The Discovery of X-Chromosome Inactivation
Mary F. Lyon
The sex chromosomes differ from all others in that they are the only ones that vary in number between different individuals within a species. The severe developmental defects that accompany rare extra copies of autosomes raised the question of how XX females and XY males can accommodate different numbers of X-chromosomes. In mammals the almost complete inactivation of one X- chromosome in each cell of a female provided an answer. The discovery of X-chromosome inactivation arose from a synthesis of three or four separate observations in different areas of genetics. Although discovered in the mouse, it proved to be a general mechanism among mammals.
The first inkling that the X-chromosome might behave strangely came with the discovery of the first mouse X-linked gene in 1953. Females heterozygous for a mutant allele of the gene had an unusual pattern of white spotting in the coat, different from the types of patterns due to autosomal genes. Males with the mutant gene died in utero. (Much later it was discovered that both the white spotting and the lethality were due to a gene affecting copper transport in the body). In the next few years other X-linked mutant genes were found which produced similarly variegated (i.e. mottled or spotted) coat patterns in heterozygous females, but not in affected males. It began to appear that the X-chromosome was unusual in this way. Further insight came when I found a single spontaneous mutant male, with a similar pattern of white spotting, in a uniformly darkly pigmented stock (Lyon, 1960). When he was bred, some but not all of his daughters had coats resembling his. Breeding studies with the spotted daughters showed that the trait was due to a new X-linked mutation which, as in some previous cases, was lethal in utero in male mice. This raised the question of how the original mutant male had survived.
A possible explanation for his survival was that the mutation which had inactivated the color gene had occurred early in his development, and in only one of the relatively small number of cells in the embryo at the time. This would allow him to survive because most of his cells would still be normal. I assumed that the pattern of spots in his coat had arisen during development by the migration and mingling of the two types of cells, normal and mutant, followed by their clonal growth to give relatively large patches of normal or mutant color. As expected on this basis, the male’s pattern was similar to that seen in mice with somatic mutations resulting from irradiation of early embryos. Thus, the male’s pattern resulted from an early event in development, which in his case was somatic mutation. As the pattern in females heterozygous for X-linked genes was similar, I suggested that it also stemmed from an early developmental event which resulted in the presence of two types of cells. However, at that time I could not suggest the nature of this early event.
Further progress towards the idea that the early event was inactivation of one X-chromosome came when I became aware of other discoveries. In 1949 Barr and Bertram (Barr and Bertram, 1949) reported the serendipitous discovery of the sex chromatin body in nuclei of somatic cells of mammals. Murray Barr and his student Ewart Bertram were engaged in a detailed study of the histology of neurons of the cat, looking for changes induced by prolonged stimulation of the nerve. They found none, but in the course of looking noticed a small body near the nucleolus of these neurons, which they called the "nucleolar satellite". After looking in a few different animals they realized that this feature was seen only in the cells of females. It was soon renamed the "sex chromatin" and found to be present generally in nuclei of female mammals. At that time they had no idea of its function or composition, the term "chromatin" being applied to it only because it stained well with the same dyes that were used to stain chromosomes.
The underlying structure of the sex chromatin remained a mystery for ten years. At the time, ideas in the field of animal genetics were dominated by results in the fly, Drosophila melanogaster, then the predominant model organism. In Drosophila the X-chromosome consists partly of heterochromatin, highly condensed chromosomal material that contains few genes and stains darkly relative to other parts of a chromosome. It was assumed that presence of heterochromatin was a property of X-chromosomes in general. Thus, one suggestion was that the sex chromatin might be formed from the heterochromatic part of both X-chromosomes of females. A breakthrough came when Susumu Ohno (Ohno and Hauschka, 1960) and colleagues showed that in prophase cells of female rats and mice a single whole chromosome was heterochromatic (ie it stained darkly and was condensed), and no such chromosome was visible in males. The heterochromatic chromosome was interpreted as an X-chromosome, and this was the first indication that the two X-chromosomes of female mammals behave differently.
A further major advance in genetic knowledge at the same time concerned mammalian sex determination. Welshons and Russell (Welshons and Russell, 1959) showed that mice with a single X-chromosome and no Y chromosome, so-called XO mice, were viable, fertile females. Similarly, human XO individuals are female, albeit malformed. Thus, in mammals the Y-chromosome is male determining. This constitutes another difference from Drosophila, in which sex is determined by the ratio of X-chromosomes to autosomal sets; a ratio of 1X:2A giving maleness, and 2X:2A femaleness. The significance of this work with respect to X-inactivation, however, was the conclusion that, in the mouse at least, a female requires the activity of only one X-chromosome for normal development.
