Screening for DMAP1 deletions

Here is a description of the male recombination screen used to target DMAP1  which my colleague Dr. Kathleen Fitzpatrick (Simon Fraser University (SFU) in Burnaby BC Canada) and I wrote together. Dr. Fitzpatrick used this screen as a part of her undergraduate genetics lab teaching program, and so this work is discussed largely by her (I made the pretty pictures). An application of the Awesome Power of Undergrads in the service of a very tricky screen!

In our class (Bisc 302W) at SFU, we spent several semesters screening for a directed deletion of the gene DMAP1 in Drosophila melanogaster. We used a P transposable element located between DMAP1 and its near neighbour CG33785, on chromosome 2. The P element is called GS10389,  but I called it P* in the class, so I’ll refer to it that way here as well. Note here that P* is closer to the gene CG33785 than it is to DMAP1.

There are two main requirements for P elements to move in the genome: first: an enzyme called transposase and second, special DNA sequences on either end of the P element called inverted terminal repeats. The transposase enzyme recognizes those repeats, and can cut the P element out of the original position and insert it into a new position somewhere else in the genome. P elements tend to have insertion preferences: places where they are more likely to insert and places where they won’t insert. But there is no way to predict exactly where a P element will go. In this sort of work, you mobilize a P element – allow it to move – and then you have to select for the result you want. You can’t just make it do exactly what you want.

The majority of transpositions are precise – we call them clean (or perfect) excisions. Sometimes transposition is not neat and tidy; sometimes it’s a bit messy – the element is not cut out and moved discretely, but some DNA to one side or another of the element is cut out too. This is referred to as imprecise excision. For our purposes, we are very interested in those specific imprecise excision events that also include a crossover event with a homologous sequence, because in this way, we might be able to select the direction in which the deletion takes place  (and remember, P* is closer to a gene we do NOT want to disrupt). In 1996, Preston et al (PMID: 8978050) published the results of their analysis of crossovers (recombination events) induced by P elements. They found a crossover rate of about 1% of the chromosomes tested (though it varied with temperature) and – here’s the important thing – deletions were generated from the insertion’s original site to one side or the other of the P element in some of the crossover chromosomes. The direction of the crossover reflected the direction of the deletion. Thus, if you induce a P element to move, and the element starts out near a gene you are interested in, careful selection any recombinants that result may net you some deletions of your gene. The mechanism is kind of complicated – the P element sequences located on sister chromatids try to pair with each other, forming a complex structure, the resolution of which may generate the deletion we need. Take a look at Movie 1 (and many thanks to Dr. Bill Engel for letting us use it).


So, here is a strange thing about fruit flies that turns out to be very useful in this context. Everybody knows that during meiosis, homologous chromosomes pair and undergo crossing over during Prophase I. This is one of those very important generators of variability upon which our evolutionary history depends. But, in male fruit flies, for reasons I think are still not well understood, there is NO meiotic recombination. Very strange. In fact, reduced rates of recombination associated with heterogametic organisms – those with different sex chromosomes like X and Y – have been observed since the early part of last century – here’s a recent paper using Zebrafish that describes it well (PMID: 11861568). Male fruit flies represent an extreme example with no meiotic recombination at all, but that turns out to be very helpful to us, because if we do our directed deletion experiment with male flies, any crossovers that result are almost certainly due to the P element “attempting” to move and making a useful mistake (for us), as opposed to a general recombination event we can’t control.

But how do we set this up so that we can select for those aberrant male recombination events that move in the direction of DMAP1? We really want to make a deletion in DMAP1 and only DMAP1. How can we distinguish between deletions in the left or right direction?

