In an article published earlier this month, researchers color-coded cells based on which of two X-chromosomes they expressed resulting in beautiful images of marbleized cells like the one above. They are undeniably beautiful; but they also rely on complicated biological pathways and illuminate processes at play in every mammalian female. How does this work, what does it show, and how would it be useful?
Females are Complicated
Mammalian sex chromosomes are treated very differently than autosomes. There are a few reasons for that, but the biggest one is that different individuals have different numbers of them, without any bad effects. You need two copies of every autosome for healthy development; deletions of more than a handful of genes are often fatal. And having too many copies of most autosomes also results in miscarriage (with notable exceptions such as trisomy 21, or Down Syndrome).
But that’s just not the case with sex chromosomes. You might disregard the Y chromosome due to its gene-poor nature, although to do so would overlook the fact that less significant deletions on autosomes result in miscarriage. But the X chromosome is every bit as large and gene-rich as an autosome, and can exist either unpaired (monosomy, generally in males) or paired (disomy, generally in females) without ill effects.
What’s more, adding in extra copies of the X chromosome – two in most males, three in females or males, and so on – produces far fewer health and developmental problems than most trisomies.
How does that work? How do you get a chromosome that doesn’t care how many copies it has?
The short answer is that mammalian cells are set up to shut down all but one of the X chromosomes they possess. So any extra X chromosomes that an organism has get “turned off”, bundled at the edge of the nucleus. They can’t run amok.
That includes females – in each cell in most females, one of the two X chromosomes is selected randomly and shut down.
This process, called “X-inactivation”, is a part of a broader pattern of dosage compensation in mammalian cells that essentially makes sure expression from a gene on a sex chromosome averages out to the same amount as expression from a gene on an autosome, no matter how many copies of that sex chromosome are present.
It also means that, for female mammals, there’s a good deal of complexity in determining how any gene on the X chromosome gets expressed. It will most likely be “on” in some cells and “off” in others, determined in a mostly-random pattern and varying significantly between individuals. Which means that a lot of X-linked disorders show a huge amount of variation in severity between women.
Seeing the Mosaic
What the researchers in question did was, essentially, make the mosaicism present in all mammalian females – the fact that different cells express different X chromosomes – visually apparent. One X chromosome had a green-glowing protein sequence on it; the other a red-glowing protein sequence on it. In the presence of another gene (so that the researchers could choose which organ to light up like a Christmas tree), those proteins would be expressed. But since only one X chromosome is “on” in any given cell, only one color would be expressed. The cell would be either red or green – not yellow.
What you get is this beautiful marbleized pattern, with swirls of red and green cells. The red cells express one X chromosome, the green cells the other.
There are a few things you can see from this right away. For any given organ, is it usually uniformly one color or is it generally a mottled combination? In the first case, it’s likely that it arose from a single founder cell and whatever silencing pattern that cell had is inherited. In the second case, it either arose from a large number of founder cells or the silencing pattern isn’t stably inherited in that line.
Are red and green cells distributed evenly, or do they form patches? If the former, cells in that organ probably move around a lot after they divide – causing the mixing you see. If the latter, cells don’t move very much, so a mother cell and a daughter cell (with the same copy of their X chromosome silenced) tend to be right next to each other.
Does an organ always tend to have more of one color, in a bunch of different animals? That could be evidence that there’s some parental bias going on: that a specific copy of the X chromosome (usually the paternally-inherited copy) is silenced rather than one chosen at random.
And so on. The point is, there are all of a sudden a lot of questions that can be asked, elegantly and simply, with this tool.
Epigenetic Inheritance: Stable Like DNA
I want to get more into how this could be used to determine something potentially interesting. In particular: is the silent X-chromosome chosen anew every cell division, or is it inherited from mother cell to daughter cells?
As an aside, we knew going into this study that the silent X chromosome tends to be inherited across cell divisions. We knew this because people tried, for years, to reactivate the inactive X – by treating cells with drugs, by growing millions of cells in culture, and even by adding poison into the media, the antidote to which was put on the inactive X chromosome – and were almost universally unsuccessful. But there’s a particularly elegant proof of the stability and clonal inheritance of X chromosome silencing that uses exactly a tool like this color-coded system.
As I said before, the cells in the images I’ve shown are color-coded based on which X chromosome they silenced: red cells silenced the maternal X, and green cells silenced the paternal X. To tell if the epigenetic silencing that causes this color-coding is inherited from mother cells to daughter cells, we’re going to add a color: blue. Blue cells will represent something that we know for sure is inherited from mother cells to daughter cells.
There are two ways to do this, although really one works a lot better in mammals. The first method is easier to understand: infect the cells with a virus which inserts a blue pigment-producing gene into their DNA sequence. If you do it at a low rate, and use a virus that can only infect once, you’ll tag a random population of cells. From there, you can see if the borders between “blue” and “not blue” cells line up with the borders between red and green cells. If they do – clonally inherited. If they don’t – not clonally inherited.
But the thing is, it’s relatively difficult to do this experiment, because early-stage embryos are hard to work with, and harder to reimplant once you’re done. What would be easier is if you could package all of that into an adult mouse, which has the advantage of not requiring a microscope to see, and just breed the mice.
Fortunately you can, if you think about it kind of backwards. If you put a sequence into the mouse that has two properties – glow blue, and at a relatively infrequent interval, cut itself out.
This is based on a “jumping gene” first identified in the petunia, which controls flower color: every so often, it cuts itself out of the genome, and you get white stripes on a purple flower (or, if it happens early enough, purple stripes on a white flower).
Once again, you have clusters of blue cells that you know are clonally inherited. You have clusters of red or green cells you aren’t sure of. If the boundaries line up, the red/green decision is inherited across cell divisions. If not, not.
That’s one of my favorite, elegant, experiments in epigenetics and molecular biology. There’s something just so beautiful about it – a way of using color to see, physically, whether complicated epigenetic traits like X chromosome silencing are inherited as consistently as DNA itself. And while it wasn’t the first proof of the stability of X chromosome silencing itself, we used techniques like that to show that a lot of other epigenetic silencing mechanisms were inherited across cell divisions, largely looking at the pigmentation of the eyes of Drosophila.