlie recessive/dominant non-syndromic hearing loss205,98 or can cause a skin

disease erythrokeratodermia variabilis142.

Even more genes have been identified for syndromic forms of hearing impairment. Currently, upto 30 such genes are known, and more information can be obtained in the Hereditary Hearing

Homepagea or in recent reviews45,53,178.Identification of genes causing hearing impairment contributes to our knowledge on themolecular mechanism of the hearing machinery, how it develops, functions and dysfunctions.

a http://dnalab-www.uia.ac.be/dnalab/hhh/

Also, it provides better molecular tools in clinical and prenatal diagnostics as well as geneticcounseling. To date, only few treatments are available: hearing aids that amplify sound, andcochlear or auditory brainstem implants that stimulate cochlear nerve or auditory nuclei.Hopefully, in a near future, new approaches for treatment will emerge from efforts to replacedefective cochlear genes using virus-mediated gene transfer as well as to induce regeneration of

inner ear sensory hair cells102.

Human monoamine oxidases

Protein studies

Monoamine oxidases (MAO) are enzymes localized in the outer mitochondrial membrane thatcatalyze the oxidative deamination of biogenic and xenobiotic amines, including severalneurotransmitters. There are two isoenzymes, MAOA and MAOB, that differ in their substratespecificity, sensitivity to inhibitors, cellular localization and are encoded by two separate genes.MAOA preferentially degrades serotonine, norepinephrine and is selectively inhibited byclorgyline, while MAOB is more efficient to metabolize phenylethylamine and benzylamine and

is selectively inhibited by deprenyl (for reviews see ref192,199,157). Both enzymes are present invarious tissues throughout the body with the highest levels in liver and brain. However, sometissues predominantly express one type of MAO: placenta expresses mainly MAOA, while

platelets and lymphocytes posses only MAOB activity176,56,158. In human brain, MAOA

predominates in catecholaminergic neurons, while MAOB prevails in serotonergic neurons,

histaminergic neurons, and astrocytes198,150,149.The ability of monoamine oxidases to metabolize neurotransmitters suggests their role in theetiology of human behavioral traits and neuropsychiatric disorders. It was found that MAOA

inhibitors are effective to treat mental depression99 and anxiety disorders35, while MAOB

inhibitors delay progression of Alzheimer’s175 and Parkinson’s diseases164. Furthermore, it hasbeen reported that deficiency of MAOA in humans leads to borderline mental retardation and

increased impulsive behavior in parallel to abnormal metabolism of neurotransmitters14,15,13.

Also, transgenic mice lacking the MAO genes manifest similar behavioral and neurochemical

disturbances as described for humans18,57.Most of the studies investigated the relationship between platelet MAOB enzyme activity levelsand certain phenotypes. Platelets were used for these analyses because they are easier to obtainthan other human tissues. Low levels of platelet MAOB enzyme activity have been associatedwith behavioral traits such as sensation seeking, monotony avoidance,

aggressiveness152,47,1,194,163, suicide109,189,190, alcoholism169,6,167,40,39, as well as psychiatric

conditions like schizophrenia and affective disorder114,204,147,200. On the other hand, increasedlevels of platelet MAOB enzyme activity were found in patients with Parkinson’s and

Alzheimer’s disease2,34,165,8. However, no causal relationships for the associations have been

established so far, and not all the subsequent studies yielded congruent results. Moreover, nocorrelation was found between MAOB enzyme activity in platelets and brain of the same

individuals202,206, indicating a distinct MAOB control in brain and in blood. Later, it has beenhypothesized that platelet MAOB enzyme activity serves as a genetic marker for the functionalcapacity of some central neurotransmitter system, most likely serotonergic, that is actually

underlying differences in certain human behaviors or mental states125.The monoamine oxidase enzyme activity levels in peripheral tissues have been extensivelystudied. The stable activity levels of the isoenzymes among control individuals vary substantially

reaching over 30-fold differences for the platelet MAOB enzyme activity115 and up to 100-fold

differences for the MAOA enzyme activity measured in skin fibroblast20. Several studies haveindicated that the variation of the enzyme activities is substantially determined by genetic

components121,11,141,132.

The identification of the MAOA and MAOB genes23,159 opened the possibility to explore thephenotypic effects caused by genetic variation of these genes among humans.

Genetic studies

The MAO genes and identified polymorphisms

The MAOA and MAOB genes reside adjacent to each other on the X chromosome (Xp11.23-

11.4) in a tale-to-tale orientation, and are separated by 50 kb of non-coding DNA134,76,91,23. At

the amino acid level, the human MAO enzymes share 70% identity4. They are postulated to have

originated from a duplication of the ancestral MAO gene more than 500 million years ago55.Both of the genes contain 15 coding exons and exhibit identical structure of exon-intron

organization55.Several DNA polymorphisms in the MAOA and MAOB genes have been described (Table 3 and

4, respectively). In contrast to the MAO repeat polymorphisms that were widely spread andhighly variable among all human populations, most of the identified SNPs were found only inparticular populations. For example, seven out of ten SNPs at the MAOB gene have beenidentified in African-American or/and Native-American populations but were not detected inCaucasians and Asians (Table 4). Only one SNP in the intron 13 (No. 9 in Table 4) was found in

all populations. The polymorphisms at the MAOA exons 4, 9 and 15 described by Tivol et al177

had a frequency of 3 to 4% in the studied geographically undefined population (see Table 3).

This indicates that only a few of the polymorphic sites found up to date at the MAOA and theMAOB genes are informative in the majority of the human populations. Moreover, in spite of theextensive efforts of many studies that screened the coding part of the genes in order to detectfunctional DNA changes, most likely none of the identified exonic polymorphisms have afunctional effect. Only one exonic substitution in the exon 15 of the MAOA gene (Table 3)

results in a neutral amino acid change (lysine to arginine), while the others are silentsubstitutions.

Table 3. Summary of published polymophisms at the MAOA gene.

No. Polymorphism Genomicpositiona

cDNApositionb

Frequency Population Ref.

1 a 30-bp VNTR Exon 1 - 1.2 kbc Highly

variable

Widespread 146

2 a 23-bp VNTR intron 1 Highly

variable

Widespread 60

3 a (CA)n

dinucleotide

repeat

intron 2 Highly

variable

Widespread 7

4 MspI RFLP 5’ non-coding

region

30% not defined 127

5 A to C exon 4: 70 bp 435 4% not defined 177

6 T to G exon 8: 96 bp 941 40%

3%

Caucasians, Asians

Africans

62,17,

5

7 Deletion AACAT Exon 9 - 139bp 40%

3%

Caucasians, Asians

Africans

5

8 A to T exon 9: 71bp 1076 3% not defined 177

9 G to A Exon 10 - 46bp 40%

8%

Caucasians, Asians

Africans

5

10 A to G Exon 12 + 69 bp 60%

70%

Caucasians, Asians

Africans

5

11 C to T exon 14: 36 bp 1460 40%

30%

Caucasians, Asians

Africans

177,5

12 A to G (Lys-Arg) exon 15: 122 bp 1609 3% not defined 177

a The genomic position of a polymorphic locus located in an exon is shown in base pair distance from the first

nucleotide of the given exon. The position of an intronic locus is given in base pair distance from the nearest exon:“-“ indicates the location upstream of the 5’ end from the given exon, while “+” indicates the location downstream

from the given exon.b According to Bach et al. (1988)4

c The distance is approximate

Table 4. Summary of published polymorphisms at the MAOB gene.

No. Polymorphism Genomic positiona Frequency Population Ref.

1 (GT)n repeat Exon3 -264 bp Highly

variable

Widespread 77

2 A to G Exon 5: 18 bp 20% Native-American 161

3 A to G Exon 6 : 130 bp 30% African-American 161

4 C to T Exon 7 -15 bp 30% African-American 161

5 A to G Exon 10: 28 bp 30% African-American 161

6 deletion of 7

consecutive C

Exon 10 +42 bp 30% African-American 161

7 A to G Exon 10 +140 bp <1% Swedish

8 T to C Exon 11 -52 bp 30% African-American 161

9 A to G Exon 14 -36 bp 25%

50%

80%

Asians

Caucasians

Africans

61,5

10 T to C Exon 15 -3 bp 30%

20%

African-American

Native-American

161

11 C to T Exon 15: 51 bp <5% diverse sample 161,1

7

The high degree of conservation of the MAO genes suggests strong functional constrain thathampered accumulation of amino acid changes during evolution. Most probably, if there is aDNA change that affects enzyme activity, it should be located in regulatory or other non-codingregions of the genes. Only one polymorphism has been proposed to have an effect on theexpression of the gene. This has been suggested for the 30 bp VNTR in the MAOA promoter,

that has alleles with 3, 3.5, 4 and 5 copies of the repeats among the general population146. Twoindependent studies conducted promoter fusion and transfection experiments where a coresequence of the MAOA promoter with variable number of the 30 bp repeats was fused to aluciferase reporter gene and transfected into several cell lines, including neuroblastoma cell lines.The experiments indicated that the construct with 3 copies of the repeats had a significantlylower luciferase activity compared to the constructs with 4 copies. These findings suggested a

regulatory role of the VNTR locus on the transcriptional activity of the MAOA promoter161,36.However, the functional effect of the VNTR locus should be assessed in human brain.

a The genomic position is indicated in the same manner as in the Table 3.

Allelic association studies of the monoamine oxidase genes

Association to enzyme activity and/or mRNA levels

One way of looking for a presence of a functional variant that alters expression of the encodedprotein, is to analyze the correlation between different alleles of the gene and different proteinlevels in a large number of samples. Figures 12 and 13 summarize this type of studies conductedfor the MAOA and MAOB genes, respectively. For example, the MAOA activity levels in skinfibroblasts were found to correlate with specific alleles of the MspI, exon 8 and exon 14

polymorphisms at the MAOA gene62. A specific haplotype formed by alleles “MspI-“, “T”, and“C”, was clearly over-represented among samples with low MAOA enzyme activity, whilesamples with high MAOA enzyme activity displayed a mixture of several haplotypes. Also, 3copies of the repeat at the VNTR locus in the promoter region was associated with lower MAOA

activity compared to 4 copies of the repeat in skin fibroblasts of 11 individuals37. This is inagreement with previous data from gene fusion and transfection experiments mentioned above.However, only one study has attempted to investigate the effect of the VNTR on the MAOAactivity in the central nervous system (CNS). In this study, the concentration of monoaminemetabolites in cerebrospinal fluid (CSF) was measured. CSF metabolite levels should reflect the

monoamine turnover in the brain69. They found that females (N=37) with at least one copy of theallele with 4 repeats displayed a higher metabolite concentration level. In contrast, a non-significant opposite tendency was detected in males (N=51). The authors suggested a possiblesexual dimorphism in the regulation of the MAOA in brain. They also cautioned for spuriousfindings due to the weak statistical power of the studied sample. For the MAOB gene (Figure13), association analysis between alleles at the (GT)n polymorphism in intron 2 and MAOB

activity measured in platelets of 41 males yielded negative results54. In contrast, another studyreported significant but weak differences in the platelet MAOB activity, associated to different

alleles of the polymorphism at the MAOB intron 13, in a sample of 55 males52. No attempts to

identify the presence of the MAOB functional polymorphism in brain were reported.