The trick is to use other genetic markers – mutations in other genes, to the left and to the right of DMAP1 that can be recombined onto the chromosome at the same time, during the same recombination event in the male germline that generates the deletion when the P element is mobilized. So this is what we did in our class (Figure 1): we crossed flies with the P element beside DMAP1 (P*) to flies with eye colour mutations on chromosome 2, and a source of transposase enzyme on chromosome 3. We collected male offspring that combined P* with the transposase source. Then we crossed those males to females with the recessive eye colour mutations cinnabar (cn) and brown (bw) on both homologues of chromosome 2.  Here’s the important thing: these eye colour genes are located on either side of DMAP1 and P*, so ultimately, we can select for deletions in one direction or the other by selecting for one eye colour phenotype or the other (this is also depicted in Movie 1). The males were heterozygous for the recessive eye colour mutations, and therefore had wild type (red) eye colour, and the females were homozygous for the mutant forms (which combine to yield a white eye colour – a subject for a later post; just accept this curiosity for now). Therefore any crossover in the males could be detected in the offspring as either mutant for cinnabar but normal for brown (bright orange eyes) or mutant for brown and normal for cinnabar (brown eyes).

DMAP1 male RC 04

Most of the progeny will have either normal red eyes (one wild type allele of each gene) or white eyes (mutant alleles of both genes). Hopefully Figure 1 will help you work through the chromosome shuffling that’s going on. Figure 2 is what it looks like when the crosses are written out in fly language:

Cross scheme

Since we wanted deletions in the direction of DMAP1 (and not CG33785) we kept any brown-eyed males and crossed them individually to a stock of flies containing the SM6a balancer chromosome (having the dominant Curly (Cy) mutation and the recessive cn mutation, along with a bunch of rearrangements that prevent further recombination, thus allowing us to maintain potentially useful deletions). Brown eyed males have a recombinant chromosome that is likely to have a deletion towards, hopefully into, possibly all the way through, the DMAP1 gene. We had to cross each brown-eyed male individually because theoretically, each male arose from the union of a paternal gamete (sperm) in which the desired recombination event took place, and an egg from the white eyed, cn,bw mother. Every F2 male, in other words, represents a gamete from the F1 father, in which the P element was induced to recombine.

Preston et al. found that about 1/3 of the recombinant chromosomes they tested had deletions. The deletions ranged in size from just a few base pairs to more than 100,000 base pairs, but the majority fell within about 2000 base pairs. This is good because the DMAP1 gene is about 1800 base pairs long, and we hope to delete most or all of the DMAP1 gene without deleting additional genes. I set up the cross to make the F1 males, and gave those males to groups of students in several successive Genetics Lab semesters. Students crossed those males en masse to cn, bw females, and then sorted through the F2 progeny for brown eyed flies. Those males, when found, were crossed INDIVIDUALLY to females with the multiply rearranged balancer for the second chromosome that was also marked with cn, and in the next generation, the F3, the recombinant chromosome bearing the brown allele and P* was recovered and balanced as a genetically uniform population of flies, or a stock. If the induced recombination event caused a lethal mutation, then the stock should remain balanced, all flies showing curly wings, with no brown-eyed homozygotes present. If the event produced a mutation that was not lethal, then brown-eyed homozygotes should show up in the stock.

There’s a lot to keep track of in this screen. Lots of different markers and mutations, understanding what balancers are and what they are good for, why some matings are en masse (lots of males and females together in a bottle or vial) and why others have to involve single males. After several semesters, we generated seven stocks carrying independently induced recombinant events that hopefully have deleted into DMAP1. Table 1At first glance (Table 1), we see that some stocks are lethal, and some are not, which could mean that DMAP1 encodes an essential function, and we have possibly generated an allelic series, ranging from complete lethality due to a large deletion that removes most or all of the coding region (a null), to perhaps a small change that simply reduces the expression of the gene (a hypomorph). But lots of other, less appealing results are also possible – the generation of an off-site hit in some other gene (that happens to be essential), or a complex rearrangement that will be awful to try an analyze. We have lots of work to do…so we had better get started!



DMAP1 Project: Preamble

DISCUSSION MAPWARNING: This gene causes many digressions. You might need a discussion map to keep track. Here it is. Figure 1.

DMAP1 is a rather interesting “under the radar” gene. I always like those kinds of genes, or projects, that no one in their right mind would tackle. They represent small spaces in a very crowded research field that people without much cash can try and work on.

