Scholar’s Week Spring 2016

Last Spring the labrats put together a couple of posters and presented them at our local Scholar’s Week showcase. Both posters related to the DMAP1 project and give me a chance to see where we are at. I’m going to put up the poster slides here, mishing them together as they overlapped on some background stuff. True to form, I let the students play with powerpoint, illustrator and photoshop, to try and assemble a poster, but told them in advance I would be control freaky about requiring a very specific style, which I prefer (since the posters go up outside my lab and they MUST be clear).  I am always amazed by how my social media and computer savvy students don’t actually know how to put a poster together in powerpoint. It is always a brand new experience (and given the amount of time I spend using powerpoint, photoshop and the like, a useful one.)

When students come up with good design ideas (and some of them are really amazing) I do my best to make it so, together with them. So while the appearance of the poster and all the illustrations are mine (the fly pics typically have 50-100 component parts!) many of the layout and design ideas come from the students.

So, five students, two posters. Group 1: Genetic analysis of recovered lesions. Group 2: single embryo PCR analysis. What they had in common of course was the gene (DMAP1) and the model (Drosophila) and the male recombination scheme performed by students in Dr. Kathleen Fitzpatrick’s class at Simon Fraser University.  But I asked one group of students to focus on the genetic complementation tests that were used to define the limits of the deficiencies we obtained, and the other group to concentrate on the single embryo PCR (AKA getting DNA out of pieces of lint). Here are the abstracts for each poster. Poster 1: the genetics poster, has a blue dominant theme, while poster 2, the single embryo PCR poster, is green dominant (because, well, GFP, as you will see).

abstractsThe links to stem cell identity  and the immune system  are discussed here. Common to both posters was a description of the genomic landscape in which DMAP1 resides, and an outline of the male recombination scheme used to generate the deletions removing DMAP1 (and downstream flanking genes), more information about which can be found here.

So, a summary of the male recombination scheme carried out at SFU under Dr. Fitzpatrick’s supervision: this was included as part of the Single Embryo PCR poster (hence the green dominant theme).


And a description of the genetic landscape where DMAP1 resides on the right arm of chromosome 2 (this slide appeared in both posters so there the blue version):


For the genetics poster, poster 1, we began by presenting evidence that DMAP1 might be essential, using transgenic RNAi.  A more detailed discussion about whether being essential matters or not can be found here.

transgenic-rnai-slidesHaving established that DMAP1 is likely essential using RNAi, we wanted to figure out how many genes were deleted by our male recombination-induced lesions. So we ordered a few overlapping deficiencies from the Drosophila stock center in Bloomington Indiana and performed complementation tests. Very cool, since some of the kids were in my genetics class at the time and so this was a practical application (“see? It does happen in the real world!”) The red lines in slide 5 below indicate two of those overlapping deficiencies (note that the red line indicates the DNA MISSING in the deficiency ) – one (BSC 402) takes out DMAP1 and MANY additional upstream and downstream genes, while the other (BSC 404) appears to break just upstream of a gene called “exu” (exuperantia). Slide 6 shows the results of the complementation test: all pairwise tests between K8-15 and 10-14 (the two male recombination deletions we made) and BSC 402 and 404 (the deficiencies of known limits). The bottom line is that deficiencies that remove DMAP1 and flanking genes fail to complement our male recombination mutants (10-14 and K8-15) whereas a deficiency that begins downstream of DMAP1 somewhere in the vicinity of exu complement our male recombination mutants. These data tell us that both lesions we made using male recombination are pretty large, and theoretically should end in the vicinity of exuperantia.

complementation-slidesBest laid plans, and all that. Sometimes, the molecular breakpoints of these deficiency fly stocks that one orders from a center like Bloomington have been mapped, sometimes not. Theoretically, Df(2R)404 has been molecularly mapped, and does in fact delete exuperantia. So based on the complementation data, our male recombination mutants should include exuperantia, right? So we’ve mapped the breakpoint, right?

Wrong. Lethality is one phenotype that results from a failure to complement. Lethality means there is an overlap between the known deficiency and our male recombination mutants. By sheer chance (and curiosity) one of Kathleen’s students tested the progeny from a cross between Df(2R)BSC404 and 10-14 or K8-15 for fertility. SURPRISE: progeny are both male AND female STERILE. So in fact, 10-14 and K8-15 do NOT complement Df(2R)BSC404. Take a look at the region deleted by this deficiency (the red bar highlighted in yellow on the map below):


Exuperantia is the tiny little blue triangle near the top of this figure on the left hand side. Our male recombination mutants in fact fail to complement Df(2R)BSC404, because progeny heterozygous for this deficiency and either K8-15 or 10-14 are sterile. Sterility, from an evolutionary point of view, is just as bad as lethality, and therefore is just as much a failure to complement. So what is exu? What does it do? Is sterility a predictable phenotype?

