There is now a very comprehensive chicken linkage map containing well over 800 loci. However, in order to use this map to identify the genes responsible for traits it is necessary to progress from linkage information to physical clones of the relevant region. The identification of chicken yeast artificial chromosome (YAC) clones and the gene of interest is the first step towards physical cloning of a candidate gene but at present the efficiency of this process would be greatly improved if the density of genetic markers in the targeted region could be increased. The initial approach to this problem was to place as many known genes as possible on the map, by using expressed sequence tags (ESTs) as markers. This placed seven novel chicken genes on the linkage map. However, this method was inefficient, so a method of saturating the genome with markers was applied. Random amplified polymorphic DNA (RAPD) analysis was investigated as a means of increasing marker density over the whole genome, but was also inefficient at producing large numbers of markers. Then the potential of representational difference analysis (RDA) was investigated to target markers to specific regions of the chicken genome. As an initial test of the method the genomes of two inbred lines were compared. This provided a relatively undemanding test of the technique and generated large numbers of polymorphic clones. As a more stringent test of the method, and to provide marker loci in a region of particular interest, a second comparison was made targeting chicken chromosome 16. Chromosome 16 contains the major histocompatibility complex (MHC), nucleolar organiser region (NOR) and Rfp-Y complex. This succeeded in targeting 10 out of 18 RDA clones to chromosome 16. It appeared possible to target RDA towards the direct identification of a trait gene or genes and as a model for this, RDA was applied to the identification of genes affecting resistance to Marek's disease. This generated four RDA loci which were located on chromosome 1 three of which showed association with both mortality and Marek's disease virus quantitative PCR product values. These RDA loci appear to identify a novel region of the chicken genome containing a trait gene conferring resistance to Marek's disease. Thus RDA appears to provide a viable solution to the generation of targeted markers, which will aid YAC identification in specific areas.

I would like to thank Professor John Bourne and the Institute for Animal Health, Compton for allowing me to undertake the research for this degree.

I would also like to thank my supervisor Dr Nat Bumstead for all his help and support, especially with the proof reading of this thesis and his belief that PCR is not controlled by the tooth fairies! My thanks also go to Dr Ayo Toye, for generating the chicken YAC library, letting me use it and giving positive encouragement in the final stages of writing up. I would also like to thank all those within the Avian Genetics lab. especially Julie Sillibourne and Nigel Salmon without whose technical support everything would have been a great deal harder.

I would also like to acknowledge the help of my external supervisor Dr Terry Burke and all those who have previously provided markers for the Compton chicken linkage map, without which a great deal of this work would have been impossible.

I am also grateful to all other staff at the Institute who have made my working time more pleasant: particularly the librarians Chris Gibbons and Diane Collins, the DID Jim White-Cooper, Bernard Clark and formerly Dave Hawkins, the Computing Dept. Roger Harrison, Fiona Wycherley, the canteen staff and Alison Thomson and everyone in Media and Wash-up. I would specifically like to thank Christine Jones for her understanding of my need for isotope and Dr Pete Kaiser for his reference checking, proof reading and bullying.

I would like to say a special thank you to Janene, Thomas and Simon Bumstead for tolerating my somewhat unexpected intrusions into their lives and Marie Quick for encouraging me to start this PhD. Another very special thank you must go to my parents Drs Chris and Bill Wain without whom I would never have got this far.

Finally I am also extremely appreciative of my husband Graham for putting up with my computer problems, long hours away from home and continuous Dial-a-pizzas for dinner.

This thesis is wholly the result of my own work. No part of it has been submitted to any other board for another qualification. The views expressed are those of the author and not of the University.

The HTML contents of this thesis vary from the original publication, due to size contraints.  For any further information please email Hester.

Title page
Abstract
Acknowledgements
Declaration
Contents
List of Tables (removed)
List of Figures (removed)
Abbreviations
Genes and Loci
Chapter One
Introduction
1.1 The Chicken
1.2 Genomic structure and organisation of the chicken
1.2.1 Repetitive Sequences
1.3 Genomic Mapping
1.3.1 Genetic Mapping
1.3.1.1 Restriction Fragment Length Polymorphism
1.3.1.2 Single-Stranded Conformational Polymorphism
1.3.1.3 VNTR-Hypervariable Minisatellites
1.3.1.4 Microsatellites
1.3.1.5 Random Amplified Polymorphic DNA markers
1.3.1.6 Amplified Fragment Length Polymorphism
1.3.1.7 Targeted Mapping
1.3.1.8 Expressed Sequence Tags
1.3.1.9 Subtractive hybridisation
1.3.1.10 RFLP subtraction
1.3.1.11 Differential Display-Reverse Transcription (DD-RT)
1.3.1.12 Genomic Mismatch Scanning
1.3.1.13 Representational Difference Analysis
1.3.2 Physical Mapping
1.3.2.1 Pulse Field Gel Electrophoresis
1.3.2.2 Radiation Hybrids
1.3.2.3 Flow Karyotyping
1.3.2.4 Chromosome microdissection
1.3.2.5 Yeast Artificial Chromosomes (YACs)
1.3.2.6 Sequence Tagged Sites (STS)
1.3.2.7 Fluorescent In-Situ Hybridisation (FISH)
1.4 Traits of interest in the chicken
1.4.1 Production Traits
1.4.2 Disease Traits
1.4.2.1 The major histocompatibility complex (MHC)
1.4.2.2 Salmonellosis
1.4.2.3 Infectious bursal disease virus
1.4.2.4 Marek's disease virus
1.4.2.5 Newcastle disease
1.4.2.6 Infectious bronchitis virus
1.4.2.7 Avian leukosis and sarcoma virus
1.4.2.8 Coccidiosis
1.4.2.9 Fowl cholera
1.5 Objectives
1.6 Approaches
1.6.1 Expressed Sequence Tags (EST)
1.6.2 Random Amplified Polymorphic DNA markers (RAPD)
1.6.3 Representational Difference Analysis
Chapter Two
Materials and Methods
2.1 Materials
2.2 Bacterial strains and plasmids
2.2.1 cDNA Library IAHchB1
2.2.2 Transformation of IAHchB1 by electroporation of
Escherichia coli MC1061/P3
2.2.3 Cloning RDA products into pGEM-T vector
2.3 Chicken Strains and Crosses
2.4 DNA Preparation
2.4.1 Genomic DNA Extraction
2.4.2 Plasmid DNA Preparation-ABI Protocol
2.4.3 Plasmid DNA Preparation-Hybaid maxiprep Protocol
2.5 Restriction Endonuclease Digestion
2.6 Electrophoresis
2.6.1 Agarose Gel Electrophoresis
2.6.2 ABI Genescan electrophoresis
2.6.2.1 Creation of a Matrix File
2.6.2.2 RAPD-PCR Electrophoresis
2.6.3 Sequence Gel Electrophoresis
2.7 PCR Amplification
2.7.1 Plasmid insert PCR Amplification
2.7.2 Colony PCR of RDA products
2.8 Gel purification of PCR products
2.9 Automated Sequencing
2.9.1 Sequence Analysis
2.10 Southern Blotting
2.10.1 Probe preparation and hybridisation
2.10.2 Nick translation system (Gibco BRL)
2.10.3 Prime-It RmT random primer labelling kit (Stratagene)
2.11 Mapping Techniques
2.11.1 Restriction Fragment Length Polymorphism Analysis
2.11.1.1 Mapping Segregation Analyses
2.11.2 RAPD Analysis
2.11.2.1 RAPD PCR Reaction
2.11.2.2 Genescan Analysis
2.12 Representational Difference Analysis
2.12.1 RDA comparison of Line N and line 15I
2.12.1.1 Screening and characterisation of RDA clones
2.12.2 RDA targeted to chromosome 16
2.12.2.1 BamHI Representation
2.12.2.2 TaqI Representation
2.12.2.3 NheI Representation
2.12.3 RDA targeted to Marek's Disease Resistance Genes
Chapter Three
Mapping Expressed Sequence Tags
3.1 Introduction
3.2 Results
3.2.1 Mapping ten novel chicken genes
3.2.2 RFLP Mapping
3.3 Discussion
3.3.1 Analysis of clones from the bursal cDNA library
3.3.2 RFLP Analysis and Mapping of clones
Chapter Four
Random Amplified Polymorphic DNA Analysis
4.1 Introduction
4.1.1 Choice of RAPD Primer
4.2 Results
4.2.1 Creation of the Matrix File
4.2.2 RAPD Analyses using fluorescent primers
4.3 Discussion
Chapter Five
Representational Difference Analysis comparison of line N
and line 15I
5.1 Introduction
5.2 Results
5.2.1 Comparison of Line N and line 15I
5.2.2 Colony PCR of RDA products
5.2.3 Screening and characterisation of RDA clones
5.2.4 Mapping RDA Clones
5.2.5 YAC Hybridisation
5.2.6 Nucleotide Sequences
5.3 Discussion
Chapter Six
Targeted mapping of chromosome 16 by Representational
Difference Analysis
6.1 Introduction
6.2 Results
6.2.1 BamHI Representation
6.2.2 TaqI Representation
6.2.3 NheI Representation
6.2.4 Screening and characterisation of RDA clones
6.2.5 Mapping RDA Clones
6.2.6 Nucleotide Sequence
6.2.7 YAC Hybridisation
6.3 Discussion
Chapter Seven
Resistance to Marek's disease targeted by Representational
Difference Analysis
7.1 Introduction
7.2 Experimental Design
7.3 Results
7.3.1 RDA using BamHI representation
7.3.2 Screening and Characterisation of RDA clones
7.3.3 Mapping RDA Clones
7.3.4 Analysis of association of RDA clones to MD resistance
7.3.5 Segregation of hV32 in the F₂ population
7.4 Discussion
Chapter Eight
General Discussion
8.1 The Chicken
8.2 Genome Mapping
8.3 Saturation Mapping
8.4 Random amplified polymorphic DNA (RAPD) analysis
8.5 Microsatellite Markers
8.6 Amplified fragment length polymorphisms (AFLP)
8.6 EST mapping
8.7 Subtractive Hybridisation
8.8 Differential display-reverse transcription (DD-RT)
8.9 RFLP subtraction
8.10 The RDA Approach
8.11 The Future of Mapping
8.12 Future Work
References
Appendices (removed)

°C degree(s) Celsius
A adenine
aa amino acid
ABI Applied Biosystems (Perkin Elmer)
AceDB A C.elegans database
AFLP Amplified fragment length polymorphism
AIDS Acquired Immunodeficiency syndrome
ALSV Avian leukosis and sarcoma viruses
ALSV-A Subgroup A Avian leukosis and sarcoma viruses
Alu-PCR PCR of inter Alu fragments
ALV Avian leukosis virus
Amp Ampicillin
APC Antigen presenting cell
App. Appendix
APS Ammonium Persulphate
ARS Autonomous Replication Sequence
BCE Before common era
bp base pair
BSA Bovine Serum Albumen
C cytosine
C band chromatin band
cDNA complementary Deoxyribonucleic acid
CH₃COONH₄ Ammonium acetate
CHEF Clamped homogeneous electrophoresis field
cM centi-Morgan(s)
cm centimetre
CMRP Compton mapping reference population
contig contiguous
COS7 Monkey kidney cell line
CpG Cytosine and guanine dinucleotide
CR1 Chicken Repeat 1
Cys Cysteine
dCTP 2'-deoxycytidine triphosphate
DD-RT Differential Display-Reverse Transcriptase
DMF Dimethyl formamide
DNA Deoxyribose nucleic acid
dNTP deoxynucleoside triphosphate
DOP-PCR Degenerate oligo. primed PCR
dsb Double stranded break
dT deoxy-Thymine
DTT dithiothreitol
dUTP 2'-deoxyuridine triphosphate
EDTA Ethylenediaminetetra acetic acid
EE EDTA and EPPS buffer
EPPS N-(2-hydroxyethy)piperazine-N'-(3-propanesulfonic acid)
EMBL European molecular biology laboratory
ES Embryonic stem cell
EST Expressed sequence tag
F₁ First cross
F₂ Intercross between two F₁
FAM 5-carboxyfluorescein (ABI fluorescent label)
Fig Figure
FISH Fluorescent In Situ Hybridisation
g Acceleration due to gravity
G guanine
GCG Genetics computer group
GCN4 DNA binding protein of ds DNA
GDB Genome database
GDRDA Genetically Directed Representational Difference Analysis
GL Suggestive linkage
GMS Genomic mismatch scanning
GTE Glucose, tris, EDTA buffer
HAT Hypoxanthine/aminopterin/thymidine
HepG2 Liver cancer cell line
HEX 6-carboxy-2',4',7'4,7-Hexachlorofluorescein
(ABI fluorescent label)
His Histidine
HPRT Hypoxanthine phosphoribosyltransferase
HSL Highly significant linkage
HVT Herpes virus of turkeys
IAH Institute for Animal Health
IBDV Infectious bursal disease virus
IBV Infectious bronchitis virus
IFGT Irradiation and fusion gene transfer
IPTG Isopropyl-[beta]-D-thiogalactopyranoside
JBAM 2^nd RDA oligo for BamHI derived fragments
