differences in gene density on chicken macrochromosomes and microchromosomes
TRANSCRIPT
Differences in gene density on chickenmacrochromosomes and microchromosomesJ Smith, C K Bruley, I R Paton, I Dunn, C T Jones, D Windsor, D R Morrice,A S Law, J Masabanda, A Sazanov, D Waddington, R Fries, D W Burt
Summary
The chicken karyotype comprises six pairs of
large macrochromosomes and 33 pairs of smal-
ler microchromosomes1. Cytogenetic evidence
suggests that microchromosomes may be more
gene-dense than macrochromosomes. In this
paper, we compare the gene densities on
macrochromosomes and microchromosomes
based on sequence sampling of cloned genomic
DNA, and from the distribution of genes
mapped by genetic linkage and physical map-
ping. From these different approaches we
estimate that microchromosomes are twice as
gene-dense as macrochromosomes and show
that sequence sampling is an effective means of
gene discovery in the chicken. Using this
method we have also detected a conserved
linkage between the genes for serotonin 1D
receptor (HTR1D) and the platelet-activating
factor receptor protein gene (PTAFR) on chicken
chromosome 5 and human chromosome 1p34.3.
Taken together with its advantages as an
experimental animal, and public access to
genetic and physical mapping resources, the
chicken is a useful model genome for studies on
the structure, function and evolution of the
vertebrate genome.
Keywords: gene-density, microchromosomes,
sequence-sampling, chicken, gene-mapping,
gene-discovery
Introduction
The chicken karyotype consists of 39 chromo-
some pairs, six large macrochromosomes (see
footnotes) (MACs) and 33 smaller microchro-
mosomes (MICs) (Pollock & Fechheimer 1976;
Fechheimer 1990). Microchromosomes are
found in all birds (TegelstroÈm & Ryttman
1981) and some reptiles (Hinegardner 1976).
The importance of the microchromosomes has
long been debated, but we now know that they
are fully functional chromosomes with centro-
meres and telomeres (Rodionov 1996), contain
genes and show conserved synteny with other
species (Jones et al. 1997). The microchromo-
somes constitute 30% of the total chicken
genome (Smith & Burt 1998), are GC-rich
(Auer et al. 1987) and have a higher CpG
content than the macrochromosomes (McQu-
een et al. 1996). Since 60±70% of known
chicken genes are associated with a CpG-
island, the microchromosomes may represent
a gene-dense fraction of the chicken genome.
This idea is supported by acetylation studies of
histone H4 and analysis of CpG-island-like
sequences in cloned genomic DNA (McQueen
et al. 1998), which indicate that microchromo-
somes are also associated with higher gene
activity.
In this paper we compare the gene densities
on the macrochromosomes and microchromo-
somes based on sequence sampling of cloned
genomic DNA and from the distribution of
genes mapped by genetic linkage and physical
mapping. From these different approaches we
estimate that the microchromosomes are twice
as gene-dense as macrochromosomes. We also
show that sequence sampling can be used as a
means of gene discovery in the relatively
compact chicken genome.
Materials and methods
Isolation of cosmid clones and preparation of
random subclones for DNA sequencing
For sequence sampling, cosmid clones were
picked at random from a commercially avail-
able chicken genomic library (Clontech, Cali-
fornia) and DNA was prepared using Qiagen
purification columns (Qiagen GmbH, Ger-
many). Cosmid clones were found to have
an average insert size of 35 kb. 5 mg cosmid
Animal Genetics,
2000, 31, 96±103
J SmithC K BruleyI R PatonI DunnC T JonesD WindsorD R MorriceA S LawD WaddingtonD W BurtDivision of Molecular
Biology, Roslin Institute(Edinburgh), Roslin,
Midlothian EH25 9PS,
UK
J MasabandaA SazanovR FriesLehrstuhl fur Tierzucht,
TUM, Freising, Weihen-stephan, Munich D-
85350, Germany
Correspondence: Dr David W. Burt.
E-mail: [email protected] are usually classed as chromo-
somes 1±8. However, for this study, we have called
chromosomes 1±6 macrochromosomes and the re-
mainder microchromosomes, as this was the classifi-
cation used by McQueen et al. (1996). We have used
this numbering system here so that direct comparisons
could be made with this earlier work (McQueen et al.
1996, 1998).
ã 2000 International Society for Animal Genetics 96
DNA was partially digested with the restric-
tion enzyme CviJI (Cambio), which cuts at the
sequence PuG/CPy. Digestion was carried out
at 37°C for 1 h using 0.2 U enzyme mg±1
DNA. Fragments in the size range of 1±3 kb
were excised from an agarose gel and recov-
ered using a Qiaex II DNA extraction kit
(Qiagen GmbH, Germany). DNA fragments
were subcloned into EcoRV digested pBlue-
scriptSK+ (Stratagene) and plasmid DNA pre-
pared using the Qiagen 48-well system
(Qiagen GmbH, Germany). Gene specific cos-
mids were isolated as described by Buitkamp
et al. (1998).