Exuperantia has an illustrious pedigree. It was isolated as a maternal effect gene necessary for the normal anterior-posterior patterning in the Drosophila embryo. Exu was one of an elite suite of genes isolated and studied in the mid 1980’s by Trudi Schüpbach and Eric Wieschaus in a screen to study pattern formation in development. This work revealed a highly conserved genetic determination of body patterning and opened up the fascinating field of developmental genetics. Oh yes, and resulted in a Nobel Prize in 1995.  Bottom line: if the mothers are homozygous for loss of function mutations in exu, the embryos fail to develop, and they show anterior patterning defects. How does the exu gene product do this? Presumably by controlling some of the other mRNA’s that mum packs into the cytoplasm of her unfertilized egg, providing the egg with patterning coordinates (head end, tail end, dorsal side ventral side etc.)


A number of mutations in exu have been isolated since then. Most are female sterile, one is male sterile. A couple are lethal (but we have determined that at least one of those lethals is due to a second side hit on the same chromosome). So, yes, theoretically, male and female sterility is a reasonable phenotype for our male recombination lesions, and tells us that they both extend at least out to exuperantia.


Can we confirm these data using PCR? Certainly we know our lesions take out DMAP1 and do not disrupt the region upstream (see here), but how far downstream do these lesions extend? More primer design for the students,  targeting genes in the vicinity of exu.

slide-7-poster-1-spring-2016And our results? Using DNA from adults means using balanced flies, with the wild type genetic region on the balancer. (Remember what a balancer is? No? Explanation to follow). Thus, the data below show the results using a primer to the P element which we know is still present upstream of DMAP1 (at least half of it is) and gene-specific primers. P-out is the P element primer, in combination with the upstream gene CG33785 we have a product, but nadda in combination with anything down stream, including exuperantia.slide-8-poster-1-spring-2016Now, these data make a bold assumption, that so much DNA has been deleted in  10-14 or K8-15, that the region between the P element and the exu gene-specific primer is now small enough to be spanned by garden-variety PCR (usually a limit of about 2 to 3 kilobases of DNA). What if that’s wrong? Also, what if the priming site in the P element isn’t there anymore, for down stream amplification? Really the only way to find that out is to try PCR on single K8-15 or 10-14 homozygous individuals. And as those are dead, we may be stuck.

Or are we? We put both K8-15 and 10-14 over a balancer chromosome that also carried a transgene expressing the Green Fluorescent Protein (GFP). What’s a balancer? Recall that if you want to maintain a recessive lethal mutation in a population of individuals, meiosis will eventually purge the population of the mutation – lethality rather interferes with fitness, after all. Drosophila geneticists have developed a useful tool called a balancer chromosome, which has essentially all the right DNA in it, but not necessarily in the right order, having been scrambled somewhat by radiation (which reminds me of the famous Morecambe and Wise sketch with Andre Previn: “I am playing all the right notes, just not necessarily in the right order” – see around 11:00 in the sketch). The balancer is usually marked by a mutation that has a dominant phenotype (curly wings, stubbly bristles, messed up eyes etc) that is recessive lethal. That way you can instantly see if your stock is balanced. What this means is that during meiosis, when a chromosome with a lethal mutation on it attempts to pair with a balancer homolog, well, it can’t, and all recombinant products are lost. Thus, the population consists entirely of individuals that are heterozygous for the balancer and the chromosome with the desired mutation. A marvellous tool, so far unique to Drosophila.

So we balanced our male recombination mutations over a balancer with GFP on it, which allowed us to do two things. First, we could identify the lethal phase, or when homozygotes for K8-15 and 10-14 died. Second, we could isolate individuals, dying, to be sure, that were homozygous for the lesions, which would be GFP minus, and extract their DNA for PCR analysis.

slide-4-poster-2-spring-2016In fact the graphic above is not entirely accurate: the GFP -/- individuals never made it to adult hood, as implied. They died as first instar larvae – fresh out of the egg.


Panel C in slide 5 above shows a first instar larva that is homozygous for 10-14 or K8-15 – both lesions behaved the same way. Slide 6 shows how we used GFP fluorescence to select against all the genotypes we did not want, and select only those that did not fluoresce. Those larvae were placed individually into eppendorf tubes for DNA extraction and subsequent single embryo PCR. We tested a number of genes in the vicinity of exu – here are some data for CG13437: clearly it is absent from the male recombination lesion 101-14.


I can tell you we also tested using primers to exu, and to galla-1, which is even further away.  The Galla-1 results are puzzling because Galla-1, which may play a role in chromosome segregation through establishment of sister chromatid cohesion (PMID 25065591), is supposed to be lethal (same paper). But our “complementing” progeny from a cross between Df(2R)BSC404 (removes galla-1 completely) and our male recombination lesions (also remove Galla-1 – enough of it to be considered gone based on where the primers target) are alive. Sterile, but alive. What could this mean? Who knows.

Buggah, buggah, these are large deficiencies we made, when what we really wanted were small lesions confined to DMAP1.

Do we have ANY idea where the break point is? Both K8-15 and 10-14 are viable but sterile over Df(2R)BSC404.  Are there any essential genes, in terms of viability, included within the genomic territory defined by this deficiency? Aside from Galla-1? Yes, but perhaps we have to stick to the brut force single embryo PCR approach at this point. The next victims in my lab will design more primers getting further and further away from exu, Galla-1…until we get a PCR result that tells us a gene is actually present in our male recombination induced lesions.