JTAQ 2^nd RDA oligo for TaqI derived fragments
kb kilobase
kV kilo volts
l litre
[lambda] Wavelength
L-agar Luria-Agar
LOD Logarithms of odds ratio
LR Long Ranger
LTR Long terminal repeat
M Molar
MAS Marker assisted selection
Mb Megabase
Mbq Mega bequerel
µCi micro Curie
MD Marek's disease
MDV Marek's disease virus
mer Oligomer
mF milli Farad
µg microgram
mg milligram
MgCl₂ Magnesium chloride
MGD Mouse genome database
MHC Major Histocompatiblity Complex
MIC Micro-chromosome
min minute(s)
µl microlitre
ml millilitre
µm micrometer
MPC Biomagnetic separator
MQ Milli-Q pure water
mRNA messenger Ribonucleic Acid
N Normal
NaCl Sodium chloride
NaClO4 Sodium perchlorate
NaOH Sodium hydroxide
NBAM 3^rd RDA oligo for BamHI derived fragments
ng nano gram
nmol nano mole
NOR Nucleolar organiser region
NTAQ 3^rd RDA oligo for TaqI derived fragments)
OD Optical Density
oligo Oligomer
p short arm of chromosome
P Probability
PB Qiagen's PB buffer
PBS Phosphate Buffered Saline
PCR Polymerase Chain Reaction
PCR-ELA PCR-Enzyme Linked Assay
PCR-SSP PCR-Sequence Specific Primer
PEG Polyethylene Glycol
pers. com. Personal communication
PFGE Pulsed field gel electrophoresis
pfu plaque forming units
pg pico grams
PI Post infection
pmol pico moles
Poly A Poly Adenylate
PVP Polyvinylpyrrolidone
q long arm of chromosome
QTL Quantitative Trait Locus
rad Unit of Radiation dose
RAPD Random Amplified Polymorphic DNA
RBAM 1^st RDA oligo for BamHI derived fragments
RDA Representational Difference Analysis
rDNA Ribosomal DNA
RE Restriction Endonuclease
RFLP Restriction Fragment Length Polymorphism
RH Radiation hybrid
RNA Ribonucleic Acid
RNaseA Ribonuclease A
RNHE 1^st RDA oligo for NheI derived fragments
rpm revolutions per minute
RPRL Regional poultry research laboratory
RT Room Temperature
RTAQ 1^st RDA oligo for TaqI derived fragments
SB Special buffer
SDS Sodium Dodecyl Sulphate
sec Second(s)
Seq Ed Sequence editor
SL Significant linkage
ss single stranded
SSC Saline sodium citrate
SSCP Single Stranded Conformational Polymorphism
STS Sequence tagged sites
T Thymine
T_a Annealing Temperature
TAE Tris-acetate EDTA
TAP2 Transporter associated with antigen processing 2
Taq Thermus aquaticus DNA polymerase
TB Terrific broth
TBE Tris-borate EDTA
TE Tris EDTA
T_e Extension Temperature
TEMED N,N,N',N'-tetramethylethylenediamine
Tet Tetracyclin
T_m Melting Temperature
Tris 2-amino-2-(hydromethyl) propane-1, 3 diol
tRNA transfer RNA
U Unit
UK United Kingdom
UTR Untranslated region
UV Ultraviolet
V Volt
VNTR Variable number tandem repeat
vs. versus
[Omega] Ohm
WG-RH Whole genome radiation hybrid
X-Gal 5-Bromo-4-chloro-indoyl-[beta]-D galactoside
YAC Yeast artificial chromosome
YS Yoshida sarcoma

Loci/prefixes
ADL Microsatellite markers from East Lansing
COM Marker generated at IAH Compton
MCW Microsatellite markers from Wageninen
28S 28S ribosomal protein
60S 60S ribosomal protein
B-F Chicken MHC Class I locus
B-L Chicken MHC Class II locus
B-G Chicken MHC Class IV locus
BAT8 (homologue of G9a) function unknown
bcl-x bcl2 family, apoptosis gene
BG32.1 Chicken MHC Class IV locus marker 32.1
Bu-1 antigen expressed on B cells (marker=CHB6)
C4 MHC class III gene
CMYC28 Chicken myelocytomatosis viral oncogene 28
CNBP Cellular nucleic acid binding protein
COL4a1 Collagen type 4, alpha 1 chain
Db Eumuelanin restrictor gene
ERV Endogenous retrovirus
ev21 endogenous virus locus
FN1 fibronectin 1 gene
G6PD Glucose 6 phosphate dehydrogenase gene
G9a (homologue of BAT8)
GDI GDP dissociation inhibitor
GDID4 GDP dissociation inhibitor for the rho (GDI) protein
GDP Guanosine diphosphate
hg high growth gene
HLA Human leukocyte antigen loci A-D
HMG1 High mobility group 1
HPRT Hypoxanthine phosphoribosultransferase
Insr Insulin receptor gene
ITPR2 Inositol tri-phosphate receptor 2
Ity Salmonella typhimurium resistance gene
jcpk Juvenile congenital polycystic kidney disease
Lec Lectin type C gene
Ly-4 receptor gene on lymphocytes
MHCII[beta] Major Histocompatiblity Complex II [beta]
Ml Eumuelanin extension gene
NF2 Neurofibromatosis 2 gene
NOR Nucleolar organiser region
Nramp Natural resistance associated macrophage protein
nude Nude locus (mouse chromosome 1)
Pg Pattern gene
pol polymerase gene
PSF PTB associated splicing factor (marker SFPQ)
PTB Polypyrimidine tract binding
pudgy Pudgy locus
RAG-1 Recombination activation gene 1
RAG-2 Recombination activation gene 2
Rfp-Y Restriction fragment pattern-Y
SFPQ PTB associated splicing factor (PSF)
SRE Sterol regulatory element
TAP2 Transporter associated with antigen processing 2
TET Tetracycline resistance gene
TFRC transferrin receptor gene
Th-1 antigen expressed on T cells
tottering tottering gene
TUBA Tubulin alpha
tv-a receptor locus for ALSV-A
unl unlinked
VIL villin gene
Xic X inactivation centre
Xist X-inactive specific transcripts

1.1 The Chicken
The domestic chicken (Gallus gallus domesticus) originated from south-western Asia as a descendant of the Red Jungle Fowl and was first introduced into China in about 1400 before the common era (BCE). Chickens are also depicted in Babylonian carvings of about 600 BCE and are mentioned by ancient Greek writers, particularly Aristophanes in 400 BCE (reviewed by Stockbridge, 1995). Since that time, chickens were kept in small flocks for home consumption until the 20^th Century when poultry farming became commercialised. A modern poultry farm may contain from several hundred thousand to over a million chickens, either layers for egg production or broilers for the meat industry. The chicken industry is now a multi-million pound business with a gross output value of £1339 million in the UK in 1991 (Law and Payne, 1996), which depends on high-production egg layers, very rapidly growing broilers and disease-free stock. These three factors have been the targets of poultry breeders who have tried to select the best genetic traits in their stock, possibly eradicating some of the main poultry pathogens.
However, disease is still a major cause of loss in production and still proves of great economic concern to poultry farmers, and from the point of view of the birds' welfare is unethical. Control of disease has been improved by good husbandry and the use of vaccines, but, it would be far more efficient if the birds could be bred to be disease resistant. Therefore it would be useful to know the identity of the gene or genes affecting the resistance traits for each disease, as with this detailed information it will be possible to directly select for the required trait. To this end linkage maps of the chicken have been developed, with the intention of generating enough marker loci to identify linkage with the resistance trait, and ultimately by positional cloning to isolate the gene concerned. Once resistance genes have been isolated and characterised it will be possible to use them to screen commercial breeding stock and selectively breed for resistance. Knowledge of the gene may also suggest improved pharmaceutical or immunological therapies.
Chickens are animals of agricultural importance as well as a valuable model organism. Their use as a model organism is due to the many inbred lines of chickens, which are well characterised for disease and production traits. The parent stocks produce large numbers of progeny which eases genetic experimentation. All commercially bred chickens have accurately detailed pedigrees and are of low cost per animal. Because it is possible to produce and assess large progeny families, the chicken is well suited to genome analysis, particularly as the red blood cells are nucleated, facilitating large scale good quality DNA extraction. Chickens have a small genome, about half the size of mammals' but three times as big as that of the pufferfish (Elgar et al., 1996). The chicken genome is therefore more easily accessible to cloning than those of mammalian species, and yet retains sufficient sequence homology to mammals to provide comparisons which could elucidate function. It is not yet clear whether this similarity also applies to gene order between these species.
Chickens can harbour a number of different types of pathogen; viruses, bacteria and parasites, all of which cause major diseases in poultry. A wide variety of resistance mechanisms operate against these pathogens. The identification and characterisation of these resistance genes in chickens could help elucidate basic mechanisms for resistance in other organisms as well as chickens.
1.2 Genomic structure and organisation of the chicken
The genome of the domestic chicken has a haploid number of 39 chromosomes; the ten largest are referred to as macro-chromosomes, and the other 29 are termed micro-chromosomes (MICs) (Yamashina, 1944). In chickens chromosomes have been numbered in size order, the biggest first. The large number of MICs is typical of avian species (Abbott and Yee, 1975). In comparison to man, the first six chromosomes are of similar size, the largest being 8 µm. However the MICs are much smaller (the smallest being about 7 Mb) than the smallest human chromosome, which contains about 50 Mb of DNA (Bloom and Bacon, 1985). There is a size difference of 23 times between the largest and the smallest chromosomes in the chicken. Chickens, like other avian species, differ from mammals in that the female is the heterogametic sex (ZW) and the male is the homogametic sex (ZZ), the Z and W chromosomes displaying heteromorphism. The chicken chromosomes are mostly euchromatic with the exceptions of a large terminal C-band (chromatin-band) on the Z chromosome and an almost totally heterochromatic W chromosome, with small C-bands on most of the MICs (reviewed by Fecheimer (1990)). The chicken genome is relatively small, about 1.2 x 10⁹ bases (Olofsson and Bernardi, 1983), less than half that of the mouse and human genomes. This makes the chicken's genomic structure and organisation particularly interesting, as evolution appears to have pruned the genome to a minimal size. Alternatively it is possible that mammalian genomes have expanded in the 300 million years since splitting from the avian lineage. However whatever the mechanism of change between the two lineages, the reason for this difference is still unknown. One theory is that small genome size was favoured by directional selection in birds in order to cope with the metabolic demands of flight (Hughes and Hughes, 1995). However, one aspect of the chicken's unique genome is the relative paucity of repetitive sequences.
1.2.1 Repetitive Sequences
The chicken genome like that of other animals contains repetitive sequences. In the genomes of many animals, for example the amphibian Xenopus, there is a short interspersion pattern of 0.3 kb repetitive sequences with 2 kb single copy sequences (Davidson et al., 1973). However, the genomes of avian species show organisation similar to that of the long period interspersion pattern of the Drosophila genome (Crain et al., 1976). Estimates vary as to the degree of repetition in the chicken genome. Eden and Hendrick (1978) concluded that 87% was single copy and 13% repetitive sequences, 42% of which contained ten to twenty copies and 58% contained about 1500 copies. Another study by Epplen et al. (1978) using reassociation kinetics obtained similar results showing that the chicken genome contained 15% highly repetitive sequences, 10% intermediate repetitive sequences and 70% single copy sequences, of which 28% are 2.3 kb single copy DNA pieces interspersed with 1.5 kb long repetitive fragments. Olofsson and Bernardi (1983) fractionated the genome using density gradient centrifugation and similarly concluded that the genome was composed of 84% unique sequences and 13% repetitive sequences. Arthur and Straus (1983), however, concluded that the genome contained 34% of unique sequences interspersed with repeated sequences, 38% of long stretches of unique sequences and 19% of foldback elements. Sequences complementary to mRNA were found to be randomly distributed with respect to the interspersion patterns, implying that the distribution of the repetitive sequences does not parallel that of the structural genes.