DNA sequencing of random subclones
Each subclone was sequenced from both DNA
strands using T3 and T7 primers. Sequencing
reactions were carried out using a dye termi-
nator kit (Applied Biosystems, Perkin Elmer,
UK) and separated on an ABI 377 automated
sequencer. Typically, 500±600 nucleotides
were obtained from each primer. Following
sequencing, vector and poor quality sequence
data were removed. Contig assembly was
carried out using the Seqman program (DNAS-
tar). Searches for gene homologies were per-
formed using the GCG (version 8.0)
implementation of the BLAST algorithm
(Altschul et al. 1990). A database `hit' was
deemed to be significant if the probability and
high scores were better than 10±6 and 150,
respectively, for nucleotide searches and better
than 10±3 and 75, respectively, for protein
searches.
Fluorescence in situ hybridisation and physical
mapping of genes
Carried out as described in Smith et al.
(1999).
Genetic mapping of chicken genes
Sequences of chicken genes were obtained from
EMBL or GenBank and named using the human
gene symbol. For loci based on random cDNA
sequences, the lab symbol with an `E' postfix
was utilised, e.g. ROS0044E. Primers were
designed to amplify introns or 39-untranslated
regions, if available. Polymorphisms were iden-
tified in PCR-amplified DNA by variation in
fragment length or as single strand conforma-
tion polymorphisms (SSCP). Primers, PCR con-
ditions and any other information relating to
mapped genes are available at the chicken
genome database ± Arkdb-chick (http://
www.ri.bbsrc.ac.uk/). All markers were scored
on the chicken backcross panels previously
described (Burt et al. 1995) and genetic maps
were constructed using the Map Manager
program (Manly 1993).
Statistical analyses
The estimate for sequence sampling was
standardised by the ratio of lengths of DNA
sequence sampled. This estimate does not
have any measure of uncertainty, such as
standard errors. These can be calculated if the
marker counts are assumed to follow a
Poisson distribution, using a regression
model (McCullagh & Nelder 1989). From
physical (FISH) mapping data, the estimate
of the ratio of microchromosomal to macro-
chromosomal gene densities was calculated as
the ratio of the number of genes mapped,
standardised for the unknown sample lengths
of macro-and microchromosomes, by dividing
the corresponding ratio for anonymous loci.
This allowed for differences in the length of
the genome being sampled by this method.
For genetic linkage data, the logarithms of the
observed counts are equated to those of their
expectations (products of marker densities and
the lengths of genome sampled), giving an
estimate of the logarithm of the relative
densities of genes in macro vs. microchromo-
somes, together with its standard error. A
95% confidence interval for the relative gene
density is the antilogarithm of the estimate on
the log scale � 1.96 times its standard error.
CpG content was analysed by a two-sided
Student's t-test assuming equal variance.
Results and discussion
We used two approaches to examine the
hypothesis that chicken microchromosomes
are more gene-dense than macrochromosomes.
The first approach was based on sequence
sampling of total genomic DNA cloned from
macrochromosomes or microchromosomes and
identification of genes based on homologies
with gene sequences in nucleotide and protein
databases. (`Gene-finder' programs were initi-
ally tested on the chicken sequences we
generated. However these proved to be unsuc-
cessful in predicting the location of genes,
probably due to the contig nature of our DNA
sequences). In the second, we compared the
distribution of genes on macrochromosomes or
microchromosomes based on genetic linkage
and physical mapping data in this paper or
published elsewhere.
97
Chicken
macrochromosomes
and
microchromosomes
ã 2000 International Society for Animal Genetics, Animal Genetics 31, 96±103
Comparison of gene densities on macrochromo-
somes and microchromosomes based on se-
quence sampling of genomic DNA
Cosmid clones were selected at random and
physically mapped to either macrochromosomes
or microchromosomes by FISH. Sixteen cosmid
clones were characterised further, 8 mapped to a
macrochromosome and 8 to a microchromosome
(Table 1). Random subclones were made for each
cosmid clone by partial digestion with the
restriction enzyme Cvi JI (Xia et al. 1987; Gin-
grich et al. 1996). Sufficient random subclones
were sequenced to provide at least 40% coverage
of each cosmid clone. High quality, vector-free
sequences were then used to search nucleotide
and protein databases for gene homologies using
the BLAST algorithm (Altschul et al. 1990). The
results are summarised in Table 1. Within the
cosmid clones that mapped to macrochromo-
somes three gene homologies were found: TGFb
receptor (TGFBR1), serotonin 1D receptor
(HTR1D) and platelet-activating factor receptor
protein (PTAFR) genes. The genes for HTR1D and
PTAFR were found in the same cosmid clone and
are also linked in human on chromosome 1, at
1p36.3±34.3 and 1p35±34.3, respectively. Four
gene homologies were found in the microchro-
mosomal cosmid clones: opioid-binding cell
adhesion molecule gene (OPCML), hydroxybuty-
rate dehydrogenase (BDH), a scavenger receptor-
like protein homologous to human M130 antigen
and a homologue of the human zinc-finger
protein KIIA0677 (Ishikawa et al. 1998). From
the number of genes found in the cosmid clones
isolated from macrochromosomes and micro-
chromosomes and the amount of DNA sequenced
we estimate that the gene-density of the micro-
chromosomes is 1.3 times that of macrochromo-
somes (Table 1).