We will have to map these breakpoints. We want to know how big these lesions are. The next step is to use them as known deficiencies over which to carry out a “local hop” – try to get the P element just upstream of DMAP1 to move out of its current location and into DMAP1, hopefully knocking the gene out. K8-15 and 10-14 may well be huge, but they are much smaller than the stock deficiencies available, like Df(2R)402 or 404. They may yet be useful. Fingers crossed.

Here are the concluding slides for the two posters my students presented last spring, and a couple of shots of students by their posters.




DMAP1 mutations: PCR analysis

So all our putative DMAP1 mutants are stably balanced (if lethal) and floating balancers (if viable – the balancer chromo is still in the stock but homozygotes abound). At this point, it becomes possible to purify genomic DNA from them and see how much we can find out about the molecular nature of any lesions in the region using PCR (I make the bold assumption all readers know how PCR works. If not, try this).

First of all, we need to generate a map of the region, and design primer pairs that are likely to be diagnostic of any changes in the DNA. One very important primer will anneal to the inverted repeat of the engineered transposable element that we used as a starting point for our male recombination scheme. There are two primers of use here: P-OUT which anneals to the 31 base pair inverted repeat sequence in such a way as to point OUT of the P element (3′ end of primer directed away from the internal P element sequence) and P-IN, simply the complement of P-OUT and of course, pointing INTO the P element.  By “pointing” into or out of, I mean, of course, that the Taq polymerase will extend from the 3′ end of the primer into or out of the P element.

So here is a map of the primers I’ve designed so far.

DMAP1 primer map

Recall that the original engineered P element is inserted into the 5′ UTR of the upstream (on the left) gene CG33785/6 (UTR = untranslated region – grey in the Flybase gene diagram above). The male recombination scheme is designed to select for events in which the P element partially excises from this location, taking with it material in the direction of DMAP1 – we want to leave the upstream region intact. If all goes according to plan, a successful event will retain the upstream (5′) portion of the P element, but should lose the downstream (3′) end, plus hopefully a nice chunk of the DMAP1 gene and nothing else. How MUCH DMAP1 DNA downstream of the P element is retained should be possible to determine by seeing if the downstream priming sites are retained.

Of course the best way to do this is to carry out a Southern analysis.  Cut up the DNA with a well-chosen restriction enzyme (well-chosen based on the sequence we know), fractionate the digest in a gel by electrophoresis, blot the pattern of fragments onto a membrane, incubate the membrane with a DMAP1 cDNA probe, labelled in some way to make it possible to visualize (usually radioactivity) and analyse the pattern of bands. Southerns are still important experiments – you see them commonly described in four point font in the materials and methods section of knock-out mouse studies. But they are not simple experiments, and they are not cheap, so PCR it is for us at SmallScienceWorks.

So without further ado, here are the results. Note that we did these PCRs at the same time as the complementation tests, so some of these data make more sense in hindsight, which I’ll discuss a bit later. For the homozygotes, I selected males only (mated females might have eggs from balanced sibs that will yield false positives – remember the region is wild type on the balancer)

In Figure 1, top row of lanes, we tested all the viable isolates and w[1118] (which has no P element and is wild type for the region) using gene specific primers in DMAP1. The bottom row, on the other hand, tests for the presence of the DOWNSTREAM (or 3′) end of the P element in all the isolates, by using P-OUT together with a series of DMAP1 gene specific primers.

Gel 1a

From these data, top row, which tests for gene structure in the mutants that are viable (as homozygotes) you can see that there appears to be no disruption in the region covered by all these primers (which is pretty much the whole gene). The band sizes for the mutants are indistinguishable from w[1118], which is wild type (the gel is frowning a bit – the bands are distorted on the edges – not a big deal). Were we disappointed? Yes. NO! Of course not! We are dispassionate scientists, who want what we get, not get what we want. Buggah.

From the data on the bottom row though, you can see that the P element is absent in all mutants EXCEPT K14, and of course in still present in GS10389, the original P stock used in the male recombination scheme (and therefore serving as a positive control). You may note some laddering in the bottom row, right hand side. That’s OK, that is non specific product, resulting from some weak false priming of the primers that shows up after 30 cycles.

The data shown in Figure 2 are even more disturbing, at least the top row is. The NGFPF primer is a forward primer that anneals to the very beginning of DMAP1, starting at the ATG. The subsequent lanes are testing for amplification across the P element insertion site. CG33785 is upstream of DMAP1 – remember the P element is inserted into the 5′ UTR of this gene.

The bottom row is another test for the downstream or 3′ portion of the P element.

Gel 2a

So the data in this gel tell us that the gene structure is probably completely unaltered in the homozygous “mutants” – I use inverted commas because they are scarcely looking like mutants at all. Any lesions, if they exist, must be so small there is no difference in band size when the “mutants” are compared with w[1118], which remember, is wild type for this region. Hmm.