The only characterised chicken repeat sequences identified so far are the chicken repeat 1 (CR1) elements, which are middle repetitive sequences discovered by Stumph et al. (1981). CR1 elements are an ancient group showing six sub-families in chickens, with four sharing a common progenitor (Vandergon and Reitman, 1994). Over 95 CR1 elements have been identified, of an average length of 300 bp, the largest being 2.3 kb (Burch et al., 1993). They have truncated 5' ends and a consensus 3' end containing two or more repeats of an eight nucleotide sequence. CR1 elements are non-long terminal repeat (LTR) retrotransposons (Burch et al., 1993), with a pol-like open reading frame encoding reverse transcriptase which is responsible for dispersal of the element throughout the genome. There are an estimated 100,000 copies of CR1 throughout the chicken genome (Vandergon and Reitman, 1994). CR1 elements are also represented in other avian genomes. Nine orders have so far been studied all of which contain CR1 elements (Chen et al., 1991). All CR1 elements have two regions of high homology (Stumph et al., 1984), which extend to other avian species. These regions appear to be conserved, the first coding for a silencer domain and the second for a nuclear protein binding domain located at the 3' end of the element, probably involved in transposition (Chen et al., 1991).
Reptiles have also been shown to possess CR1 elements (Vandergon and Reitman, 1994), showing that CR1 elements existed before the divergence of birds and reptiles. CR1 elements are not randomly distributed but occur mostly in G-C rich regions of the genome (Olofsson and Bernardi, 1983). These regions are also the most gene-rich and at least 16% of the [beta]-globin gene cluster contains CR1 elements (Reitman et al., 1993). CR1 elements are associated with a number of genes, particularly near transition regions of chromatin structure, for example the ovalbumin gene (Stumph et al., 1983).
In mammals repeat elements are highly useful as, like the Alu repeats of man, they can be used to characterise the genome by hybridisation or PCR. PCR amplification of inter-Alu fragments (Alu-PCR) has been used for high resolution physical mapping of radiation hybrids, cosmids and yeast artificial chromosomes (Monaco et al., 1991; Aburatani et al., 1996). However, CR1 repeats in chickens are less common and also much less conserved than the repeats in mammals and it seems unlikely that they can be used for these purposes. Similarly although chickens possess simple microsatellite repeats, there are far fewer of these than in humans and mice. For example in chickens estimates of the numbers of (AC)n repeats range as low as 7,000, almost 10 fold less than in mammals and they appear not to be uniformly distributed throughout the genome (Toye, 1993).
1.3 Genomic Mapping
The first published genetic map was of the Drosophila X chromosome (Sturtevant, 1913). Since then, genome mapping technology has greatly improved and genetic maps for over one hundred species have been generated.
Two types of genomic maps have been established for many animals. Linkage maps are determined from the frequency of meiotic exchange among linked polymorphic markers in sexual crosses. Physical maps relate traits to actual chromosomal location. Ultimately, the aim is to merge these two maps into one giving both physical and linkage data.
1.3.1 Genetic Mapping
Genetic or linkage maps are necessary for the location of genes responsible for monogenic and polygenic traits. Linkage analysis can be applied to backcrosses of the same breed, or interspecies crosses can be used to generate more polymorphisms. The murine genome has been analysed using interspecific crosses (Avner et al., 1988). The advantages of interspecific crosses are that they maximise the number of possible genetic polymorphisms, and that pedigree analysis can be aligned with linkage analysis in order to place loci on a particular chromosome (Copeland and Jenkins, 1991).
In chickens, genome mapping was first achieved by the analysis of phenotype and limited maps established (Hutt, 1949). The first genetically studied crosses were generated by Bateson and Saunders (1902) but it was not until 1936 that Hutt generated the first chicken linkage map. Chickens of differing phenotypes were crossed and the linkage between different genes established. For example, the eumelanin extension gene (Ml) lies between the loci of the eumelanin restrictor (Db) and the pattern gene (Pg) (Carefoot, 1987). This system is inherently limited as these markers require expression as phenotypes and are only present in rare lines of chickens. These morphological polymorphisms are also infrequently observed, as only one or two occur in any one line. This makes segregation analysis difficult as no single population is capable of giving scores for each polymorphism, thus leading to ambiguity in the results.
Bumstead and Palyga (1992) published the first map of the chicken genome derived from Restriction Fragment Length Polymorphism (RFLP) analysis.
68 White leghorn progeny were generated from a backcross between an F₁ (Line 15I x line N) female and her male line 15I parent. An initial map of 100 markers was developed from segregation analyses, covering a minimum of 585 cM of the chicken genome and identifying 18 linkage groups.
More recently, Crittenden et al. (1993) crossed a Red Jungle Fowl and a White Leghorn in order to maximise the chance of finding DNA polymorphisms in the backcross. It was determined that RFLP were observed between the parental lines with in the order of 91% of those enzymes used, compared to 50% in the inbred White Leghorn lines used by Bumstead and Palyga (1992). This gave a polymorphic population providing a good resource for collaborative mapping. A variety of marker types were used, including linkage mapping of feather colour and blood group loci, RFLP analysis and RAPD (Random Amplified Polymorphic DNA) analysis. Further analysis of this intercross has provided a linkage map (Levin et al., 1994b). Recent comparisons of the Compton and East Lansing maps has shown that there is in fact greater recombination in the Compton population, containing 2700 cM in comparison to 2100 cM in the East Lansing population (Mariani et al., In Press). Within the Compton Linkage map there are over 100 genes and 300 other loci, and now fewer than 2% of markers do not show linkage (of LOD>3.0).
There are two main strategies for genetic mapping; saturation or random mapping and targeted mapping. Saturation or random mapping entails the generation of large numbers of markers in an attempt to cover the whole genome densely, thus ensuring that any trait of interest must lie near to at least one marker. Examples of techniques which achieve this are: microsatellite analysis, RAPD STS, AFLP, RFLP and SSCP. Targeted mapping is the isolation of markers in a specific region, or for a specific trait. This is achieved by a variety of methodologies including EST, which can be targeted by the isolation of cDNAs from specific tissues, as well as more refined methods of subtractive hybridisation and representational difference analysis.
Markers generated by these methods can be used for comparative mapping, but are defined as two classes based on their polymorphic nature. Comparative mapping is the positioning of identical markers on maps of different species or strains. These markers are termed anchor or index loci and are mostly used to define new linkage maps. Type I anchor loci are coding sequences which are relatively conserved across different species. However, this sometimes means that they contain no identifiable polymorphisms. So, although they are particularly useful in interspecific mapping for the identification of genes they cannot always be used. Type II anchor loci are defined as high resolution polymorphisms. These are typified by microsatellite and minisatellite polymorphisms which have traditionally not been used as loci for interspecific mapping, due to their non-coding nature. However, O'Brien and Graves (1991) have shown that these type II anchor loci can also play an important role in the identification of syntenic regions between maps of different species.
1.3.1.1 Restriction Fragment Length Polymorphism
RFLPs are generated by the digestion of DNA with restriction endonucleases. RFLP was used by Donis-Keller et al. (1987) in the construction of the first genetic linkage map of the human genome. RFLP are often visualised by hybridisation of radioactive cDNA probes to the digested genomic DNA fragments. RFLP analysis of the chicken B-F and B-L genes has been used to confirm serological B-typing (Juul-Madsen et al., 1993).
There are disadvantages with RFLP analysis, as often a number of probe and enzyme combinations have to be tested in order to generate significant numbers of RFLP. This method generally uses radioactive probes, and requires a large amount of target DNA. In fact, RFLPs detect about 1 in 10000 polymorphic nucleotides in the human genome (Soller and Beckmann, 1986). The problem with searching for direct associations between RFLP and traits of economic value is the low likelihood of finding one; at best 1:200, but probably 1:20000 (Soller and Beckmann, 1986). RFLP markers are also difficult to transfer between different linkage maps, as what is polymorphic in one population may well not be in another. This, however, is true for any randomly selected type of marker, but RFLPs do generate a reasonable number of markers which can be easily placed onto a map. This technique has been extensively employed by Bumstead and Palyga (1992) to place new markers on the chicken linkage map.
1.3.1.2 Single-Stranded Conformational Polymorphism
SSCP involves the digestion of genomic DNA by restriction endonucleases, denaturation with NaOH and EDTA and electrophoresis on a non-denaturing polyacrylamide gel. Polymorphisms are detected as a mobility shift of the single stranded DNA, due to conformational changes in alternative secondary structures (Nakabayashi and Nishigaki, 1996). Even single-base substitutions can be detected as shifts in electrophoretic mobility (Orita et al., 1989a). This technique has been successfully used to detect three point mutations in highly conserved residues found in mildly-affected cystic fibrosis patients (Dean et al., 1990).
SSCP-PCR is a development of this technique, where the sequences to be examined are amplified by PCR using labelled primers for the clones of interest; these could include fluorescently-labelled primers. The samples are then denatured and analysed by gel electrophoresis (Orita et al., 1989b). This technique is usually used for genetic screening, as it requires sequence data for the design of the primers. However, if RFLP analysis shows no polymorphism with a particular clone, this method can be used once the genomic sequence has been determined permitting design of primers to the most variable region possible, usually introns or 3' UTR. But it is costly in primer synthesis and often many primer pairs have to be tested before a polymorphism is detected.
1.3.1.3 VNTR-Hypervariable Minisatellites
Minisatellites or VNTR (Variable number tandem repeats) are regions of DNA containing simple tandem repeats of 10-15 bp segments. They are detected by hybridisation of a probe to a core sequence within the repeat. This generates a multi-band pattern or fingerprint which can vary between individuals (Jeffreys et al., 1985). These band patterns are analysed by radioactive labelling of the products and electrophoresis on a polyacrylamide gel. This technique is mainly used in forensic and paternity investigations but can also be used for mapping. Fingerprinting usually detects a large number of loci from a single probing but these loci are mostly unlinked and scattered throughout the chicken genome; however, hybridisation with cloned single locus minisatellites shows that some could be clustered (Bruford et al., 1994). However, this necessitates knowledge of the core sequence and the identification of polymorphisms between the parents of the cross before it can be used for linkage mapping. VNTR analysis is also time consuming and involves the use of large amounts of DNA and radioactive label.
1.3.1.4 Microsatellites
Microsatellite markers consist of di-, tri-, or tetra-nucleotides repeated up to 40 times. Oligonucleotides are synthesised to sequences flanking these repeats and can be used in PCR, to amplify loci from genomic DNA (Holmes et al., 1993). Microsatellites are now being used as markers for linkage analysis as they can be highly polymorphic, polyallelic, very abundant and usually behave as single locus markers (Weber, 1990). Disadvantages include the large amount of labour required and initial expense of the oligonucleotide primers necessary to identify enough microsatellite markers (Rafalski and Tingey, 1993). Microsatellites are useful type II anchor loci as they are often polymorphic between more than one linkage cross. However, within chickens less than 50% of microsatellites are polymorphic between two inbred lines (Khatib et al., 1993). This is particularly noticeable in the Compton mapping reference population where microsatellite polymorphisms occur in less than 40% of primers (Bumstead pers. com.). However, many microsatellites have been used in the chicken linkage maps and recently a third map has been generated, based entirely on microsatellites (Crooijmans et al., 1996). With increased use of fluorescence technology, using the ABI Genescan system, many fluorescent primers can be used in each PCR reaction and the products separated by electrophoresis. While re-loading gels three or four times greatly increases the productivity, cloning and screening the initial microsatellite is still a time-consuming process.
1.3.1.5 Random Amplified Polymorphic DNA markers
RAPD markers can be generated using short arbitrary primers to amplify genomic DNA, giving a genotype-specific pattern of bands. This is discussed later in section 1.6.2.
1.3.1.6 Amplified Fragment Length Polymorphism
AFLP is based on selective amplification of digested genomic DNA by a series of extended primers. The genomic DNA is digested with two enzymes, a frequent and a rare cutter, e.g. MseI and EcoRI. Adaptors, with a core sequence and the appropriate overhangs for the enzymes used are ligated to the ends of the fragments. Pre-amplification of these fragments by PCR is achieved using primers with a core sequence, an enzyme specific sequence and a random-extension nucleotide. This generates a pool of fragments which are re-amplified by two primers of core sequence, enzyme specific sequence and this time further random-extension nucleotides. The primer with the rare cutting enzyme sequence is end-labelled before the reaction so that those fragments only will be detected by electrophoresis and autoradiography (Vos et al., 1995). Fragments generated by MseI digestion will be small enough to amplify by PCR, while those with one MseI and one EcoRI end will be much rarer but will also easily amplify. During the second amplification, further selection of fragments occurs using the extension nucleotides and finally only those fragments generated by the EcoRI-labelled primer will be resolved visually by autoradiography following polyacrylamide gel electrophoresis. The use of two enzymes also enables the AFLP reaction to be fine-tuned for the production of the maximum number of bands. This technique was originally developed by plant geneticists but can easily be transferred to the analysis of the chicken genome. The development of many polymorphic bands for each reaction should enable large numbers of marker loci to be mapped for a relatively small amount of work and input DNA. However, all these markers will be anonymous, and to clone one requires the extraction of DNA from a gel matrix. This method can be visualised by silver staining rather than by radioactive labelling, but the complexity would not then be further reduced, and might need a further PCR step, which will ultimately be more expensive.