If only the euchromatic portion of the genome
were analysed, the results would be very similar
as the chicken is known to contain a very low
percentage of heterochromatin (12% compared
Table 1. Cosmid sequence sampling data
Cosmid
name
Physical
location Gene homology
Length of
match (bp)
P-value
of match CpG/kb
No. of
sequences
Total
sequence
read (bp)
Average
sequence
read (bp)
Assembled
sequence
(bp)
COS 08 MACS None 15 52 24076 463 16423
COS 16 MACS None 15 72 38808 539 26788
COS 27 MACS Serotonin 1D receptor (HTR1D) 157/251 3.6e-24 19 80 27600 345 18262
MACS Platelet activating factor receptor (PTAFR) 158/233 7.2e-33
COS 28 MACS Transforming growth factor beta receptor
Type 1 (TGFBR1)
191 2.8e-69 9 106 34132 322 13537
COS 30 MACS None 15 61 22509 369 14785
COS 33 MACS None 10 66 22704 344 14226
COS 34 MACS None 12 77 29953 389 15069
COS 35 MACS None 14 90 29970 333 12199
ALL 14 604 229752 380 131289
COS 01 MICS None 24 80 41600 520 21480
COS 07 MICS Hydroxybutyrate dehydrogenase (BDH) 91/109 7.9e-22 17 49 17738 362 13570
COS 14 MICS M130 antigen-like 201/287 8.1e-45 21 46 22540 490 12248
COS 20 MICS Opioid binding cell adhesion molecular
(OPCML)
2.4e-51 27 114 48678 427 24010
COS 21 MICS None 21 72 37512 521 19354
COS 31 MICS None 29 68 23936 352 11710
COS 32 MICS KIAA0677 protein 103/128 3.1e-24 14 88 25256 287 15577
COS 36 MICS None 21 64 21696 339 13266
ALL 22 581 238956 411 131215
Table 2. Gene density on macrochromosomes and microchromosomes based on genetic and physical mapping
Genetic linkage* Physical linkagey
Chr. No. Sizez (Mb) Anonymous Gene Relative densityx Anonymous Gene Relative density
MAC (5A + Z) 843 193 58 20 14
MIC (33A + W) 357 85 56 2.3 20 18 1.3
*See Table 3.
ySee Table 4.
zSmith & Burt 1998.
xUsing physical sizes (Mb) as shown in this Table.
98
Smith, Bruley, Paton
et al.
ã 2000 International Society for Animal Genetics, Animal Genetics 31, 96±103
to 36% in human), most of which is located on the
Z and W sex chromosomes (Bloom et al. 1993).
We used the average CpG content of cosmid
clones from macrochromosomes and microchro-
mosomes as an indirect measure of the abun-
dance of CpG islands (Table 1). It was not
possible to count the number of CpG islands
directly because sequencing of cosmid clones
was not complete and we were also unable to
tell from sequencing data alone, whether the
DNA was methylated or not in vivo, and thus a
true CpG-island. From our sample of cloned
chicken genomic DNA we estimate that micro-
chromosomal DNA has a CpG content 1.6 times
that of macrochromosomal DNA (P < 0.001 and
95% confidence intervals of 1.2, 2.1). Whether
the differences in CpG content are related to
gene content or a structural feature of micro-
chromosomes has still to be determined.
On macrochromosomes we find a gene homol-
ogy every 44 kb and once every 33 kb on
microchromosomes. Previous sequencing of the
genomes of C. elegans and S. cerevisiae have
shown that only 50% of all genes are found by
database searches alone (Jones 1995). A similar
rate is found in the chicken based on partial
cDNA sequences (Bumstead and Burt, personal
communication). Therefore these estimates,
based on database searches alone, are likely to
be underestimated by a factor of two. The true
gene densities may therefore approach a gene
every 22 kb on macrochromosomes and one gene
every 17 kb on microchromosomes. From the
size of the chicken macrochromosomes and
microchromosomes (Smith & Burt 1998;
Table 2) we estimate that these chromosome
fractions contain 38000 and 21000 genes, respec-
tively. Our estimate of 59000 genes as the total
number of genes in the chicken genome is similar
to the number estimated for other vertebrates.