Association to behavioral and neuropsychiatric human phenotypes

Identification of the DNA variants boosted studies attempting to investigate the associationbetween DNA polymorphisms in MAO genes and candidate human phenotypes. The MAOAgene gained more popularity among these types of studies due to several reasons. MAOAdeficiency leads to more prominent behavioral and neurochemical alterations compared to the

lack of MAOB, both in humans and in transgenic mice90. Also, a more consistent support for the

presence of a functional DNA variant, which influences enzyme activity, was presented for the

MAOA gene62,37. And most importantly, some evidence suggests a functional role for the

VNTR polymorphisms at the MAOA promoter146,36,37. The MAOA gene has been studied for

its association with phenotypes such as alcoholism, mood disorders, panic attacks, andobsessive-compulsive disorders (Figure 12). The majority of the studies investigating the MAOBgene focused on the relationship of the gene to Parkison’s disease (Figure 13). As seen in thefigures, conflicting association findings were reported for almost all of the phenotypes studied.For example, several different studies suggested a link between (CA)n polymorphism at theMAOA gene and alcoholism (Figure 13). However, the longer alleles were associated with

alcoholism in a sample of Euro-American and Afro-American males187, but only a single

specific allele was linked to alcoholism in another cohort of Euro-Americans129 as well as in

Chinese64. Moreover, an investigation of the MAOA gene exon 14 variants among alcoholic

subjects living in Taiwan led to contradictory findings in different sub-populations studied64.

Figure 12. Summary of allelic association studies investigating the MAOA gene. Only phenotypes that havebeen reported to be influenced by the MAOA gene variants are shown. The arrows indicate positive findings to the

analyzed phenotype, while the crossed arrows indicate lack of association. The numbers indicate the polymorphism

that yielded the results, and they correspond to the numbers listed in the Table 3. Different symbols representdifferent type of DNA markers: “���“ indicates a VNTR, “��“ denotes dinucleotide polymorphism, “�“

represents single nucleotide polymorphism.

MAOAgene

Bipolardisorder

Obsessive-compulsive

disorder 1

Alcoholism

Panic disorder

2

13

63

Parkinson’sdisease 3

MAOA activity

In fi

brob

last In brain

Monoamine Metabolites

in CSF11

46

11

Aggression

11

6

11

1143

3 11

114

31

1

36

11

1

Figure 13. Summary of allelic association studies investigating the MAOB gene. Only phenotypes that have

been reported to be influenced by the MAOB gene variants are shown. The arrows indicate positive findings to theanalyzed phenotype, while the crossed arrows indicate lack of association. The numbers indicate the polymorphism

that yielded the results, and they correspond to the numbers listed in the Table 4. Different symbols representdifferent type of DNA markers: “��“ denotes dinucleotide polymorphism, “�“ represents single nucleotide

polymorphism.

Studies that investigated the relationship between the MAOA genetic variation and mooddisorders (bipolar and unipolar) have also provided conflicting observations. Most of the positive

association were observed for the (CA)n polymorphism in the MAOA94,73,145,51,95,136.

However, there were opposite reports as well33,123,113,130,179. A meta-analysis performed on acombined sample of Caucasian and Japanese populations discovered significant differences in

the (CA)n allele distribution among bipolar cases compared to controls51. However, the twopopulations showed different alleles associated with a higher risk of the disease.

The association of the VNTR polymorphism and panic disorder was proposed in one study36 but

not detected in another58. Similar discrepancies emerged between different studies investigating

MAOBgene

Parkinson’sdisease

9

1Alcoholism

1

MAOB activity

1

1

9

1

9

1

9

allelic association of polymorphisms at the MAOA gene with aggression186,103 and Parkinson’s

disease82,63,119,135,30.Studies that analyzed the relationship between polymorphisms at the MAOB locus and

Parkinson’s disease have also reported antagonistic findings. Allelic association of susceptibilityto Parkinson disease with "A" allele at the polymorphic site in intron 13 of the MAOB gene was

suggested in a study performed among Caucasians82. However, it was not replicated in a

Japanese population nor in an Australian cohort110,107, and more surprisingly, the "G" allele was

related to Parkinson’s disease in another sample of Caucasians61,29. Another marker in theMAOB gene, the (GT)n microsatellite in the MAOB intron 2, also resulted in opposite findingsconcerning the association to the susceptibility to Parkinson disease in different populations

studied119,118,107,108. Similar controversies were reported for association of the MAOB

polymorphisms and alcoholism129,166.A major statistical weakness for the association studies is that most of them do not adjust thesignificance threshold for the number of statistical tests performed (Bonferroni correction). Mostlikely, after correction for multiple testing, many of the reported association results would not besignificant. Moreover, the low statistical power in some of them may be due to small samplesize.On the other hand, the contradictory association findings are not so surprising, if they are viewedin the context of the differences in genetic composition of different populations, the age of thepolymorphisms and their frequencies in different populations. These are the factors determiningthe success for allelic association studies as discussed earlier.Dinucleotide repeat polymorphisms have been used in many association studies (Figure 12 and

13), but as it was discussed in a previous section, this type of markers may be unfit to discoveran unknown functional variant because of their high mutation rate. Reports where association isfound to only a few markers out of several linked markers tested are also anticipated, given thedifferent ages of the polymorphims tested and the different physical distances in relationship tothe unknown functional variant. Moreover, studies reporting association of the same phenotypeto different alleles at a certain polymorphism are also possible (e.g. Parkinson’s disease andintron 13 polymorphisms in MAOB gene) given that the polymorphism used is not functional,and that the tested samples come from different populations that are likely to have distinctpattern of linkage disequilibrium.

Aims

Part I. Hearing disabilities

� To genetically map gene(s) causing hereditary hearing loss in two large Swedish pedigrees.

� To perform a mutation screening of candidate genes for hearing disabilities that map withinthe identified genomic location(s) for hearing loss in the Swedish families.

Part II. Monoamine oxidase activity

� To identify a set of DNA polymorphisms in the genes for Monoamine Oxidases that arefrequent in the Swedish population.

� To investigate the association between the characterized DNA polymorphisms and MAOmRNA levels and /or protein activity in human brain.

Present investigations

Part I. Hearing disabilities

Family I. Results

A Swedish family I (Figure 14) with non-syndromic progressive bilateral sensorineural hearing

loss was ascertained. We observed that affected family members differed in the severity ofhearing impairment and age of onset for the disease. One group of individuals manifested asevere hearing loss with onset in the first decade of life (Type I), while another group ofindividuals expressed a mild hearing decrease with the onset in the second or third decades oflife (Type II). Meticulous description of the hearing loss in the family I was presented

elsewhere9.

Figure 14. A Swedish pedigree I with non-syndromic hearing loss.

?

?

UnavailableModerate to severe hearingloss with high frequenciesmost impaired

Mild hearing loss atfrequencies 4-6 kHzUnaffected

1 2

1 1098765432 131211 14 15 16 17

6 7 8 9 10 11 12 13 141 2 3 4 5

?Abnormalaudiogram

I

II

III

Paper I prensents the results of a linkage study, where 13 candidate loci for autosomal dominantnon-syndromic hearing loss (NSHL) have been tested for linkage to hearing loss in the family.Surprisingly, we found significant linkage (Z>3) for markers at candidate locus DFNA12(11q22-q24) and suggestive linkage (Z>2) for markers at locus DFNA2 (1p32). These results

suggested digenic inheritance of hearing loss in the family. For the chromosome 11 locus, wenarrowed down the reported candidate region for the DFNA12 locus from 36 to 12 cM. Thedisease-associated haplotype on chromosome 1 spanned over 6 cM region and overlapped withthe candidate region delimited in some other families with hearing loss mapped to DFNA2.Moreover, we observed that individuals with severe hearing loss inherited both disease-associated haplotypes, on chromosome 1 and chromosome 11, while mildly affected individualspossessed only one of the haplotypes. This finding suggests that two genes segregate togetherwith the hearing impairment in the family and the presence of both of them results in aggravatedhearing loss compared to the hearing loss caused by each of the genes separately.

DFNA12 locus

Paper II describes the results of a mutation screening in a candidate gene residing at one of thecandidate regions (DFNA12), reported for the Swedish family in the Paper I.The human α-tectorin (TECTA) gene was cloned and mapped within the DFNA12 candidate

region at the same time as we were conducting our investigations66,188. α -Tectorin gene is

expressed only in the inner ear138. It encodes a noncollagenous protein of extracellular matrixcalled tectorial membrane that participates in the mechano-transduction of a sound (Figure 11).TECTA protein is very conserved among different species. Moreover, different domains of theprotein show strong homology to other known proteins: entactin G1, zonadhesin protein or D-

type repeats of the von Willebrand factor and zona pellucida protein (Figure15)143,87.During our investigations, several missense mutations were reported by others to segregate with

deafness in some families188. We performed a mutation analysis of the TECTA gene in theSwedish family I and identified four amino acid changes (Figure 15). One of the changesresulted in a cysteine to serine (C1057S) replacement, in the zonadhesin domain of TECTA,which segregated with the disease haplotype on chromosome 11 and was not present in a controlpopulation. This results in the replacement of a cysteine that is conserved between the mouse andthe human proteins. Moreover, the replaced Cys was also conserved among four internal repeatsof the zonadhesin/von Willebrand factor domain of the TECTA protein. Based on findings from

functional biochemical studies of von Willebrand factor88,106,191, we hypothesize that thereplacement of the Cys in TECTA can cause a change in the cross-linking of the protein. The

Figure 15. Schematic structure of α-Tectorin protein. α-Tectorin protein consists of several domains that are

homologous to other known proteins. Locations of the reported amino acid are shown. The bold arrowheads indicate

substitutions that are associated to a hearing loss in the different families tested188,3,116, while the thin arrowheadsindicate possibly neutral amino acid changes.

identification of the mutation in the Swedish family provides additional evidence for theinvolvement of TECTA in hearing disabilities.