1.3.1.7 Targeted Mapping
Recently, a number of strategies have been directed towards the isolation of candidate genes. In comparison with the saturation mapping approach, these techniques would not necessarily place other informative loci on the map, but would significantly reduce the amount of work needed to isolate a candidate gene.
1.3.1.8 Expressed Sequence tags
EST are partial sequences of cDNA clones. These will be discussed later in section 1.6.1.
1.3.1.9 Subtractive hybridisation
Subtractive hybridisation has been used to find candidate tumour suppressor genes. This entails the positive selection of cDNA expressed only in non-tumour cells of interest, when radioactively-labelled cDNA is hybridised with tumour cell mRNA. Unhybridised labelled cDNA can be recovered and used as a probe to screen a cDNA library constructed from non-tumour cells. New suppressor genes were identified in this way (Lee et al., 1991). This method also takes up considerably more time than that of, for example, differential display reverse transcriptase (DD-RT) see below, although, in comparison, it does have the advantage of being a positive selection technique. However, it does rely on the selection of a small amount of unique DNA from a large amount of initial DNA. For this technique to work effectively for the isolation of candidate genes, the pattern of tissue expression must be known for each gene. This is very difficult to achieve with anything but tumours or highly expressed tissue specific genes.
1.3.1.10 RFLP subtraction
RFLP subtraction (Rosenberg et al., 1994) purifies small restriction fragments from one genome (the tracer) from sequences which reside on fragments in a related genome (the driver). The genomic DNA is digested with HindIII, and then the driver is ligated to biotinylated adaptors, whilst the tracer is ligated to unlabelled adaptors. Both are PCR-amplified separately and then mixed together, the driver in 100 times excess of the tracer. After hybridisation, the biotinylated driver and its heteroduplexes are removed by streptavidin-coated beads, leaving the tracer-specific DNA. This technique can be used to generate DNA for library construction or to obtain markers near to a gene of interest. However, only tiny amounts of DNA will be generated and small under-represented sequences may well be lost as there is no enrichment. Sequences of interest may also be lost in the removal of heteroduplexes from the mixture as the tracer-specific DNA is not positively selected.
1.3.1.11 Differential Display-Reverse Transcription (DD-RT)
DD-RT is a method of analysing changes in gene expression in cells at different stages of differentiation, thus isolating genes expressed at a particular stage (Bauer et al., 1993). The mRNA is systematically amplified and the 3' termini visualised by electrophoresis on a denaturing polyacrylamide gel. An anchored oligo dT primer is used to anneal the mRNA polyA tails, in conjunction with a decamer oligodeoxynucleotide in solution. The decamer is arbitrary in sequence, so that it can anneal at different positions in relation to the first primer for PCR amplification. Thus, when the RNAs from two or more relevant cell types are compared, a small fraction of differentially expressed mRNAs are revealed. Probes can be made from the products, and the genes isolated by genomic library screening (Liang et al., 1993).
DD-RT is useful as it displays subsets of mRNA as short cDNA bands, enabling visualisation of cell mRNA composition. Alterations in gene expression can be simply detected between parallel samples as variable band patterns. The cDNA can easily be sequenced and compared to sequences in databases, or used for probing (Liang and Pardee, 1992). This technique has the advantage of being quick to apply. It is also non-selective, isolating all genes present at any given stage. Again this technique relies on knowledge of tissue expression patterns and times during which the gene of interest could be isolated, which may not be known or fully understood. For the isolation of genes associated with development during particular stages this technique can be very useful, but as a general technique it is too complex and technically demanding.
1.3.1.12 Genomic Mismatch Scanning
Genomic mismatch scanning (GMS) enables mapping in two related individuals. The genomic DNA of one individual only is methylated at GATC sites, while the other remains untreated. The two samples of DNA are then denatured, mixed, and hybridised in solution. The samples are digested at fully methylated and fully unmethylated sites leaving only whole heterohybrids. Base-mismatch hybrids are nicked using a single mismatch repair system. The DNA is then degraded by digestion, leaving mismatch-free heterohybrids. These hybrids should be from regions of identity by descent. This DNA can be labelled and used for probing DNA from the whole genome, yielding positive results with identity by descent and negative results at sites of meiotic recombination (Nelson et al., 1993). This methodology is very prone to failure as there are many separate enzymatic steps, which must initially be optimised in order to generate the complete methylation of one genome. GMS is performed on the whole genome, without reducing its complexity, thus at hybridisation not all homoduplexes will have time to reanneal. The under-represented sequences which differ between the two genomes are also unlikely to be detected as there is no enrichment for target sequences.
1.3.1.13 Representational Difference Analysis
RDA is a new technique developed by Lisitsyn et al. (1993) and will be discussed later in section 1.6.3.
1.3.2 Physical Mapping
Physical mapping relates genes and markers to their actual location on a specific chromosome. Physical mapping of chicken genes is difficult as cytological discrimination between the 29 MICs is difficult on chromosome spreads. Physical mapping in other species is now achieved using techniques such as fluorescent in situ hybridisation (FISH) of single copy molecular clones to metaphase chromosomes (Korenberg et al., 1992).
1.3.2.1 Pulse Field Gel Electrophoresis
PFGE is a technique for the separation of large segments of DNA, up to 1 Mb, using alternating changes in the angle of an electric field to separate the DNA on an agarose gel (Schwartz and Cantor, 1984). There are now many variations of the basic concept, the most common being CHEF (Clamped homogeneous field electrophoresis), in which the agarose gel is surrounded with a series of electrodes (Lognonne, 1993). Modern systems can resolve fragments from 10 kb to 10 Mb, helping to close the gap between the standard cloning systems in plasmid, phage and cosmid vectors and genetic linkage maps. However, PFGE still cannot resolve the chromosomes of larger genomes like the chicken, as even the smallest MIC is over 7 Mb in size (Bloom et al., 1993). Since the development of yeast artificial chromosome clones, however, PFGE has become essential to all mapping labs.
1.3.2.2 Radiation Hybrids
Initially hybrids were made by fusion of somatic cells with either single or total alien chromosomes. However, a more useful method of irradiation and fusion gene transfer (IFGT) was initiated by Goss and Harris (1975) to produce a resource for mapping genes, and was further refined by Cox et al. (1990). Goss and Harris (1975) discovered that genes could be rescued from cells killed by X-rays if the dose was increased so that double stranded breaks (dsb) occurred along the chromosome. A dose of 1500 rad is adequate to kill the cells whilst 4000 rad was sufficient to give a dsb between two loci. Cox et al. (1990) applied this technique to the production of radiation hybrids (RH) from a Chinese hamster-human somatic cell hybrid, containing a single copy of human chromosome 21. The somatic cell hybrid was irradiated with 8000 rad to kill the cell and produce dsb and then rescued by non-irradiated enzyme-deficient hamster cells. The dsb recombine with the chromosomes in the healthy cell to give RH clones containing fragments of human chromosome 21.
Walter et al. (1994) took this technique further to develop whole genome radiation hybrids (WG-RH). Human fibroblasts were irradiated with 3000 rad and rescued by thymidine kinase-deficient hamster cells. These WG-RH were grown in HAT (hypoxanthine/aminopterin/thymidine) medium to select for the human chromosomes and were initially analysed for the presence of chromosome 14 markers by FISH. In the forty-four cell lines produced, a marker retention of 20-50% was attained, so in theory a set of about 100 WG-RH should provide good coverage of the whole genome. WG-RH are a powerful resource, as they enable the mapping of YACs and ESTs which are non-polymorphic within the linkage map. The amount of recombination can be controlled by the radiation dosage (rad), as increased dose leads to increased dsb and therefore increased recombination. However, good coverage of the chicken genome in WG-RH might be difficult, again due to the large numbers of MICs which need large doses of radiation to induce enough dsb for significant recombination. It is possible that some of the MICs would remain whole within the RH clones and would quickly disappear. There is an added complication in that the RH clones will eject the alien chromosomes fairly quickly, so once made aliquots of each clone should be stored frozen while large amounts of DNA are prepared as a future resource.
1.3.2.3 Flow Karyotyping
Flow karyotyping is a technique used to separate individual chromosomes. A large number of chromosomes are isolated as a suspension, and stained with fluorescent dyes. The chromosomes are passed singly through a laser beam and sorted by their fluorescence pattern. The fluorescence pattern depends on DNA content, as the different dyes have differing affinities for certain sequences. The dye Hoechst 33258 binds preferentially to A/T sequences, whilst chromomycin A3 prefers G/C regions. This technique is very sensitive and has to be carefully optimised, only using the most pure chromosomes in order to prevent loss of resolution (Gray et al., 1975). Once single chromosomes are obtained they can be used as a resource for PCR amplification, cloning into single chromosome libraries and direct selection of yeast artificial chromosome clones (YACs) into single chromosome pools. However, although flow karyotyping has been used with great success for man and plants, the large numbers of MICs in the chicken genome prevent resolution beyond the first ten chromosomes. Flow cytometry can also be highly variable and needs to be verified by conventional cytogenetic methods, which can be problematical in the chicken.
1.3.2.4 Chromosome microdissection
Chromosome microdissection has been achieved with improved laser technology (Greulich, 1992). Chromosomes are held in an optical trap, whilst they are dissected by laser microbeam, and then removed by glass needle. This has the advantage of removing a specific piece of a specific chromosome. Microdissected fragments could be used in the construction of a contiguous YAC library.
Unfortunately, there is little information to aid selection of the correct chromosome in the chicken, as after the ten largest chromosomes the MICs are cytologically indistinguishable. This technique yields miniscule amounts of DNA (a few hundred femtograms to a picogram) which makes cloning of the dissected fragments difficult. However, PCR amplification of the fragments by DOP-PCR (Degenerate oligonucleotide primed-PCR) in a general and non-specific manner, using 22-mer degenerate primers in a dual annealing temperature protocol, can generate enough DNA for cloning the dissected chromosome (Arumuganathan et al., 1994). The actual region of the dissected fragment is still not that small, in the order of tens of megabases, which still leaves a significant gap from the 700 kb sequences achievable by conventional molecular techniques. This technique can generate enough DNA to be used as a probe for FISH to specific chromosomes.
1.3.2.5 Yeast Artificial Chromosomes (YACs)
YACs are constructed by inserting yeast chromosomal fragments into plasmid vectors. These yeast fragments include an autonomous replication sequence (ARS), centromere and telomeres, as well as selectable markers. These constructs are linearised and transformed into yeast cells where they replicate as synthetic chromosomes (Burke et al., 1987). Large fragments of exogenous DNA up to 2 Mb in length (Little, 1992) can be cloned faithfully into these vectors, although on average in the available chicken library they contain about 630 kb of inserted DNA (Toye et al., In Press).
YACs have been used in many cases to isolate genes, for example the Duchenne muscular dystrophy gene (Kunkel et al., 1985), where the gene is greater than 10⁶ bp in length. YACs have also been used for fine mapping around the murine Xist (X-inactive specific transcripts) sequence, which could not be achieved by interspecies backcross mapping (Heard et al., 1993). YAC clones for the Xic (X inactivation centre) region were isolated from the total genomic DNA library by probing with Xist. These isolates were then mapped by RFLP, giving localised marker loci, spanning a 1 Mb region. Construction of a chicken YAC library has been completed and this library is now being characterised in this laboratory (Toye et al., In Press).
Characterisation of YAC clones initially involves their alignment into a contig of a particular chromosomal region using hybridisation or PCR amplification of markers known to be present in that region. After verification of these results, end cloning of the YAC clones and their subsequent hybridisation can be used to check the contig alignment. Once a panel of YAC clones have been identified, these are subject to a number of techniques to refine the search for the candidate gene. This is initially achieved by targeting coding regions using CpG island mapping and exon trapping. CpG islands are short unmethylated CpG-rich sequences often positioned at the 5' end of vertebrate genes and can be identified by restriction mapping. They have been shown to be present in 14 out of 20 chicken genes selected at random from the EMBL DNA database (McQueen et al., 1996). Exon trapping utilises RNA splice sites to remove exons and insert them into vectors containing an intron. These constructs are transfected into COS7 cells and if the original splice sites can pair with those in the intron, mRNA is produced. This is then PCR amplified and cloned to isolate the original exon which can be investigated more fully (Buckler et al., 1991).