The bottom row tells us what we already know. The downstream portion of the P element is absent in all mutants except K14. Perhaps, in the homozygotes, we had a perfect excision of the P element, plus a recombination event further down the chromosome, to put the brown eye color mutant onto the same chromosome. Who knows.

So what about the P element upstream of DMAP1? According to the scheme, the 5′ portion of the P element should still be there, though this is hard to believe with the homozygotes (top row Figure 2) – PCR across the insertion site shows no size difference with w[1118]. But of course, we did the PCR to prove it to ourselves. Here it is, Figure 3.

Gel 3a

So all peculiarities solved, right? Everything makes perfect sense, the end.

Alas no. According to the top row of data, the upstream portion of the P element is present, (as it should be in theory) EXCEPT for 20-13 homozygotes, and 22-13/SM6a heterozygotes. Great. There is a little bit of false priming in the negative lanes but the upshot is, for three out of four homozygous viable “mutants”, the upstream or 5′ portion of the P element is there, and yet, not there, when PCR across the site is performed. The P element is there and not there. How very Heisenberg.

Huh? I didn’t get all that. And I wrote it.

For the mutants that are homozygous viable, the upstream portion of the P element is present (except in 20-13). The smallest amount of P element that could be left is 31 bp – the annealing site of the P-OUT primer. Unlikely – if the priming site is there, a chunk of the P element must be there. So when you do PCR using primers on either side of that chunk, then the band should be BIGGER than it is in w[1118] which doesn’t have the P element. BUT all the bands representing PCRs across the site are identical in size to the w[1118] stock. BUT the upstream P element is still there. According to the PCR with P-OUT and the upstream gene CG33785/6. So for these homozygous viable mutants, the P element is there, but not there. See? It is a headache inducing contradiction. But there must be a solution, because this is real matter, this is Nature, this is real DNA. We just have to figure it out.

Here is a summary of the data in Table 1:

Table 1

Note K8-15* – this is a new mutant, generated by my colleague Kathleen Fitzpatrick this year up at Simon Fraser University in BC Canada. It looks good – fails to complement 10-14, which fails to complement a deficiency for the region, and it behaves just as 10-14 in the PCR assays as well.

Now we know from the complementation tests that of the lethals, 22-13 and K-14 complement deficiencies for DMAP1 obtained from the Bloomington stock center. So these are lethal hits elsewhere in the genome, and therefore should be chucked (which is hard because Drosophila geneticists tend to be hoarders of genotypes). Of the viables, 20-11 shows a weak failure to complement with 10-14 so I am inclined to keep it, in case it turns out to be a hypomorph (partial loss of function). Though, like the other viables, it shows this bizarre P element there-and-not-there molecular phenotype. BUT: there is an outside chance that these viables might in fact be duplications for the region, which, if you examine the original paper (PMID: 8978050) indicates that duplications can occur with approximately equal frequency to the deletions. We might be able to detect these by PCR – a forward primer pointing downstream in DMAP1 and a reverse primer upstream of the forward primer might amplify a product, if a local duplication as occured. Worth a try.

We have no idea about the gene structure for DMAP1 in the lethals. These stocks are balanced, and so primer pairs NOT involving P-OUT will amplify from the wild type sequence present on the balancer. I need homozygotes for whatever the lesion is in 10-14 (or K8-15) in order to test for gene structure using solely gene specific primers. I tried rebalancing the lethal 10-14 over a CyO-GFP balancer and looking for GFP-minus larvae but saw none, suggesting this genoytpe has a very early lethal phase (dying as an embryo perhaps, or a first larval instar, where the GFP doesn’t show up).

So plan B: note from the complementation tests that 10-14 complements the original P strain GS10389. If I can purify DNA from flies transheterozygous for the P element-bearing chromosome, and the 10-14 deletion chromosome (or K8-15 for that matter), I might have a chance at mapping the gene structure, using primers that span the insertion site. For the GS10389 chromosome, the region is too large for conventional PCR (the engineered P element insertion is about 7000+ base pairs of DNA – conventional PCR is good to about 3000 bp), and so if I find a primer pair that anneal to the remaining DNA, I should be able to amplify a product. If that happens, I will certainly cut out the amplified band from the gel, extract the DNA and clone the fragment, and stick it into my freezer until I can find some $$$$ to sequence it. At that point, we will know the precise molecular nature of the lesion. Hurrah!


DMAP1 complementation tests

And now time for Ye Olde Fashionede Geneticse. We have 7 lines of flies, each a potential mutation in DMAP1. Four produce homozygous flies with variable levels of viability and fertility and three are lethal (see here). Are they all lesions in the same gene? We expect so, perhaps even more than we might from a standard mutagenesis screen using a chemical mutagen like EMS, because we used a mutator P element so close to our target. But still, all pairwise combinations of crosses between all the lines are necessary, resulting in a grid that is symmetrical on either side of a line separating reciprocal crosses – that is, the same cross repeated with the contributing parents’ gender reversed.

