You know, we’re all mutants. No, not just those of us reading (or writing!) this web site, I mean all of us. We all have regions of our genome that are highly variable, of course – the sequences (often based on number of repeat markers) that are used in forensic DNA analysis or the mitochondrial regions used in
paternity tests, for example – but we also have variations in all sorts of common DNA sequences as well. Single-nucleotide polymorphisms are everywhere, and if you sequence at enough depth you’ll find plenty of them. Most of these are silent, down in the difference-that-apparently-makes-no-difference category. But some of them aren’t.
Here’s an example of the latter: there’s a human developmental disorder called Floating-Harbor Syndrome (FHS) that leads to a number of phenotypic changes – facial abnormalities such as a small chin, large nose, deep-set eyes, along with general growth effects, language problems, high-pitched voice and others. It’s been traced back to a protein called SRCAP, which is involved in remodeling chromatin during gene expression. Various truncating mutations in that gene have been associated with FHS, and this new paper goes even further into the details.
An important function of SRCAP (and its partners in its chromatin remodeling complex) is to change histone interactions: the common H2A-H2B dimer is switched over to the H2A.Z-H2B dimer. That’s a rendering at right (via Wikipedia) of DNA wound around H2A.Z, and gives you a picture of how this works. There’s an awful lot of that sort of thing going on during gene expression, and it’s safe to say that we don’t understand enough of the details, what with the genome being the size it is, and being wound and packaged at the level of complexity it is. There’s plenty of protein machinery dedicated to winding and unwinding particular regions to make them accessible (or inaccessible) for transcription, and this cannot be anything other than ferociously complex. Consider the number of signals that can lead to differential gene expression under reasonable normal conditions in the cells of an adult organism (such as that oxygen-sensing system whose discovery just won the Nobel, for one), and then think of what must be going on during a complex organism’s development and embryogenesis. It’s pretty terrifying: at one end you have genes whose expression is flipping on and off constantly every hour of the day, and at the other you have particular combinations of genes that are critically important to be expressed just so at Day X and Day Y of embryo development and then must never be switched on in that way again after that time is done.
The H2A.Z variant of the “classic “H2A histone protein is already known to be important in a number of gene-expression programs. And just to add to the enjoyment, there are two varieties of H2A.Z, namely H2A.Z.1 and H2A.Z.2. These seem to have come about from a gene-duplication event way back when in evolutionary time and (as is customary) there has been some drift between the two since then and our cells seem to have found a way to make use of the two variants. They differ only in three amino acids, and while most of the older literature considers them to be the same thing (all just H2A.Z), it’s becoming clearer that they can have different roles.
And this paper certainly nails that down! In both Xenopus (frog) and human cells, the authors (from Stanford) show that the H2A.Z.1 and 2 subtypes are quite different when it comes to FHS. The effects of SRCAP truncations, for example, appear to come from a deficiency of SRCAP (rather than it exerting a negative effect), and these can be at least partially rescued by supplying more H2A.Z.2. Even further, it appears that this rescue is mediated by just one of the three different amino acid residues: mutating a serine at position 38 to a threonine is enough by itself to show the effect. And that is a pretty conservative change! Threonine has an extra methyl group in the side chain, on the carbon next to the hydroxy, and that’s it. But that’s enough.
Which takes me back to what I started with in the first paragraph: the nonlinearity of biology. There are so many places that you could make a single-residue change like this in the human proteome and nothing noticeable would happen – but not here. The analogy I used several years ago (while talking about mutations that lead to early-onset Alzheimer’s) was that if you could reduce molecular biology to a huge equation, it would have thousands and thousands of terms in it, with many of the coefficients in front of them being (most of the time) close to zero and thus not important. But some of those, if they ever do get an actual coefficient that lets them pop up out of the noise, have huge exponents hiding in them that can make them suddenly, wildly important. It need hardly be added that this situation is an absolute nightmare for anyone who might actually be trying to produce such a mathematical/computational model (!) But that’s the reality. . .