Karyotype Analysis and Chromosome Banding

Chromosome Banding

Methods of Chromosome Banding

Nearly all methods of chromosome banding rely on
harvesting chromosomes in mitosis. This is usually
achieved by treating cells with tubulin inhibitors, such as
colchicine or demecolcine (colcemid), that depolymerize
the mitotic spindle and so arrest the cell at this stage.
Excessively long incubations with Colcemid result in
overcondensed chromosomes that band poorly and more￾over some cell types, especially those from the mouse,
eventually escape the Colcemid block and proceed through
the cell cycle.
Chromosome banding methods are either based on
staining chromosomes with a dye or on assaying for a
particular function. The most common methods of dye￾based chromosome banding are G-(Giemsa), R-(reverse),
C-(centromere) and Q-(quinacrine) banding. Bands that
show strong staining are referred to as positive bands;
weakly staining bands are negative bands. However the
staining patterns are not black and white, different bands
stain to different intensities (Francke, 1994). G-positive
bands are usually just called G-bands and likewise for R￾positive (R-) bands. Positive C-bands contain constitutive
heterochromatin. Q-bands are considered equivalent to G￾bands.
The most widely used function-based banding method is
replication banding and is based on the fact that different
bands replicate their DNA at different times during S phase
of the cell cycle. Generally, R-band DNA is replicated
earlier than G-bands (Dutrillaux et al., 1976). G-bands also
correspond to the condensed chromomeres of meiotic
chromosomes and R-bands to the interchromomeric
regions.

Uses of Chromosome Banding

G-and R-banding are the most commonly used techniques
for chromosome identification (karyotyping) and for
identifying abnormalities of chromosome number, trans￾locations of material from one chromosome to another,
and deletions, inversions or amplifications of chromosome
segments. This has had an invaluable impact on human
genetics and medicine and the power of this approach has
been augmented by combining cytogenetics with fluores￾cence in situ hybridization (FISH). The detection of
chromosome deletions associated with disorders, very
often contiguous gene syndromes, provided some of the
first disease gene localizations in humans. Similarly,
translocations have been important in pinpointing the
location of disease-associated genes and the characteristic
translocations associated with some leukaemias is impor￾tant, not only for understanding the molecular basis of
these cancers, but also for their diagnosis and prognosis.
One of the best examples of this is the translocation
between human chromosomes 9 and 22 – t(9;22)(q34:q11)
– or the Philadelphia chromosome diagnostic of chronic
myelogenous leukaemia (CML).
Comparisons of chromosome banding patterns can
confirm evolutionary relationships between species and
also reveal changes in karyotype that may have been
important in speciation. The banding patterns of human,
gorilla and chimpanzee chromosomes are almost identical,
though human chromosome 2 is the result of a fusion
between two great ape chromosomes. There are also
extensive similarities between human chromosome bands
and those of lower primates.

Number and Size of Bands

Idealized diagrams (ideograms) of G-banded chromo￾somes are published as standard reference points for
chromosome banding. The G-bands are usually portrayed
in black and the R-bands in white. Bands are numbered
consecutively away from the centromere on both the short
(p) and long (q) arms (Figure 1). The total number of bands
or ‘resolution’ in the human karyotype depends on how
condensed the chromosomes are, and at what stage of
mitosis they are in. A 350-band resolution corresponds to
chromosomes late in metaphase. High-resolution ideo￾grams (approximately 1250–2000 bands) have also been
produced for human chromosomes in mid-prophase (Yunis, 1981). When a low-resolution band is subdivided,
the number of each subband is placed behind a decimal
point following the first band designation. For example the
most distal low-resolution band on the short arm of human
chromosome 11 (11p15) can be subdivided into bands
11p15.1, 11p15.2, 11p15.3, 11p15.4 and 11p15.5 at higher
resolution (Figure 1). A 2000-band resolution chromosome
band may contain 1.5 Mb of DNA, while a 300-band
resolution band will contain 7–10Mb of DNA. A skilled
cytogeneticist may be able to spot a deletion of 5–10 Mb of
DNA depending on its location, but at a molecular level the
human genome probably comprises >2000 ‘bands’.