Let’s consider an example of a simple complementation cross assuming five independent mutant lines that are all recessive lethal (so they are all balanced – remember the balancer carries a dominant visible marker that is recessive lethal). A PLUS sign indicates the cross produces transheterozygous progeny that are therefore not balanced: mutant 1/mutant 2. This means the mutations are in different genes, and so the progeny have wild type alleles for the two different genes, and so they complement each other. The genotype of the progeny is written as:

  m1+ m2 / m1 m2+

 or more simply (if obliquely):

 + m2 / m1 +

(If an allele is wild type we tend to make the symbol vanish and just leave the + sign, like the Cheshire cat’s smile).  

If the alleles fail to complement, indicated by a MINUS sign, then only balanced progeny are obtained, arguing the independently isolated mutant alleles are in fact in the same gene. In Figure 1, note that mutant 1 fails to complement mutant 2, but complements all the other mutants. So mutant 1 and mutant 2 are in two different genes; mutant 2 has only one allele, and mutant 1 has 4 alleles (mutant 1, 3, 4 and 5.)

Comp test ex

Another way to visualize this is by using colour. Yellow for failure to complement (alleles of the same gene) and blue for complementation (alleles of different genes) as shown in Figure 2. I chose these colors partly because I get to make what looks like a reverse Swedish flag, and partly to help those with red green colour blindness (but not people with blue yellow colour blindness which is far more rare – highly unlikely that the well over three people reading this blog are afflicted….)ex comp full shade b&yNotice the symmetry. Usually, only half the crosses are done, in one direction only, if no parent-of-origin effects are suspected. What are parent-of-origin effects? An example is imprinting, discussed in an earlier post. Something that causes the expression of otherwise identical alleles to change, depending on the sex of the parent from which they were inherited. If no parent-of-origin effects are suspected (often a rather rash conclusion), only half the cells in the table are provided with data, the undescribed cells on the flip side of the line of symmetry are assumed to be the same, as shown for our hypothetical example in Figure 3.

ex comp half shade b&y

So what about our DMAP1 mutants? The first and most important crosses to carry out use a different genetic background with a known lesion in the DMAP1 region. Why? Look here for an explanation of the importance of genetic backgrounds.

My colleague Kathleen Fitzpatrick ordered in some fly stocks with deficiencies in the region – molecularly or cytologically defined deletions of genetic material. Since these deficiencies remove many essential genes, they are usually recessive lethal, and so are balanced. Kathleen ordered two different deficiencies (two different genetic backgrounds) that remove DMAP1, and one that does not, but which is located very near DMAP1. Take a look at this map from Flybase: the regions DELETED  by the deficiencies are indicated by red bars.

Df map detail 02

The three relevant deficiencies (indicated with asterisks) are Df(404) (DMAP1 still present), Df(403) and Df(701), (both of which delete DMAP1, but were generated from different screens and so have different genetic backgrounds). I crossed all the putative DMAP1 lethals to each other; Kathleen introduced the deficiencies. Figure 5:

Df comp y&b

So here is the first surprise (and inevitable disappointment). Remember that YELLOW means failure to complement, and BLUE complements. Yellow all down the diagonal as we expect – all the stocks involved in these crosses have recessive lethal mutations or deletions. First look at the inter se crosses between the putative DMAP lesions. Notice that 10-14 complements the other two putative DMAP1 lesions – 22-13 and K-14 – which fail to complement each other, so there must be hits in at least two different complementation groups, or genes, here. Already we have information showing that at least a subset of our putative deletions are unlikely to be in DMAP1. Buggah.

The cross to the deficiencies tells the tale, however. All three putative DMAP1 lesions we have made complement the deficiency that does NOT remove DMAP1 (+DMAP1) but only 10-14 fails to complement the deficiencies that remove DMAP1 (-DMAP1). Notice I have blue hatching to indicate results with K-14. These crosses were not actually done, but since K-14 complements 10-14, and fails to complement 22-13, I suspect K-14 carries the same background lethal as 22-13 (which is odd, because they were isolated in different years, so are bound to be independent events). But it looks like only 10-14, of the three lethal putative DMAP1 male recombination mutants, is actually likely to be a lesion in DMAP1. Which is better than nothing.

What about parent-of-origin effects? DMAP1 stands for DNA methyltransferase 1-associated protein, and DNA methylation plays a crucial role in imprinting, which is an evolutionarily conserved parent-of-origin effect. So we developed a more complicated complementation grid – somewhat incomplete – that shows data from crosses in both directions. Also, since this test includes potential DMAP1 mutants that produce homozygous progeny, some with fertility issues, I have altered the colouring scheme to reveal potential hypomorphic effects, assuming a reduction in DMAP1 activity, as opposed to a complete loss on function – which we assume to be lethal (see here for argument). Remember that yellow means failure to complement, and blue means full complementation. Use Table 1 to follow how variations in these colours mean variations in viability. The cut-offs are arbitrary, simply based on my experience. For every cross, I scored (counted) at least 100 progeny flies. If you are unsure of where the ratios come from,  study Figure 6.