YACs are very important as cloning vectors because they can also be used for germ-line transmission of DNA into mouse embryonic stem (ES) cells. YAC clones containing the human hypoxanthine phosphoribosyltransferase (HPRT) gene (size 670 kb) were fused as yeast spheroplasts with an HPRT-deficient ES cell line and selected in HAT medium for the expression of HPRT. These transformed ES cells were then injected into mouse blastocysts and the subsequent chimaeric mice were shown to express the human HPRT gene (Jakobovits et al., 1993). Copy number dependent and position independent expression has also been shown by expression of the YAC derived tyrosinase gene in transgenic mice (Schedl et al., 1993). This technique has been adapted by Choi et al. (1993), by co-lipofection of the YAC clone into the ES cells in order to reduce the amount of yeast chromosomal DNA being introduced. Choi et al. (1993) also succeeded in creating chimaeric mice with a human heavy chain immunoglobulin gene which should lead to mice producing fully human monoclonal antibodies, a significant improvement on current technologies.
The use of YAC clones to produce transgenic animals has a further advantage as gene mutations can be engineered within the YAC by homologous recombination to create deletion mutants or knockouts. Constructs are designed, containing the mutant gene with an antibiotic resistance gene inserted in the middle, edged with antibiotic susceptibility genes. Once introduced into the yeast this construct will insert into the YAC. If homologous recombination has taken place the mutated gene with the antibiotic resistance will have inserted in the place of the original gene. In contrast, randomly inserted genes would still carry the ends encoding antibiotic susceptibility.
1.3.2.6 Sequence Tagged Sites (STS)
STS were proposed Olson et al. (1989) as an ideal way of integrating all markers into physical maps, by making sets of PCR-based markers. STS are 200-500 bp tracts of unique single-copy sequence which can be amplified by two 20-mer primers. Details of the primers, reaction conditions and product are available to all researchers, so any DNA clone can be tested for STS content. STS are particularly useful in the construction of YAC contigs, although they cannot be relied upon in isolation (reviewed by Ward and Davies, (1993)). However, Olson's original concept of an STS map, with a coverage of one marker per 100 kb of the human genome by 1995, has proved unrealistic. STS provide a relatively expensive procedure as they rely on the production of primer pairs and the sequencing of large numbers of clones. Although STS give some information about the genomic sequences they are not useful as anchor loci, and are only really useful in the characterisation of YAC or cosmid clones.
STS are currently being employed for sequence scanning in the pufferfish (Fugu rubripes) genome. This entails random sequencing of the well characterised cosmid library to give at least 50 STS per cosmid. The small size of this genome, 400-500 Mb means that there are a maximum of 8 genes per cosmid. This approach should lead to STS which are within the genes, and when compared to other genome databases these genes will be identified by homology. Once genes are identified, conservation of synteny can be examined (Elgar et al., 1996). Although the Fugu genome is considerably smaller than those of other vertebrates, including the chicken, it should still contain the same number of genes, estimated at 100,000. Thus randomly sequencing this genome should lead to the isolation of many genes, and comparative mapping will enable these to be mapped in other organisms.
1.3.2.7 Fluorescent In-Situ Hybridisation (FISH)
FISH enables the localisation of regions of DNA to a particular chromosome or chromosomes in metaphase or interphase cells following their detection by fluorescent label. Probes of any size can be used, from cDNAs or cosmids to YACs, but the bigger the clone the more fluorescence can be detected. The probe is usually labelled with biotin-16-dUTP and hybridised overnight to pre-treated slides of chromosome spreads. After washing at the required stringency, fluorescein-avidin is added and after a further washing biotinylated anti-avidin is added. This acts as a FISH "sandwich" to increase the number of binding sites for fluorescein-avidin, which is again added before the slides are sealed. The slides are viewed by fluorescence microscopy and the results photographed (Korenberg et al., 1992). A single copy gene should hybridise to both chromatids of an homologous pair of chromosomes, usually visualised as two fluorescent dots. However it is difficult to obtain good data from this technique in birds, due to the large number of chromosomes and the small amount of spread achieved with the metaphase chromosomes. It is also impossible to cytologically distinguish between the 29 pairs of MICs, so if a FISH probe hybridises to MIC, there is no way of deciding which chromosome it is. However, this technique is very useful for determining the chimaerism of the YAC libraries, and is being employed in our laboratory for this purpose. FISH can also show to which size of chromosome the probe hybridises.
1.4 Traits of interest in the chicken
Chickens exhibit a number of traits differing between the commercial lines available. These include associations of particular lines with increased productivity and disease resistance. This gives an initial starting point for the selection of birds in which a trait of interest is expressed.
1.4.1 Production Traits
The main production traits of interest in chickens are rapid weight gain and increased egg-lay. To this end, two types of commercial chickens are produced, broilers and egg-layers. These traits are often polygenic but may also be affected by other non-hereditary factors, which makes the identification of genes difficult. In order to map the genes controlling production traits, markers which are linked to the quantitative trait, termed Quantitative Trait Loci (QTL), are identified within a large population (about 1000 per sire). Quantitative traits are measurements of a phenotypic characteristic such as body weight, shank length (Dunnington et al., 1993) or abdominal fat deposition (Plotsky et al., 1993). The marker loci used are often derived from DNA fingerprints which are randomly distributed throughout the genome. QTLs can subsequently be used directly in the screening of commercial flocks for marker assisted selection (MAS), as they segregate with the trait of interest. The identification of genes controlling these complex production traits requires very large numbers of progeny and the QTL linkage is only as good as the number of markers in those regions. The final outcome of QTL mapping is usually identification of a number of highly significant areas of chromosomes which then need to be studied in greater detail using YAC contigs to identify putative candidate genes, for the trait under investigation.
1.4.2 Disease Traits
Chickens suffer from a number of diseases, some of which are fatal, all of which affect the growth, production and welfare of the birds. Thus disease resistance is an important economical consideration to both the poultry farmer and breeder. The most important diseases of chickens are those for which vaccination is now proving useless against very virulent strains and those which could affect man. Resistance to disease is fundamental to survival for all organisms. However, two species will not necessarily share the same mechanism of resistance to the same disease. Disease resistance will vary across a population, and is particularly clear in comparisons between the inbred chicken lines at the IAH, which have been studied by Bumstead et al. (1991). There are three basic mechanisms of genetic resistance: recognition of the pathogen by the immune system, under the control of the MHC complex; metabolic variants which affect the penetration or reproduction of the pathogen; and incorporation of pathogen DNA into the host genome which prevents infection.
1.4.2.1 The major histocompatibility complex (MHC)
A major element of disease resistance is the major histocompatibility complex (MHC) which in chickens lies on chromosome 16. Chromosome 16 is a MIC which has been shown by silver-staining to contain the nucleolar organiser region (NOR) containing the 5.8S, 18S and 28S ribosomal RNA genes (rDNA) for protein synthesis. Chickens only have one copy of rDNA consisting of a cluster of 145 40 kb repeats, which represents about 0.5% of the chicken genome (Bloom et al., 1993). The NOR was shown to exist on the same chromosome as the chicken MHC by trisomy mapping of a bird expressing B⁶, B¹³ and B¹⁵ antigens (Bloom and Bacon, 1985). Estimates of the size of the microchromosome range from 8 Mb upwards; however, the NOR occupies a large portion of the chromosome, perhaps as much as 6 Mb (Bloom and Bacon, 1985). The MHC region is therefore perhaps 2 Mb or 0.17% of the genome.
The MHC B complex in chickens was first identified by the ability of leukocytes to give strong graft rejection (Schierman and Nordskog, 1961). It was subsequently discovered to contain three loci, B-F, B-L and B-G (Pink et al., 1977), producing class I, II and IV antigens respectively. It has been shown that frequent recombination occurs between B-F and B-G regions, although none has been detected between B-F and B-L (Hála et al., 1976). Class I antigens are expressed on the surface of virtually all cells including leukocytes and erythrocytes, as a single transmembrane polypeptide chain folded into three extracellular domains associated non-covalently with [beta]₂ -microglobulin. Class I presents antigens for recognition by cytotoxic T cells. This region is homologous to HLA-A, B, and C genes in man. Class II antigens are expressed on antigen presenting cells (APC), B-cells and macrophages, as heterodimers with two non-variable immunoglobulin-like domains near the membrane and two variable domains, furthest from the cell. The class II antigen is recognised by T helper cells, which become activated to stimulate T cell proliferation. This region is homologous to the HLA-D genes in man. Class IV antigens are expressed on the cell surface of erythrocytes and are unique to avian species. They consist of two glycoproteins which are highly polymorphic, but of unknown function. However, they are distantly related to butyrophilin and myelin-oligodendrocyte glycoprotein (Kaufman et al., 1995). At present, no regions have been identified for class V loci (Sander, 1993). MHC haplotypes have traditionally been determined by serotyping each bird by haemagglutination of the B-G antigens (and in some cases B-F antigens), but give no information for the other MHC loci.
The chicken MHC region has now been more extensively explored and there are chromosome 16 linkage maps for both the Compton mapping reference population (Bumstead and Palyga, 1992) and the East Lansing mapping reference population (Levin et al., 1994b), as well as a resource of cosmid clusters (Guillemot et al., 1988). The class II genes are interspersed with the class I genes along a relatively small region of chromosome 16, the B-F/B-L region, represented in a cosmid cluster of 130 kb. There are at least 5 class II [beta] genes in total, two of which are known to be expressed and lie within the B-F/B-L region about 8 kb apart. The intron/exon structure for class II is similar to that in mammals, but with much smaller introns. The total gene sizes are less than 2 kb in comparison to 8-20 kb in man (Guillemot and Auffray, 1989). At present, only one class II [alpha] gene has been found (Kaufman et al., 1993), about 5 cM away from the B-F/BL region. There are four class I genes, two within the B-F/B-L region (Kaufman et al., 1993) which lie in a 20 kb segment of the chromosome. Between these two genes, in the B-F/B-L region, lies the TAP2 (Transporter associated with antigen processing-2) gene (Bumstead et al., 1994b). In the same B-F/B-L region are also located one of many B-G genes and the gene encoding the [beta] subunit of a GTP binding protein (Guillemot and Auffray, 1989). There are many B-G genes, a number of which are transcribed. However, their positions on the chromosome are not as yet defined, although one clone maps to the same linkage position as TAP2 (Mariani et al., In Press). Another gene mapping to this linkage position is the chicken homologue of G9a (BAT8), which is linked to the class III genes in man (Spike and Lamont, 1995). One of the few identified class III genes, C4, also lies within this region (Kaufman pers. com.).
There is a second MHC region, termed the Rfp-Y system (Briles et al., 1993), which was originally unlinked to the B-complex. This region contains two closely-linked class I and two class II genes (Miller et al., 1994a). Recent work (Miller et al., 1996) using trisomy mapping has shown that the Rfp-Y system contains two class I and three class II genes and lies on chromosome 16, in the same cosmid cluster as a class II MHC gene and the rDNA genes (NOR). Between the two class I [alpha] genes lies a lectin, type-C gene, Lec (Bernot et al., 1994).
Thus the MHC region in chickens is significantly smaller than that in mammals, although it appears to contain two regions of class I and class II genes, these are interspersed while the mammalian regions are separate.
The association of the MHC region, or genes associated with it, with disease resistance is shown in a number of diseases some of which are discussed below.
1.4.2.2 Salmonellosis
Salmonellosis is caused by the colonisation of the gut following ingestion of the Salmonella bacterium. This can lead to a systemic infection of the chicken with high morbidity and mortality by species such as S. typhimurium, S. gallinarum and S. pullorum. Alternatively chickens can be infected by S. enteriditis which causes little clinical infection in the birds, but is excreted into the eggs which when eaten uncooked by man can result in infection. At present control is by oral administration of antibiotics. Salmonellosis resistance in mice has been shown to be provided by the presence of a glycine at amino acid position 105 in a variant of the Nramp1 protein, formerly known as Ity (Vidal et al., 1993). The Nramp protein is expressed in macrophages as a putative transmembrane protein. However, although expressed in chickens, Nramp1 does not control resistance to Salmonellosis in the cross studied (Hu et al., 1996). Experiments have shown that chicken lines 6₁ ,WL and N were resistant to all serotypes tested, whereas lines 15I, 7₂ and C were susceptible (reviewed by Bumstead et al. (1991)). Crosses and backcrosses between resistant and susceptible birds showed that resistance is dominant and not linked to the sex or the MHC chromosomes, and is probably controlled by a single gene.
