AS A CHILD, I ALWAYS PREFERRED GAMES INVOLVING PHYSICAL activity to board games. In part, this was because I found board games boring; in part, because my sister was clearly superior. Monopoly was particularly unpleasant in both regards. But a broken leg at age eight severely restricted my physical activities. Still untempted by board games, I focused much of my attention on a game called Caroms, a sort of billiardish affair played on a small, square wooden surface, with wooden sides and four corner pockets. The wooden cues were about eighteen inches long. The object was to pocket the small doughnut-like objects, also made of wood, which came in two colors: red and green. You were responsible for knocking in all of the reds, or greens, as the case may be—like the solids and stripes in pool. Whoever knocks in all of the appropriate wooden disks wins.
Caroms had at least two virtues from my perspective: it required some physical skills, and I completely trounced my sister at it. In fact, I pretty much beat everyone. Most of my friends presented a bigger challenge than my sister, who in any case soon developed an aversion to the game. But one of my friends, Steve, was even worse than my sister, for reasons that I initially could not comprehend. It certainly wasn’t for lack of hand-eye coordination. Steve was fine in that regard. The problem was that Steve was quite indiscriminate as to which caroms he pocketed. At first I attributed this to boredom, as Steve was even more physically active than me. I thought he might just be trying to get the game over with. But he actually seemed to enjoy it, and when I called his attention to the fact that he had just pocketed one of mine, he would just smile. I found his smile disconcerting, because Steve was much more worldly than I; it was he who first informed me that there was no Santa Claus and that the Tooth Fairy was in fact my mom. So initially, I suspected that he had a motive for “losing” that was unavailable to the uninitiated. This greatly reduced any pleasure I derived from victory. Eventually, I grew so frustrated with Steve that, when he was obviously aiming at a carom of mine, I would call his attention to the fact. Steve would just smile and proceed to knock it in anyway.
At some point, I got my mother involved in order to get to the bottom of Steve’s seeming perversity. It didn’t take her long to discover that Steve was color-blind. He was pocketing my reds with his greens because he couldn’t see the difference. Steve was aware of this at some level, but he was too proud to admit it, hence his disconcerting smile. The fact of Steve’s color blindness did not jar my worldview quite as much as the fictitiousness of Santa Claus, but it did initiate some protophilosophical reflections in my young mind. What does the world look like to Steve—flowers? trees? traffic lights? Especially traffic lights. How would he know when to cross the street when he was alone? Color blindness was fascinating stuff.
Over the years, I have found myself returning at odd intervals to the subject of color blindness, most recently because of an epigenetic connection, the subject of this chapter.
Why Males Are Truly the Weaker Sex
Let’s begin with the fact that the red-green color blindness that Steve exhibited is much more common in boys than girls. In that respect, it resembles a host of other developmental defects, from dyslexia to certain forms of heart disease. Such defects are said to be sex-linked. Sex linkage occurs when the relevant genes reside on the sex chromosomes, overwhelmingly, the X chromosome. The X chromosome is the largest and most gene-rich of our chromosomes, so lots of our traits reflect some degree of sex linkage—or X linkage, to be more precise. The Y chromosome, on the other hand, is a tiny little thing.
Sex-linked mutations have a signature pattern of inheritance. This is particularly true of recessive mutations—those that must be present on both chromosomes—the one you got from your mother and the equivalent one that you got from your father—to have any effect.1 To varying extents, this pattern applies to genes on all other chromosomes, collectively referred to as autosomes, but not to the sex chromosomes—at least, not the male sex chromosomes. Females are blessed with two X chromosomes, one from each parent. Males, on the other hand, only inherit one X chromosome—from the mother—along with the diminutive Y chromosome from the father. As such, any recessive mutation on the maternally derived X chromosome is effectively dominant in males and causes problems. Hence, males are affected at a much higher rate by these recessive mutations than females. The male X deficit is certainly part of the explanation for the fact that at any age or developmental stage, from before birth to dotage, a male has a higher risk of mortality than a female.2
Among the many genes on the X chromosome are two that specify opsins, the color-sensitive proteins in cone cells, our color detectors in the retina. There is a third opsin gene, but it resides on chromosome 7, not on the X chromosome.3 Since only one opsin gene is expressed per cone cell, there are three distinct types of cone cells, which we can call red, green, and blue cones. The red and green opsin genes reside on the X chromosome, the blue opsin on chromosome 7. When inherited in a male such as Steve, a recessive mutation in either the red or green opsin gene results in defective red or green cone cells, and hence in red-green color blindness. But even if Steve’s sister had inherited the same mutation from her mother, she would not be color-blind unless she also inherited the mutation on the X chromosome bequeathed by her father, in which case her father must have been color-blind.
