Chapter 2:  Chromosomes in cells

2.1.1. Prokaryotes and eukaryotes represent a fundamental division of living cells

All cellular organisms can be subdivided into two major classes, prokaryotes and eukaryotes, on the basis of the architecture of their cells (Figure 2.1).

Prokaryotes lack a defined nucleus and have a relatively simple internal organization. Under the electron microscope they appear relatively featureless. They comprise two kingdoms of life: eubacteria which include most of the bacteria; and the archaea, rather poorly understood organisms that superficially resemble bacteria and often grow in unusual environments, such as in acid hot springs, saturated brines, etc. The genome of a prokaryote typically consists of a single small circular chromosome in which the DNA is not packaged in any obviously organized way. Prokaryotes may be simple, but they are not primitive - they have been through far more generations of evolution than we have.

Eukaryotes have a much more complex intracellular organization with internal membranes, membrane-bound organelles including a nucleus, and a well-organized cytoskeleton. The general features are summarized briefly in Box 2.1. Eukaryotic cells have several linear chromosomes in their cell nuclei, in each of which a single very long DNA molecule is elaborately packaged by histone and other proteins. The number and DNA content of the chromosomes vary greatly between species (Table 2.1). In general the genome size tends to parallel the complexity of the organism, but there are many exceptions. Most mammals have much the same size genome. Humans do not have specially large genomes, while the cells of an onion and a lily contain respectively about five and 30 times as much DNA as a typical human cell. Eukaryotes are thought to have first apeared about 1.5 billion years ago.

2.1.2. Cell size and shape can vary enormously, but rates of diffusion fix some upper limits

Bacteria and some other simple organisms, such as the yeast Saccharomyces cerevisiae, consist of a single cell. Such cells are necessarily able to carry out all the functions that are required to sustain the organism. Multicellular organisms begin life as a single cell but then undergo repeated cell division, cell differentiation and cell turnover. They may end up containing huge numbers of cells. Cell differentiation ensures that individual multicellular organisms are composed of a variety of cell types that can vary greatly in size and shape.

Cells depend on diffusion to coordinate their metabolic activities. As they grow larger, the surface-to-volume ratio decreases. It is thought that the simple internal structure of prokaryotic cells limits their maximum size - typically bacterial cells are 1 μm in diameter. The complex internal membranes and compartmentalization of eukaryotic cells may be important in allowing them to grow larger. Nevertheless, metabolically active internal regions are seldom more than 15–25 μm from the cell surface, so that the limit of cell size is typically 30–50 μm. The average diameter of cells in a multicellular organism falls within the range of 10–30 μm. Some individual human nerve cells can be as long as 1 metre, but the long projections are very thin. However, the ostrich egg warns us not to over-generalize.

2.1.3. Genetic content of cells: ploidy, the cell cycle and the life cycle

Most cells of humans are diploid. They contain two copies of the human genome. The DNA content and chromosome number of a genome are designated C and n respectively. For humans C = 3.5 × 10^-12 g, approximately, and n = 23. The DNA content of diploid cells is 2C and they have 2n chromosomes. Almost all mammals are diploid, but among other organisms there are many examples of species that are normally haploid (n chromosomes, DNA content C), tetraploid (4n) or polyploid. Triploidy (3n) is less common because triploids have problems with meiosis (see below).

The diploid cells of our body are derived from the original diploid fertilized egg by repeated rounds of mitotic cell division. Each round can be summarized as one turn of the cell cycle (Figure 2.2). This comprises a short stage of cell division, the M phase (mitosis; see Figure 2.10) and a long intervening interphase. Interphase can be divided into S phase (DNA synthesis), G₁ phase (gap between M phase and S phase) and G₂ phase (gap between S phase and M phase). From anaphase of mitosis right through until DNA duplication in S phase, a chromosome of a diploid cell contains a single DNA double helix and the total DNA content is 2C. G₁ is the normal state of a cell, and the long-term end state of nondividing cells. Cells enter S phase only if they are committed to mitosis; nondividing cells remain in a modified G₁ stage, sometimes called G₀. The cell cycle diagram can give the impression that all the interesting action happens in S and M phases - but this is an illusion. A cell spends most of its life in G₀ or G₁ phase, and that is where the genome does most of its work.