DFNA2 locus

Two genes, KCNQ4 and GJB3, have been cloned and shown to be responsible for hearing loss insome of the families linked to the DFNA2 locus. KCNQ4 is a member of the potassium channel

gene family81 and GJB3 is a gap junction protein ß-3 (connexin 31)205. A mutation in KCNQ4

was found in a family with autosomal dominant hearing loss. It was shown that the mutation

abolished the potassium currents and exerted a strong dominant-negative effect81. Later, more

ENT ZPPartial entactin G1 domain Zona pellucida domain

Von Willebrand factor D repeats/Zonadhesin like domains

Hydrophobicregion

Ser/pro/thre rich domain

ENT 0 1 2 3 4 ZPG

ly37

1Arg

Ala

932V

al

Cys

1057

Ser

Ser2

100T

hr

Cys

1619

Ser

972

ST

OP

Tyr

1870

Cys

Leu

1820

Phe

Gly

1824

Asp

Swedish family

Other families

families with mutations in the KCNQ4 were reported32,170,184. Similarly, several families with

hearing loss have been found to harbor different mutations in the GJB3 gene205,98.To find out whether GJB3 and/or KCNQ4 are responsible for hearing loss in the Swedish

pedigree, the coding regions of the genes were sequenced in two members (severely affected andhealthy) of the family (for KCNQ4: Kubisch et al., unpublished data; for GJB3: Balciuniene etal., unpublished data). However, we did not find any mutations in the coding sequences of thegenes. Moreover, the KCNQ4 gene was later localized outside the region of linkage onchromosome 1. Genes for two additional gap junction proteins, GJB4 and GJB5, were found tobe located in a vicinity of the GJB3 gene. However, sequence analysis of the coding regions forthese two genes did not reveal any mutation in the Swedish family. The possibility remains thatadditional gene(s) involved in hearing disabilities are clustered in this region.

Figure 16. A Swedish pedigree II with non-syndromic hearing loss.

?

?

?

?

?

?

?

?

?

? Abnormal audiograms

Severe to profound hearingloss with higherfrequencies most affected

Severe hearing loss starting atfrequencies >2000Hz

Mild to severe hearingloss with higherfrequencies most affected

1 2

1 2 3 4 5 6 7 8 9

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 2322

19181514 161312

3332313029282726252423222119181716151413121110987654321 34

176 7 8 9 10 111 2 3 4 5

V

I

III

II

IV

?

?

? ?

Unaffected

Suspected environmental damage

Dimmedsymbols Unavailable

** * *

*

Family II. Results

In parallel to the analysis of the family I, we have investigated another Swedish family (familyII, Figure 16) with sensorineural bilateral post-lingual hearing loss. Phenotypic examinationrevealed three different types of hearing loss segregating among the affected family members: asevere hearing loss over all frequencies (Figure 16, black filled symbols), a severe loss atfrequencies higher than 2 kHz (Figure 16, black filled symbols with white dots), a mild to severehearing loss most pronounced at higher frequencies (Figure 16, white filled symbols with blackdots). The onset of the disease was around 10 years of age.We have also performed a linkage analysis for DNA markers residing at thirteen candidate

regions for autosomal dominant hearing impairment reported by then. However, none of thegenotyped markers showed linkage to the hearing loss in this family.

Discussion: three puzzles

I. Phenotypic variation

As it was mentioned earlier, hearing impairment is a very variable phenotype among differenthuman pedigrees. Moreover, a considerable phenotypic variation inside families can also beobserved, particularly for post-lingual hearing loss. The two Swedish families presented in this

work are good examples of such intra-familial variability. In addition, an Indonesian family wasalso reported where some of the affected members displayed a severe hearing loss at allfrequencies, some of them manifested hearing loss only for high frequencies, while some of the

persons had hearing threshold decrease only for certain frequencies31. A similar pattern of

phenotypic variation was observed in a French family21.What factors can contribute to this phenotypic variation?Various environmental factors, such as ototoxic drugs, exposure to high noise and variousinfections play a significant role in shaping individual hearing capabilities. On the other hand,the phenotypic variation may have a genetic reason. The same gene may respond differently to a

different genetic background, or there may be an additional major genetic factor that interactswith the causative gene. As we have shown, the Swedish family I probably exemplifies a casewhere additive interaction of mutations in two genes produces the severe form of hearing loss.The presence of more than one defective gene contributing to the phenotypic variation in thesame family would not be surprising, taking into account a large number of genes expected to beinvolved in hearing process, and a high prevalence of the disorder among humans. Also,assortative mating among deaf individuals in earlier generations leads to an enrichment ofdefective genes in later generations.

A different gene interaction hypothesis has been proposed in a consanguineous Pakistanianfamily with recessive deafness linked to 4q28. There were 15 individuals coming from differentbranches of the pedigree that were homozygous for the disease haplotype. However, seven ofthem did not manifest hearing impairment. Later it was found that these non-penetrantindividuals shared a haplotype on 1q23 and it was hypothesized that this region contains a

modifier gene that suppresses the hearing loss phenotype140.Nevertheless, identification of the gene aggravating hearing loss in the Swedish family, and thegene suppressing the deafness in the Pakistanian family, will help to resolve the puzzle of thephenotypic variation in these families and encourage similar kind of detailed studies in others.

II. α-Tectorin

Identification of the Cysteine to Serine mutation in TECTA, that segregated with hearing loss inthe Swedish family, but not in a control population, provides good support for the association ofthis gene to the disease. However, the proof that the mutation is causing the disease can beachieved only if the functional effect of the mutation is demonstrated. α-Tectorin is a very

conserved protein among different species. It has been estimated that human and mouse α-

Tectorin share 95% identity at the amino acid sequence level188. This suggests that the gene hasa very conserved evolutionary function and similar changes may produce similar phenotypicconsequences among different species. Moreover, all mutations identified in several humanpedigrees result in a change of a conserved amino acid between mouse and human TECTAproteins. Therefore, a transgenic mouse with a specific human mutation in the TECTA gene orwith the entire mutant human gene would be a very promising model to unravel the functionaleffect of the TECTA mutations. It can also provide clues about why different TECTA mutations

result in different hearing loss phenotypes among humans188,116.

III. DFNA2

The DFNA2 locus contains two genes: KCNQ4 and GJB3 that have been implicated in hearingimpairment in some families, but none of them is mutated in the Swedish family. This suggeststhat there may be a third gene present. The possibility of the third gene is also supported by thepresence of an Indonesian family with a hearing disability locus that maps to the same candidate

region as the Swedish family and does not harbor a mutation in any of the two genes183. Inaddition, we have sequenced the coding regions of two other gap junction proteins, GJB4 and

GJB5, lying in a vicinity of the GJB3 gene, and failed to detect mutations in these genes. It maybe that the functional DNA mutation resides in the non-coding part of the GJB genes. We havenot investigated the non-coding regions of these genes.

Recently, a draft assembly over the clones sequenced by the Human genome project was

released and is constantly being updated75. The investigation of the genomic sequencessurrounding the DFNA2 region delimited in the Swedish family resulted in several findings. Firstof all, I discovered that the GJB genes lie outside the delimited region. This fact, provided that

the order of the clones is correct, unequivocally excludes them as susceptibility candidates in theSwedish family. Also, the region contains several known genes such as glutamate receptor 7(GRIK3), nucleolar GTPase, axonemal dynein light chain (hP28), and some predicted gene-likesequences. The region has several gaps, not yet sequenced, that may contain additional genes.

Future directions

Family I. The typing of additional individuals belonging to the family I would be extremelyhelpful in clarifying the issue on the possible involvement of two genes in the hearing loss trait.However, no additional members of the family are available to date. Therefore, the alternativeapproach is to identify other genes that reside in the candidate region, and screen them formutations in the family I. However, the implication of the same gene in the Indonesian familyand/or other families would be necessary to substantiate the evidence for a third candidate genefor hearing disabilities in the DFNA2 locus. In parallel, functional studies using animal modelsdefective for both genes: TECTA and the potential DFNA2 gene, may shed light on theinteraction mechanism.

Family II. As we discussed earlier, there is a considerable amount of phenotypic variation amongthe affected members of the family. Also, some individuals married into the family (Figure 16,individuals III:8, IV:10 and III:4) have a decreased hearing sensitivity. This poses the risk thattheir genes could underlie the hearing loss in their children, and therefore, their children mightnot be suitable subjects for linkage analysis. Considering all these limitations, we concluded thatthe currently available family material is not sufficient to obtain significant results in a genome-wide scan. On the other hand, the material may be sufficient to follow-up already reportedcandidate regions, or regions that will be identified in the future.

Part II. Monoamine oxidase activity

The goal was to investigate whether genes encoding MAOA and MAOB have an effect on thevariation of MAO mRNA levels and/or enzyme activities observed in humans. We identifiedDNA polymorphisms in the MAO genes and investigated their association with MAO mRNAand/or enzyme activity levels.

Results

Paper III describes novel DNA variants in the MAOA and MAOB genes. Furthermore, thenewly identified variants provide new information on the evolutionary history of modernhumans.The sequences of exonic and flanking intronic regions of these two genes were determined in agroup of Swedish males. The sequence analysis revealed several novel bi-allelic polymorphismsin the MAO genes. Since both of the genes reside on the X chromosome, and only males werestudied, we were able to define the haplotypic relationship between the identified allelic variants.We investigated the diversity and geographical distribution of haplotypes that contained five

closely located and very frequent MAOA polymorphisms, in a sample of Caucasians, Asians,and Africans. We found only two common haplotypes with similar frequencies in Caucasian andAsian populations. However, only one of the two haplotypes was also the most frequent inAfricans, while the other haplotype was present in only one African male. In general, Africansshowed greater diversity of haplotypes and displayed 5 of 6 possible non-recombinanthaplotypes for these 5 positions. This profound change in haplotype number and frequenciesfrom African to non-African populations supports a possible bottleneck during the dispersion ofmodern humans from Africa.

Paper IV reports a study investigating the involvement of MAO genes in the regulation of MAOmRNA expression levels and enzyme activity in human brain.