1.4.2.3 Infectious bursal disease virus
IBDV (Infectious bursal disease virus) is a birnavirus which mostly affects chicks 3-6 weeks old, causing lesions within the bursa, and subsequent immunosuppression. This usually leads to secondary infection by opportunistic pathogens, which causes increased suffering for the birds and great expense for the farmers. There are a number of different strains, some of which cause high mortality and are termed very virulent. Bacon (1987) found some resistance associated with the MHC; as haplotypes B² or B²¹ were more resistant than haplotypes B⁵,B¹⁵ or B¹². However, Bumstead et al. (1993) examined 11 inbred and partially inbred lines of chickens and determined by F₁ and F₂ crosses that resistance was partially dominant and did not involve the MHC. Although there is a vaccine against this virus, in 1989 a very virulent strain of the virus reached Britain and, despite vaccination with "hot" strains, has increased mortality by 1.5-2%. In 1994 the National Farmers' Union estimated that IBDV had cost the industry £15 million (Law and Payne, 1996). It would be very useful to identify genes for resistance to this disease as they could be used directly to screen birds for resistance or used to initiate a more informed approach to vaccine development.
1.4.2.4 Marek's disease virus
MDV (Marek's disease virus) is a herpes virus, causing a lymphoproliferative disease in chickens leading to paralysis and tumours. The initial disease phase is acute and occurs about 3 days PI involving virus replication in B lymphocytes, and as a consequence of an immune reaction to antigen particles the T lymphocytes become activated. These T lymphocytes disperse around the body giving a persistent viraemia. Some lymphocytes are transformed and proliferate as lymphomas in visceral organs. After two weeks a second cytolytic infection occurs in the feather follicle epithelia resulting in the shedding of cell-free virus into the environment (Payne, 1996). MDV can be controlled by administration of the Rispens vaccine or the HVT vaccine, which is a live, attenuated herpes-virus of turkeys. However, there are now a number of very virulent strains of MDV which are unaffected by vaccination and are increasingly causing high mortality, up to 80% in comparison to 10% associated with the classical from of MDV (Payne, 1996). In 1994 the National Farmers' Union estimated that MDV had cost the industry £3 million (Law and Payne, 1996), however this is a gross under-estimate.
The genetic differences within the B-complex have been associated with resistance to a number of diseases; the initial association of the MHC was with MDV resistance. The B²¹ haplotype has been shown by a number of experimenters to confer resistance to MDV (Hansen et al., 1967; Longnecker et al., 1976; Briles et al., 1977; Powell, 1984; Hedemand et al., 1993). However, the B¹⁹ and B² haplotypes tend to confer susceptibility. The B²¹ haplotype resistance is not associated with the bursa, but these birds have a higher virus-neutralising-antibody titre than susceptible birds which are immunosuppressed. At 7 days post-infection the virus titre falls in the blood and spleen whilst there is development of humoural antibody and cell-mediated immunity.
Non-MHC genes also have an effect on the bird's resistance to disease. This area has again been most intensively studied for MDV, where two lines of birds, from the Regional Poultry Research Laboratory (RPRL) line 6 and 7, have identical MHC haplotypes of B² , yet show very different resistances to MDV. Line 6 was shown by Stone (1969) to suffer from only 6% mortality in comparison to 100% of line 7 birds with MDV. It has been shown by thymic transplant from line 7 to thymectomised line 6 birds that lymphomas will occur in cells derived from line 7. However, no resistance is shown if the transplant is from line 6 to line 7. Transplants of bursa or spleen show no alteration in resistance, indicating that line 6 resistance is associated with its T cells (Powell et al., 1982). Line 6 birds show a high level of immune response which correlates with viral disappearance not seen in line 7 birds. It has also been shown that macrophages in uninfected line 6 birds have a higher phagocytic index than those of line 7 birds (Powell et al., 1983). The genes which control these differences between lines 6 and 7 are not fully understood, but a number of candidate genes have been suggested. These include Ly-4 (Payne, 1991), Bu-1, and Th-1 (Gilmour et al., 1976), alkaline phosphatase (Yotova et al., 1990), polyesterase (Bachev and Lalev, 1990), thermal regulatory genes (Yotova, 1988) and immune response and competence genes (Okado and Yamamoto, 1987). If the genes controlling resistance can be identified, breeders will be able to select genetically resistant stock and it should be possible to target vaccines or chemical prophylaxes to both the very virulent and normal forms of the virus. At present these candidate genes appear to have some effect on resistance but do not appear to be the major genes involved.
1.4.2.5 Newcastle disease
Newcastle disease, originally imported from the far east, is caused by a myxovirus which can infect chickens as well as turkeys and occasionally man. It has a 2-7 day incubation period culminating in lassitude of the birds, respiratory distress, diarrhoea and drooping wings. Egg production is greatly affected and there is high mortality in young chicks. Oral administration of live vaccine has significantly reduced the incidence of this disease.
1.4.2.6 Infectious bronchitis virus
IBV (Infectious bronchitis virus) is a highly contagious respiratory disease of chickens caused by a coronavirus. The respiratory infection reduces egg production and quality and can lead to secondary infections with other micro-organisms and thus death. This is controlled to some degree by vaccination. Resistance to IBV is also non-MHC derived. However, it does not appear to be controlled by a single gene as the levels of infection and organs affected are considerably different between the inbred lines of birds (Bumstead et al., 1991).
1.4.2.7 Avian leukosis and sarcoma virus
ALSV (Avian leukosis and sarcoma virus) are a group of retroviruses which are characterised by autonomous proliferation of blood cell precursors. They are acquired by infection with exogenous virus or vertical transmission from the hen. The infection leads to erythroblastosis at a few months of age or to lymphoid leukosis in the more mature birds. This can also result in tumours, as in Rous sarcoma virus infection. Resistance to infection induced by the MHC has also been observed for the ALVS group (reviewed by Schierman and Collins (1987)). Rous sarcoma virus-induced tumours have regressed dramatically in birds of haplotype B⁶ or B²;however, haplotypes B⁵ or B¹³ are susceptible. It has been determined that the genes influencing regression lie within the B-F region of the MHC and are also associated with inhibition of the development of metatastic tumours. There is a small amount of resistance to ALV shown in birds with B² haplotype whereas those with B⁵ haplotype have increased erythroblastosis and a higher incidence of haemangiocarcinomas. ALSV resistance is due to a lack of a cell surface receptor for the specific sub-group of virus reviewed by Bumstead et al. (1991). Recently the chicken gene tv-a, which confers susceptibility to infection by ALSV subgroup A, was isolated by Young et al. (1993).
1.4.2.8 Coccidiosis
Coccidiosis is caused by an apicomplexan protozoan, Eimeria, which inserts itself intra-cellularly into the gut wall, where it reproduces. It is spread by ingestion of contaminated faeces, and so is prevalent in litter-kept chickens. Infected chickens show a reduction in growth rate and feeding efficiency, which can lead to increased fatality. There has been some control of infection using attenuated vaccines such as Paracox, a vaccine based on an initial infection with small numbers of parasites which produces a strong protective immune response. MHC haplotype affects the resistance of birds to coccidiosis, where those with B²¹ haplotype are more resistant to Eimeria maxima than those with B¹⁵ haplotype. However this differs for the various species of Eimeria (Bumstead et al., 1991). Resistance to coccidiosis is also determined by non-MHC as well as MHC genes. Although these have not yet been defined, it has been shown that different lines of birds show differing resistance to coccidiosis. However, it again seems that the genes involved act differently for each parasite as, for example, those birds with resistance to E. maxima show susceptibility to E. tenella, (reviewed by Bumstead et al. (1991).
1.4.2.9 Fowl cholera
Fowl cholera is caused by the gram-negative, facultative aerobe Pasturella multocida. This disease has a sudden onset with fever leading to haemorrhagic septicaemia. Oral antibiotics can be administered, but this has little effect on those birds already showing the disease. Lamont et al. (1987) have also shown that MHC haplotype affects the resistance of birds to fowl cholera. Lines of birds homozygous for MHC haplotype and the F₁ and F₂ progeny were tested by intra-nasal administration of P. multocida. Homozygote B¹ haplotype birds demonstrated significantly higher resistance to the bacteria than B¹⁹homozygote birds.
It should be noted that resistance to different diseases associated with the MHC, is not generated by a single haplotype, thus suggesting an important role for non-MHC genes in the determination of resistance (Lamont, 1993).
1.5 Objectives
The genetic map of the chicken is now essentially complete, containing more than 800 loci at an average spacing of around 3 cM (Bumstead and Palyga, 1992; Levin et al., 1994b). A YAC library has now been constructed which contains 16,000 clones of an average insert size of 0.58 Mb giving a coverage equivalent to 10 copies of the genome (Toye et al., In Press). However, although it is now possible in principal to move from mapping information to the equivalent YAC clones, to do this effectively requires a means of efficiently increasing the density of markers in relevant regions of the map. Once this is achieved, it will be possible to isolate specific genes using the framework of the markers. In conjunction with the generation of markers, it is now possible to directly select for candidate genes, by positional cloning, which could code for traits of interest. Positional cloning is now used as the established approach to determine disease genes. It proceeds from an existing disease to identifying the responsible gene's chromosomal locus, then to recovering the gene, and finally to discovering the mutational alteration that accounts for the disease (Collins, 1992). Variants at a locus defined by a candidate gene could help to determine their effects on a trait of interest.
The aim of this project was to develop techniques for generating this high density of marker loci. Possible ways to approach this include adding known genes to the map whose position or possible function can be compared with that of other organisms; the addition of large numbers of markers to the map, by saturation or random mapping leading to enough markers such that there is one every centi-Morgan; and targeting markers to regions of the genome which contain genes of interest, in order to increase the marker density just within that region.
1.6 Approaches
In order to isolate the genes responsible for inherited traits it will be necessary to clone the chromosomal region containing the trait locus. In practice, this means isolating YAC clones containing the marker loci used to map the trait gene. In order to do this, refined mapping of the chicken genome using a large number of markers is needed. Three major areas of investigation have been undertaken:
i. Isolation of cDNAs which might code for candidate traits by ESTs.
ii. Total saturation of the genome with large numbers of anonymous markers by RAPD.
iii. Targeted isolation of markers, which could include candidate genes, in specific regions by RDA.
1.6.1 Expressed Sequence Tags (EST)
ESTs of genes have been used as an effective way of generating further loci on genomic maps which provide useful genetic information rather than just being anonymous markers. ESTs are the sequences generated from cDNA clones and act as type I anchor loci within the linkage map (O'Brien and Graves, 1991). The chance of cloning a candidate gene can to some extent be controlled by the choice of tissue for mRNA extraction which will generate the ESTs. Candidate genes can be isolated as ESTs and homology with known genes in other species identified and the inference of possible association to traits pursued. Comparative mapping of ESTs can determine possible regions in which a gene of interest may lie, as it is already mapped in that area in a different species (Adams et al., 1991). However, beyond using a specific tissue for mRNA extraction, there is no other method for targeting the required candidate gene.
In order to genetically map ESTs in chickens it was necessary to find polymorphisms between the parents of the Compton reference mapping population, this is usually achieved by RFLP analysis. However, this is time consuming and not always productive. Ultimately with the construction of a WG-RH library all chicken ESTs will be mapped, as there is no need for polymorphisms between the two lines. Using ESTs derived either from tissue known to express the gene product of interest, or from comparative mapping synteny, it is unlikely that the candidate gene will be isolated, but this will at least lead to further placement of informative loci on the chicken linkage map.
1.6.2 Random Amplified Polymorphic DNA markers (RAPD)
RAPD analysis should lead to the saturation of the genome with markers to a level of at least one marker every 0.5 cM, without the requirement of previous genetic information and using few expensive oligonucleotides. RAPD markers are usually generated by the amplification of random DNA segments with single short arbitrary primers. An oligo is used as both forward and reverse primer in PCR amplification, at a low annealing temperature (in the region of 36°C) (Williams et al., 1990). This procedure generates large numbers of bands for each individual's genomic DNA (resolved by gel electrophoresis) and may identify multiple polymorphisms between individuals. Polymorphisms detected by RAPD are due to point mutations, which affect the binding of the primer to the DNA and also to insertions and deletions between primer binding sites. Although these markers are anonymous, once a large number have been generated, high density genetic maps can be constructed. This system should rapidly generate large numbers of closely spaced markers. Each primer should be capable of detecting a number of polymorphisms leading to the identification of multiple informative loci for a small amount of initial resource. The benefits of RAPD technology include its simplicity and ease of use in the laboratory. It does not require any previous sequence data, but will give easily interpreted results. It is also less costly in time than other methods, for example RFLP. However, changes in enzyme manufacturer, primer or enzyme concentration, or thermal cycling equipment, can give inconsistent results (MacPherson et al., 1993; Meunier and Grimont, 1993). RAPD markers have been used for gene cloning, medical diagnostics and trait introgression in breeding programs (Williams et al., 1990). Levin et al. (1993) used RAPD markers to generate new markers on the Z chromosome of the chicken in order to identify sex-linked traits. RAPD markers have also been useful in determining phylogenetic relationships between species as demonstrated by Barral et al. (1993) with the Shistosoma genome.