A subset of the diploid body cells constitute the germ line (see Figure 2.12). These give rise to specialized diploid cells in the ovary and testis that can divide by meiosis to produce haploid gametes (sperm and egg). In humans (n = 23) each gamete contains 22 autosomes (nonsex chromosomes) plus one sex chromosome. In eggs the sex chromosome is always an X; in sperm it may be an X or a Y. After fertilization the zygote is diploid (2n) with the chromosome constitution 46,XX or 46,XY (Figure 2.3). The other cells of the body, apart from the germline, are known as somatic cells. Somatic cells have no input to succeeding generations - unless animal cloners intervene.

Although most somatic cells are diploid, there are exceptions. Some terminally differentiated cells, red blood cells for example, have no nuclei and others, such as keratinocytes, are devoid of organelles altogether. Such cells are nulliploid. Other cells are polyploid as a result of DNA replication without cell division (endomitosis). Regenerating cells of the liver and other tissues are naturally tetraploid because of endomitosis, while the giant megakaryocytes of the bone marrow usually contain 8C, 16C or 32C, and individually give rise to thousands of nulliploid platelet cells (Figure 2.4A). Other naturally occurring cells have multiple diploid nuclei as a result of cell fusion (Figure 2.4B).

2.2.1. Only a small percentage of the cells in the early embryo give rise to the mature organism

As shown in Box 2.2, the early stages of human development are largely concerned with establishing the future placenta and extraembryonic membranes, and implanting the conceptus in the uterine endometrium. About 2 weeks of development and many rounds of mitosis pass before the group of cells that will give rise to the embryo has any separate identity. Embryonic development proper begins with gastrulation in the third week after fertilization.

2.2.2. Cells differ in their potential to divide and differentiate

Multicellular animals begin life as a single cell following fertilization of an egg cell (oocyte) by a sperm cell. The fertilized egg proceeds to undergo a series of cell divisions. At the early stages of development, individual cells in the embryo are totipotent: each cell retains the capacity to differentiate into all the different types of cell in the body. As development proceeds cells become more restricted in their capacity to generate different types of descendant cells and are said to be pluripotent. Progenitor cells that can only develop into a single cell type are unipotent cells. The processes of cell differentation lead to individual cells acquiring specialized forms and functions (Box 2.3).

Histology textbooks recognize about 210 different types of cell in the human body. Some, such as epithelial cells and fibroblasts, perform essentially the same role in a variety of organ systems. Others, such as hepatocytes, may be restricted to an individual organ. Some mature terminally differentiated cells do not undergo cell division. Other cells (often distinguished by the suffix -blast, as in osteoblasts, chondroblasts, myoblasts, etc.) divide actively and act as precursors of terminally differentiated cells. In some cases, the precursor cells are also capable of undergoing self-renewal and are known as stem cells. Stem cells are an important target of gene therapy (Chapter 22). As an example, Figure 2.5 illustrates the successive commitment of cells in the hemopoietic lineage.

2.2.3. X inactivation is a special feature of female development

Chromosomally-based sex-determination systems, like the human XX/XY system, create a problem for development. Having the wrong number of chromosomes almost always makes an organism develop abnormally, and yet the two sexes must develop normally with different chromosome constitutions. In humans and other mammals the solution is different for the X and the Y chromosomes. The Y chromosome contains very few genes, and these are mostly genes governing male sexual function, so that females can get by perfectly well without a Y chromosome. The X chromosome, however, contains many genes that play vital roles in both sexes, and so some method of dosage compensation is required, to ensure that cells function normally with either one or two X chromosomes.

Mammals achieve X-chromosome dosage compensation by the mechanism of X-inactivation, often called lyonization, after Dr Mary Lyon who first suggested this mechanism (Lyon, 1999, Heard et al., 1997; Figure 2.6). Early in embryonic development (at late blastocyst stage in the mouse, and probably also in humans), cells somehow count their X chromosomes and then inactivate all but one of them. 46,XY cells (and also 45,X cells) leave their single X active; 46,XX cells (and also 47,XXY cells) inactivate one X, while a 47,XXX cell would inactivate two Xs. Thus, regardless of the number of X chromosomes in the karyotype, there is only one set of active X-linked genes in a cell. Males are constitutionally hemizygous for X chromosome genes (that is, they have only a single copy of each gene), while females become functionally hemizygous: they have only a single functional copy of each gene.