Three polymorphic variants were investigated: a VNTR located in the MAOA promoter, and twoneutral single nucleotide polymorphisms located in the exon 8 of the MAOA (A-SNP8) and inthe intron 13 of the MAOB (B-SNP13). mRNA expression levels and enzymatic activities of theMAO were measured in post-mortem human brain samples. We studied whether the presence ofa certain allele, or a certain haplotype, correlated with different mRNA expression levels and/orprotein activities of the MAO found in these brain samples. In addition, we studied the linkbetween MAOB alleles and MAOB protein activity in platelets. We found a significantassociation (p=0.02) between B-SNP13 alleles and MAOB protein activity in human brain butnot in blood. These results suggested a different MAOB control in brain compared to blood.Also, we show that MAOB activity is significantly higher (p=0.006) in the brain of patients withAlzheimer’s disease compared to a control population. Moreover, we provide evidence for a

functional DNA variant that lies on a specific MAOA haplotype, and that is likely to have aninhibitory effect on the MAOA protein activity in human brain.

Discussion: MAO genetic variation

I. Implications for human evolution and linkage disequilibrium studies

We identified five frequent polymorphic loci at the coding and flanking intronic regions of theMAO genes. These variants spanned over a relatively short distance (approximately 12 kb) andwere present in all major human populations tested. Recombination events within short stretchesof DNA are less likely to occur, thus the evolutionary relationship among such DNA sequencescan be determined more reliably than for long DNA sequences reshuffled by recombination. A

distinct distribution of the non-recombinant haplotypes among the three major populations wasobserved. The African population presented several non-recombinant haplotypes, including aputative ancestral haplotype, while Asians and Caucasians displayed only two types ofhaplotypes. This haplotype distribution pattern supports earlier suggestions that the oldest human

population originated in Africa (for a review, see ref70). In addition, we observed that thefrequency of the two non-recombinant haplotypes found in Asians and Caucasians was similarbetween these two populations, while it differed dramatically in Africans. This implies thatAsian and Caucasian populations originated from the same ancestor, and probably suffered apopulation bottleneck soon after the split from the ancestral African population. However, ourdata do not allow to distinguish between the types of bottleneck that could have occurred. Two

possible types are demographic or selective27. Demographic bottlenecks, such as epidemics,famines, migration etc., would leave imprints on many loci of the genome. In contrast, abottleneck caused by selection for the tested loci or closely linked loci would be restricted to asmall genomic region. Some of genetic loci may have undergone both types of bottlenecks, andthis would make the inference about their evolutionary history even more difficult. Thepossibility of demographic events shaping the genetic diversity of the MAOA locus is alsosupported by another study that investigated the DNA variation of a non–coding 10 kb sequence

on Xq13.371. Similar to the MAOA gene, this region is also close to the X chromosomecentromere, but lies on the opposite arm than the MAOA locus. Representatives from all major

linguistic groups were sequenced. The authors also reported a higher genetic diversity inAfricans than in other populations. Moreover, they observed one haplotype that was specific toAsians, and they speculated that this haplotype could be due to some past demographic events.However, their findings can also be explained by the action of natural selection.The description of haplotype variation existing among different populations can also revealimportant information for the proper design of allelic association studies. For example, thedistinct population distribution of the non-recombinant haplotypes observed in the MAOA locusindicates a different linkage disequilibrium pattern among the three populations. Consequently, it

Figure 17. Evolutionary relationship among current non-recombinant African haplotypes for five MAOApolymorphisms. The horizontal bars represent 5 different haplotypes for five variable positions in the MAOA gene(No. 6, 7, 9, 10, and 11, in Table 3). “T”,”G”,”A”,”C” denote nucleotides found in the variable positions, “ins” and

“del” are abbreviations for “insertion” and “deletion” of an AACAT sequence, respectively. The haplotype 1 is theputative ancestral haplotype. The two most frequent haplotypes among Africans are framed in bold. The asterisk “*”

denotes the two main haplotypes found in Asians and Caucasians. A part of the “T,ins,G,A,T” chromosomes carry a

putative functional variant indicated as solid triangle “�”.

suggests that alleles of the same closely located polymorphisms would not yield similarassociation findings in different populations.In Figure 17, the possible evolutionary relationship among identified African non-recombinanthaplotypes is indicated. Let's assume that a hypothetical functional variant occurred on one of thechromosomes carrying haplotype 2. In this case, a case-control study performed on Caucasiansand Asians would show clear allelic association of the functional polymorphism with any of thefive polymorphic loci. However, among Africans, the allelic association would strongly dependon the sample size selected for the study and the variant tested. Moreover, the presence ofrecombinant haplotypes will make the association picture even more blurred.

T ins G C G

G del T A

T ins T A

T ins C G A

*

* T ins T G

Cases

Controls

A

A

A

1

2 3

2

4

5

T ins T G A

The power of association studies may be improved by analyzing haplotypes comprising severalclosely located polymorphic markers. However, the number of DNA variants described for theMAO genes is relatively small, and some of the variants are rare and restricted to a specificpopulation. This suggests that some populations will contain only few and most probablyinfrequent haplotypes for the given polymorphisms. Therefore, a panel of new DNA

polymorphisms, that are common in the studied populations, would be advantageous. Also, largesets of samples should be investigated to increase the chance to detect smaller effects of theMAO in complex human phenotypes.

II. Implications for functional variation

The lack of clear evidence for functional variants influencing MAOA and MAOB activity inhuman brain is the main caveat for allelic association studies attempting to decipher thecontribution of the MAO genes to complex behavioral and neuropsychiatric human traits. We

investigated the functional effect of MAO polymorphims in human brain by examining a set ofpost-mortem male brain samples. We studied the relationship of the variation in MAO genes totwo phenotypes: mRNA levels and enzyme activity. Theoretically, this strategy can provideclues on two questions: are there functional DNA variants that control the MAO gene function?If so, do they act at the transcriptional or post-transcriptional level? DNA variants showingassociation to mRNA expression levels, but not to protein activity, may seem less important forthe variability of candidate human phenotypes. Indeed, proteins are the final determinants for theexpression of most phenotypes. However, it has been reported that different pharmacological,physiological or other non-genetic agents can influence substantially MAO activity in

humans199. For instance, a 20-40% reduction of both MAO activities was reported in the brain of

living smokers compared to non-smokers or ex-smokers10,48,49. This would suggest that MAOactivity measured at a certain time point may not reflect the regulatory effects exerted on thegenes. Consequently, the association between enzymatic activity and DNA polymorphismswould be missed. Similarly, an allelic association to protein activity, but not mRNA expressionlevels, may indicate the presence of a functional DNA variant influencing post-transcriptionalregulation of the genes.The most important finding presented in our study is the association between the neutralpolymorphism in the intron 13 of the MAOB gene and MAOB activity in brain but not in blood.However, the information on smoking habits was not available for the individuals that were thedonors of the brain autopsies. Therefore, we could not correct our association findings for the

effect of smoking. On the other hand, the difference between the MAOB activity levels found inindividuals with different alleles is unlikely to be greatly affected by smoking, unless thesusceptibility to smoking is associated to the tested intron 13 polymorphism.

The positive association between the MAOB polymorphism and MAOB activity in brain, but notin platelets, suggest the presence of a DNA variant that has functional importance in brain. Thisis not unexpected, given a distinct regulation of the MAOB activity in platelets and in brain as it

was proposed earlier202,206,42. Moreover, it stresses the importance to investigate the functional

effect of a given DNA variant in the relevant tissue for the phenotype of interest.A functional effect of the VNTR in the MAOA promoter was clearly demonstrated in cellcultures with transfected reporter constructs. However, our study performed on human brainsamples, together with another study that investigated the link between MAOA polymorphisms

and MAO metabolite levels in CSF could not detect the same effect69. Several explanations arepossible. It may be that the VNTR does not have such a profound effect on MAOA activity inhuman brain and an additional functional variant is present in this gene as we have suggested(Paper IV). An alternative is that the MAOA activity measurements in these two studiesperformed in brain autopsies and in cerebrospinal fluid do not reflect the activity in the livingbrain. Another explanation may be that due to a small sample size, the two studies did not have

enough power to reach statistically significant results for small changes in activity.In any case, for more definitive conclusions about the presence of the functional MAO variationsin human brain, additional investigations performed on larger sample sets and in different humanpopulations are needed. A specific sampling strategy that would decrease the effect of non-genetic factors on MAO variability would be an advantage.

Future directions

The experimental evidence supports the presence of at least one functional polymorphismaffecting MAO activities in brain, but the sequence and position of this variant is not yet known.A large sequencing enterprise can be launched to find DNA differences between individuals withextreme values in the levels of brain MAO activity. Depending on the length of the sequencedfragments, many polymorphic variants will be found. As functional MAO variants are likely tobe located in non-coding regions of the genes, the next difficult task will be to prove thefunctional effect of any of them. In vitro experimental approaches such as DNase footprintingand gel retardation assays can be used to investigate the possibility that the tested polymorphismis located within a protein-binding site. Northern blots would reveal the presence of alternative

transcripts that may be caused by the polymorphism. Gene fusion experiments performed in cellcultures may be useful to investigate the involvement of the given polymorphism intranscriptional regulation of the gene. Transgenic animal models may also complement theevaluation of the phenotypic effect of the candidate functional polymorphism at the wholeorganism level. Finally, allelic association studies between a given polymorphism and alteredprotein expression levels are necessary to evaluate the functional significance of the testedpolymorphism(s) in human tissues.

Acknowledgements

I would like to thank all the past and present members of the Department of Genetics andPathology, Section of Medical Genetics for creating an enjoyable and pleasant atmosphere.Especially, I would like to express my gratitude to:

Professor Ulf Pettersson for opening the gate to Medical Genetics, for your remote but very

supportive and effective supervision, for scientific enthusiasm and inspiration.

Elena Jazin, my latest and closest supervisor, a tough woman with a passion for science andsuccess, for your scientific open-mindedness and inspiration, for your patience, support andunderstanding, for great brainstorming minutes and for the well-timed and very pleasantoutdoors-indoors bowling and hamburger parties.

To all people in the Psychiatric Genetics group:Eva Lindholm for sharing the last never ending meters to the finish, for extremely interestingphilosophical discussions where reality and insanity fuses, for your sincere and analytic

personality and for inventing a new way to change radio channels. Do not analyze this time!..Lina Emilsson for helping me to straighten working relationship with Taqman, for being a veryenergetic and cheerful person and for showing how easy it could be to keep things in order; Anja

Castensson for statistical discussions, growing-up recollections and for verbal enthusiasm to“gympa” passes; Paula Jalonen for teaching me tricks of sequencing and for the help in myprojects.