Basis for G-/R-banding

G-banding involves staining protease-treated chromo￾somes with Giemsa dye and is thought to result from
interactions of both DNA and protein with the thiazine
and eosin components of the stain. The most common R￾banding method involves heat denaturing chromosomes in
hot acidic saline followed by Giemsa staining. This method
is thought to preferentially denature AT-rich DNA and to
stain the under-denatured GC-rich regions. T-banding
identifies a subset of R-bands – the most intensely staining
ones – by employing either a more severe heat treatment
than R-banding. It is thought to identify the GC-richest R￾bands, of which approximately half occur at telomeres in
the human genome, hence the name.
The need to combine chromosome banding with
fluorescence in situ hybridization has meant that banding
techniques using fluorescent dyes has become more
popular. Q-banding involves staining with quinacrine
which reacts specifically with certain bases. Quinacrine
intercalates into chromosomal DNA irrespective of
sequence, but fluoresces brighter in regions of AT-rich
DNA. There are a number of other molecules whose
fluorescence is influenced by the base composition of the
DNA to which they are bound. In addition to quinacrine,
other commonly used fluorochromes with a specificity for
AT-rich DNA include Hoechst 33258, DAPI (4’-6-
diamidino-2-phenylindole) and daunomycin. The fluores￾cence of Hoechst and DAPI is not quenched by guanine
and so they give less distinct bands than those produced by
quinacrine; however, daunomycin fluorescence is greatly
quenched by DNA with a GC content of 432%. DAPI
staining has the advantage that it is very resistant to fading
and that its excitation and emission spectra are compatible
with reporter molecules and filters commonly used in FISH
(Figure 2).
Fluorochromes with a preference for GC-rich DNA
include chromomycin and 7-amino actinomycin D. These
dyes give an R-band-like pattern.

Bands Reflect the Domain Organizationof the Genome

Techniques used to reveal chromosome bands enhance aninherent pattern of chromosome organization. A chromo￾some band is a manifestation of a chromatin domain with

functional and structural characteristics that are homo￾geneous and distinctive over a long enough stretch to be
seen down the microscope. G-and R-banding reflect
differences in chromatin structure and base composition
between different regions of the genome. Fluorochrome
banding also reflects variation in base composition along
chromosome length.
A rather different sort of banding is that seen after
staining of polytenized (interphase) chromosomes from
some tissues of Dipteran insects. The basis for this banding
is not well understood but arises through the alignment, in
register, of many thousands of copies of the chromosome.

Patterns in DNA sequence

Banding patterns can arise as a consequence of differences
in the DNA sequence along chromosomes. R-and G-banding patterns are revealed on human chromosomes
by FISH with Alu and LI interspersed repeats, respectively
(Korenberg and Rykowski, 1988). A similar distribution
has also been reported for SINEs (short interspersed
elements) and LINEs (long interspersed elements) on
mouse chromosomes. However, molecular studies of the
human genome including sequencing show that both SINE
and LINE repeats can be found in close proximity to each
other. To reconcile these differences it is suggested that the
SINEs located in R-bands are those that have retroposed
most recently and hence are closest in sequence to the
progenitor copy. These will therefore hybridize better to
the SINE probes that are based on the consensus repeat
sequence than to those that are more diverged, and hence
will produce stronger FISH signals than their G-band
counterparts.
Fractions of DNA of differing base composition
(isochores) produce banding patterns on human meta￾phase chromosomes with the most GC-rich fractions
highlighting T-bands (Saccone et al., 1993). Molecular
analyses on a finer scale reveal that, although there is a
general tendency for G-bands to be quite AT-rich, R-bands
can contain both AT-rich isochores and GC-rich iso￾chores.
A fraction of DNA (CpG islands) that contains the 5’
ends of approximately 50% of mammalian genes also
produces a T-and R-banding pattern on chromosomes
telling us that genes are not uniformly distributed along the
chromosome’s length. Most islands appear to reside in T￾and R-bands (Craig and Bickmore, 1994) (Figure 3). A
similar clustering of CpG islands into early-replicating R￾band compartments is seen in rodent genomes (Cross et al.,
1997). Chromosomes of a nonmammalian vertebrate (the
chicken) show a striking concentration of CpG islands in
distinctive parts of the karyotype – the microchromosomes
– rather than on the macrochromosomes.
Examination of genome databases suggests that CpG
island clustering can be extrapolated to gene density itself.Hence the highest densities of genes in our genomes are
located in the T-and R-bands. This explains why human
chromosomes with a high G-band content, e.g. 13, 18 and
21, are seen as viable trisomies in the population, whereas
trisomies for small but T-/R-band-rich chromosomes (e.g.
chromosome 22) are lost early in embryonic development.
Whole genome sequencing has confirmed the low gene
density of human chromosome 21 compared to that of 22
(Dunham et al., 1999; The Chromosome 21 Mapping and
Sequencing Consortium, 2000).
There has been a conservation of chromosome organi￾zation at the chromosome band level in terms of relative
gene density, time of DNA replication and banding
characteristics over 100 million years of mammalian
chromosome evolution. This suggests either the influence
of a strong selective pressure to maintain these character￾istics together in particular chromosome bands or the
action of common mechanisms linking the properties of
gene density, replication time and band type.