I say “under the radar” but that is not true for research using mammalian models. DMAP1 was originally discovered in 2000 (PMID: 10888872) and ever since then has been quite extensively studied in mice and mammalian cell culture. DMAP1 interacts with a variety of other proteins to form transcriptionally repressive complexes that remodel chromatin by modifying histones and DNA (TIP60-p400 AKA NuA4 PMID12963728).  A mouthful of acronyms. In fact, this protein crops up in both signal transduction papers and chromatin papers, both subjects that are THE WORST acronymic offenders. Perhaps it is worth spelling some of them out (hence, the digressions).

The TIP60-p400 complex is essentially the same thing as the NuA4 complex – a nucleosome remodeler. (TIP60 nomenclature from flies and humans, NuA4 from yeast). A kind of chromatin snow plough that uses ATP hydrolysis to shift nucleosomes around the DNA, in order to expose sequences that are important for gene regulation. (Always remember that when controlling gene expression in eukaryotes, you have to contend with the occluding effects of chromatin. Here is an excellent review about the eukaryotic chromatin paradigm: an oldie but a goodie PMID: 10412974) Chromatin remodelers are remarkable molecular machines, very complex with multiple subunits (“rabbit dropping” diagrams often used to explain them), and intriguing interactions with actin molecules IN THE NUCLEUS (PMID: 11880634; highly relevant for discussions relating to a putative nucleoskeleton….another story). Here is a nice, comprehensive review about chromatin remodeling in general: PMID: 21358755. Anyhow, this TIP60/NuA4 complex interacts with DMAP1 and another acronym, DNMT1, to function primarily in transcriptional repression by histone modification (PMID: 14978102), heterochromatin formation (PMID: 14665632), imprinting and stem cell pluripotency (PMID: 21383065) and repair of double stranded DNA breaks via the homologous recombination pathway (PMID: 19845771, PMID: 20864525).

So, DMAP1 is a busy molecule. DMAP1 stands for “DNMT1 associated protein 1” (an acronym that stands for another acronym).  DNMT1 is a DNA methyltransferase – if there are two halves to epigenetic regulation (i.e., regulation of gene expression by chromatin structure) – one would be histone modifications (acetylation, phosphorylation, methylation etc) and the other would be DNA modification – exclusively at this time, methylation. Methyl groups are added to the 5’ carbon of the cytosine ring (we are talking vertebrates here). The enzymes that transfer these methyl marks are called methyltransferases, and they are either de novo methyltransferases (creating a new pattern of methyl marks on the DNA) or maintenance methyltransferases (that maintain an existing pattern). DMAP1 was identified as a molecule that interacts with DNA methyltransferase 1 (DNMT1) – a maintenance methyltransferase. So all that epigenetic stuff the NuA4/Tip60 complex does – histone modifications and the like – is presumably engaged in cross talk with DNA modifications via DMAP1 and DNMT1.

Oh dear this is a huge subject which threatens to go off the rails so let me just refer you to an interesting review about DNA methylation patterns across species, what it’s good for, how it might have evolved etc. (PMID: 18463664).

At last, to flies. If you peruse the fly lit, (via Flybase for example) there isn’t much. In fact, until very recently, DMAP1 really only showed up as a putative potential possible something-or-other in one of those interminable (literally) excel spreadsheets in the submerged supplemental iceberg that characterize most publications these days….maybe DMAP1 is involved in cell cycle regulation, maybe Notch signaling, maybe this, maybe that…trouble with these high through-put insert-subject-here papers is that everything gets connected to everything eventually…it is often hard to find the line that separates useful (though expensive) candidate-finding preliminaries from meaninglessness…but I digress again…

