Sex Linkage, Dosage Compensation and Non-Mendelian Inheritance

In 1910 a white-eyed male fruit fly appeared in a laboratory where every other fly had red eyes, and its crosses gave different answers depending on which parent was white.

Mendel’s peas gave the same result whichever plant was the mother. The white-eyed fly did not. This lesson asks why, then follows the answer, a gene riding on the X chromosome, to how sex is decided, how animals balance their X genes, and which other genes care about the parent of origin.

One white-eyed male: a gene that follows the X

Thomas Hunt Morgan’s laboratory at Columbia University bred Drosophila melanogaster. The white-eyed male turned up in 1910. Morgan crossed it to red-eyed females and got red-eyed offspring of both sexes, so white is recessive: an allele that shows only when no dominant allele is present. He published the work that year in Science as “Sex limited inheritance in Drosophila”. The term sex-linked means carried on a sex chromosome.

You already know, from chromosomes and meiosis, that a female fly is XX and a male XY. The white gene sits on the X, and the Y carries no copy. So a male is hemizygous for white: he has one copy, not two, and whatever allele his X carries is what you see. A male also gets his only X from his mother. Write X+ for the red allele and Xw for white.

Worked problem. Assume ordinary segregation, equal survival of all genotypes, half the offspring of each sex, and complete penetrance (every fly with the genotype shows the phenotype).

Cross A. Red female X+X+ by white male XwY.

1. The mother makes only X+ eggs. The father makes Xw and Y sperm, half each.
2. F1: daughters X+Xw, red. Sons X+Y, red. All red.
3. Intercross the F1: X+Xw female by X+Y male. Her eggs are X+ or Xw, half each. His sperm are X+ or Y, half each.
4. F2 daughters all get X+ from the father: half X+X+, half X+Xw, all red.
5. F2 sons get their X from the mother: half X+Y (red), half XwY (white).
6. Overall: 3 red to 1 white. Every white fly is a male.

Cross B, the reciprocal. White female XwXw by red male X+Y.

1. The mother makes only Xw eggs. The father makes X+ and Y sperm.
2. F1: daughters X+Xw, red. Sons XwY, white. The father’s phenotype goes to his daughters and the mother’s to her sons. This is criss-cross inheritance.
3. Intercross the F1: X+Xw female by XwY male. Her eggs are X+ or Xw; his sperm are Xw or Y.
4. The four combinations each have probability 1/2 x 1/2 = 1/4:

X+Xw red female : XwXw white female : X+Y red male : XwY white male = 1 : 1 : 1 : 1

5. Among 400 F2 flies, expect 100 in each class. Check: 4 x 100 = 400; 200 of each sex; 200 white.

Conclusion. An autosomal gene would give the same F1 in both crosses, since each parent passes one copy to every child. These reciprocal crosses differ, and the differences follow the X: sons always take the mother’s allele. So the gene is on the X. Segregation still works as Mendel said. Real broods are noisier than these expectations. In Morgan’s own F2 from Cross A the white-eyed males made up less than a quarter of the flies.

The same logic covers human X-linked conditions: glucose-6-phosphate dehydrogenase deficiency (pentose phosphate pathway) and the HPRT deficiency of Lesch-Nyhan syndrome (nucleotide metabolism) are both seen far more often in males. Pedigrees shows the pattern in a family tree.

How sex is decided: mammal versus fly

Mice and flies are both XX female and XY male, by different mechanisms.

In mammals, the Y decides. A gene on the Y, called SRY in humans and Sry in mice, starts testis development. Sinclair and colleagues reported SRY in 1990. In 1991 Koopman and colleagues added a 14-kilobase DNA fragment containing Sry to chromosomally female mouse embryos, and it was sufficient to make them develop testes. The consequence shows in people with unusual sex chromosomes: XO individuals (Turner syndrome) develop as females and XXY individuals (Klinefelter syndrome) as males, because one Y carrying SRY is the switch.

In Drosophila, the Y does not decide. In a diploid fly, two X chromosomes ordinarily specify female development and one specifies male. The classic description is a ratio of X chromosomes to sets of autosomes. Erickson and Quintero argued in 2007 that the embryo reads the combined dose of a few X-linked signal genes rather than measuring a ratio. Either way, a high X dose switches on a gene called Sex-lethal in females. Once on, Sex-lethal keeps itself on by controlling the splicing of its own RNA, and it steers the splicing of the genes below it. In flies, XO is male and XXY is female, the mirror of the human rule. The fly Y is needed for male fertility, not for maleness.