Patterns in DNA replication

Different regions of the genome replicate at different times
during S phase. The relationship between timing of
replication and chromosome banding is usually studied
by incorporating pulses of the thymidine analogue 5-
bromo-2’-deoxyuridine (BrdU) into cells during defined
stages of S phase and then examining chromosomes in the
subsequent metaphase. Sites of BrdU incorporation can be
detected with antibodies that detect the presence of BrdU
in denatured DNA (Figure 3). T-bands replicate on average
earlier than ordinary R-bands, and DNA in G-bands is
replicated even later (Dutrillaux et al., 1976).Aspects of the primary DNA sequence or chromatin
structure in different types of band could influence their
replication time. Also differences in replication time could
influence some characteristics of chromosome bands, e.g.
base composition or chromatin structure. Sites of tran￾scription at the G1/S boundary may seed the assembly of
the first replication factories in early S phase and hence the
most transcriptionally active regions of the genome, and
the regions with the highest concentrations of genes, would
tend to be the ones to be replicated first.

Patterns in chromatin structure

Several banding techniques, especially G-banding, suggest
that there are both qualitative and quantitative variations
in the interaction of DNA and proteins along the length of
metaphase chromosomes. The chromatin of active genes is
generally considered to be more accessible to nuclease
attack than is inactive chromatin. Consistent with this
nucleases preferentially digest R-bands and T-bands of
intact mitotic chromosomes, with G-bands and C-bands
refractive to digestion. The extent of chromatin packaging
in the interphase nucleus also differs between chromosome
bands. C-band positive heterochromatin remains visibly
condensed through interphase. FISH has shown that over
the 150 kb to 1Mb size range G-band chromatin is more
tightly packaged than that of R-bands.
The N-terminal tails of core histones H3 and H4 can be
modified by acetylation of lysine residues. A consequence
of this acetylation may be to facilitate access of proteins
such as transcription factors to the DNA. Histones can be
deacetylated or acetylated throughout the cell cycle by
nuclear acetyltransferase and deacetylase enzymes. Im￾munofluorescence of mammalian metaphase chromo￾somes with antibodies raised against each of the
acetylated forms of H4 has shown that the differently
modified forms are found preferentially in different regions
of the chromosome. Mono-acetylated (Lys16) H4 is found
throughout euchromatin, whereas acetylation at Lys8 and
Lys12 occurs mainly in R-bands. Acetylation at Lys5 is
found in the most highly modified (tri-and tetra￾acetylated) forms of histone H4. Antibodies to this H4
isoform produce a good banding pattern on metaphase
chromosomes, especially from cells that have been briefly
exposed to histone deacetylase inhibitors. Bright immuno-
fluorescence is seen over T-/R-band regions of the
karyotype and only faint staining is seen in G-bands
(Jeppesen and Turner, 1993). Hence the distribution of
histone acetylation on mammalian metaphase chromo￾somes mirrors that of genes.
The radial loop/scaffold model of chromosome struc￾ture proposes that higher order chromosome packaging
arises through the arrangement of the 30-nm chromatin
fibre into loops tethered to a proteinaceous chromosome
scaffold that runs centrally down the length of the
chromosome. The two major protein components of the
mitotic scaffold are topoisomerase IIa Sc I (170 kDa) and
Sc II (135 kDa). Sc II is a member of the structural
maintenance of chromosomes (SMC) family of proteins.
Immunofluorescence analysis shows that topoisomerase
IIa is not uniform along the chromosome length. It is
particularly concentrated in regions of centromeric hetero￾chromatin and there is stronger staining of G-bands than
R-bands. This might indicate that chromosome loop
anchoring sites to the scaffold are the least frequent in R￾bands. The pattern of immunofluorescence with anti-ScII
antibodies on vertebrate mitotic chromosomes is very
similar to that of topoisomerase II – the chromosome axes
are lit up and centromeres are particularly brightly stained.
Specific sites along the DNA associate with the
chromosome scaffold. Such sequences are generally
referred to as SARs (scaffold attachment regions). They
contain oligo(dA)-oligo(dT) tracts and fluoresce brightly
with daunomycin. The path of aligned SARs along the
scaffold at the core of metaphase chromosomes, as defined
by daunomycin staining is G-band like and colocalizes
with topoisomerase II immunofluorescence (Saitoh and
Laemmli, 1994). This might mean that G-bands have
smaller loops, and hence more frequent scaffold attach￾ments than R-bands and chromosome painting with SARs
suggests that there are indeed more of these in G-bands
than in R-bands (Craig et al., 1997). However, these
staining patterns could also result from differences in the
path of the SARs between different types of band, with the
chromosome scaffold being more tightly coiled within G￾bands and straighter within R-bands.
Variations in the density of meiotic chiasmata along
mammalian chromosomes are apparent when physical and
genetic maps are compared. T-bands have the highest rates
of exchange, followed by ‘mundane’ R-bands, then G￾bands. Heterochromatin (C-bands) shows the lowest rates
of recombination. R-bands are both the sites of synaptic
initiation and the preferred sites of crossing-over in
mammals and other vertebrates.