Although a reasonable method for finding
genes in the chicken, the sequence sampling
method has not provided us with an accurate
estimate of gene density due to the small sample
of gene homologues. Many more clones would
have to be examined in order for a more realistic
evaluation of gene density by this method. We
then went on to examine physical and genetic
mapping data in order to establish a more
confident assessment of the gene density on
the microchromosomes compared to the macro-
chromosomes.
Comparison of gene densities on macrochromo-
somes and microchromosomes based on physi-
cal mapping and genetic linkage
The likelihood of mapping a gene to a
Table 3. Genes which have been mapped by FISH
Locus Chr. type R Locus Chr. type R Locus Chr. type R
ASCL1 MACS 0 SCD* MICS 1 IGF1R MICS 0
CCND2* MACS 1 ALDH MICS 0 MC1R MICS 0
CCND2P* MACS 0 NRAMP1* MICS 1 NCAM1* MICS 1
DCN MACS 0 RPL37A* MICS 1 NGFB* MICS 1
GAPD* MACS 0 RPL5* MICS 1 OPCML MICS 0
H5* MACS 0 ABL1 MICS 0 OVM* MICS 1
HISA@* MACS 0 ACACA* MICS 1 PGA@ MICS 0
IGF1* MACS 1 ADORA1 MICS 0 PPY MICS 0
PGR* MACS 1 ADORA3 MICS 0 PRNP MICS 0
UCP2 MACS 0 AK1* MICS 0 RAF1* MICS 0
TGFBR1 MACS 0 ANX2 MICS 0 RARB MICS 0
ACTB* MACS 0 B2M* MICS 0 RNR* MICS 1
CCNC* MACS 1 BBC1 MICS 1 RPL7A* MICS 1
MYB* MACS 0 BMP7 MICS 0 SLC6A4 MICS 0
HMG14* MACS 1 CAMLG MICS 0 SUV3 MICS 1
IL8 MACS 0 CCNE* MICS 1 TAX1 MICS 0
IRF2* MACS 1 CD3E MICS 0 TF* MICS 1
KIT MACS 0 CDC2L1* MICS 1 TRAF1 MICS 0
PGK1* MACS 1 COS0032* MICS 0
CCND1* MACS 1 CRABP1 MICS 0
HTR1D MACS 0 DCM11* MICS 1
MAX* MACS 1 DMD* MICS 1
TGFB3* MACS 1 FASN* MICS 1
TH* MACS 1 FES* MICS 0
AVDL@ MACS 1 FLN2 MICS 0
IFN1* MACS 1 FMOD MICS 0
TRKB* MACS 1 H3F3B* MICS 0
99
Chicken
macrochromosomes
and
microchromosomes
ã 2000 International Society for Animal Genetics, Animal Genetics 31, 96±103
macrochromosome or microchromosome by
physical or genetic methods depends on the
underlying gene density. We therefore made a
comparison of the relative number of genes on
these chromosomes from physical mapping
experiments (Table 3) and genetic linkage
data (Table 4) described in this paper and
elsewhere. Only markers, which had been
Table 4. Genes which have been mapped by genetic linkage analysis
Locus Chr. No. Type R Locus Chr. No. Type R Locus Chr. No. Type R
SMOH 1 SSCP 1 FMR1 4 MICR 0 RYR1 E25C31 SSCP 1
NRCAM* 1 SNP 1 SPP1* 4 HET 1 GNRHR E29C09W09 SSCP 1
GNRH 1 SSCP 1 COM0117E* 4 RFLP 1 IGF1R* E29C09W09 RFLP 1
NAGA* 1 SNP 1 MSX1 4 RFLP 1 AGC1* E29C09W09 SSCP 1
LGALS4* 1 SNP 1 CD8A* 4 RFLP 1 HMX3* E29C09W09 RFLP 1
COM0119E* 1 RFLP 1 CAPN1* 5 HET 1 CYP19 E29C09W09 SSCP 1
IGF1* 1 RFLP 1 RYR3* 5 HET 1 CCNE E30C14W10 SSCP 1
LDHB* 1 RFLP 1 HTR1D 5 SSCP 1 MYH@* E31E21C25W12 MICR 0