1.6.3 Representational Difference Analysis
Representational difference analysis (RDA) uses subtractive and kinetic enrichments to purify restriction endonuclease fragments present in one population of DNA fragments but not in another (Lisitsyn et al., 1993). It does not rely on knowledge of the sequences involved but can be used to compare complex DNAs elucidating the sequence differences between them. In comparison to subtractive hybridisation, this method is far quicker and more successful, as the complexity of DNA is first reduced by the representation step, and the regions of interest are enriched. A representation of the genome is made by digestion with a specific endonuclease, ligation of enzyme-specific initial adaptors (termed R) and amplification of fragments by PCR. This generates a pool of fragments referred to as amplicons. Two species of DNA are used; DNA containing the target sequence, the "tester" and the non-target-containing DNA, the "driver". The adaptors are subsequently removed from the amplicons, and new adaptors (termed J) ligated only to the tester amplicons which are then mixed, melted and reannealed with excess non-target DNA (driver) amplicons. During treatment with Taq DNA polymerase, only those amplicons with adaptors can be filled in at the 3' ends. When the samples are amplified by PCR, exponential amplification only occurs with the tester homoduplex DNA. Single-stranded DNA is not amplified, and any heteroduplexes are only amplified linearly. This hybridisation and amplification step can be repeated many times alternating between two sets of adaptors (termed N and J) to greatly increase the enrichment of the target DNA. Disadvantages of this system mainly concern the resource of target and non-target DNA, which particularly in some experiments, should come from the same individual, or at least a related individual.
More recently, Lisitsyn et al. (1994) have refined this technique by using either congenic strains or two-generation crosses to generate genetic markers linked to a trait of interest. They used Genetically Directed Representational Difference Analysis (GDRDA) to target three polymorphisms to within less than 1 cM of the mouse nude locus on chromosome 1. Thus, RDA can be used to isolate candidate genes directly or to target specific regions of genomic maps, in order saturate these regions with markers for ease of YAC contig alignment.
RDA has been used to isolate a number of candidate genes from a variety of diseases which may be caused by viruses, including the identification of human herpes virus 6 from patients with multiple sclerosis (Challoner et al., 1995), a flavivirus from patients with hepatitis strain GB (Simons et al., 1995) and a herpes-like virus from AIDS patients with Kaposi's sarcoma (Chang et al., 1994). RDA has also been applied to the isolation of candidate genes for different cancers; Schutte et al. (1995) identified a tumour suppressor gene deleted in pancreatic carcinoma, while Hino et al. (1995) found 4 candidate genes for renal carcinomas in the Eker rat. However when Tsuchiya et al. (1994) tried to use RDA to isolate genes specific to the Yoshida sarcoma (YS), not expressed in the long term survival Yoshida variant, they discovered that the two were not as closely related as had been believed and generated nine genes specific to YS.
RDA has also been adapted to the isolation of candidate differentiation sequences from tissue using cDNA, initially by finding the two genes RAG-1 and RAG-2 which are known to be responsible for activating V(D)J recombination (Hubank and Schatz, 1994). More recently the technique has been applied to the isolation of sequences present on a single chromosome. This has been successful in isolating sequences specific to the bovine-Y chromosome (Wigger et al., 1996), mouse chromosome 10 (Baldocchi et al., 1996) and wheat chromosome 6 (Delaney et al., 1995).
RDA is a powerful tool which can be used to isolate sequences on specific chromosomes, candidate genes and previously undetected viruses. This technique depends on a number of enzymatic steps, which could lead to failure. However, with the decrease in DNA complexity and enrichment of target sequences, RDA could also prove highly successful for chickens as well as mammals and plants.

2.1 Materials
Chemicals, stock solutions, media, oligonucleotides and the thermal profiles used for PCR reactions are detailed in Appendix A.
2.2 Bacterial strains and plasmids
Details of bacteria and plasmids used are shown in Table 2.1.
Table 2.1 Descriptions of bacterial strains and plasmid.

2.2.1 cDNA Library IAHchB1
The cDNA library, IAHchB1 in pCDM8, was used as a source of clones for sequencing and mapping. The library was constructed as detailed in Tregaskes and Young (1997), from RNA extracted from a 17-day embryonic whole bursa and has insert sizes of 0.5 to 3.5 kb. Thus any random clone could contain part or all of any gene expressed in the bursa. This library would therefore include ubiquitously expressed genes, as well as those specific to the tissue. The Bursa of Fabricius is a major organ in the immunological system of the bird mainly concerned with the multiplication and development of B lymphocytes, which are involved in the control of disease through antibody production. Thus expressed genes specific to this tissue could be of great interest as they could code for proteins associated with the immune system or the development of B lymphocytes. Any of these genes could code for proteins involved in host resistance, and so should be fully characterised.
2.2.2 Transformation with IAHchB1 by electroporation of Escherichia coli MC1061/P3
A single colony of Escherichia coli MC1061/P3 was picked and grown up overnight in 5 ml L-broth (160 rpm, 37°C). Two ml of this overnight culture were used to seed 200 ml of L-broth which was then incubated (160 rpm, 37°C) until the OD₆₀₀ measured 0.5. The culture was placed on ice for 30 min, with occasional swirling, and then transferred to a pre-cooled sterile 250 ml centrifuge bottle. The cells were pelleted by centrifugation at 2500 g for 15 min. The supernatant was discarded, and the pellet resuspended in 200 ml pre-cooled (4°C) sterile distilled water. The cells were replaced on ice for 30 min, and then centrifuged as before. The cell pellet was resuspended in 100 ml cold sterile distilled water, and returned to the ice for another 30 min. The cells were centrifuged as before, the supernatant removed and the pellet resuspended in 5 ml of pre-cooled sterile filtered (using a 0.2 µm pore filter) 10% glycerol. This cell suspension was transferred to a 50 ml Falcon tube and left on ice for 30 min. The cells were then centrifuged at 700 g, 15 min. The supernatant was removed and the pellet resuspended in 500 µl of 10% glycerol and stored on ice until used.
Aliquots of 100 µl of competent cells were added to varying amounts (10 ng, 50 ng and 100 ng) of the IAHchB1 library plasmid DNA ( initial concentration 500 µg/ml) in pre-cooled tubes and electroporated, using a Biorad Gene Pulser apparatus with pulse controller at 2.5 kV, 200 [Omega] and 25 mF. One ml of SOC medium was immediately added to the cuvette, and the entire contents were transferred to a 5 ml Sterilin tube containing 1.5 ml SOC. The tubes were incubated (160 rpm, 37°C) for one hour before plating onto pre-warmed L-agar plates, containing 7.5 µg/ml tetracycline and 12.5 µg/ml ampicillin. The plates were incubated overnight at 37°C. Colonies were selected at random, and preparations of plasmids made from these.
2.2.3 Cloning RDA products into pGEM-T vector
All Representational Difference Analysis (RDA) products were cloned into the pGEM-T (Promega) TA cloning vector following the manufacturer's protocol. Approximately 8 ng of un-purified RDA re-amplification product was ligated to 50 ng of pGEM-T at 15°C for 3 hours. Once ligated, 2 µl (11.6 ng) of the product was transformed into 50 µl of E. coli JM109 competent cells (Promega). Initially, the tubes were kept on ice for 20 min before heat-shock treatment at 42°C for 45 sec, and then returned to ice for 2 min. 900 µl of SOC broth was added with subsequent incubation with agitation at 37°C for one hour. The cells were then plated onto L-agar plates containing 100 µg/ml Amp, 40 µg/ml X-GAL and 46 µg/ml IPTG. After incubating the plates overnight at 37°C they were placed at 4°C for one hour to enhance blue/white colony identification.
2.3 Chicken Strains and Crosses
The inbred lines of chickens used for all the experiments are white leghorn chickens kept at Compton. The individual lines described here, lines N, 15I, 6₁ and 7₂ were all imported from the Regional poultry research laboratory (RPRL), East Lansing. The coefficients of inbreeding for all the lines used are detailed in App. A. All the inbred lines produce two generations a year, the female siblings are artificially inseminated by pooled semen samples from the male siblings at two weeks after the start of lay, in order to synchronise the flock production. Line 15I has been closed since its importation in 1962 and had been inbred for 17 generations by 1978 . Lines 7₂ and 6₁ have been closed since their importation in 1972 and had been inbred for 7 generations by 1978 . Line N has been closed since its importation in 1982, it is not known how many generations of inbreeding have occurred.
Chickens from the Compton Mapping Reference Population were derived from a backcross from line N and line 15I birds, as described by Bumstead and Palyga (1992). These lines are inbred and differ in their resistance to salmonellosis, line N being resistant, and line 15I susceptible. One of the F₁ progeny (B989), a line N x line 15I female, was backcrossed to a line 15I male parent (B988). This mating produced 113 backcross progeny, which have been extensively used to construct a genetic map in this population.
A second population of birds, the Marek's Disease (MD) reference population, has been established using line 6₁ and line 7₂ . These two White Leghorn lines show differences in their susceptibility to MD: line 6₁ is resistant to MD, whilst line 7₂ is susceptible. In this population F₁ birds were backcrossed to the line 7₂ parent to give 85 progeny for the generation of a new linkage map and the mapping of resistance. F₁ birds were also mated to give an F₂ population of 42 birds for further analysis of disease resistance. All the birds used were infected with MD virus and the infection characterised in terms of viraemia by quantitative PCR and plaque assays, as well as mortality and tumour development
Quantification of Marek's disease virus by PCR has been developed (Bumstead et.al. In press). The PCR assay used two MDV specific primers to amplify a 279 bp product, which was fluorescently labeled to enable quantifiable detection by ABI Genescan gel electrophoresis. Primers to chicken interferon were used as a control assay for cell number and PCR effectiveness. The PCR assay was calibrated and its linear range determined by comparing the amount of PCR product for a range of dilutions of an MDV-transformed cell line. The relationship of PCR value to the number of viral genomes was also determined by applying the PCR assay to a series of dilutions of a plasmid containing the template sequence. This was carried out in addition to the usual viral plaque assay, with which it correlated well. The PCR assay was then tested on samples of lymphocytes from pure line 6₁ and 7₂ birds as well as F₂ birds of a cross between these two lines. The PCR assay gave repeatable and consistent results which in fact correlated better with the fate of the birds than did the plaque assay results.
There is no data for the phenotypes of the F₁ birds with respect to Marek's disease resistance, as all available birds were used to generated the F₂ population. However, the phenotype data for the pure lines and the backcross birds is shown in App. A. Approximately half the backcross birds show some resistance to Marek's disease according to the quantitative PCR values.
2.4 DNA Preparation
2.4.1 Genomic DNA Extraction
DNA was extracted from 0.5 ml of chicken blood collected from the wing vein using 3% sodium citrate as an anti-coagulant. The blood was treated with 1.5 ml of 1% saponin in PBS, mixed and left at room temperature (RT) for 10 min before washing with PBS. The nuclei were then pelleted by centrifugation at 1000 g for 5 min. This was repeated twice to ensure that the nuclei were free of cytoplasmic contamination. After centrifugation, the pellet was resuspended in 2 ml TE with 10% SDS. Protein and RNA were removed from the pellets by proteinase K digestion (20 mg/ml) for two hours at RT, followed by addition of RNase A (100 mg/ml) and mixed by rotation at RT overnight. The samples were phenol-extracted twice and chloroform:iso-amyl alcohol extracted once, before precipitation with 0.5 ml 2 M sodium acetate and 8.75 ml ethanol at -20°C overnight. The DNA was pelleted by centrifugation at 7800 g for 15 min, washed with 2 ml 70 % ethanol, air-dried and resuspended in 2 ml TE.