Then, recently – about six months ago – DMAP1 shows up in a much more convincing context, as a regulator via chromatin structure of signaling pathways in innate immunity in Drosophila. Yey! Which of course necessitates a brief description of innate immunity. What is innate immunity? Let’s start with immunity. Most of us recognize this to mean our humoral and cellular immune system which detects antigenic substances like viral coat proteins, bacterial cell wall components, pet dander, food allergens (for those of you with peanut sensitivity check out this recent discovery: PMID: 25592987)…this remarkable system is adaptive…when presented with an antigen (literally, an antibody-generating substance) the immune system can recognize it, send cells out to engulf and destroy it, and stick bits of the antigen onto the surface of specific cells that provide a long term reminder of the antigen so the response is quicker and bigger the next time the body is exposed to that antigen. Sometimes the response needs priming (hence, booster shots). Sometimes, people produce an exaggerated response which can be lethal (allergies to bee stings, or certain foods for example). Sometimes people can be attacked by their own immune systems (as in autoimmune disorders like Lupus).

The vertebrate immune system is very complicated. There are a large number of specialized cell types involved, which are part of the blood system, itself built on antibody/antigenic relationships. The genetics of the immune system are wild and crazy…something I really enjoy teaching in my advanced classes, research that garnered a Nobel Prize (Susumu Tonegawa 1987, check out A deeper and more mysterious question relates to the evolution of the immune system…connected perhaps to the unfolding stages of development in metazoans (multicellular organisms) where within one individual, tissues become increasingly differentiated from each other.

What then is innate immunity? In short, whereas the adaptive system I’ve just described is highly specific, innate immunity is non-specific – a first internal line of defense (underneath the skin and mucosal epithelia) that produces a range of antimicrobial peptides and chemicals that enhance an inflammatory response. Mediators of this inflammatory response include the cytokines you might have heard of, which are a little bit like hormones – small molecules that diffuse quickly through the blood stream and into cells. Adaptive immunity and innate immunity work together to mediate a response to injury or infection – innate immunity works fast (matter of hours – think how quickly bacteria reproduce!) while the adaptive system takes longer (days). The word “innate” is also meant to describe the fact that this early response system is shared across major phyla – both vertebrates and invertebrates. But only vertebrates have adaptive immunity. Thank goodness we’re vertebrates.

So fruit flies are protected by an innate immune response. There are two parts to this response. First, a cellular response, where cells circulating in the hemolymph (fly blood) recognize and engulf invading pathogens. Secondly, a humoral response from the fat body (serves as a lymphatic organ among other things) where a signaling cascade releases certain transcription factors from inhibition in the cytoplasm so they can migrate into the nucleus and promote the expression of genes that result in an increased synthesis of immunity-generating peptides that can destroy the invading pathogens.

This business of a ligand-binding-to-a-receptor-and-releasing-a-transcriptional-regulator-from-inhibition is a common backbone for pathways that regulate cytokine synthesis in vertebrates AND innate immunity in flies AND…drumroll please….egg polarity in early fly development. Pardon?  In one of those wonderful twists of evolutionary fate, the same pathway that specifies patterning in the Drosophila egg is used again to regulate immunity. This kind of retooling is not uncommon. It reminds me of two important quotes: one comes from Theodosius Dobzhansky, who famously pointed out that nothing in evolution makes sense except in the light of evolution (American Biology Teacher 35 (3): 125–129). Another quip comes from a colleague who shall remain nameless unless I am instructed otherwise (you know who you are RSH) but I modified it – evolution is not a German engineer making a Mercedes, but in fact a teenager with a broken car and no money. Just gotta get to the next generation, and if you have to jerry-rig it, so be it, deal with it later. Signaling fig smallI have tried to depict the relationships between these signaling pathways in Figure 2, but this is of course grotesquely simplified. Also these pathways probably talk to each other as well – which I didn’t show in my figure: here is a decent review: PMID: 21209287.