When chromosomes of a female cell are observed at metaphase of mitosis, the active and inactive Xs look the same - but this is because all chromosomes at metaphase of mitosis are condensed and largely inactive. After the end of cell division, the inactive X remains condensed while the other chromosomes decondense and resume transcriptional activity. In some cells the inactive X can be seen as a Barr body or sex chromatin body near the membrane of the interphase nucleus. This allows a simple but not very reliable method of sexing interphase cells.

Which X in a 46,XX cell is inactivated is random (with a few exceptions discussed later), so that in a female embryo, some cells will inactivate the paternal X and some the maternal X. Once the choice is made, it is remembered. When the cell divides, the daughter cells inactivate the same X as the mother cell. An adult female is a mosaic of clones derived from different embryonic cells. Within a clone, all the cells inactivate the same X, but between clones the choice is random. If she happens to be a carrier of an X-linked recessive disease, this can have major implications (Section 3.1.2).

X inactivation is a fascinating and imperfectly understood phenomenon. One would like to know how the cell counts its X chromosomes, how the inactivation works, how it is perpetuated, and how it is reversed during gametogenesis (see Figure 2.6). We know that there is not blanket inactivation of all genes on the inactive X; certain regions (see Section 2.4.3), and certain genes in other regions, escape inactivation. The mechanisms are discussed in more detail in Section 8.5.2

2.3.1. Packaging of DNA into chromosomes requires multiple hierarchies of DNA folding

In the cell the structure of each chromosome is highly ordered (Manuelidis, 1990). Even in the interphase nucleus the 2 nm DNA double helix is subject to at least two levels of coiling (Figure 2.7).

• The most fundamental unit of packaging is the nucleosome. This consists of a central core of eight histone proteins, small highly conserved basic proteins of 102–135 amino acids. Each core comprises two molecules each of histones H2A, H2B, H3 and H4, around which a stretch of 146 bp of double-stranded DNA is coiled in 1.75 turns. Adjacent nucleosomes are connected by a short length of spacer DNA. Electron micrographs of suitable preparations show a ‘string of beads’ appearance.

• The string of beads, approximately 10 nm in diameter, is in turn coiled into a chromatin fiber of 30 nm diameter. The interphase chromosome seems to consist of these chromatin fibers, probably organized into long loops as described below.

During cell division, the chromosomes become ever more highly condensed. The DNA in a metaphase chromosome is compacted to about 1/10 000 of its stretched-out length. Loops of the 30 nm chromatin fiber, containing 20–100 kb of DNA per loop, are attached to a central scaffold. This consists of nonhistone acidic proteins, notably topoisomerase II, an enzyme which has the interesting ability to pass one DNA double helix through another by cutting a gap and repairing it. Topoisomerase II and some other chromatin proteins are known to bind to AT-rich sequences, and the chromatin loops may be attached by stretches of several hundred base pairs of highly AT-rich (>65%) DNA (scaffold attachment regions). In the chromatids of a metaphase chromosome the loop-scaffold complex is compacted yet further by coiling (see Figure 2.7).

2.3.2. Chromosomes as functioning organelles: the centromere

Normal chromosomes have a single centromere that is seen under the microscope as the primary constriction, the region at which sister chromatids are joined. The centromere is essential for segregation during cell division. Chromosome fragments that lack a centromere (acentric fragments) do not become attached to the spindle, and so fail to be included in the nuclei of either of the daughter cells.

During late prophase of mitosis, a pair of kinetochores forms at each centromere, one attached to each sister chromatid. Multiple microtubules attach to each kinetochore, linking the centromere of a chromosome and the two spindle poles (see Figure 2.11). At anaphase, the kinetochore microtubules pull the two sister chromatids toward opposite poles of the spindle. Kinetochores play a central role in this process, by controlling assembly and disassembly of the attached microtubules and, through the presence of motor molecules, by ultimately driving chromosome movement.

Specific DNA sequences presumably specify the structure and function of centromeres. In simple eukaryotes, the sequences that specify centromere function are very short. For example, in the yeast Saccharomyces cerevisiae the centromere element (CEN) is about 110 bp long, comprising two highly conserved flanking elements of 9 bp and 11 bp and a central AT-rich segment of about 80–90 bp (Figure 2.8). The centromeres of such cells are interchangeable - a CEN fragment derived from one yeast chromosome can replace the centromere of another with no apparent consequence. In mammals, centromeres comprise hundreds of kilobases of repetitive DNA, some nonspecific and some chromosome-specific (Section 7.4.1).