GAPD* 1 SNP 1 TGFB3* 5 RFLP 1 ROS0022E E31E21C25W12 MICR 0
HSD3B* 1 SNP 1 COM0089E* 5 RFLP 1 H3F3B* E31E21C25W12 RFLP 1
COM0155E* 1 RFLP 1 DNCL* 5 RFLP 1 FASN* E31E21C25W12 SSCP 1
ROS0044E 1 MICR 0 CKB* 5 HET 1 COM0093E* E31E21C25W12 RFLP 1
ROS0081E 1 MICR 0 BMP4 5 RFLP 1 HLF E31E21C25W12 MICR 0
LAMP1* 1 MICR 0 PRLR* Z SSCP 1 BMP7* E32 RFLP 1
COM0092E* 1 RFLP 1 MSU0068E* Z RFLP 1 FZF* E32 SNP 1
RB1* 1 SNP 1 ROS0072E Z MICR 0 ROS0078E E36C06W08 MICR 0
FUCT4* 1 SNP 1 PTCH* Z SSCP 1 EIF4A2* E36C06W08 RFLP 1
ROS0025E 1 MICR 0 ROS0017E Z MICR 0 SNON* E36C06W08 SNP 1
ROS0055E 1 MICR 0 CHRNB3* Z SNP 1 ROS0073E E38 MICR 0
WNT11* 1 RFLP 1 LPL* Z SSCP 1 CPRR* E38 CLAS 1
SHH 2 SSCP 1 ALDOB* Z SNP 1 COM0152E* E38 RFLP 1
NPY 2 SSCP 1 GGTB2* Z SNP 1 ROS0020E E41W17 MICR 0
ROS0018E 2 MICR 0 ROS0028E 6 MICR 0 RPL7A* E41W17 RFLP 1
CP49* 2 RFLP 1 PDE6C* 6 HET 1 ABL1* E41W17 SNP 1
TGFBR1 2 RFLP 1 ACTA2* 6 MICR 0 CD39L1* E41W17 SNP 1
PRL 2 SSCP 1 SCD* 6 SSCP 1 AMBP* E41W17 SNP 1
BMP6 2 RFLP 1 PSAP* 6 RFLP 1 HSF1* E46C08W18 MICR 0
BCL2* 2 MICR 0 CPII* 7 CLAS 1 ITGAM* E46C08W18 MICR 0
ZNF5* 2 SNP 1 ROS0019E 7 MICR 0 POU4F3* E48C28W13 RFLP 1
ROS0023E 2 MICR 0 CD28* 7 RFLP 1 CAMLG* E48C28W13 SNP 1
ROS0074E 2 MICR 0 EEF1B2* 7 RFLP 1 CDX1* E48C28W13 SNP 1
PENK* 2 SNP 1 ROS0021E 8 MICR 0 ROS0083E E48C28W13 MICR 0
CA2* 2 SNP 1 GGTB1* 8 HET 1 MSX2 E48C28W13 RFLP 1
CALB1* 2 RFLP 1 VTG2* 8 HET 1 OPCML E49C20W21 SSCP 1
BMP2 3 RFLP 1 ROS0026E 8 MICR 0 POU2F3* E49C20W21 RFLP 1
TGFB2* 3 RFLP 1 PLA2G2A* 8 HET 1 APOA1* E49C20W21 SNP 1
ACTN2* 3 SNP 1 B@* 16 RFLP 1 W* E49C20W21 CLAS 1
HMX1* 3 RFLP 1 RFP-Y@* 16 RFLP 1 ACACA* E52W19 SSCP 1
T 3 RFLP 1 CPAA* E04 CLAS 1 CRK* E52W19 SSCP 1
IGF2R* 3 SNP 1 CPEE* E04 CLAS 1 AMH E53C34W16 MICR 0
VIP 3 SSCP 1 PGA@* E04 HET 1 TVA* E53C34W16 RFLP 1
ESR* 3 RFLP 1 ARF4* E16C17W22 RFLP 1 CDC2L1* E54 SNP 1
PLN* 3 MICR 0 MIF* E18C15W15 RFLP 1 AGRN* E54 SNP 1
BMP5 3 RFLP 1 IGL@* E18C15W15 MICR 0 ENO1* E54 SNP 1
GSTA2* 3 SNP 1 CRYBB1* E18C15W15 MICR 0 PLOD* E54 SNP 1
ODC1 3 RFLP 1 I* E22C19W28 CLAS 1 SLC2A1* E54 RFLP 1
CPPP* 3 CLAS 1 ROS0054E E22C19W28 MICR 0 TP53* E57 SNP 1
COM0094E* 4 RFLP 1 GLI E22C19W28 SSCP 1 COL1A1* E59C35W20 MICR 0
HMG14A* 4 RFLP 1 TGFB1 E25C31 RFLP 1 ROS0071E E59C35W20 MICR 0
Bold, new gene assignments in this paper.
SSCP-single strand conformation polymorphism; SNP-single nucleotide polymorphism; RFLP-restriction fragment length polymorphism;
MICR-microsatellite; CLAS-classical. R-random; 1 ± clone was mapped with no a priori knowledge of map location; 0 ± clone was mapped
with prior knowledge of its position in the genome and cannot therefore be used in this analysis.
*References in Chicken genome database (http://www.ri.bbsrc.ac.uk/).
100
Smith, Bruley, Paton
et al.