To other present and past office-mates:Ludmila Elfineh for being a very supportive, warm and caring friend, for frequent talks about thereal life outside the lab and for sharing many moments of happiness and distress;Niklas Nordqvist for relaxed and positive but constructive attitude in all stress-threatening

situations; Hongxing Zhao for teaching me the secret of Chinese kitchen: “it is so simple”;Haitao Yang and Tesfai Emahazion for being helpful and nice persons and for sharing the“radioactive age” of genotyping; Mats Sundvall for sharing his expertise on linkage analysis;Hamid Darban for surviving the room remodeling chaos that I initiated.

My collaborators: Professor Niklas Dahl and Professor Erik Borg for fruitful collaboration onthe hearing loss project, Professor Ann-Christine Syvänen for making the mini-sequencing areality, Professor Lars Oreland for initiating the monoamine oxidase project and Sigrid

Sandberg for the coordination of the collection of "tons" of blood samples.

Martha Alarcón-Riquelme for optimism and enthusiasm in every day life and in science; Bo

Johannesson for fun in anatomizing any scientific problem; Lucia Cavelier for multiple tips onsurviving in foreign countries; Charlotta Olsson, Anette Hagberg, Ludmila Prokunina, Dimitri

Tentler, Anna-Karin Lindqvist, Delal Gemici, Patrik Magnusson, Maritha Mendel-Hartvig, Anh-

Nhi Tran, Dan-Oscar Antson and Kicki Lagerstedt for occasional but pleasant chats.

Elsy Johnsen and Inger Jonasson for keeping the department safe and in order; Britt-Marie

Carlberg for the great parties at your place and for cute “tomtes”; William Schanong forcomputer help; Eddie Lundh, Stig Söderberg, Marcus Lundqvist and Eva Ferngren for all thepractical help; Jeanette Backman, Elisabeth Sandberg and Anette Uselius for arranging the fluentcourse of administrative matters.

The Beijer Foundation for financial support.

All members and associate-members of HUGO’94 group at Clinical Genetics Department in

Umeå University, in particular Kristina Forsman-Semb for introducing me to linkage analysisand for the great time in permanent darkness of Umeå.

Professor Vaidutis Kucinskas, the head of Human Genetics Centre at Vilnius University forintroducing me to the field of human genetics and for creating the opportunity for me to come toSweden. Also, to Ramune Vilkaite, my very first supervisor.

My former and present Lithuanian friends in Uppsala and Stockholm, in particular Arvydas

Kanopka, Jurga Laurencikiene, Evaldas Laurencikas, Egle Miniotiene, Andrius Miniotas andVita Dauksaite for once in a while creating the illusion of being “at home” and for making the“exile” to Sweden more enjoyable.

My parents for the immeasurable love and support and for their devotion to the grandchildrenthat helps them to survive with two scientists as parents...My parents-in-law, especially my mother-in-law for being so helpful when needed.

Finally, to my two treasures – Aiste and Aurimas – for uncovering the secrets of true happiness,and to my husband Darius for love, encouragement and tolerance, for being the perpetuum

mobile in our life.

References

1. af Klinteberg, B., Schalling, D., Edman, G., Oreland, L. & Asberg, M. (1987) Personality correlates of plateletmonoamine oxidase (MAO) activity in female and male subjects. Neuropsychobiology 18, p.89-962. Alexopoulos, G.S., Young, R.C., Lieberman, K.W. & Shamoian, C.A. (1987) Platelet MAO activity in geriatric patientswith depression and dementia. Am J Psychiatry 144, p.1480-14833. Alloisio, N. et al. (1999) Mutation in the zonadhesin-like domain of alpha-tectorin associated with autosomal dominantnon-syndromic hearing loss. Eur J Hum Genet 7, p.255-2584. Bach, A.W. et al. (1988) cDNA cloning of human liver monoamine oxidase A and B: molecular basis of differences inenzymatic properties. Proc Natl Acad Sci U S A 85, p.4934-49385. Balciuniene, J., Syvänen, A.-C., McLeod, H.L., Pettersson, U. & Jazin, E.E. (2001) The Geographic Distribution ofMonoamine Oxidase Haplotypes Supports a Bottleneck During The Dispersion Of Modern Humans From Africa. J Mol Evol 52,6. Belmaker, R.H., Bracha, H.S. & Ebstein, R.P. (1980) Platelet monoamine oxidase in affective illness and alcoholism.Schizophr Bull 6, p.320-3237. Black, G.C., Chen, Z.Y., Craig, I.W. & Powell, J.F. (1991) Dinucleotide repeat polymorphism at the MAOA locus.Nucleic Acids Res 19, p.6898. Bongioanni, P., Mondino, C., Boccardi, B., Borgna, M. & Castagna, M. (1996) Monoamine oxidase molecular activity inplatelets of parkinsonian and demented patients. Neurodegeneration 5, p.351-3579. Borg, E., Samuelsson, E. & Dahl, N. (2000) Audiometric characterization of a family with digenic autosomal, dominant,progressive sensorineural hearing loss. Acta Otolaryngol 120, p.51-5710. Boulton, A.A., Yu, P.H. & Tipton, K.F. (1988) Biogenic amine adducts, monoamine oxidase inhibitors, and smoking.Lancet 1, p.114-11511. Breakefield, X.O. & Edelstein, S.B. (1980) Inherited levels of A and B types of monoamine oxidase activity. SchizophrBull 6, p.282-28812. Brinkmann, B., Klintschar, M., Neuhuber, F., Huhne, J. & Rolf, B. (1998) Mutation rate in human microsatellites:influence of the structure and length of the tandem repeat. Am J Hum Genet 62, p.1408-141513. Brunner, H.G. (1996) MAOA deficiency and abnormal behaviour: perspectives on an association. Ciba Found Symp 194,p.155-164; discussion 164-15714. Brunner, H.G., Nelen, M., Breakefield, X.O., Ropers, H.H. & van Oost, B.A. (1993) Abnormal behavior associated witha point mutation in the structural gene for monoamine oxidase A. Science 262, p.578-58015. Brunner, H.G. et al. (1993) X-linked borderline mental retardation with prominent behavioral disturbance: phenotype,genetic localization, and evidence for disturbed monoamine metabolism. Am J Hum Genet 52, p.1032-103916. Burge, C. & Karlin, S. (1997) Prediction of complete gene structures in human genomic DNA. J Mol Biol 268, p.78-9417. Cargill, M. et al. (1999) Characterization of single-nucleotide polymorphisms in coding regions of human genes. NatGenet 22, p.231-23818. Cases, O. et al. (1995) Aggressive behavior and altered amounts of brain serotonin and norepinephrine in mice lackingMAOA. Science 268, p.1763-176619. Castensson, A., Emilsson, L., Preece, P. & Jazin, E.E. (2000) High-resolution quantification of specific mRNA levels inhuman brain autopsies and biopsies. Genome Res 10, p.1219-122920. Castro Costa, M.R., Edelstein, S.B., Castiglione, C.M., Chao, H. & Breakefield, X.O. (1980) Properties of monoamineoxidase in control and Lesch-Nyhan fibroblasts. Biochem Genet 18, p.577-59021. Chaib, H. et al. (1994) A gene responsible for a dominant form of neurosensory non-syndromic deafness maps to theNSRD1 recessive deafness gene interval. Hum Mol Genet 3, p.2219-222222. Chakravarti, A. (1999) Population genetics--making sense out of sequence. Nat Genet 21, p.56-6023. Chen, Z.Y., Powell, J.F., Hsu, Y.P., Breakefield, X.O. & Craig, I.W. (1992) Organization of the human monoamineoxidase genes and long-range physical mapping around them. Genomics 14, p.75-8224. Clayton, D. (1999) A generalization of the transmission/disequilibrium test for uncertain- haplotype transmission. Am JHum Genet 65, p.1170-117725. Collins, A., Frezal, J., Teague, J. & Morton, N. The genetic location database.(http://cedar.genetics.soton.ac.uk/public_html/ldb.html).26. Collins, A., Frezal, J., Teague, J. & Morton, N.E. (1996) A metric map of humans: 23,500 loci in 850 bands. Proc NatlAcad Sci U S A 93, p.14771-1477527. Collins, A., Lonjou, C. & Morton, N.E. (1999) Genetic epidemiology of single-nucleotide polymorphisms. Proc NatlAcad Sci U S A 96, p.15173-1517728. Comings, D.E. (1998) Polygenic inheritance and micro/minisatellites. Mol Psychiatry 3, p.21-3129. Costa, P. et al. (1997) Association of a polymorphism in intron 13 of the monoamine oxidase B gene with Parkinsondisease. Am J Med Genet 74, p.154-15630. Costa-Mallen, P. et al. (2000) The EcoRV genetic polymorphism of human monoamine oxidase type A is not associatedwith Parkinson's disease and does not modify the effect of smoking on Parkinson's disease. Neurosci Lett 278, p.33-36