Evolution of chromosome bands

Whereas Q-, G- and R-banding patterns have only been
observed in some eukaryotes, replication banding is almost
universal among living organisms possessing chromo￾somes large enough to see by microscopy, suggesting that it
is a fundamental consequence of, or requirement for, the
compartmentalization of complex genomes.
Chromosomes from most mammals and birds can be G￾and R-banded. In addition, most reptilian chromosomes
band with G-and R-banding techniques to some extent.
With amphibia, fish and plants, some species band whereas
others do not. The lowest vertebrates with reported good Evolutionary analysis of chromosome banding patterns
suggests that the first cytogenetically detectable compart￾mentalization that arose in the genomes of eukaryotes was
the temporal control of replication and differences in
chromatin packaging and the segregation of some chro￾mosomal domains into heterochromatin. Ability to be G￾banded (and we will assume here that this is a reflection of
differences in chromatin structure on mitotic chromo￾somes) followed later. Fluorochrome banding seems to
have appeared on the scene last of all.

Extreme Chromosome Bands

All groups of eukaryotes that have chromosomes large
enough to be visualized contain a proportion of hetero￾chromatin, visualized by C-banding, and even simple
eukaryotes with microscopic chromosomes have hetero￾chromatin at a molecular level. It seems unlikely that there
is a completely strict dividing line between heterochroma￾tin and euchromatin – we might consider the very gene-rich
T-bands and gene-free constitutive heterochromatin as the
opposite extremes of a continuum of chromatin flavours.
In humans, the main C-bands are on the long arm of the
Y-chromosome and close to the centromeres of chromo￾somes 1, 9 and 16 (pericentric). The size of these C-bands
differs between different individuals. Smaller C-bands are
found at the centromere of each chromosome and on the p
arms of the five acrocentric chromosomes (13, 14, 15, 21,
22). In the mouse, the main blocks of visible heterochro￾matin are found close to the centromere of each chromo￾some (Figure 2). The amount of heterochromatin at these
sites also varies between different strains of Mus musculus.
In Drosophila the main sites of constitutive heterochroma￾tin are found at the chromocentre, telomeres and on the
fourth/Y-chromosome. At the molecular level all of these
sites are characterized by the presence of middle and highly
repetitive sequences, often in tandem arrays of satellites.
Classical satellites (I, II and III) locate to the pericen￾tromeric C-bands of human chromosomes 1, 9, 16 and the
long arm of the Y; a-satellite is located at the centromeres.
It is the minor satellite repeat that is found at mouse
centromeres, with major satellite being found beyond this,
between the centromere and telomere. In Drosophila,
satellite DNA at the chromocentre is interspersed with
stretches of more complex DNAs (both unique and
moderately repetitive DNA, including transposable ele￾ments).
Heterochromatin is normally associated with repression
of transcription and recombination, and with late replica￾tion. In Drosophila, the constitutive heterochromatin is
underrepresented in polytene chromosomes. No endogen￾ous genes or CpG islands have yet been mapped to regions
of C-banded constitutive heterochromatin in the human or
mouse genomes, and transgenes that integrate close to
constitutive heterochromatin are frequently silenced. A
few endogeneous genes in Drosophila are found in a
heterochromatic location and moreover they cease to
function when moved to euchromatic locations.
C-band-positive constitutive heterochromatin has a
distinctive chromatin structure. It is the site of most
methylation in human and mouse chromosomes and so the
5MeCpG-binding protein MECP2 is highly concentrated
over constitutive heterochromatin in vertebrates in most
cell types. Constitutive heterochromatin is also associated
with very low levels of histone acetylation; with the
exception that the chromocentre of Drosophila is enriched
in histones acetylated at Lys12, and that in mammalian
embryonic stem cells prior to differentiation histone H4
within the pericentromeric heterochromatin is acetylated.
A set of distinctive chromosomal proteins can also be
found concentrated in heterochromatin. Many of these
proteins have motifs, such as the chromobox, that are
characteristic of proteins involved in the formation of
multiprotein heterochromatin complexes. Antibodies
against M31, a chromobox containing protein that is a
homologue of Drosophila heterochromatin protein HP1,
highlight the pericentric heterochromatin of mouse and
human chromosomes.