ã 2000 International Society for Animal Genetics, Animal Genetics 31, 96±103
mapped at random, i.e. having no a priori
knowledge of physical map location, were
used in this analysis.2
Physical mapping
We have cytogenetically mapped 31 more genes
onto the physical map of the chicken genome
(Table 3) and when combined with published
results brings the total number of genes mapped
at random to 72. It should be noted however,
that the anonymous clones distribute them-
selves between the macro-and microchromo-
somes in a ratio of 50 : 50. This is not the 70 : 30
split which would be expected if the clones
were distributing according to the size of the
genome represented by the two sets of chromo-
somes. It must be assumed therefore, that there
is a cloning bias inherent in the genomic
libraries used for obtaining clones for these
physical mapping studies. The number of
physical markers mapped with no a priori
knowledge of gene content were used as a
correction factor for the true physical size of
chromosomes. This allowed for differences in
the length of the genome being sampled by this
method (see Materials and methods).
Twenty anonymous and 14 genes (random)
have been assigned to the physical maps of
macrochromosomes, and represent a ratio of
0.70 genes/anonymous loci. On the microchro-
mosomes, 20 anonymous and 18 gene markers
(random) have been mapped, and represent a
ratio of 0.90 genes/anonymous loci (Tables 2
and 3). From this physical mapping data, we
estimate that the microchromosomes are 1.3
times (with a 95% confidence interval of 0.5,
3.3) as dense as the macrochromosomes. So far,
108 genes have been mapped by FISH but many
target specific chromosomes or were selected
based on comparative maps with human
(unpublished data). Unfortunately the small
sample size of randomly mapped genes and
the skewed distribution of cloned genomic DNA
is a limitation of this approach. Genetic linkage
data was therefore analysed.
Genetic linkage
In this paper we have mapped a further 46 genes
on the genetic map of the chicken (Table 4) and
together with published data, this brings the
total number of genes mapped at random to 147.
A wide variety of genetic markers (i.e. micro-
satellites, RFLP, RAPD, AFLP, SSCP and CR1)
were used to construct the genetic map, repre-
senting markers on all chicken chromosomes
(Burt & Cheng 1998). We estimate that the
genetic map is almost complete (Smith & Burt
1998) with a new genetic marker having a
greater than 95% chance of linkage to any
other previously mapped marker.
Previous reports by Primmer et al. (1997)
suggested that microsatellite sequences may
not be randomly distributed and may be present
at a lower density on the microchromosomes.
To test this we compared the distribution of
microsatellites and other genetic markers on the
genetic map for both anonymous and gene
sequences (Table 5). For anonymous markers
we expected 70% to map to the macrochromo-
somes and 30% to the microchromosomes based
on their physical size (Smith & Burt 1998).
Contrary to expectations from the work by
Primmer et al. (1997), we found more micro-
satellite markers than expected on the micro-
chromosomes (157 out of 368, which is 43%).
This discrepancy may be explained as a techni-
cal problem with the PRINS technique used by
Primmer et al. (1997) on chicken microchromo-
somes. The distribution of other genetic markers
appears to be at random and is proportional to
Table 5. Distribution of genetic markers on macrochromosomes and microchromosomes*
Chromosomes Microsatellites % Other %
Anonymous genetic markers
MAC (5A + Z) 211 57.3 193 69.4
MIC (33A + W) 157 42.7 85 30.6
Total 368 100.0 278 100.0
Genetic markers of genes
MAC (5A + Z) 13 39.4 58 50.9
MIC (33A + W) 20 60.6 56 49.1
Total 33 100.0 114 100.0
*Based on the East Lansing genetic map (Burt et al. 1995).
2Although `randomness' of cloning cannot be accurately quantified for every gene that has been mapped, we tried
to get as near to this as possible be eliminating genes which we knew had definitely been mapped with prior
knowledge or estimate of map position.
101
Chicken
macrochromosomes
and
microchromosomes
ã 2000 International Society for Animal Genetics, Animal Genetics 31, 96±103
the physical size of each chromosome
(70% : 30%). Consequently, the distribution of
genes based on these kinds of genetic markers
are more likely to mirror the true distribution of
genes over the chicken genome.
Excluding microsatellite-based markers, 58
gene markers have been assigned randomly to
the genetic maps of macrochromosomes
(Table 2). On the microchromosomes, 56 gene
markers have been mapped. When this genetic
mapping data is compared to the physical size
of the genome represented by the two sets of
chromosomes, we estimate that the microchro-
mosomes are 2.3 times as gene-dense as the
macrochromosomes (with a 95% confidence
interval of 1.6, 3.3).
Conclusion
Accurate estimates for the relative gene density
were not possible in earlier studies (McQueen
et al. 1996; McQueen et al. 1998) but suggested a
6-fold difference in gene density (an estimate
based mostly on CpG-island-like sequences in
cloned genomic DNAs) on macrochromosomes
and microchromosomes, and that 75% of all
chicken genes were located on the microchro-
mosomes. If true, this should have been reflected
in significantly more gene homologies being
found by the sequence sampling approach,
possibly 16 gene homologies on the cosmids
isolated from the microchromosomes and only 2
or 3 from the eight macrochromosomal cosmids.