2.3.3. Chromosomes as functioning organelles: origins of replication

The DNA in most diploid cells normally replicates only once per cell cycle. The initiation of replication is controlled by cis-acting sequences that lie close to the points at which DNA synthesis is initiated. Probably these are sites at which trans-acting proteins bind. Eukaryotic origins of replication have been most comprehensively studied in yeast, where the presence of a putative replication origin can be tested by a genetic assay. To test the ability of a random fragment of yeast DNA to promote autonomous replication, it is incorporated into a bacterial plasmid together with a yeast gene that is essential for growth of yeast cells. This construct is used to transform a mutant yeast that lacks the essential gene. The transformed cells can only form colonies if the plasmid can replicate in yeast cells. However, the bacterial replication origin in the plasmid does not function in yeast, therefore the few plasmids that transform at high efficiency must possess a sequence within the inserted yeast fragment that confers the ability to replicate extrachromosomally at high efficiency - that is an autonomously replicating sequence (ARS) element.

ARS elements are thought to derive from authentic origins of replication and, in some cases, this has been confirmed by mapping a specific ARS element to a specific chromosomal location and demonstrating that DNA replication is indeed initiated at this location. ARS elements extend for only about 50 bp and consist of an AT-rich region which contains a conserved core consensus and some imperfect copies of this sequence (Figure 2.8). In addition, the ARS elements contain a binding site for a transcription factor and a multiprotein complex is known to bind to the origin.

Mammalian replication origins have been much less well defined because of the absence of a genetic assay. Some initiation sites have been studied, but such studies have not been able to identify a unique origin of replication. This has led to speculation that replication can be initiated at multiple sites over regions tens of kilobases long. Mammalian artificial chromosomes seem to work without specific ARS sequences being provided. Computer analysis of regions encompassing several eukaryotic origins of replication, including some human and other mammalian examples, identified a consensus DNA sequence WAWTTDDWWWDHWGWHMAWTT where W = A or T; D = A or G or T; H = A or C or T; and M = A or C (Dobbs et al., 1994).

2.3.4. Chromosomes as functioning organelles: the telomeres

Telomeres are specialized structures, comprising DNA and protein, which cap the ends of eukaryotic chromosomes. They have several likely functions:

• Maintaining the structural integrity of a chromosome. If a telomere is lost, the resulting chromosome end is unstable. It has a tendency either to fuse with the ends of other broken chromosomes, to be involved in recombination events or to be degraded. The loop structure of human telomeres (see below) means that natural chromosomes have no free DNA end.

• Ensuring complete replication of the extreme ends of chromosomes. During DNA replication, synthesis of the lagging strand is discontinuous and requires the presence of some DNA ahead of the sequence which is to be copied to serve as the template for an RNA primer (see Figure 1.9). However, at the extreme end of a linear molecule, there can never be such a template, and a different mechanism is required to solve the problem of replicating the ends of a linear DNA molecule (see below).

• Helping establish the three-dimensional architecture of the nucleus and/or chromosome pairing. Chromosome ends appear to be tethered to the nuclear membrane, suggesting that telomeres help position chromosomes.

Eukaryotic telomeres consist of a long array of tandem repeats. One DNA strand contains TG-rich sequences and terminates in the 3′ end; the complementary strand is CA-rich. Unlike centromeres, the sequence of telomeres has been highly conserved in evolution - there is considerable similarity in the simple sequence repeat, for example TTGGGG (Paramecium), TAGGG (Trypanosoma) TTTAGGG (Arabidopsis) and TTAGGG (Homo sapiens) (see also Figure 2.8).

The problem of replicating the ends of a chromosome has been solved by extending the synthesis of the leading strand using a specialized enzyme, telomerase. This RNA-protein complex carries within its RNA component a short sequence which will act as a template to prime extended DNA synthesis of telomeric DNA sequences on the leading strand. Further extension of the leading strand provides the necessary template for DNA polymerase α to complete synthesis of the lagging strand (Figure 2.9). This mechanism leaves the telomere itself with a protruding 3′ end. In mammalian chromosomes, the single-stranded end is believed to loop round and invade the double helix several kilobases proximally, producing a triple-stranded structure resembling the mitochondrial D-loop (Figure 7.2), which is stabilized by binding telomere-specific proteins (Greider, 1999). However, the actual nature of the telomere sequence may not be important. The telomere length is known to be highly variable and is subject to genetic control.