That, at least, is the standard textbook explanation for sex-linked traits that I learned in Introductory Genetics. But there must be more to this sex difference, if only for this startling fact: some female carriers of these mutations actually have enhanced color vision.4 These mutant women see color differences that no mortal man can. Let’s call them X-women.
In this chapter, we will get to the bottom of the X-women phenomenon. This will require that we explore a new epigenetic mechanism, one that involves a high degree of randomness. It is appropriate that we use the X chromosome in this regard, because it was in exploring its mysteries that much of the foundation for the science of epigenetics was laid.5
A Dosage Problem
As bad as things are for males X-wise, they would be a lot worse were it not for a process called dosage compensation, which helps level the physiological playing field. Without dosage compensation females would have twice as much of every X chromosome–derived protein as males. This would require that the characteristics of male and female diverge well beyond even those imagined by the most dyed-in-the-wool evolutionary psychologists. And males would be downright feeble compared with females. (Think of those deep-sea anglerfish in which the tiny male attaches to the first giant female that happens by, at which point he degenerates into a physiologically parasitic sperm purveyor: a wartlike testis and little else.)
The evolutionary solution to this dosage problem is called X inactivation,6 in which one of the two X chromosomes in every female cell is inactivated. As a result of X inactivation in females, both males and females have one functional X chromosome per cell. But if both males and females are working with one functional X chromosome, why do males have so many more X-linked problems than females? It turns out that even though females are essentially operating with one X chromosome tied behind their backs, they still derive many of the benefits of having two.
Part of the explanation is that not all of the genes on the inactivated X chromosome are inactivated. In humans, 15 to 25 percent of the genes on the inactivated X chromosome escape inactivation.7 Many of these genes that escape inactivation are referred to as housekeeping genes, which participate in basic cellular processes required by all cells, from skin cells to neurons to cone cells.
There is another reason why females have many of the benefits of two X chromosomes despite the fact that one is largely inactivated. In most mammals, including humans, X inactivation is random with respect to the maternal X and the paternal X. And this random inactivation occurs independently in each cell lineage. That means that in a given population of cells—such as, say, red cone cells—roughly half will have paternal X inactivation, and half maternal X inactivation. Females are essentially X chromosome mosaics. If a female inherits a recessive mutation of the red opsin gene, from either father or mother, only half of her cone cells will be affected, whereas all of a male’s red cone cells would be affected by any such mutation. Operating with half the amount of normal red cone cells is sufficient to avoid color blindness as defined by standard tests, but as we shall see, there can be subtle deficits in the color perception of such females.
In marsupial mammals (kangaroos, koalas, wombats, and such) X inactivation is not random. Instead, the paternal X is always inactivated.8 Hence, male and female kangaroos both operate solely with the maternal X, and are therefore physiologically equivalent with respect to the X chromosome.
Epigenetics of X Inactivation
X inactivation is initiated at a site called the X-inactivation center (Xic). There are several genetic elements within Xic, one of which is particularly crucial for X inactivation: X-inactive-specific transcript (Xist). Sometimes bits of one chromosome get dislodged and land on a different chromosome, a process called translocation. If the bit of the X chromosome containing Xist is translocated to one of the autosomes in this way, the X chromosome cannot be deactivated. The autosome on the receiving end gets (partially) inactivated instead.9 So Xist is absolutely essential for X inactivation.