31. Coucke, P. et al. (1994) Linkage of autosomal dominant hearing loss to the short arm of chromosome 1 in two families. NEngl J Med 331, p.425-43132. Coucke, P.J. et al. (1999) Mutations in the KCNQ4 gene are responsible for autosomal dominant deafness in fourDFNA2 families. Hum Mol Genet 8, p.1321-132833. Craddock, N. et al. (1995) No evidence for allelic association between bipolar disorder and monoamine oxidase A genepolymorphisms. Am J Med Genet 60, p.322-32434. Danielczyk, W., Streifler, M., Konradi, C., Riederer, P. & Moll, G. (1988) Platelet MAO-B activity and thepsychopathology of Parkinson's disease, senile dementia and multi-infarct dementia. Acta Psychiatr Scand 78, p.730-73635. Davidson, J.R. (1998) Pharmacotherapy of social anxiety disorder. J Clin Psychiatry 59, p.47-5336. Deckert, J. et al. (1999) Excess of high activity monoamine oxidase A gene promoter alleles in female patients with panicdisorder. Hum Mol Genet 8, p.621-62437. Denney, R.M., Koch, H. & Craig, I.W. (1999) Association between monoamine oxidase A activity in human male skinfibroblasts and genotype of the MAOA promoter-associated variable number tandem repeat. Hum Genet 105, p.542-55138. Denoyelle, F. et al. (1998) Connexin 26 gene linked to a dominant deafness. Nature 393, p.319-32039. Devor, E.J., Abell, C.W., Hoffman, P.L., Tabakoff, B. & Cloninger, C.R. (1994) Platelet MAO activity in type I and typeII alcoholism. Ann N Y Acad Sci 708, p.119-12840. Devor, E.J., Cloninger, C.R., Hoffman, P.L. & Tabakoff, B. (1993) Association of monoamine oxidase (MAO) activitywith alcoholism and alcoholic subtypes. Am J Med Genet 48, p.209-21341. Dunham, I. et al. (1999) The DNA sequence of human chromosome 22. Nature 402, p.489-49542. Ekblom, J. et al. (1998) Transcription factor binding to the core promoter of the human monoamine oxidase B gene in thecerebral cortex and in blood cells. Neurosci Lett 258, p.101-10443. Ewing, B. & Green, P. (2000) Analysis of expressed sequence tags indicates 35,000 human genes. Nat Genet 25, p.232-23444. Fields, S. & Sternglanz, R. (1994) The two-hybrid system: an assay for protein-protein interactions. Trends Genet 10,p.286-29245. Fischel-Ghodsian, N. (1999) Mitochondrial deafness mutations reviewed. Hum Mutat 13, p.261-27046. Fishman, P.M., Suarez, B., Hodge, S.E. & Reich, T. (1978) A robust method for the detection of linkage in familialdisease. Am J Hum Genet 30, p.308-32147. Fowler, C.J., von Knorring, L. & Oreland, L. (1980) Platelet monoamine oxidase activity in sensation seekers. PsychiatryRes 3, p.273-27948. Fowler, J.S. et al. (1996) Inhibition of monoamine oxidase B in the brains of smokers. Nature 379, p.733-73649. Fowler, J.S. et al. (1996) Brain monoamine oxidase A inhibition in cigarette smokers. Proc Natl Acad Sci U S A 93,p.14065-1406950. Fox, K.R. (1997) DNase I footprinting. Methods Mol Biol 90, p.1-2251. Furlong, R.A. et al. (1999) Analysis of the monoamine oxidase A (MAOA) gene in bipolar affective disorder byassociation studies, meta-analyses, and sequencing of the promoter. Am J Med Genet 88, p.398-40652. Garpenstrand, H., Ekblom, J., Forslund, K., Rylander, G. & Oreland, L. (2000) Platelet monoamine oxidase activity isrelated to MAOB intron 13 genotype. J Neural Transm 107, p.523-53053. Gasparini, P., Estivill, X. & Fortina, P. (1999) Vestibular and hearing loss in genetic and metabolic disorders. Curr OpinNeurol 12, p.35-3954. Girmen, A.S. et al. (1992) Relationship between platelet monoamine oxidase B activity and alleles at the MAOB locus. JNeurochem 59, p.2063-206655. Grimsby, J., Chen, K., Wang, L.J., Lan, N.C. & Shih, J.C. (1991) Human monoamine oxidase A and B genes exhibitidentical exon-intron organization. Proc Natl Acad Sci U S A 88, p.3637-364156. Grimsby, J., Lan, N.C., Neve, R., Chen, K. & Shih, J.C. (1990) Tissue distribution of human monoamine oxidase A andB mRNA. J Neurochem 55, p.1166-116957. Grimsby, J. et al. (1997) Increased stress response and beta-phenylethylamine in MAOB-deficient mice. Nat Genet 17,p.206-21058. Hamilton, S.P. et al. (2000) No genetic linkage or association between a functional promoter polymorphism in themonoamine oxidase-A gene and panic disorder. Mol Psychiatry 5, p.465-46659. Harpending, H.C. et al. (1998) Genetic traces of ancient demography. Proc Natl Acad Sci U S A 95, p.1961-196760. Hinds, H.L., Hendriks, R.W., Craig, I.W. & Chen, Z.Y. (1992) Characterization of a highly polymorphic region near thefirst exon of the human MAOA gene containing a GT dinucleotide and a novel VNTR motif. Genomics 13, p.896-89761. Ho, S.L., Kapadi, A.L., Ramsden, D.B. & Williams, A.C. (1995) An allelic association study of monoamine oxidase B inParkinson's disease. Ann Neurol 37, p.403-40562. Hotamisligil, G.S. & Breakefield, X.O. (1991) Human monoamine oxidase A gene determines levels of enzyme activity.Am J Hum Genet 49, p.383-39263. Hotamisligil, G.S. et al. (1994) Hereditary variations in monoamine oxidase as a risk factor for Parkinson's disease. MovDisord 9, p.305-31064. Hsu, Y.P. et al. (1996) Association of monoamine oxidase A alleles with alcoholism among male Chinese in Taiwan. AmJ Psychiatry 153, p.1209-1211

65. Hudspeth, A.J. (1997) How hearing happens. Neuron 19, p.947-95066. Hughes, D.C., Legan, P.K., Steel, K.P. & Richardson, G.P. (1998) Mapping of the alpha-tectorin gene (TECTA) tomouse chromosome 9 and human chromosome 11: a candidate for human autosomal dominant nonsyndromic deafness.Genomics 48, p.46-5167. Jareborg, N. & Durbin, R. (2000) Alfresco--a workbench for comparative genomic sequence analysis. Genome Res 10,p.1148-115768. Jeffreys, A.J., Royle, N.J., Wilson, V. & Wong, Z. (1988) Spontaneous mutation rates to new length alleles at tandem-repetitive hypervariable loci in human DNA. Nature 332, p.278-28169. Jonsson, E.G. et al. (2000) A promoter polymorphism in the monoamine oxidase A gene and its relationships tomonoamine metabolite concentrations in CSF of healthy volunteers. J Psychiatr Res 34, p.239-24470. Jorde, L.B., Bamshad, M. & Rogers, A.R. (1998) Using mitochondrial and nuclear DNA markers to reconstruct humanevolution. Bioessays 20, p.126-13671. Kaessmann, H., Heissig, F., von Haeseler, A. & Paabo, S. (1999) DNA sequence variation in a non-coding region of lowrecombination on the human X chromosome. Nat Genet 22, p.78-8172. Kalatzis, V. & Petit, C. (1998) The fundamental and medical impacts of recent progress in research on hereditary hearingloss. Hum Mol Genet 7, p.1589-159773. Kawada, Y., Hattori, M., Dai, X.Y. & Nanko, S. (1995) Possible association between monoamine oxidase A gene andbipolar affective disorder [letter; comment]. Am J Hum Genet 56, p.335-33674. Kelsell, D.P. et al. (1997) Connexin 26 mutations in hereditary non-syndromic sensorineural deafness. Nature 387, p.80-8375. Kent, J. Draft Human Genome Browser. (http://genome.ucsc.edu/goldenPath/hgTracks.html).76. Kochersperger, L.M., Parker, E.L., Siciliano, M., Darlington, G.J. & Denney, R.M. (1986) Assignment of genes forhuman monoamine oxidases A and B to the X chromosome. J Neurosci Res 16, p.601-61677. Konradi, C., Ozelius, L. & Breakefield, X.O. (1992) Highly polymorphic (GT)n repeat sequence in intron II of the humanMAOB gene. Genomics 12, p.176-17778. Kruglyak, L. (1999) Prospects for whole-genome linkage disequilibrium mapping of common disease genes. Nat Genet22, p.139-14479. Kruglyak, L., Daly, M.J., Reeve-Daly, M.P. & Lander, E.S. (1996) Parametric and nonparametric linkage analysis: aunified multipoint approach. Am J Hum Genet 58, p.1347-136380. Kruglyak, L. & Lander, E.S. (1995) Complete multipoint sib-pair analysis of qualitative and quantitative traits. Am JHum Genet 57, p.439-45481. Kubisch, C. et al. (1999) KCNQ4, a novel potassium channel expressed in sensory outer hair cells, is mutated indominant deafness. Cell 96, p.437-44682. Kurth, J.H., Kurth, M.C., Poduslo, S.E. & Schwankhaus, J.D. (1993) Association of a monoamine oxidase B allele withParkinson's disease. Ann Neurol 33, p.368-37283. Landegren, U., Nilsson, M. & Kwok, P.Y. (1998) Reading bits of genetic information: methods for single-nucleotidepolymorphism analysis. Genome Res 8, p.769-77684. Lander, E. & Kruglyak, L. (1995) Genetic dissection of complex traits: guidelines for interpreting and reporting linkageresults. Nat Genet 11, p.241-24785. Lange, K., Weeks, D. & Boehnke, M. (1988) Programs for Pedigree Analysis: MENDEL, FISHER, and dGENE. GenetEpidemiol 5, p.471-47286. Lathrop, G.M., Lalouel, J.M., Julier, C. & Ott, J. (1984) Strategies for multilocus linkage analysis in humans. Proc NatlAcad Sci U S A 81, p.3443-344687. Legan, P.K., Rau, A., Keen, J.N. & Richardson, G.P. (1997) The mouse tectorins. Modular matrix proteins of the innerear homologous to components of the sperm-egg adhesion system. J Biol Chem 272, p.8791-880188. Legaz, M.E., Schmer, G., Counts, R.B. & Davie, E.W. (1973) Isolation and characterization of human Factor VIII(antihemophilic factor). J Biol Chem 248, p.3946-395589. Lench, N., Houseman, M., Newton, V., Van Camp, G. & Mueller, R. (1998) Connexin-26 mutations in sporadic non-syndromal sensorineural deafness. Lancet 351, p.41590. Lenders, J.W. et al. (1996) Specific genetic deficiencies of the A and B isoenzymes of monoamine oxidase arecharacterized by distinct neurochemical and clinical phenotypes. J Clin Invest 97, p.1010-101991. Levy, E.R. et al. (1989) Localization of human monoamine oxidase-A gene to Xp11.23-11.4 by in situ hybridization:implications for Norrie disease. Genomics 5, p.368-37092. Liang, F. et al. (2000) Gene index analysis of the human genome estimates approximately 120, 000 genes. Nat Genet 25,p.239-24093. Liang, P. & Pardee, A.B. (1998) Differential display. A general protocol. Mol Biotechnol 10, p.261-26794. Lim, L.C., Powell, J.F., Murray, R. & Gill, M. (1994) Monoamine oxidase A gene and bipolar affective disorder. Am JHum Genet 54, p.1122-112495. Lin, S. et al. (2000) Association analysis between mood disorder and monoamine oxidase gene. Am J Med Genet 96,p.12-14