Just internal to the essential telomeric repeats, eukaryotic chromosomes also have a more complex set of repeats called subtelomeric or telomere-associated repeats. Their sequences are not conserved in eukaryotes and their function is unknown.

2.3.5. Heterochromatin and euchromatin

In the interphase nucleus most of the chromatin (euchromatin) exists in an extended state, dispersed through the nucleus and staining diffusely. However, some chromatin remains highly condensed throughout the cell cycle and forms dark-staining regions (heterochromatin). Genes located in euchromatin may or may not be expressed, depending on the cell type and its metabolic requirements, but genes that are located within heterochromatin, either naturally or as the result of a chromosomal rearrangement, are very unlikely to be expressed. There are two classes of heterochromatin:

• Constitutive heterochromatin is always inactive and condensed. It consists largely of repetitive DNA and is found in and around the centromeres of chromosomes and in certain other regions (see Figure 2.18).

• Facultative heterochromatin can exist in either a genetically active (decondensed) or an inactive and condensed form, as in the case of mammalian X-chromosome inactivation (Section 2.2.3).

In euchromatin, the G bands (Section 2.5.2) partake of some of the properties of heterochromatin, but to a lesser degree. G band chromatin in metaphase chromosomes is more condensed than R band chromatin, and data on CpG island distribution (Section 7.1.2; Figure 7.4) show that G bands are relatively poor in genes. The subset of R bands that are revealed by T-banding have a particularly high density of genes. Section 1.3.5 discusses the different structures of chromatin in transcriptionally active and inactive chromosomal regions.

2.4.1. Mitosis is the normal form of cell division

As a person develops from an embryo, through fetus and infant to an adult, cell divisions are needed to generate the large numbers of cells required. Additionally, many cells have a limited lifespan, so there is a continuous requirement to generate new cells in the adult. All these cell divisions occur by mitosis. Mitosis is the normal process of cell division, from cleavage of the zygote to death of the person. In the lifetime of a human there may be something like 10¹⁷ mitotic divisions (Section 9.2.1).

The M phase of the cell cycle (Figure 2.2) consists of the various stages of nuclear division (prophase, prometaphase, metaphase, anaphase and telophase of mitosis), and cell division (cytokinesis), which overlaps the final stages of mitosis (Figure 2.10). In preparation for cell division, the previously highly extended chromosomes contract and condense so that, by metaphase of mitosis, they are readily visible under the microscope. Even though the DNA was replicated some time previously, it is only at prometaphase that individual chromosomes can be seen to comprise two sister chromatids, attached at the centromere.

The mitotic spindle (Figure 2.11) is formed from tubulin-based microtubules and microtubule-associated proteins. Polar fibers, which extend from the two poles of the spindle towards the equator, develop at prophase while the nuclear membrane is still intact. Kinetochore fibers do not develop until prometaphase. These fibers attach to the kinetochore, a large multiprotein structure attached to the centromere of each chromatid, and extend in the direction of the spindle poles. The interaction between the different spindle fibers pulls the chromosomes towards the center, and by metaphase each chromosome is independently aligned on the equatorial plane (metaphase plate). Paternal and maternal homologs do not associate at all during mitosis. Following centromere division at anaphase, the spindle fibers pull the separated sister chromatids of each chromosome to opposite poles (Figure 2.11). The DNA of the two sister chromatids is identical, barring any errors in DNA replication. Thus the effect of mitosis is to generate daughter cells that contain precisely the same DNA sequences.

2.4.2. Meiosis is a specialized form of cell division giving rise to sperm and egg cells

Primordial germ cells migrate into the embryonic gonad and engage in repeated rounds of mitosis (many more in males than in females, which may be a significant factor in explaining sex differences in mutation rates - see Figure 9.4) to form oogonia in females and spermatogonia in males. Further growth and differentiation produces primary oocytes in the ovary and primary spermatocytes in the testis. These specialized diploid cells can undergo meiosis (Figure 2.12). Meiosis involves two successive cell divisions but only one round of DNA replication, so the products are haploid. In males, the product is four spermatozoa; in females, however, the cytoplasm divides unequally at each stage: the products of meiosis I (the first meiotic division) are a large secondary oocyte and a small cell (polar body). The secondary oocyte then gives rise to the large mature egg cell and a second polar body.