At this point all the facts were in place for the discovery of X-chromosome inactivation. I realized that if XO mice developed normally, then normal XX mice need only have one X-chromosome active, and that the "early event" previously suggested to underlie the variegation of heterozygous mutant females could be inactivation of one X-chromosome. The inactive X-chromosome would take on the heterochromatic properties of condensation and dark staining which Ohno had recognised, and form the sex chromatin. If either of the two X-chromosomes could be inactive in any cell, this would give rise to the two types of cells that I had earlier suggested to occur in heterozygous females. I then put the idea forward in a letter to Nature (Lyon, 1961). I suggested that the heteropyknotic (or heterochromatic) X-chromosome could be either of paternal or maternal origin in different cells of the same animal, and that it was genetically inactive. The inactivation was postulated to occur early in embryonic development. Figure 1 shows that this would result in the adult having large patches of cells with the same X-chromosome active, resulting in mutant and normal colored patches in heterozygotes.
The first paper was limited to the mouse only, but raised the possibility that X-inactivation might occur in mammals generally, since the sex chromatin is found in other species. In the second paper (Lyon, 1962) I extended the idea to man and mammals generally, on the basis of variegated or mosaic phenotypes that had been observed in females heterozygous for some mutant genes. In humans there were at that time no X-linked mutant genes known which affected the color of hair or skin, but the X-inactivation theory predicted that there should be two types of cells present regardless of the cell type in which a mutant gene acted. The pattern in the affected tissue would depend on the manner in which the cells migrate and mingle during development. Thus, two genes were discussed that affect the pigmentation of the retina, causing heterozygous females to have light and dark retinal patches. Other genes, including glucose-6-phosphate dehydrogenase (G6PD) deficiency, affected red blood cells, and in heterozygotes a mixture of normal and mutant cells were present. In yet other cases, including color blindness and hemophilia, the mingling of cells was so complete that the two cell types could not be seen, but the heterozygous females were mildly abnormal, indicating that mutant cells were in fact present. Among other species of mammals the tortoiseshell cat was notable. The tortoiseshell pattern is seen in female cats heterozygous for an X-linked gene causing orange color, X-inactivation giving rise to patches of orange and black. However, despite the fact that so many X-linked genes showed such variegated patterns, these observations did not prove that the whole chromosome was inactivated. Independent inactivation of genes on one X-chromosome or the other was also a possibility.
The X-inactivation hypothesis explained many observations, but still needed to be tested by experiment. One way to do this was to ask if two genes on different copies of the X-chromosome could ever be inactivated in the same cell. To do this experiment I bred mice which could potentially show three colors in their coats, as shown in Figure 2. They were homozygous for a recessive allele of an autosomal gene (called "pink-eye", symbol p) which could give their coats a sandy color, but their two X-chromosomes carried different genes which would mask this. On one X-chromosome was an X-linked gene called "mottled" (Mo), giving a white color, and on the other the normal counterpart of the pink-eye gene, which had been transferred to the X-chromosome by a translocation, and which would result in a dark color. The two genes for white and dark were at different points on the X-chromosome. If the inactivation affected a whole X-chromosome then in every cell either the white gene or the dark gene would be active, and the mice would have no sandy patches. If on the other hand each gene was being affected independently, then the mice would show three colors. In fact, all the mice with the two test X-chromosomes showed only white and dark patches. I did a similar test also with two other X-linked genes that affected the hair texture of the coat rather than its color, and this test gave a similar result. Thus, these results were as expected if one or the other whole chromosome was being inactivated in every cell. Later, Davidson et al (Davidson and Childs, 1963) obtained crucial evidence of the irreversibility of X-inactivation by growing colonies of cultured cells from single cells of human females heterozygous for the X-linked enzyme glucose-6-phosphate dehydrogenase (G6PD) and showing that each clone expressed one allele only of G6PD.
X-chromosome inactivation is an example of a discovery made when the time became ripe for it. Knowledge of the variegated phenotype of heterozygotes for X-linked genes, the normality of XO mice, and the fact that the sex chromatin was a single, complete X chromosome, all essential for the idea, came together in a short time. Not surprisingly, others had similar ideas. Beutler (Beutler and Fairbanks, 1962) made a similar proposal to explain the two types of red blood cell present in women heterozygous for G6PD deficiency, and Russell (Russell, 1961) made a less complete suggestion to explain the variegation seen in female mice carrying X-autosome translocations, saying that it was "presumably a heterochromatic effect".
Some further developments in ideas concerning X-inactivation took place early on. The phenomenon of X-inactivation was regarded as a mechanism for equalizing the effective gene dosage of X-linked genes in XX females and XY males (called "dosage compensation"). Therefore, it was suggested that since X-linked genes with copies on the Y would not require dosage compensation they would escape X-inactivation (Lyon, 1962). This is now known to be indeed the case, but some genes without copies on the Y also escape, particularly in the human (Disteche, 1995), and the basis of this escape is not understood.
Another early modification of the idea came from findings in humans with supernumerary X-chromosomes. In these individuals the number of sex chromatin bodies per cell is always one less than the number of X-chromosomes. Thus, the hypothesis was modified to say not that one X-chromosome is inactivated but rather that a single X-chromosome remains active (Lyon, 1962). There is thought to be a counting mechanism maintaining one X-chromosome active per two autosome sets, but the mechanism of this counting is not yet known. This single X-chromosome activity applies to somatic cells only, with different effects in germ cells. Figure 3 shows that in oocytes of females the inactive X-chromosome is reactivated as the cells enter meiosis, and in males the X-chromosome becomes inactive during spermatogenesis.