Xist is not actually a gene in the traditional sense of the term. A gene, you will recall, acts as an indirect template for a protein. But there is no Xist protein, only Xist RNA. That is why it is called X-inactive specific transcript, not X-inactive specific protein (or Xisp). The Xist RNA is quite long and it attaches to the X chromosome from which it is made. As more and more copies of Xist RNA are produced, the X chromosome becomes covered with the stuff, the first stage of inactivation. Next, the Xist RNA attracts histones (see Chapter 5)—which further coat the inactive X—as well as methylating factors. Finally comes the big crunch, when the inactivated X is compacted like a scrapped automobile. Under a microscope, the compacted X chromosome is a little bloblike structure called a Barr body, which doesn’t look anything like the active X chromosome. The compacted X chromosome is still much larger than the Y chromosome, however.
Earlier, I stated that X inactivation is random. That is not strictly true, for two reasons. The first has to do with the timing of X inactivation. We do not know, with any precision, when exactly X inactivation occurs during early development, but it occurs long before birth. There are many cell divisions subsequent to X inactivation, and a given cell lineage retains the X-inactivation pattern of its original X-inactivated founder. So it would be more accurate to say that X inactivation is random with respect to cell lineages, a particular population of hair cells, or cone cells, say. This is easiest to discern in certain patterns of hair (coat) coloration in mammals like cats. Calico and tortoiseshell patterns are particularly useful in this regard, since both colorations are X linked and confined to females. A calico cat reveals in great detail the random X inactivation of hair cell lineages, in the distribution of light and dark and orange areas. It is ironic, therefore, that the first cloned cat was a calico. The owner wanted to recreate her beloved cat, Rainbow. The cloning procedure was successful, but the clone, named Cc (for “carbon copy”) was not even close to a carbon copy.10 She developed a completely different distribution of colors than Rainbow. Given the randomness of X inactivation, this should have been expected. (Cc also displayed a much different personality than Rainbow, but that’s a different story.)
X inactivation is also not random with respect to the maternal tissue that sustains the fetus. Instead, only the paternal X chromosome is inactivated, as in kangaroos and other marsupials.11 The selective inactivation of the X chromosomes from one sex is a form of imprinting, a phenomenon I will discuss in the next chapter. For now it is sufficient to note that in kangaroos, X-chromosome imprinting is pervasive, extending to most cells; for cats and humans, the imprinted X chromosome is confined to cells in the placenta and some other extraembryonic tissue.
The marsupial form of X-inactivation is considered the primitive condition for mammals. The random X inactivation of cats, humans, and other mammals in the more modern mammal lineage evolutionarily diverged from that of the marsupials. The signal event in this regard was the advent of Xist. Marsupials lack Xist and therefore the benefits of random inactivation. Indeed, Xist RNA may constitute the most important single difference between marsupials and more “advanced” mammals like us.12
X Inactivation and Cone Cells
Cc, the calico cat clone, is testimony to the fact that, due to random X inactivation, we should expect female clones to be more variable than male clones for any X-linked trait.13 The greater female variability should extend to nontwins as well. That certainly seems to be the case for color vision. Within the normal range of color vision, that is, excluding the color-blind like Steve, human females are more variable than males in color discrimination tests.14
At the low end of the range for normal color vision, some female carriers of the red-green mutation are less sensitive to red-green differences than noncarriers and normal males.15 We can attribute this to the fact that they have fewer normal cone cells. On the other hand, some females can make finer red-green distinctions than normal males. Paradoxically, these ultrasensitive females may also harbor red-green mutations that cause color blindness in males. These are the ones I refer to as X-women. How do we explain these X-women?