There are two crucial differences between mitosis and meiosis (Table 2.2).

• The products of mitosis are diploid; the products of meiosis are haploid.

• The products of mitosis are genetically identical; the products of meiosis are genetically different.

Mitosis involves a single turn of the cell cycle (Figure 2.2). The DNA is replicated in S phase and the two copies are divided exactly equally between the daughter cells in M phase. Meiosis is also preceded by one round of DNA synthesis, but then there are two cell divisions without intervening DNA synthesis, so that the products end up haploid. The second division of meiosis is identical to mitosis, but the first division has important differences whose purpose is to generate genetic diversity between the daughter cells. This is done by two mechanisms, independent assortment of paternal and maternal homologs (Figure 2.13), and recombination.

2.4.3. X-Y pairing and the pseudoautosomal regions

In female meiosis, each chromosome has a fully homologous partner, and the two Xs synapse and cross over just like any other pair of homologs. In male meiosis there is a problem. The human X and Y sex chromosomes are very different from one another. Nevertheless, they do pair in prophase I in males, thus ensuring that at anaphase I each daughter cell receives one sex chromosome, either the X or the Y. X-Y pairing is end-to-end rather than along the whole length, and it is made possible by a 2.6 Mb region of homology between the X and Y chromosomes at the tips of their short arms. Pairing is sustained by an obligatory crossover in this region. Genes in the pairing segment have some interesting properties:

• they are present as homologous copies on the X and Y chromosomes;

• they are not subject to X-inactivation (as expected since each sex has two copies);

• because of the crossing over, alleles at these loci do not show the normal X-linked or Y-linked patterns of inheritance, but segregate like autosomal alleles.

Because of this behavior, this region is known as the major pseudoautosomal region. A second smaller pseudoautosomal region of 320 kb is located at the tips of the long arms of both chromosomes, but pairing and crossing-over in this minor pseudoautosomal region is not an obligatory feature of male meiosis.

2.5.1. Mitotic chromosomes can be seen in any dividing cell, but most conveniently in lymphocytes; meiotic chromosomes are hard to study in humans

Chromosomes can only be seen in dividing cells, and obtaining dividing cells directly from the human body is difficult. Bone marrow is a possible source, but it is much easier all round to take an accessible source of nondividing cells and culture them in the laboratory. Blood is the material of choice - most people don't mind giving a few millilitres, and the T lymphocytes in blood can be easily induced to divide by treatment with lectins such as phytohemagglutinin. Other common sources include fibroblasts grown from skin biopsies, and (for prenatal diagnosis) chorionic villi or fetal cells shed into the amniotic fluid.

Although chromosomes were described accurately in some organisms as early as the 1880s, for many decades all attempts to prepare spreads of human chromosomes produced a tangle that defied analysis. The key to getting analyzable spreads was a new technique, growing cells in liquid suspension and treating them with hypotonic saline to make them swell. This allowed the first good quality preparations to be made in 1956. White cells from blood are put into a rich culture medium laced with phytohemagglutinin and allowed to grow for 48–72 hours, by which time they should be dividing freely. Nevertheless, because M phase occupies only a small part of the cell cycle, few cells will be actually dividing at any one time. The mitotic index (proportion of cells in mitosis) is increased by treating the culture with a spindle disrupting agent such as colcemid. Cells reach M phase of the cycle, but are unable to leave it, and so cells accumulate in metaphase of mitosis. Often it is preferable to study prometaphase chromosomes, which are less contracted and so show more detail. Cell cultures can be prevented from cycling by thymidine starvation; when the block is released the cells progress through the cycle synchronously. By trial and error, the time after release can be determined when a good proportion of cells are in the desired prometaphase stage.

Meiosis can only be studied in testicular or ovarian samples. Female meiosis is especially difficult, as it is active only in fetal ovaries, whereas male meiosis can be studied in a testicular biopsy from any post-pubertal male who is willing to give one. The results of meiosis can be studied by analyzing chromosomes from sperm, although the methodology for this is cumbersome. Meiotic analysis is used for some investigations of male infertility.