This contrasts with our more conservative
estimate of around 50% of all chicken genes
being located in the microchromosome fraction,
thus explaining the lower number of sequence
homologies found by sequence sampling. Our
analysis of physical and genetic mapping data
gives us an estimate of around a 1.5-fold to 3.5-
fold difference in gene density, which is not
consistent with the original estimate of a 6-fold
difference. The difference between these two
predictions is probably explained by the differ-
ent approaches taken to identify genes. In this
paper our estimate is based on a direct gene
approach: genetic mapping of previously char-
acterised genes. The work by McQueen et al.
(1996, 1998) is based on differences in CpG
content and other indirect measures of gene
activity, and so may overestimate the true gene
content. For example, the initial FISH method
used is not quantitative and the authors admit to
analysing CG-like fragments, thus conceding
that it may not be just true CpG islands which
were studied.
Recent work by Clark et al. (1999) used a
similar approach to our study and also
attempted to answer the question of gene density
differences by using a sequence sampling
approach. They identified 12 gene homologies
from database searches after sequence scanning
19 chicken cosmids ± 3 hits to macrochromo-
somes and 9 to microchromosomes. Although
contig assembly was not carried out and the total
amount of DNA sequenced was not given, the
gene density difference from this work can be
estimated. Given that there were 8 macro-and 10
microchromosomal clones analysed, and assum-
ing the same amount of sequencing was carried
out on each clone, the gene density difference
can be calculated to be around 2.4 times higher
on the microchromosomes ± a figure which is
very close to our own evaluation.
How did the two-fold difference in gene
density (and CpG islands?) between avian
macrochromosomes and microchromosomes
evolve? In mammalian chromosomes gene den-
sity differences are correlated with chromosome
banding patterns, the R bands having higher
gene densities than the G bands (Craig &
Bickmore 1994). The avian microchromosomes
share many characteristics in common with
these mammalian R bands: high gene density,
high CpG island content (McQueen et al. 1996),
early replication in S phase (McQueen et al.
1998) and an increased rate of recombination of
2±4 fold when compared to the macrochromo-
somes (Rodionov et al. 1992). Why should
microchromosomes differ from macrochromo-
somes in these ways? It has been suggested
(Craig & Bickmore 1994) that meiotic recombi-
nation may tend to initiate in the accessible
chromatin at the promoters of genes. This close
correlation between gene density and recombi-
nation rate is therefore to be expected. Rodionov
(1996) also suggest that the higher recombina-
tion rate on these small chromosomes is to
ensure pairing of chromosomes occurs during
mitosis. This correlation between higher recom-
bination rate and higher gene density, thus
suggests an evolutionary pressure for an
increase in gene density on the microchromo-
somes. Why the microchromosomes arose in the
first place, however, is a question which still
has to be answered.
Acknowledgements
Many thanks to Dr Darren Griffin for invaluable
assistance with the FISH method. This work was
supported by the Biotechnology and Biological
Science Research Council (BBSRC), Ministry of
Agriculture, Fisheries and Food (MAFF) and the
EC Biotechnology programme BIO4-CT95±0287,
as part of the European ChickMAP project.
102
Smith, Bruley, Paton
et al.
ã 2000 International Society for Animal Genetics, Animal Genetics 31, 96±103
GenBank Accession Numbers: The nucleotide
sequence data has been submitted to GenBank
under accession numbers: Cosmid 1 (AJ231686±
231708), Cosmid 7 (AJ231709±231736), Cosmid
8 (AJ231737±231764), Cosmid 14 (AJ231765±
231785), Cosmid 16 (AJ231786±231815), Cos-
mid 20 (AJ231816±231844), Cosmid 21
(AJ231845±231871), Cosmid 27 (AJ231872±
231918), Cosmid 28 (AJ231919±231950), Cos-
mid 30 (AJ231951±231979), Cosmid 31
(AJ231980±232004), Cosmid 32 (AJ232005±
232047), Cosmid 33 (AJ232048±232078), Cos-
mid 34 (AJ232079±232106), Cosmid 35
(AJ232107±232137) and Cosmid 36 (AJ232138±
232169).
References
Altschul, S.F., Gish, W., Miller, W., Myers, E.W. &
Lipman, D.J. (1990) Basic local alignment search
tool. Journal of Molecular Biology 215, 403±10.
Auer, H., Mayr, B., Lambrou, M. & Schleger, W. (1987)
An extended chicken karyotype, including the NOR
chromosome. Cytogenetics and Cell Genetics 45,
218±21.
Bloom, S.E., Delany, M.E. & Muscarella, D.E. (1993)
Constant and variable features of avian chromo-
somes. In: Manipulation of the Avian Genome, 1st
edn (Etches, R.J. & Gibbins, A.M., eds). Boca Raton:
CRC Press, 39±50.
Buitkamp, J., Ewald, D., Schalkwyk, L., Weiher, M.,
Masabanda, J., Sazanov, A., Lehrach, H. & Fries, R.
(1998) Construction and characterisation of a
gridded chicken cosmid library with four-fold
genomic coverage. Animal Genetics 29, 295±301.