Let’s begin with a closer look at normal cone cells. The three types of cone cells are distinguished by the wavelengths of light to which they are most sensitive, which in turn depends on the type of opsin they express.16Red cones are sensitive to long wavelengths, green to intermediate wavelengths, and blue to short wavelengths. Color perception occurs when the brain integrates the inputs from these three cone cell types. We will confine ourselves to the red-green (long to intermediate wavelength) part of the spectrum. Normally, red and green opsins have different peak sensitivities with respect to wavelength. The brain therefore receives two distinct inputs. Red-green color blindness occurs when a mutation causes the peak wavelength sensitivities in red and green cone cells to converge, due to a mutation in the red or green opsin. In essence, the red and green cone cells become more similar with respect to the wavelength of light that causes them to fire their signal to the brain. It therefore becomes more difficult to distinguish red from green. This was Steve’s condition.
Red-green color blindness is particularly common for two reasons. First, in normal individuals, the peak sensitivities of red and green cones are not that different. Blue cones, on the other hand, have a much different peak sensitivity than green cones, and of course, even more so with respect to red cones. Second, the genes encoding the green and red opsins are arranged in tandem on the X chromosome. Adjacent genes are more likely to exchange bits when they are replicated during the process of making sperm and eggs. In this case, such exchanges of gene bits often result in red and green opsins that are more similar to each other than they should be.
Peak sensitivities for red, green, and blue cones. Fig. 5 (p. 370) in Deeb, 2005.
But now consider what happens when a similar mutation occurs in a female. Not only is she largely protected from the effects of the mutation because of random X inactivation; she should also end up with four, rather than the typical three, types of cone cells. The blue cone cells, of course, remain unaltered. She also has normal red and green cone cells, albeit half as many as normal. But in addition, she has cone cells with the mutation-induced hybrid opsin. If the peak sensitivity of this hybrid opsin is about halfway between that of normal red and green cones, such a female should theoretically be able to make finer distinctions at the red-green part of the spectrum, and perhaps even between green and blue. Some experiments on color discrimination in such women suggest as much.17 Ergo X-women.
The X-woman phenomenon has a precedent in other primates. There are two main primate branches on the mammal tree: Old World and New World. Old World primates include African and Asian species such as baboons, macaques, and langurs, along with the great apes and humans (our origin was in Africa); New World primates include spider monkeys, howler monkeys, and capuchins. Color vision in Old World primates, like ourselves, is said to be trichromatic (from the Greek for “three color”).18 We see more than three colors, of course, but all the colors we see are some combination of our three types of cone cells. New World primates, in contrast, have only two types of cone cells, so they are dichromatic.19 As such, New World primates cannot discriminate colors as well as us Old World primates. But there is a common X-linked mutation in New World primates, which, though it impairs male color vision, provides the females with three functional cone cell types, rendering them trichromatic.20 The enhanced color vision of these females, again, depends on random X inactivation. A similar mutation in a kangaroo could only impair the color vision of a doe.
An Epigenetic Boon
Color blindness is only one of the ways in which human males suffer disproportionately, for lack of a second X chromosome. Dosage compensation through X inactivation is not sufficient to compensate entirely if it is random, as in humans. Random X inactivation is a boon to females and a huge advance over the kangaroo condition of imprinted X inactivation. Random X inactivation was made possible not by the evolution of a new gene, but rather by a new piece of noncoding DNA that acts as a template for a new kind of RNA, called Xist, which made possible a new form of epigenetic regulation of the X chromosome.
Xist-mediated X inactivation is but one form of RNA-based epigenetic regulation. Most forms of RNA-based gene regulation are more widely distributed throughout plants, animals, and fungi. That is a subject I will take up later. In the next chapter, I will focus on the epigenetic process responsible for the form of X inactivation found in marsupials, for it, too, is an evolutionary advance of sorts. Among vertebrates, it is largely confined to mammals. And this epigenetic process, called imprinting, is not confined to the X chromosome; it occurs throughout our genome, albeit sporadically. Moreover, genomic imprinting occurs even earlier in development than X inactivation does in X-women. In fact, it occurs before sperm meets egg.