2.5.2. Chromosomes are identified by their size, centromere position and banding pattern

Until the 1970s chromosomes were identified on the basis of their size and the position of the centromeres. This allowed chromosomes to be classified into groups (Table 2.3) but not unambiguously identified. The introduction of banding techniques (Box 2.4) finally allowed each chromosome to be identified, as well as permitting more accurate definition of translocation breakpoints, subchromosomal deletions, etc. Banding resolution can be increased by using more elongated chromosomes, for example chromosomes from prometaphase or earlier, rather than metaphase. Typical high-resolution banding procedures for human chromosomes can resolve a total of 400, 550 or 850 bands (Figures 2.16, 2.18). The chromosome constitution is described by a karyotype that states the total number of chromosomes and the sex chromosome constitution. Human females and males are 46,XX and 46,XY respectively. When there is a chromosomal abnormality the karyotype also describes the type of abnormality and the chromosome bands or sub-bands affected. See Box 2.5 for details of chromosome nomenclature. Chromosomes are displayed as a karyogram (often loosely described as a karyotype). Karyograms such as Figure 2.17 are prepared by cutting up a photograph of the spread, matching up homologous chromosomes and sticking them back down on a card - or nowadays more often by getting an image analysis computer to do the job.

2.6.1. Types of chromosomal abnormality

A chromosomal abnormality may be present in all cells of the body (constitutional abnormality), or may be present in only certain cells or tissues (somatic or acquired abnormality). Constitutional abnormalities must have been present very early in development, most likely the result of an abnormal sperm or egg, or maybe abnormal fertilization or an abnormal event in the early embryo. By contrast, an individual with a somatic abnormality is a mosaic (see Figure 3.9), containing cells with two different chromosome constitutions, with both cell types deriving from the same zygote. Chromosomal abnormalities, whether constitutional or somatic, mostly fall into two categories: numerical and structural abnormalities (Box 2.6). Occasionally, abnormalities have been identified in which chromosomes have the correct number and structure, but represent unequal contributions from the two parents (Section 2.6.4). Perhaps unexpectedly, correct parental origin matters.

2.6.2. Numerical chromosomal abnormalities involve gain or loss of complete chromosomes

Three classes of numerical chromosomal abnormalities can be distinguished: polyploidy, aneuploidy and mixoploidy.

2.6.3. Structural chromosomal abnormalities result from misrepair of chromosome breaks or from malfunction of the recombination system

Chromosome breaks occur either as a result of damage to DNA (by radiation or chemicals, for example) or as part of the mechanism of recombination. In G₂ phase of the cell cycle (Figure 2.2) chromosomes consist of two chromatids. Breaks occurring at this stage are manifest as chromatid breaks, affecting only one of the two sister chromatids. Breaks occurring in G₁ phase, if not repaired before S phase, appear later as chromosome breaks, affecting both chromatids. Cells have enzyme systems that recognize and if possible repair broken chromosome ends. Repair can be either by joining two broken ends together, or by capping a broken end with a telomere. Cell cycle checkpoint mechanisms (Section 18.7.3) normally prevent cells with unrepaired chromosome breaks from entering mitosis; if the damage cannot be repaired, the cell commits suicide (apoptosis).

Structural abnormalities arise when breaks are repaired incorrectly. Provided there are no free broken ends, the cell cycle checkpoints can be negotiated satisfactorily, but in fact the wrong broken ends may have been joined together. Any resulting chromosome that has no centromere (acentric) or two centromeres (dicentric) will not segregate stably in mitosis, and will eventually be lost. Chromosomes with a single centromere can be stably propagated through successive rounds of mitosis, even if they are structurally abnormal. Meiotic recombination between mispaired chromosomes is a common cause of translocations, especially in spermatogenesis. Table 2.5 summarizes the main stable structural abnormalities and Figures 2.20 and 2.21 illustrate how they arise.

An additional rare class of structural abnormality not shown in Table 2.5 are isochromosomes. These are symmetrical chromosomes consisting of either two long arms or two short arms of a particular chromosome. They are believed to arise from an abnormal U-type exchange between sister chromatids just next to the centromere of a chromosome. Isochromosomes are rare except for i(Xq); i(21q) are an occasional cause of Down syndrome.

Structural chromosomal abnormalities are balanced if there is no net gain or loss of chromosomal material, and unbalanced if there is net gain or loss. In general, balanced abnormalities (inversions, balanced translocations) have no effect on the phenotype, although there are important exceptions to this:

• a chromosome break may disrupt an important gene;

• the break may affect expression of a gene even though it does not disrupt the coding sequence. It may separate a gene from a control element, or it may put the gene in an inappropriate chromatin environment, for example translocating a normally active gene into heterochromatin;

• balanced X-autosome translocations cause problems with X-inactivation (see Figure 15.9).