So NOW can we please get back to DMAP1? Yes! Because about six months ago, a paper was published that I think convincingly points to a biological function for this gene in flies. Here is the paper: PMID: 24947515. In sum, these researchers showed that DMAP1 interacts with Relish which if you take a look at my Figure 2 is in one of these related pathways that regulates innate immunity. The same researchers had earlier published a high through-put proteomic paper in which they identified hundreds of potential interactors with components of these signaling pathways and this DMAP1 paper represents a detailed confirmation of this specific interaction (by other methods – QPCR and coimmunoprecipitation analyses) and takes a stab at a mechanism by using a genetic approach – knocking down DMAP1 expression first in cell culture and then in living flies. They also discover connections with other chromatin remodelers – notably the Brahma complex (NO NOT ANOTHER DIGRESSION) and suggest that DMAP1 may be involved in modulating the specificity of the response from these pathways in development. This is a signal transduction thing – multiple inputs get integrated and turned into specific responses. DMAP1 may play a role in the specification of response. The mechanism is still pretty vague (isn’t it always…don’t you love it when a reviewer sends back a 23 page manuscript with seventeen expensive experiments and asks “please establish a mechanism beyond all reasonable doubt. Have a nice day).

The notable thing for ME about this paper concerns the lengths they had to go through in order to knock down DMAP1 in flies. What they needed was a nice clean mutation in the gene. There isn’t one. Well, there are P element hits but they are all in the wrong place. These researchers used transgenic RNAi and the UAS GAL4 system combined with an inducible inhibitor to control when and where DMAP1 was knocked down (I’ll explain the UAS GAL4 system later – another nose bleeder). I always liked this gene because…well it’s under the radar as I said, but also because I am curious about what a DNA methylation associated protein is doing in flies, since flies do not have the same degree of methylation as vertebrates (see digressions above).

I want to make a mutation in this gene. But LOOK at the configuration of the one potentially useful P hit that’s available (highlighted in yellow): it is in the 5’ UTR of the upstream gene (Figure 3). If I try to make an excision, I’ll probably knock out the upstream gene CG33785, not DMAP1. No wonder there is no mutant. Ah! The perfect project for a masochistic geneticist!

Genomic region for DMAP1

I guess I could do a screen with a mutagen but I have no idea of phenotype. Well that’s not strictly true – like the authors of this paper, I used an RNAi stock from Vienna (PMID:17625558) and a ubiquitous driver and it was lethal. I also had three different RNAi lines for the upstream gene and with the same ubiquitous driver, the flies were perfectly viable. So lethality is likely a phenotype. So something like an EMS screen over a deficiency is a possibility, but there is another way too, which makes use of the P element. And for this I have enlisted the collaborative support of a colleague at SFU, also a masochist, who also teaches genetics, but she teaches a LAB as well as a lecture, so can enlist the awesome power of undergraduates to do the boring stuff (undergrads: it is good for you).  In fact, she has been doing this for some time now– using the P element in the upstream gene in a male recombination scheme to select for imprecise excisions in the direction of DMAP1. She has incorporated this experiment into several semesters of her genetics lab course. We now have several lines that we are going to characterize. I will let my colleague talk about the screens, and the results, but I can tell you that there are viability AND fertility issues, and just this weekend I yanked out the ovaries from heterozygous females and their escaper (homozygous mutant) sibs and as you can see (Figure 4), the ovaries are pretty messed up. Apparently this line doesn’t even lay eggs – I am not surprised.

ovaries mutant and wt 02

The liberating thing about putting all this into a blog that well over three people might read is that I can make wild leaps in possibly the wrong direction! Here is a wild leap. The microscopy supports a role for DMAP1 in these retooled pathways for immunity and the specification of egg polarity! Sounds GREAT doesn’t it!

Leaps in the wrong direction should be enjoyed briefly, but then, bubbles must be burst (to mix metaphors). We have yet to do the following VERY basic things:

  1. Confirm lesions in DMAP1 (PCR)
  2. Confirm no effect on upstream gene (QPCR? Rescue?)
  3. Determine if the original P mutant has a background lethal or sterile that is causing this effect

I should point out that I also pulled out testes from escaper males and they don’t look right either. So once we have a better idea about the lesion, we also have to:

  1. Determine penetrance and expressivity of sterility
  2. Lethal phase analysis
  3. At this point the sky becomes the theoretical limit but cash flow may bring that down a tad. Should also point out that the chances of doing any of this in order are pretty remote.