Burt, D.W., Bumstead, N., Bitgood, J.J., de Ponce Leon,
F.A. & Crittenden, L.B. (1995) Chicken genome
mapping: a new era in avian genetics. Trends in
Genetics 11, 190±4.
Burt, D.W. & Cheng, H.H. (1998) Bird gene maps. ILAR
Journal 39, 185±92.
Clark, M.S., Edwards, Y.J.K., McQueen, H.A., Meek,
S.E., Smith, S., Umrania, Y., Warner, S., Williams,
G. & Elgar, G. (1999) Sequence scanning chicken
cosmids: a methodology for genome screening. Gene
227, 223±30.
Craig, J.M. & Bickmore, W.A. (1994) The distribution
of CpG islands in mammalian chromosomes. Nature
Genetics 7, 376±81.
Fechheimer, N.S. (1990) Chromosomes of chickens.
In: `Domestic Animal Cytogenetics, 1st edn' (Ed. by
by , R.A.McFeely), pp. 170±207, Academic Press,
San Diego.
Gingrich, J.C., Boehrer, D.M. & Basu, S.B. (1996)
Partial Cvi JI digestion as an alternative approach
to generate cosmid sublibraries for large scale
sequencing projects. Biotechniques 21, 99±104.
Hinegardner, R. (1976) The evolution of genome size.
In `Molecular Evolution' (Ed. by F.J.Ayala), pp. 179±
99, Sinaur Associates. Massachusetts.
Ishikawa, K., Nagase, T., Suyama, M., Miyajima, N.,
Tanaka, A., Kotani, H., Nomura, N. & Ohara, O.
(1998) Prediction of the coding sequences of uni-
dentified human genes. X. The complete sequences
of 100 new cDNA clones from brain which can code
for large proteins in vitro. DNA Research 5, 169±76.
Jones, C.T., Morrice, D.R., Paton, I.R. & Burt, D.W.
(1997) Homologues of genes on human chromosome
15q21-q26 and a chicken microchromosome show
conserved synteny and gene order. Mammalian
Genome 8, 436±40.
Jones, S.J.M. (1995) An update and lessons from
whole-genome sequencing projects. Current Opi-
nions in Genetic Development 5, 349±53.
Manly, K.F. (1993) A Macintosh program for storage
and analysis of experimental genetic mapping data.
Mammalian Genome 4, 303±13.
McCullagh, P. & Nelder, J.A. (1989) `Generalized
Linear Models' (Ed. by by Chapman & Hall), London.
McQueen, H.A., Fantes, J., Cross, S.A., Clark, V.H.,
Archibald, A.L. & Bird, A.P. (1996) CpG islands of
chicken are concentrated on microchromosomes.
Nature Genetics 12, 321±4.
McQueen, H.A., Siriaco, G. & Bird, A.P. (1998)
Chicken microchromosomes are hyperacetylated,
early replicating and gene rich. Genome Research
8, 621±30.
Pollock, D.L. & Fechheimer, R.N.S. (1976) The
chromosome number of. Gallus Domesticus. British
Poultry Science 17, 39±42.
Primmer, C.R., Raudsepp, T., Chowdhary, B.P., Mol-
ler, A.P. & Ellegren, H. (1997) Low frequency of
microsatellites in the avian genome. Genome Re-
search 7, 471±82.
Rodionov, A.V. (1996) Micro versus macro: a review of
structure and functions of avian micro- and macro-
chromosomes. Russian Journal of Genetics 32, 517±
27.
Rodionov, A.V., Yu, A., Myakoshina, A., Chelysheva,
L.A., Solovei, I.V. & Gaginskaya, E.R. (1992) Chias-
mata on lampbrush chromosomes of Gallus gallus
domesticus: a cytogenetic study of recombination
frequency and linkage group lengths. Genetika 28,
53±63.
Smith, J. & Burt, D.W. (1998) Parameters of the
chicken genome (Gallus gallus). Animal Genetics
29, 290±4.
Smith, J., Paton, I.R., Bruley, C.K., Windsor, D., Burke,
D., de Ponce Leon, F.A. & Burt, D.W. (1999)
Integration of the genetic and physical maps of the
chicken macrochromosomes. Animal Genetics. in
press.
TegelstroÈm, H. & Ryttman, H. (1981) Chromosomes in
birds (Aves): evolutionary implications of macro-
and microchromosome numbers and lengths. Her-
editas 94, 225±33.
Xia, Y.N., Burbank, D.E., Uher, L., Rabussay, D. Van
Etten, J.L. (1987) IL-3A virus infection of a Chlorella
like green alga induces a DNA restriction endonu-
clease with novel sequence specificity. Nucleic
Acids Research 15, 6075±90.
103
Chicken
macrochromosomes
and
microchromosomes
ã 2000 International Society for Animal Genetics, Animal Genetics 31, 96±103