Robertsonian translocations are sometimes called centric fusions, but this is misleading because in fact the breaks are in the proximal short arms. The translocation chromosome is really dicentric, but because the two centromeres are very close together they function as one, and the chromosome segregates regularly. The distal parts of the two short arms are lost as an acentric fragment. Short arms of acrocentric chromosomes contain only arrays of repeated ribosomal RNA genes, and the loss of two short arms has no phenotypic effect. Because there is no phenotypic effect, Robertsonian translocations are regarded as balanced, even though in fact some material has been lost.

Unbalanced abnormalities can arise directly, through deletion or, rarely, duplication, or indirectly by malsegregation of chromosomes during meiosis in a carrier of a balanced abnormality. Carriers of balanced structural abnormalities can run into trouble during meiosis, if the structures of homologous pairs of chromosomes do not correspond:

• A carrier of a balanced reciprocal translocation can produce gametes that after fertilization give rise to an entirely normal child, a phenotypically normal balanced carrier, or various unbalanced karyotypes that always combine monosomy for part of one of the chromosomes with trisomy for part of the other (Figure 2.22). It is not possible to make general statements about the relative frequencies of these outcomes. The size of any unbalanced segments depends on the position of the breakpoints. If the unbalanced segments are large, the fetus will probably abort spontaneously; unbalance for smaller segments may result in liveborn abnormal babies.

• A carrier of a balanced Robertsonian translocation can produce gametes that after fertilization give rise to an entirely normal child, a phenotypically normal balanced carrier, or a conceptus with full trisomy or full monosomy for one of the chromosomes involved (Figure 2.23).

• A carrier of a pericentric inversion may produce unbalanced offspring because when the inverted and noninverted homologs pair they form a loop so that matching segments pair along the whole length of the chromosomes. If a crossover occurs within the loop, the result is a chromosome carying an unbalanced deletion and duplication. Paracentric chromosome inversions form similar loops, but any crossover within the loop generates an acentric or dicentric chromosome, which is unlikely to survive. For details of meiosis in carriers of inversions, see the book by Gardner and Sutherland (Further reading) or any other cytogenetics text.

2.6.4. Apparently normal chromosomal complements may be pathogenic if they have the wrong parental origin

The rare abnormalities described below demonstrate that it is not enough to have the correct number and structure of chromosomes; they must also have the correct parental origin. 46,XX conceptuses in which both genomes originate from the same parent (uniparental diploidy) never develop correctly. For some individual chromosomes, having both homologs derived from the same parent (uniparental disomy) also causes abnormality. A small number of genes are imprinted with their parental origin (Section 8.5) and are expressed differently according to the origin. It is assumed that the abnormalities of uniparental disomy and uniparental diploidy are caused by abnormal expression of such imprinted genes.

Uniparental diploidy is seen in hydatidiform moles, abnormal conceptuses with a 46,XX karyotype of exclusively paternal origin. Molar pregnancies show widespread hyperplasia of the trophoblast but no fetal parts, and have a significant risk of transformation into choriocarcinoma. Genetic marker studies show that most moles are homozygous at all loci, indicating that they arose by chromosome doubling from a single sperm. Ovarian teratomas are the result of maternal uniparental diploidy. These rare benign tumors of the ovary consist of disorganized embryonic tissues, without extraembryonic membranes. They arise by activation of an unovulated oocyte.

Uniparental disomy (UPD), affecting a single pair of homologs, goes undiagnosed if the result is not abnormal, but is detected for chromosomes for which it produces characteristic syndromes (see Box 16.6). UPD can be isodisomy, where both homologs are identical, or heterodisomy, where they are derived from both homologs in one parent. The usual cause is thought to be trisomy rescue: a conceptus that is trisomic and would otherwise die, occasionally loses one chromosome by mitotic nondisjunction or anaphase lag from a totipotent cell. The euploid progeny of this cell form the embryo, while all the aneuploid cells die. If each of the three copies has an equal chance of being lost, there will be a two in three chance of a single chromosome loss leading to the normal chromosome constitution and a one in three chance of uniparental disomy (either paternal or maternal). Uniparental isodisomy may possibly arise by selection pressure on a monosomic embryo to achieve euploidy by selective duplication of the monosomic chromosome.