Ratio tableCrosses and ratiosAnd now here are the actual data (Figure 7):

DMAP1 full interse y&b

So it looks like 10-14 is the closest thing we have to a lesion in DMAP1. The other lethals are suspect, and should be chucked. Since the original P hit in DMAP1, used for the male recombination scheme, was not itself a lethal, the lethal lines that complement the deficiencies which take out DMAP1 must have incurred a second hit during the male recombination scheme. That means that the lethal lines  22-13 and K-14 share the same newly induced background lethal. As I mentioned, and you might suspect from the naming system for these mutants, they were isolated in different years (2013 and 2014). So the same lethal was generated twice! I am not a big fan of coincidence, so something interesting (but annoying) happened here. But since DMAP1 wasn’t apparently involved, we really should abandon it. Buggah.

What about the non lethals? 20-11 and K2-13 sort of fail to complement each other. Moreover, 20-11 and 10-14 – our putative DMAP1 lethal – also sort of fail to complement, but only in one direction, when the lethal comes in from the male parent. So maybe we do have a hypomorph here (20-11) that supports an argument for a DMAP1 function in a parent-of-origin effect (like imprinting). To make things messier (why not), note that K2-13 also partly fails to complement 10-14, but in the other direction that we see for 20-11. (K2-13 also partly fails to complement 12-14, which itself pretty much complements 10-14. This is the sort of shop talk that drives normal people nuts). What does that mean? Dunno. I scored 100 flies total for each cross  – the numbers (and therefore shading) might be slightly different if a larger number had been scored. Sample size, statistics etc.

So if I were in a chucking mood (which I rarely am, being something of a fly hoarder), which stocks should I chuck, and which should I keep?


Of course I should simply keep 10-14 and possibly 20-11. Alas, all the interesting sterility issues must now be revisited….that wonderful messed-up ovary picture I took was for 20-13, which steadfastly complements everything.

Well it was nice while it lasted.

Now what? We have pretty much reached the limits of classical genetics. We could cross our 10-14 lesion to increasing numbers of deficiencies to narrow down cytologically where our lesion is. We could do some cytology to look at chromosomes, to try to visualize the actual nature of the lesion. All these things we might have done twenty, thirty years ago. But at this point, we have to go molecular, which is great because we can drill down in much more detail, but also not so great because it costs money.

The best thing would be to do a Southern blot at this point. A Southern is a procedure where genomic DNA is fractionated by cutting it up in a defined manner (using a restriction enzyme that cuts at a specific DNA sequence) and separating out all the fragments by size in a gel. The gel is then blotted with a nylon membrane and the pattern of fragments is transferred to the nylon membrane by capillary action. The membrane is then washed with a labeled piece of DNA (a probe) that matches the region containing DMAP1 (for instance, the probe could be a DNA copy of the mRNA). The probe is labeled with something that can be visualized (colorimetric, radioactivity etc) and the pattern of bands analyzed. This method could tell us (1) if the region containing DMAP1 has been disrupted at the molecular level and (2) what the nature of that disruption might be. It could be a complete deletion of the coding sequence for DMAP1, or partial. Look here for a description from Ed Southern who developed the method in 1975. Initially he couldn’t get the work published, so scribbled it on an envelope for colleagues who really wanted to use it. No blogs then.

But Southerns are expensive. You need labeling materials, special nylon membranes, masses of chemicals etc.  Fortunately, there is a cheaper alternative (though it does not give as much information) and that is PCR – polymerase chain reaction.

PCR is the one molecular acronym my freshman biology students usually recognize (outside of DNA), because it shows up in CSI etc. Yey, for science in crime shows (a great source of boo boo material for my exams) but it takes a solid understanding of how DNA replication works to fully appreciate the power of this highly efficient technique. It is also a method that doesn’t have to cost a ton of money. So that’s where we will head. Time to get down to the DNA, and characterize the nature of the 10-14 lesion in DMAP1!

Is DMAP1 an essential gene in Drosophila?

….and does it matter?

Yes, and no. On the one hand, if DMAP1 is essential, then that means important, right? Essential means you can’t live without it. On the other hand, if only our lethal potential DMAP1 mutants are really in DMAP1, implying the gene is essential, well then that’s a problem, because there is not much you can do with a dead fly. On the third hand, if the gene is lethal when completely knocked out, but has an interesting phenotype (like sterility) when partially knocked out,  that might be the best of all possible worlds! When geneticists want to mutate a gene, they usually want A NUMBER of mutants, not just one. They want an ALLELIC SERIES – from weak hypomorphs (partial loss of function) to complete nulls (complete loss of function). And maybe a few weirdos like gains of function, dominant negatives or antimorphs.

But of course you have to want what you get, not get what you want. And we don’t know yet what we have with DMAP1. But there are some clues. We can use transgenic RNAi to tell us if DMAP1 might mutate to lethality. Huh?

Let’s start by taking a stab at the word transgenic. Most model organisms used for genetic analysis can be made transgenic these days. A text book definition usually describes a transgenic organism as one into which “foreign” genes have been inserted. And that is basically it – transgenic organisms are also widely used in agriculture. The specter of GMO frankenfoods includes transgenic soybean, corn, canola, cotton, rice etc. This is a huge, technical and emotional subject of course. I like this blog which is very eclectic and I have only really read the GMO part but it is very balanced.