96. Liu, X.Z. et al. (1997a) Mutations in the myosin VIIA gene cause non-syndromic recessive deafness. Nat Genet 16,p.188-19097. Liu, X.Z. et al. (1997b) Autosomal dominant non-syndromic deafness caused by a mutation in the myosin VIIA gene.Nat Genet 17, p.268-26998. Liu, X.Z. et al. (2000) Mutations in connexin31 underlie recessive as well as dominant non- syndromic hearing loss. HumMol Genet 9, p.63-6799. Livingston, M.G. & Livingston, H.M. (1996) Monoamine oxidase inhibitors. An update on drug interactions. Drug Saf14, p.219-227100. Lobe, C.G. & Nagy, A. (1998) Conditional genome alteration in mice. Bioessays 20, p.200-208101. Long, A.D. & Langley, C.H. (1999) The power of association studies to detect the contribution of candidate genetic locito variation in complex traits. Genome Res 9, p.720-731102. Malgrange, B. et al. (1999) Growth factor therapy to the damaged inner ear: clinical prospects. Int J PediatrOtorhinolaryngol 49 Suppl 1, p.S19-25103. Manuck, S.B., Flory, J.D., Ferrell, R.E., Mann, J.J. & Muldoon, M.F. (2000) A regulatory polymorphism of themonoamine oxidase-A gene may be associated with variability in aggression, impulsivity, and central nervous systemserotonergic responsivity. Psychiatry Res 95, p.9-23104. Marazita, M.L. et al. (1993) Genetic epidemiological studies of early-onset deafness in the U.S. school-age population.Am J Med Genet 46, p.486-491105. March, R.E. (1999) Gene mapping by linkage and association analysis. Mol Biotechnol 13, p.113-122106. Marti, T., Rosselet, S.J., Titani, K. & Walsh, K.A. (1987) Identification of disulfide-bridged substructures within humanvon Willebrand factor. Biochemistry 26, p.8099-8109107. Mellick, G.D. et al. (1999) Variations in the monoamine oxidase B (MAOB) gene are associated with Parkinson'sdisease. Mov Disord 14, p.219-224108. Mellick, G.D. et al. (2000) The monoamine oxidase B gene GT repeat polymorphism and Parkinson's disease in aChinese population. J Neurol 247, p.52-55109. Meltzer, H.Y. & Arora, R.C. (1986) Platelet markers of suicidality. Ann N Y Acad Sci 487, p.271-280110. Morimoto, Y. et al. (1995) Association analysis of a polymorphism of the monoamine oxidase B gene with Parkinson'sdisease in a Japanese population. Am J Med Genet 60, p.570-572111. Morton, N.E. (1991) Genetic epidemiology of hearing impairment. Ann N Y Acad Sci 630, p.16-31112. Mullikin, J.C. et al. (2000) An SNP map of human chromosome 22. Nature 407, p.516-520113. Muramatsu, T. et al. (1997) Monoamine oxidase genes polymorphisms and mood disorder. Am J Med Genet 74, p.494-496114. Murphy, D.L., Belmaker, R. & Wyatt, R.J. (1975) Monoamine oxidase in schizophrenia and other behavioral disorders.in In: Matthysse SW, Kety SS, ed. Catecholamines and schizophrenia. Oxford, Pergamon Press, p. 221-247115. Murphy, D.L. et al. (1976) Platelet and plasma amine oxidase activity in 680 normals: sex and age differences andstability over time. Biochemical Medicine 16, p.254-265116. Mustapha, M. et al. (1999) An alpha-tectorin gene defect causes a newly identified autosomal recessive form ofsensorineural pre-lingual non-syndromic deafness, DFNB21. Hum Mol Genet 8, p.409-412117. Nachman, M.W. & Crowell, S.L. (2000) Estimate of the mutation rate per nucleotide in humans. Genetics 156, p.297-304118. Nakatome, M., Tun, Z., Shimada, S. & Honda, K. (1998) Detection and analysis of four polymorphic markers at thehuman monoamine oxidase (MAO) gene in Japanese controls and patients with Parkinson's disease. Biochem Biophys ResCommun 247, p.452-456119. Nanko, S., Ueki, A. & Hattori, M. (1996) No association between Parkinson's disease and monoamine oxidase A and Bgene polymorphisms. Neurosci Lett 204, p.125-127120. Nickerson, D.A. et al. (1998) DNA sequence diversity in a 9.7-kb region of the human lipoprotein lipase gene. Nat Genet19, p.233-240121. Nies, A., Robinson, D.S., Lamborn, K.R. & Lampert, R.P. (1973) Genetic control of platelet and plasma monoamineoxidase activity. Arch Gen Psychiatry 28, p.834-838122. Nobili, R., Mammano, F. & Ashmore, J. (1998) How well do we understand the cochlea? Trends Neurosci 21, p.159-167123. Nothen, M.M. et al. (1995) Association analysis of the monoamine oxidase A gene in bipolar affective disorder by usingfamily-based internal controls. Am J Hum Genet 57, p.975-978124. O'Connell, J.R. & Weeks, D.E. (1995) The VITESSE algorithm for rapid exact multilocus linkage analysis via genotypeset-recoding and fuzzy inheritance. Nat Genet 11, p.402-408125. Oreland, L. & Hallman, J. (1995) The correlation between platelet MAO activity and personality: short review offindings and a discussion on possible mechanisms. Prog Brain Res 106, p.77-84126. Ott, J. & Hoh, J. (2000) Statistical approaches to gene mapping. Am J Hum Genet 67, p.289-294127. Ozelius, L., Gusella, J.F. & Breakefield, X.O. (1989) MspI RFLP for human MAOA gene. Nucleic Acids Res 17, p.10516128. Paabo, S. (1999) Human evolution. Trends Cell Biol 9, p.M13-16129. Parsian, A. et al. (1995) Monoamine oxidases and alcoholism. I. Studies in unrelated alcoholics and normal controls. AmJ Med Genet 60, p.409-416

130. Parsian, A. & Todd, R.D. (1997) Genetic association between monoamine oxidase and manic-depressive illness:comparison of relative risk and haplotype relative risk data. Am J Med Genet 74, p.475-479131. Peck, J.E. (1994) Development of hearing. Part I: Phylogeny. J Am Acad Audiol 5, p.291-299132. Pedersen, N.L., Oreland, L., Reynolds, C. & McClearn, G.E. (1993) Importance of genetic effects for monoamineoxidase activity in thrombocytes in twins reared apart and twins reared together. Psychiatry Res 46, p.239-251133. Petit, C. (1996) Genes responsible for human hereditary deafness: symphony of a thousand. Nat Genet 14, p.385-391134. Pintar, J.E. et al. (1981) Gene for monoamine oxidase type A assigned to the human X chromosome. J Neurosci 1, p.166-175135. Plante-Bordeneuve, V. et al. (1997) Evaluation of four candidate genes encoding proteins of the dopamine pathway infamilial and sporadic Parkinson's disease: evidence for association of a DRD2 allele. Neurology 48, p.1589-1593136. Preisig, M. et al. (2000) Association between bipolar disorder and monoamine oxidase A gene polymorphisms: results ofa multicenter study. Am J Psychiatry 157, p.948-955137. Rabionet, R., Gasparini, P. & Estivill, X. (2000) Molecular genetics of hearing impairment due to mutations in gapjunction genes encoding beta connexins. Hum Mutat 16, p.190-202138. Rau, A., Legan, P.K. & Richardson, G.P. (1999) Tectorin mRNA expression is spatially and temporally restricted duringmouse inner ear development. J Comp Neurol 405, p.271-280139. Reddy, P.S. & Housman, D.E. (1997) The complex pathology of trinucleotide repeats. Curr Opin Cell Biol 9, p.364-372140. Riazuddin S. et al. A dominant modifier, DFNM1, protects seven individuals from DFNB26 hearing impairment. in 50thannual meeting of the American Society of Human Genetics (, Pennsylvania, Philadelphia, USA, 2000).141. Rice, J., McGuffin, P., Goldin, L.R., Shaskan, E.G. & Gershon, E.S. (1984) Platelet monoamine oxidase (MAO) activity:evidence for a single major locus. Am J Hum Genet 36, p.36-43142. Richard, G. et al. (1998) Mutations in the human connexin gene GJB3 cause erythrokeratodermia variabilis. Nat Genet20, p.366-369143. Richardson, G.P., Russell, I.J., Duance, V.C. & Bailey, A.J. (1987) Polypeptide composition of the mammalian tectorialmembrane. Hear Res 25, p.45-60144. Risch, N. & Zhang, H. (1995) Extreme discordant sib pairs for mapping quantitative trait loci in humans. Science 268,p.1584-1589145. Rubinsztein, D.C. et al. (1996) Genetic association between monoamine oxidase A microsatellite and RFLP alleles andbipolar affective disorder: analysis and meta-analysis. Hum Mol Genet 5, p.779-782146. Sabol, S.Z., Hu, S. & Hamer, D. (1998) A functional polymorphism in the monoamine oxidase A gene promoter. HumGenet 103, p.273-279147. Sandler, M., Reveley, M.A. & Glover, V. (1981) Human platelet monoamine oxidase activity in health and disease: areview. J Clin Pathol 34, p.292-302148. Sataloff, R.T. & Sataloff, J. (1997) Audiologic testing: an overview for occupational physicians. Occup Med 12, p.433-447149. Saura, J. et al. (1996) Molecular neuroanatomy of human monoamine oxidases A and B revealed by quantitative enzymeradioautography and in situ hybridization histochemistry. Neuroscience 70, p.755-774150. Saura, J., Kettler, R., Da Prada, M. & Richards, J.G. (1992) Quantitative enzyme radioautography with 3H-Ro 41-1049and 3H-Ro 19- 6327 in vitro: localization and abundance of MAO-A and MAO-B in rat CNS, peripheral organs, and humanbrain. J Neurosci 12, p.1977-1999151. Schena, M. et al. (1998) Microarrays: biotechnology's discovery platform for functional genomics. Trends Biotechnol 16,p.301-306152. Schooler, C., Zahn, T.P., Murphy, D.L. & Buchsbaum, M.S. (1978) Psychological correlates of monoamine oxidaseactivity in normals. J Nerv Ment Dis 166, p.177-186153. Schork, N.J., Nath, S.K., Fallin, D. & Chakravarti, A. (2000) Linkage disequilibrium analysis of biallelic DNA markers,human quantitative trait loci, and threshold-defined case and control subjects. Am J Hum Genet 67, p.1208-1218154. Schwab, S.G. et al. (1995) Evaluation of a susceptibility gene for schizophrenia on chromosome 6p by multipointaffected sib-pair linkage analysis. Nat Genet 11, p.325-327155. Schwartz, S. et al. (2000) PipMaker--a web server for aligning two genomic DNA sequences. Genome Res 10, p.577-586156. Scott, V., Clark, A.R. & Docherty, K. (1994) The gel retardation assay. Methods Mol Biol 31, p.339-347157. Shih, J.C., Chen, K. & Ridd, M.J. (1999) Monoamine oxidase: from genes to behavior. Annu Rev Neurosci 22, p.197-217158. Shih, J.C., Grimsby, J. & Chen, K. (1990) The expression of human MAO-A and B genes. J Neural Transm Suppl 32,p.41-47159. Shih, J.C., Grimsby, J., Chen, K. & Zhu, Q.S. (1993) Structure and promoter organization of the human monoamineoxidase A and B genes. J Psychiatry Neurosci 18, p.25-32160. Sill, A.M. et al. (1994) Genetic epidemiologic study of hearing loss in an adult population. Am J Med Genet 54, p.149-153161. Sobell, J.L., Lind, T.J., Hebrink, D.D., Heston, L.L. & Sommer, S.S. (1997) Screening the monoamine oxidase B gene in100 male patients with schizophrenia: a cluster of polymorphisms in African-Americans but lack of functionally significantsequence changes. Am J Med Genet 74, p.44-49