The concept of the X-chromosome inactivation center was another early development. The idea was put forward to explain findings in X-autosome translocations (Lyon, 1963; Russell, 1963). In such translocations inactivation can spread from the X-chromosome into the attached autosomal segment, but only one of the two X-chromosome segments created by the translocation can cause this. The segment from which inactivation spreads is thought to include a special site from which inactivation is initiated, called the X-inactivation center. X-chromosome segments lacking this site remain active.
Not at all foreseen was the role of imprinting in X-inactivation. Imprinting is defined as a difference in the expression of two alleles within a cell depending on their parental origin. This is now a well-established phenomenon among some mammalian autosomal genes. It is also now known that imprinting occurs in X-inactivation, in that there is preferential inactivation of the paternally derived X-chromosome in extraembryonic cells of eutherians and in all cells of marsupials ( Figure 3).
At the time of its discovery X-inactivation was of both medical and scientific importance. Medically it provided understanding of the phenotypes of women heterozygous for X-linked genes, who typically show two types of cells. Further, it explained the viability of individuals with abnormal numbers of X-chromosomes, such as XO, XXY or XXXX. Comparable anomalies of autosomes are either lethal or result in severe abnormalities, eg Down’s syndrome. The much milder phenotype of X-chromosome anomalies clearly results from the inactivity of all save one X-chromosomes. The fact that there is any abnormality is ascribed to the escape of some genes from inactivation, and consequent wrong dose of the products of these genes. Scientifically X-inactivation provided a very interesting and unusual system of gene regulation. In addition, together with the findings on sex determination, the discovery established that Drosophila is not a universal model for genetic phenomena. There is now a third major animal genetic model organism, Caenorhabditis elegans, and all three genetic models show dosage compensation for X-linked genes. In each case, however, the mechanism is different.
At present (2001) there is much interest in the mechanism of X-inactivation as an example of a system of gene regulation. For a whole, or almost whole, chromosome to be transcriptionally inactive while its homologue in the same cell is normally active is still a very unusual phenomenon. The idea of the X-inactivation center has been well borne out. The position of the center has been identified in both human and mouse, and from the relevant region a gene known as XIST or Xist ( X-inactive specific transcript) has been cloned ( Figure 3). The gene is unique in being active only on the otherwise inactive X-chromosome. It codes for an untranslated RNA which coats the inactive X-chromosome. Specific deletions of parts of Xist by gene knockouts have shown that its RNA is essential for inactivation to occur. When Xist is introduced as a transgene into an autosome it is also sufficient to suppress transcription of genes on that chromosome (reviewed in Heard, Clerc, and Avner, 1997). However, the inactive X is now known to have a set of other unusual properties. As well as condensation to form the sex chromatin body, by which it was first discovered, the inactive X-chromosome replicates its DNA late in the synthesis phase of the cell cycle, its histones are hypoacetylated, the cytosines of its dense clusters of CpG nucleotide pairs (known as CpG islands) are methylated ( Figure 3), and it is associated with an unusual histone, macroH2A1. The work of Wutz and Jaenisch (Wutz and Jaenisch, 2000) has shown that Xist RNA alone is not sufficient to bring about these other properties. Some unknown developmental factor is needed also. These special properties are thought to be involved in locking in the inactive state, which in eutherians is highly stable. Thus, Xist alone can initiate the transcriptional silencing of genes but other unknown factors are needed to make this silencing permanent. Some of the features of the inactive X-chromosome are also seen in other genetically inactive chromatin, such as heterochromatin blocks near centromeres or elsewhere. However, the mechanisms involved are not understood (for review see GENES 2000: 22.15 X chromosomes undergo global changes).
X-inactivation is likely to be a rewarding subject for research for some time yet while the mechanisms bringing about its special features, and their role, are worked out. A particularly fascinating question, that has intrigued scientists from the first, is how one X-chromosome can be singled out to remain active, while all others present are inactivated. It is suggested that there is a limited quantity of some blocking factor, which blocks one inactivation centre. However, this is no more than a formal explanation, since the nature of the blocking factor is unknown. It need not be a substance; it could be a specific site within the cell at which the X-chromosome must attach. If it is a substance, what could it be? The race is on to find out.
Mary Lyon was born in Norwich, England in 1925, and received her higher education at Cambridge University (B.A. 1946; Ph.D. 1950; ScD 1968). She then joined a group in Edinburgh set up to study the genetic hazards of radiation, using mutagenesis experiments in mice. In 1955 she moved with this group to the MRC Radiobiology Unit, Harwell, where she headed the Genetics Section from 1962-86. It was while working on radiation hazards in 1961 that she discovered X-chromosome inactivation, for which she is best known. She has also done extensive work on the mouse t-complex, and made many other contributions to mammalian genetics. She is a Fellow of the Royal Society, a Foreign Associate of the US National Academy of Sciences, and a Foreign Honorary Member of the American Academy of Arts and Sciences. Among her awards is the Wolf Prize for Medicine in 1997.