Making a transgenic fly takes several months, and the genes that are inserted are not necessarily “foreign”, i.e., non-Drosophila genes. You can add additional copies of native Drosophila genes (endogenous genes). Or normal functioning copies of a mutated gene – gene therapy, if you like, for the fly (also called a rescue experiment). You can add a gene that includes all its own regulatory signals – a so-called genomic construct, so that if it lands in a conducive chromatin environment, it can be expressed just as it would in its native (endogenous) location. You can also target the gene you insert (let’s call it a transgene) – i.e., get it to integrate in a site in the genome of your choosing – using the same tools of homologous recombination the cell uses during prophase I of  meiosis. There are lots of bells and whistles in this technique.

Here is a recipe for making a transgenic fly.

  1. Obtain a cDNA (DNA copy of a mRNA) of a gene of interest (let’s call it gene X).
  2. Cut and paste the gene X  into a special DNA plasmid that carries signals that allow it to be integrated into a chromosome. This special plasmid is called a vector, because it will “carry” your gene of interest into the genome. The signals include a gene that expresses the red eye color as a reporter indicating successful integration (a wild type version of the white, or w+ gene), a promoter region upstream of gene X that can be induced in some way, and the inverted repeats of the P transposable element which in the presence of a transposase source, will make integration into DNA possible. The vector also contains bacterial sequences like antibiotic resistance markers and an origin of replication, because this part of the experiment, this cutting and pasting, is the cloning part, the recombinant DNA work, done largely in bacteria. The whole thing – vector plus insert – is often called a construct. It is a critical part of the experiment – you better sequence your construct multiple times to make sure there are no boo boos. Otherwise masses of time gets wasted down the line.
  3. Obtain a stock of white eyed flies (for example, w[1118] – mutant for the white gene on the X chromosome).
  4. Rear large numbers of young w[1118] females (mated with w[1118] males but get rid of them because they only take up space)  and get them laying masses of eggs.
  5. Gently organize the eggs on double sided tape with their butt ends hanging over a bit.
  6. Mix your cloned construct with a “helper” plasmid (which produces transposase but cannot itself integrate) and load a very fine needle with the mix and inject the embryonic posteriors VERY CAREFULLY and VERY QUICKLY.

During the early development of Drosophila, the embryos are essentially a bag of nuclei (syncytium), no cell membranes. This means that the nuclei all share the same cytoplasm. There is some organization; the nuclei nearest to the posterior end of the embryo will become the germ line, so injecting the mix at this site increases the likelihood that the construct will integrate into DNA that will become the germline, and thus be stably inherited. The injection has to be relatively fast, however, because the embryos are alive of course, and will begin to cellularize, (the nuclei will become surrounded by cell membranes) after which integration of your construct is all but impossible.

  1. Transfer the injected embryos to vials with comfort food and wait. Those that survive this terrible insult will grow and hatch into adults – but if you are expecting to see red eyed flies indicating a successful germline transformation at this point, you will be disappointed. Some may have mottled red and white eyes, like Muller’s eversporting mutants, and that is a good thing, because it means that DNA has integrated into some nuclei, that have divided and multiplied and become incorporated into specific tissues, like the components of the compound eye. If you see this, you are observing a somatic mosaic, and you can then assume that if the soma (body) cells are mosaic for your construct, the germline probably is too.
  2. Mate your survivors to the original w[1118] stock and wait with great expectations for the next generation. Your survivors will produce gametes (sperm or eggs – usually easier to use males, so sperm) that may or may not have your construct – if a sperm from your survivor that HAS the construct fertilizes a w- egg, then hurrah! The progeny will be heterozygous for your construct in every cell!
  3. Use Ye Olde Fashionde Geneticse to map your construct to a particular chromosome and make the stock homozygous. You might have hit an essential gene, in which case the transgenic fly will carry a lethal – I usually chuck those.


transgenic 02That is the simple version. You can find the detailed protocol here – together with many beautiful images of different fly species…

So that is how you make a transgenic fly. It is the first step in a very commonly used scheme in the Drosophila world called the modularized miss-expression scheme – also known as the UAS-GAL4 method. I told you that the transgene is often induced to express in some way. It carries sequences that allow the experimenter to control when or where a particular transgene is expressed. Commonly, sequences from yeast are used – these ARE foreign DNA elements, tacked upstream of the gene of interest in the construct, called Upstream Activating Sequences – UAS – recognized specifically by the yeast GAL4 transcription factor. Yeast galactose metabolism is not happening as a rule natively in flies, thus if the experimenter can control when and where the GAL4 transcription factor is expressed, she can control when and where her transgene construct is expressed. See? The method is summarized in Figure 2A below.


The construct that responds to GAL4 because it has the UAS sequences is called, not surprisingly, the responder. When you want to express your gene X in a specific place or developmental stage, you would create the responder using the recipe described above. The construct that produces the GAL4 in some defined manner is called the driver. Yes, the driver is yet another transgenic fly that someone else has made. The main difference is that the construct was made by pasting a promoter region from a specific gene upstream of the yeast GAL4 transcription factor coding sequence. The promoter can come from anywhere. If you want to make an eye-specific driver – so you can drive and confine gene expression specifically in the eye (very useful, if you want to assess a mutant effect of an otherwise essential gene – flies can live with screwed up eyes) just tack on the eye-specific promoter region (taken from some gene you know is only expressed in the eye) to the GAL4 coding sequence.