162. Spielman, R.S., McGinnis, R.E. & Ewens, W.J. (1993) Transmission test for linkage disequilibrium: the insulin generegion and insulin-dependent diabetes mellitus (IDDM). Am J Hum Genet 52, p.506-516163. Stalenheim, E.G., von Knorring, L. & Oreland, L. (1997) Platelet monoamine oxidase activity as a biological marker in aSwedish forensic psychiatric population. Psychiatry Res 69, p.79-87164. Stern, G. (1998) Neuroprotection by selegiline and other MAO inhibitors. J Neural Transm Suppl 52, p.99-107165. Steventon, G.B. et al. (1989) Platelet monoamine oxidase-B activity in Parkinson's disease. J Neural Transm Park DisDement Sect 1, p.255-261166. Suarez, B.K., Hampe, C.L., Parsian, A. & Cloninger, C.R. (1995) Monoamine oxidases and alcoholism. II. Studies inalcoholic families. Am J Med Genet 60, p.417-423167. Sullivan, J.L. et al. (1990) Platelet MAO in subtypes of alcoholism. Biol Psychiatry 27, p.911-922168. Sundstrom, R.A., Van Laer, L., Van Camp, G. & Smith, R.J. (1999) Autosomal recessive nonsyndromic hearing loss. AmJ Med Genet 89, p.123-129169. Takahashi, S., Tani, N. & Yamane, H. (1976) Monoamine oxidase activity in blood platelets in alcoholism. FoliaPsychiatr Neurol Jpn 30, p.455-462170. Talebizadeh, Z., Kelley, P.M., Askew, J.W., Beisel, K.W. & Smith, S.D. (1999) Novel mutation in the KCNQ4 gene in alarge kindred with dominant progressive hearing loss. Hum Mutat 14, p.493-501171. Templeton, A.R. (1997) Out of Africa? What do genes tell us? Curr Opin Genet Dev 7, p.841-847172. Terwilliger, J.D. & Goring, H.H. (2000) Gene mapping in the 20th and 21st centuries: statistical methods, data analysis,and experimental design. Hum Biol 72, p.63-132173. Terwilliger, J.D. & Ott, J. (1992) A haplotype-based 'haplotype relative risk' approach to detecting allelic associations.Hum Hered 42, p.337-346174. Terwilliger, J.D. & Weiss, K.M. (1998) Linkage disequilibrium mapping of complex disease: fantasy or reality? CurrOpin Biotechnol 9, p.578-594175. Thomas, T. (2000) Monoamine oxidase-B inhibitors in the treatment of Alzheimer's disease. Neurobiol Aging 21, p.343-348176. Thorpe, L.W., Westlund, K.N., Kochersperger, L.M., Abell, C.W. & Denney, R.M. (1987) Immunocytochemicallocalization of monoamine oxidases A and B in human peripheral tissues and brain. J Histochem Cytochem 35, p.23-32177. Tivol, E.A., Shalish, C., Schuback, D.E., Hsu, Y.P. & Breakefield, X.O. (1996) Mutational analysis of the human MAOAgene. Am J Med Genet 67, p.92-97178. Tseng, C.J. & Lalwani, A.K. (2000) Cracking the auditory genetic code: part II. Syndromic hereditary hearingimpairment. Am J Otol 21, p.437-451179. Turecki, G. et al. (1999) MAOA: association and linkage studies with lithium responsive bipolar disorder. PsychiatrGenet 9, p.13-16180. Uberbacher, E.C. & Mural, R.J. (1991) Locating protein-coding regions in human DNA sequences by a multiple sensor-neural network approach. Proc Natl Acad Sci U S A 88, p.11261-11265181. Van Camp, G. & Smith, R.J.H. Hereditary Hearing Loss Homepage. (http://dnalab-www.uia.ac.be/dnalab/hhh/).182. Van Camp, G., Willems, P.J. & Smith, R.J. (1997a) Nonsyndromic hearing impairment: unparalleled heterogeneity. Am JHum Genet 60, p.758-764183. Van Hauwe, P. et al. (1999) Deafness linked to DFNA2: one locus but how many genes? [letter; comment]. Nat Genet21, p.263184. Van Hauwe, P. et al. (2000) Mutations in the KCNQ4 K+ channel gene, responsible for autosomal dominant hearingloss, cluster in the channel pore region. Am J Med Genet 93, p.184-187185. Van Laer, L., McGuirt, W.T., Yang, T., Smith, R.J. & Van Camp, G. (1999) Autosomal dominant nonsyndromic hearingimpairment. Am J Med Genet 89, p.167-174186. Vanyukov, M.M., Moss, H.B., Yu, L.M. & Deka, R. (1995) A dinucleotide repeat polymorphism at the gene formonoamine oxidase A and measures of aggressiveness. Psychiatry Res 59, p.35-41187. Vanyukov, M.M., Moss, H.B., Yu, L.M., Tarter, R.E. & Deka, R. (1995) Preliminary evidence for an association of adinucleotide repeat polymorphism at the MAOA gene with early onset alcoholism/substance abuse. Am J Med Genet 60, p.122-126188. Verhoeven, K. et al. (1998) Mutations in the human alpha-tectorin gene cause autosomal dominant non- syndromichearing impairment. Nat Genet 19, p.60-62189. Verkes, R.J., Pijl, H., Meinders, A.E. & Van Kempen, G.M. (1996) Borderline personality, impulsiveness, and plateletmonoamine measures in bulimia nervosa and recurrent suicidal behavior. Biol Psychiatry 40, p.173-180190. Verkes, R.J. et al. (1998) Platelet serotonin, monoamine oxidase activity, and [3H]paroxetine binding related toimpulsive suicide attempts and borderline personality disorder. Biol Psychiatry 43, p.740-746191. Voorberg, J., Fontijn, R., van Mourik, J.A. & Pannekoek, H. (1990) Domains involved in multimer assembly of vonwillebrand factor (vWF): multimerization is independent of dimerization. Embo J 9, p.797-803192. Waldmeier, P.C. (1987) Amine oxidases and their endogenous substrates (with special reference to monoamine oxidaseand the brain). J Neural Transm Suppl 23, p.55-72193. Wang, D.G. et al. (1998) Large-scale identification, mapping, and genotyping of single- nucleotide polymorphisms in thehuman genome. Science 280, p.1077-1082

194. Ward, P.B., Catts, S.V., Norman, T.R., Burrows, G.D. & McConaghy, N. (1987) Low platelet monoamine oxidase andsensation seeking in males: an established relationship? Acta Psychiatr Scand 75, p.86-90195. Weeks, D.E., Sobel, E., O'Connell, J.R. & Lange, K. (1995) Computer programs for multilocus haplotyping of generalpedigrees. Am J Hum Genet 56, p.1506-1507196. Weil, D. et al. (1995) Defective myosin VIIA gene responsible for Usher syndrome type 1B. Nature 374, p.60-61197. Weiss, K.M. & Terwilliger, J.D. (2000) How many diseases does it take to map a gene with SNPs? Nat Genet 26, p.151-157198. Westlund, K.N., Denney, R.M., Rose, R.M. & Abell, C.W. (1988) Localization of distinct monoamine oxidase A andmonoamine oxidase B cell populations in human brainstem. Neuroscience 25, p.439-456199. Weyler, W., Hsu, Y.P. & Breakefield, X.O. (1990) Biochemistry and genetics of monoamine oxidase. Pharmacol Ther47, p.391-417200. White, K. et al. (1983) Platelet MAO activity in depression. Compr Psychiatry 24, p.453-458201. Willems, P.J. (2000) Genetic causes of hearing loss. N Engl J Med 342, p.1101-1109202. Winblad, B., Gottfries, C.G., Oreland, L. & Wiberg, A. (1979) Monoamine oxidase in platelets and brains of non-psychiatric and non-neurological geriatric patients. Med Biol 57, p.129-132203. Woodruff, T.K. (1998) Cellular localization of mRNA and protein: in situ hybridization histochemistry and in situ ligandbinding. Methods Cell Biol 57, p.333-351204. Wyatt, R.J., Potkin, S.G. & Murphy, D.L. (1979) Platelet monoamine oxidase activity in schizophrenia: a review of thedata. Am J Psychiatry 136, p.377-385205. Xia, J.H. et al. (1998) Mutations in the gene encoding gap junction protein beta-3 associated with autosomal dominanthearing impairment. Nat Genet 20, p.370-373206. Young, W.F., Jr., Laws, E.R., Jr., Sharbrough, F.W. & Weinshilboum, R.M. (1986) Human monoamine oxidase. Lack ofbrain and platelet correlation. Arch Gen Psychiatry 43, p.604-609207. Zhao, Z. et al. (2000) Worldwide DNA sequence variation in a 10-kilobase noncoding region on human chromosome 22.Proc Natl Acad Sci U S A 97, p.11354-11358