Balancing the X: three solutions in three animals

With two X chromosomes in females and one in males, females would make about twice as much of each X-linked product. Dosage compensation is any mechanism that evens this out.

Mammals switch one X off in females. Mary Lyon proposed in 1961 that in each cell of a female mouse one X is inactivated early in development, maternal or paternal, and that all descendants of that cell keep the choice. A female heterozygous for an X-linked gene is therefore a mosaic, a patchwork of cells expressing one allele or the other. The calico or tortoiseshell cat is the visible case: an X-linked coat-colour gene with an orange and a black allele gives patches of each. Calico males are rare and are usually XXY. The inactive X is the condensed Barr body seen in female nuclei. It is coated by Xist, an RNA that appears not to code for protein and that, in mouse experiments where the gene was deleted, is required for inactivation. Silencing acts in cis, on the chromosome that makes the Xist RNA. Mosaicism also matters in disease: in Rett syndrome, which affects mostly girls and involves the X-linked gene MECP2 (synapse disorders), a girl’s cells differ in which MECP2 allele they express.

Two corrections. The choice is not always random: in mouse it is random in the embryo proper, but the paternal X is the one inactivated in the extraembryonic tissues, and in marsupials the paternal X is inactivated. And the inactive X is not completely silent. A 2017 survey of human tissues found incomplete inactivation at no fewer than 23 percent of the X-linked genes it could assess. The mechanism of silencing is part of eukaryotic regulation.

Flies turn the male X up. In male Drosophila, no X is switched off. A protein complex called MSL, for male-specific lethal, coats the single X with the help of two noncoding RNAs, roX1 and roX2, and raises transcription of active X-linked genes roughly twofold. Its enzyme MOF acetylates histone H4 at lysine 16, a mark linked to more open chromatin. Females avoid this because Sex-lethal blocks translation of the mRNA for MSL2, the subunit that starts assembly. Force MSL2 into females and the complex gathers on their X chromosomes and female viability drops.

Worms turn both X chromosomes down. In the nematode C. elegans, XX hermaphrodites reduce expression from each of their two X chromosomes by about half, matching the single X of XO males.

Genes that remember their parent: organelles and imprinting

Organelle genes travel with the egg. Mitochondria and chloroplasts carry their own small genomes outside the nucleus. Because an egg has abundant cytoplasm and a sperm or pollen nucleus brings little or none, these genomes are generally inherited from the mother. In human families, Giles and colleagues showed in 1980 that children carried their mother’s pattern of mitochondrial DNA variants rather than their father’s: both sons and daughters inherit it, but only daughters pass it on. Human sperm mitochondria do enter the egg, but a 2023 study concluded that mature human sperm are essentially devoid of mitochondrial DNA. “Generally maternal” is the right phrase: Baur’s Pelargonium showed plastid inheritance from both parents, and paternal and biparental transmission are documented in other species.

Imprinted genes are nuclear but remember their parent. In 1984 two groups, McGrath and Solter, and Surani, Barton and Norris, built mouse embryos with two maternal or two paternal genomes. Neither kind completed development, though control embryos handled the same way did. The two parental genomes are not equivalent. The explanation is genomic imprinting: a subset of genes is expressed from only one parental allele, set by marks laid down during egg or sperm formation. The marks are erased in the primordial germ cells of each new generation and rewritten according to the sex of the animal, which is why the mark depends on the parent and not on the grandparent. In mouse, Igf2, a growth factor gene, is expressed from the paternal allele in most tissues, and H19, a nearby gene, from the maternal allele. A mouse that inherits a disrupted Igf2 from its father is small; one that inherits the same allele from its mother is normal.

In humans, the region 15q11-q13 holds imprinted genes. Loss of the paternal contribution of this region causes Prader-Willi syndrome; loss of the maternal contribution causes Angelman syndrome. Keep the two apart: an imprinted gene is a nuclear gene passed on by both parents, with only one copy used.

Design problem

When Calvin Bridges, one of Morgan’s students, crossed white-eyed females (XwXw) to red-eyed males (X+Y), roughly 1 in 2,000 offspring were exceptions: white-eyed daughters or red-eyed sons, which Cross B says should not exist. Propose an explanation that keeps the white gene on the X, and design a test of it.