2.4.2 Plasmid DNA Preparation-ABI Protocol
Plasmid DNA was prepared using the following modified mini alkaline-lysis/PEG precipitation procedure. Overnight cultures of 5 ml of each clone were grown up in Terrific broth containing 7.5 µg/ml tetracycline and 12.5 µg/ml ampicillin. After removal of 850 µl for glycerol stock preparation, the remaining culture was pelleted by centrifugation at 1000 g for 3 min. Glycerol stocks were made by the addition of 150 µl glycerol to the removed cultures, which were then vortexed briefly and snap frozen in a dry-ice and ethanol mixture before storage at -70°C. The cell pellet was resuspended in 200 µl of GTE buffer, and lysed by addition of 300 µl of 0.2 N NaOH and 1% SDS. The tube was inverted and incubated on ice for 5 min. The solution was then neutralised with 300 µl of 3 M potassium acetate (pH 4.8) and inverted, before being placed on ice for 5 min. Cellular debris was pelleted by centrifugation for 10 min at 7000 g. The supernatant was removed and digested with 20 µg/ml RNase A at 37°C for 20 min and then extracted twice with 400 µl chloroform. The extracted DNA was precipitated with an equal volume of isopropanol at RT and immediately pelleted by centrifugation for 10 min at 7000g. After washing with 70% ethanol, the pellet was dissolved in 32 µl Milli-Q water and further precipitated by the addition of 8 µl 4 M NaCl and 40 µl 13% PEG₈₀₀₀, with incubation on ice for 20 min. The DNA was pelleted by centrifugation at 4°C for 15 min. After washing in 70% ethanol the pellet was dried and resuspended in deionised water at 1 µg/µl. This preparation was stored at -20°C.
2.4.3 Plasmid DNA Preparation-Hybaid maxiprep Protocol
Individual bacterial colonies were picked into two 10 ml of cultures of L-Broth, containing 100 µg/ml ampicillin and incubated with agitation overnight at 37°C. The following day 850 µl of each was removed and a glycerol stock prepared as previously described. The rest of the culture was centrifuged for 5 min at 1000 g to pellet the cells before resuspension in 250 µl Milli-Q water per tube. The combined 500 µl was then placed in a 1.5 ml microfuge tube and spun at 7000 g for 1 min, the supernatant removed and the tube re-spun to remove any remaining liquid. The pellet was then resuspended in 200 µl of pre-lysis solution and vortexed before addition of 400 µl of alkaline lysis solution. The tubes were inverted 15 times and left to stand at RT for 5 min before addition of 300 µl of ice-cold neutralising solution and re-vortexing for 5 sec. The tubes were left on ice for 5 min and then centrifuged for 5 min at 7000 g. The supernatant was transferred to a clean 2 ml tube and 900 µl of Hybaid binding buffer (containing guanidine thiocyanate) added. The tubes were again inverted 15 times and left to stand for 3 min at RT. The upper aqueous layer was discarded and the binding beads transferred to the spin filter and centrifuged for 3 min at 7000 g. The eluent was removed and 500 µl wash solution added to the spin filter before centrifugation as before. The eluent was again removed and the filter centrifuged for a final 5 min to ensure all liquid was removed. The filter was then transferred to a new tube and 100 µl TE was carefully stirred into the beads to elute the DNA followed by centrifugation for 5 min at 7000 g. This gave an approximate DNA concentration of 0.4 µg/µl.
2.5 Restriction Endonuclease Digestion
DNA samples were digested with various restriction endonucleases according to the manufacturer's instructions (of 10 U of enzyme/µg of DNA). See App. A for details of enzymes used.
2.6 Electrophoresis
2.6.1 Agarose Gel Electrophoresis
Plasmid DNA and PCR products were resolved by electrophoresis at 4 V/cm for one hour in 1 x TBE, 1% agarose gels with 10 ng/ml of ethidium bromide. Genomic DNA digests were resolved by electrophoresis at 1 V/cm overnight in 1 x TAE, 0.6% agarose gels with 30 ng/ml of ethidium bromide. Bands were visualised using a UV transilluminator, and the resulting pattern recorded using a GDS5000 digital camera system (Mitsubishi).
2.6.2 ABI Genescan electrophoresis
2.6.2.1 Creation of a Matrix File
The ABI Genescan system enables the electrophoresis and detection of fluorescent products. A matrix file was made so that background fluorescence could be reduced in the Genescan gel to enable good resolution of the fluorescent products. Prior to using the Genescan for analysis, the matrix file was created to ensure optimum fluorescent detection. An identical gel to those used for Genescan analysis, a 12 cm, 5.5 % "Long ranger", 1 x TBE gel, was set up and run for 5 ¼ hours at 800 V, on filter set B ([lambda]= 531, 545, 560, 580 nm) in the ABI electrophoresis apparatus. The dye primer matrix standard kit P/N 401114 was used, loading two samples for each dye colour. The samples were loaded, 5 µl per lane, alongside samples of 0.5 µl of Genescan ROX 2500 standard which were either native, or denatured with the addition of an equal amount of deionised formamide and heated before loading.
Raw data was collected using the Genescan (ABI) collection and analysis programs. The matrix file was constructed according to ABI instructions. Once the gel had run, the tracking was checked and a new matrix created for all four dye colours. This matrix was entitled "5.5% LR 1 x TBE Matrix" and was used in all subsequent Genescan reactions.
2.6.2.2 RAPD-PCR Electrophoresis
All the RAPD-PCR samples were denatured with the addition of an equal volume of deionised formamide, mixed with 0.5 µl of Genescan ROX 2500 standard, heated to 94°C for 2 min and kept on ice until loading. The samples were loaded on a 12 cm, 5.5 % Long ranger gel which was run in the ABI electrophoresis apparatus for 6-8 hours at 800 V on filter set B. The Genescan collection software was set up according to ABI instructions.
2.6.3 Sequence Gel Electrophoresis
Samples were loaded onto a 6% polyacrylamide sequencing gel (ABI 373A Sequencer), and electrophoresed in 1 x TBE for 14 hours at 2500 V. The apparatus was set up according to ABI protocols using Filter set A ([lambda]= 531, 560, 580, 610 nm) and the data collected using the 373A collection package.
2.7 PCR Amplification
2.7.1 Plasmid insert PCR Amplification
PCR amplification was carried out using diluted plasmid DNA (1:10) approximately 100 ng, in a 50 µl PCR reaction mix of 10 x buffer, 100 µM dNTPs, 2 mM MgCl₂, 25 pmol of each primer (oligonucleotides 12 and 16 for pCDM8 and oligonucleotides 147 and 151 for pGEM-T, see App.A). The reaction mix was overlaid with 50 µl of mineral oil before the addition of 2.5 U Taq DNA polymerase after the tubes had reached the denaturation temperature. Amplification was for 25 cycles with an annealing temperature of 55°C. The thermal profile in program 1 was followed (App. A).
2.7.2 Colony PCR of RDA products
Colonies were picked and screened for inserts by colony PCR. Each colony was picked directly into a 50 µl PCR reaction mix of 10 x buffer, 100 µM dNTPs, 2 mM MgCl₂, 1U Taq DNA polymerase, 25 pmol of each primer (147 and 151), and amplified for 25 cycles with an annealing temperature of 55°C. PCR products were analysed by agarose gel electrophoresis. Each picked colony was also streaked onto L-agar plates, containing 100 µg/ml ampicillin, in an ordered array and also inoculated into 1 ml of L-Broth, containing 100 µg/ml ampicillin, for incubation with agitation at 37°C overnight. The overnight cultures were spun down at 7000 g for 1 min and resuspended in L-Broth with 20% glycerol, followed by snap-freezing in ethanol and dry-ice, and storage at -70°C.
2.8 Gel purification of PCR products
Clones containing two plasmids were initially PCR amplified by colony PCR described above using primers 147 and 151. A 40 µl sample of each product was separated by electrophoresis on a 2% low melting point agarose gel. The insert DNA was excised as a band from the gel and placed into a 0.5 ml tube. Once the weight was obtained, 3 M NaCl was added to a final concentration of 30 mM and the agarose melted at 68°C for 10 min. The temperature was then reduced to 39°C and agarase (Sigma) added to a concentration of 50 U/ml. After incubation for one hour 60 µl of water was added along with 100 µl 4 M ammonium acetate and 200 µl propan-2-ol. The tubes were centrifuged for 30 min at 7000 g, the supernatant discarded and the pellet washed with 70% ethanol. The DNA pellet was resuspended in 40 µl of water and reamplified as detailed in the pGEM-T sequencing protocol.
2.9 Automated Sequencing
Clones were partially sequenced, from each end of the insert, using the Applied Biosystems 373A DNA automated sequencing system. Dye-deoxy terminator chemistry was used to label the products of Taq DNA polymerase cycle sequencing. Two different ABI kits were used as the technology was developed from the PRISM^TM ready reaction dyedeoxy^TM terminator cycle sequencing kit to the PRISM^TM ready reaction dyedeoxy^TM terminator cycle sequencing FS kit.
The PRISM^TM ready reaction dyedeoxy^TM terminator cycle sequencing kit was used following the ABI protocol to sequence clones in pCDM8. One microgram of double-stranded DNA template was used in each 20 µl cycle sequencing reaction. Oligonucleotides 12 and 16 (see App. A) were used as primers for the reaction. In each sequencing reaction 1 µl of cDNA was mixed with 9.5 µl of terminator premix, 3 µl of one primer and Milli-Q water up to a total volume of 20 µl. This was then overlaid with 50 µl mineral oil and placed into a Hybaid Omnigene PCR machine. The thermal cycling reaction was as described in App. A for ABI Seq. The samples were then extracted with a phenol: H₂O:chloroform mix (68:18:14) at room temperature, and precipitated with 2M sodium acetate and ethanol, to remove excess dye.
Clones in pGEM-T were sequenced using ABI PRISM^TM dyedeoxy^TM terminator cycle sequencing ready reaction FS kit (Perkin Elmer). The protocol was similar to that used previously except that only 8 µl of reaction mix was added. To obtain 1 µg of double stranded DNA each clone was initially PCR amplified in a standard reaction, using primers 147 and 151 before precipitation of the product with an equal volume of 4 M ammonium acetate and propan-2-ol at RT. This precipitated product was centrifuged for 30 min at 7000 g, resuspended in 5 µl Milli-Q water and used directly for sequencing. The PCR product was sequenced using 7 pmol of one of two nested primers: oligonucleotides 156 and 155. Each 20 µl reaction was precipitated on ice for 10 min with the addition of 2 µl of 3 M sodium acetate (pH 4.6) and 50 µl 95% ethanol, centrifuged at 7000 g, and air-dried. The pellet was resuspended in 4 µl of a 5:1 deionised formamide: 50 mM EDTA mixture and denatured at 90°C for 2 min. These samples were loaded and analysed as described earlier.
2.9.1 Sequence Analysis
The 373A software enables direct determination of the sequence. Sequences were edited using SeqEd (ABI) and converted into GCG format. Previously sequenced genes with homology to the clones screened in this project were identified by GCG (Program manual for the Wisconsin package, version 8, September 1994, Genetics Computer Group, 575 Science Drive, Madison, Wisconsin, USA 53711) FASTA analysis, in the EMBL database.
2.10 Southern Blotting
Agarose gels were blotted using a method based on that of Southern (1975). Gels were washed in 500 ml blotting solution A (1.5 M NaCl, 0.5 M NaOH) twice for 30 min and then 500 ml of blotting solution B (0.02 M NaOH, 1 M CH₃COONH₄) twice for 30 min. The gel was placed on a 3 mm filter paper wick on a platform over a reservoir of 1 l of blotting solution B. Hybond-N (Amersham) membrane was carefully laid over the gel, followed by three layers of filter paper. A stack of paper towels were placed over this, and held in position with a weight. Blotting took place overnight. The filter paper and towels were then removed, the nylon membrane marked and air-dried before baking under vacuum for two hours at 80°C.
2.10.1 Probe preparation and hybridisation
Southern blots were soaked overnight in Milli-Q water, and placed into Hybaid tubes. Blots of less than 30 cm² were probed in 20 ml plastic universals. Prehybridisation buffer (50% formamide) was added at 42°C, 0.02 ml/cm² for blots in Hybaid tubes and 0.25 ml/cm² for blots in 20 ml plastic universals. Three universals were placed into each Hybaid tube during hybridisation. The tubes were placed in a preheated (42°C) Hybaid oven and rotated overnight. Two methods of probe preparation were used; Nick translation (Gibco BRL) and Prime-It RmT random primer labelling kit (Stratagene). Probes were labelled with [[alpha]-³²P] dCTP (3000 Ci/mmol) (Dupont NEN), 10 µCi/µl (0.37 MBq/µl).
2.10.2 Nick translation system (Gibco BRL)
Blots were probed either with labelled PCR products (4 µl) or labelled whole plasmids (1 µl) containing about 1 µg DNA. The following basic protocol was used; a mixture of 2.5 µl solution A, 19.5 µl water, 2 µl of [[alpha]-³²