There are hundreds of drivers available in stock centers nowadays. You don’t have to make your own. You just order them in. So if I want to drive expression of DMAP1 in, say, just the nervous system, I would make a UAS transgenic responder line that can be induced to express DMAP1, and then cross flies from that line to a nervous system-specific GAL4 driver line I would order from a stock center like Bloomington. Hey presto, the progeny from this cross would have both the driver AND responder, and would show…whatever the result would be. Here is an excellent review: PMID:12324939.

OK so we have covered transgenic flies, and how to control transgene expression. Now for RNAi. Oh heck let’s just let the Nobel prize winners earn their keep:

If you perused this nice slide show, you will now know that small double stranded RNA molecules are an evolutionarily conserved trigger for potent gene silencing. Who knew? The process of RNA interference, which is what RNAi stands for, is still revealing its elaborate secrets, and like anything that has evolved for millions of years, is…quirky. But the bottom line for experiments – that you can introduce into your model organism double stranded RNA molecules containing a sequence of a gene you wish to silence – has immense methodological power. Especially for organisms where classical genetics isn’t possible – no mutants, no way to access the DNA for mutagenesis. And there are implications for medicine too – silencing mutations that might be causing disease because of a GAIN of function mutation rather than a loss of function.

The fact is, RNAi is widely used in medicine and basic research, even while we are discovering how complex it is. Not all portions of a given coding sequence are equally potent at eliciting the silencing response. The knock down is not necessarily complete, which could be useful, if it were possible to predict accurately (which it often is not). There are other pathways in the cell that produce double stranded RNAs for developmental purposes (the micro RNA pathway). In mammals, long double stranded RNA molecules can bugger up the cell cycle and trigger cell death – a disappointing result if you are using a cell culture model. And so on.

But that shouldn’t stop us. How to make RNAi transgenic. If you have read this far, the solution should be easy. Make a transgenic responder line of flies that instead of expressing a gene of interest when induced by a driver, expresses instead a double stranded RNA molecule targeting a gene of interest (Figure 2B). There are a few ways of doing this: sewing a coding sequence in an inverted repeat manner, or putting promoters on either side of it…but it is quite possible. There are two main transgenic RNAi projects – one in Vienna, Austria, and the other in Harvard, USA. The Viennese project (I hear waltzes) does not target their RNAi responders, i.e., does not control where their RNAi responder transgenes integrate. This means  any given responder might show position effects – that is, may variably upregulate genes in the vicinity of the insertion site (the double stranded RNA is being induced by GAL4 after all) or have landed in an inhospitable chromatin environment (making access to the UAS by GAL4 difficult). On the other hand, the Viennese project has A LOT of RNAi responders, so if you get the same results with multiple responders for the same gene, that is promising.

The other project – the Transgenic RNAi Project or TrIP in Harvard uses targeting technology to control for position effects. But the chosen sites themselves may have position effects, so it is all a compromise really.

Anyway, how is this relevant to DMAP1? Because we have no mutants for DMAP1, RNAi is the only way at present to assess whether or not the gene might mutate to lethality, therefore indicating that it is essential. In the preamble to the DMAP1 project, I described the results of a paper in which the biological effect of DMAP1 was assessed using RNAi. Also, in Flybase, data from a number of sources indicates that RNAi targeting DMAP1 causes lethality. But the P element we are using to make our DMAP1 mutants is really in the upstream gene, CG33785…for which there is very little information. It looks like an RNA polymerase subunit, specifically RNA pol III, which transcribes things like tRNAs, some rRNA and many of the burgeoning small RNA molecules complicating the story of RNA interference among other things (PMID: 21540878). There are other peculiarities about this gene – it is bicistronic (huh? For now, a weirdo gene structure which I seem condemned to study) and no one seems interested in it so who knows what it is actually doing.

I obtained RNAi responders for DMAP1 and CG33785/CG33786 (two gene number designations because of the bicistronic structure).  They came from the Vienna center, because Harvard had none for these genes. Hard to believe it, but the point of this blog entry is right here: when using a ubiquitous driver (ACT5 GAL4) DMAP1 is lethal, and CG33785/6 is not. I only had one DMAP1 RNAi responder, but had THREE for CG33785/6. If I wanted to be sure, I would do QPCR (quantitative PCR) on pools of mRNA from all these experiments, to confirm gene knock down etc. Possibly, the three CG33785/6 RNAi stocks are non-functional (meaning the RNAi doesn’t work) and the one DMAP1 stock is lethal because of something in the background. But I am inclined to get what I want here. I am going to assume that DMAP1 can be mutated to lethality, whereas CG33785 – probably not. A summary of the results is shown below: the driver is balanced (marked with the dominant mutation CyO – more about balancers here) so if a particular driver/responder combination results in lethality (presumably because knock down of the gene in question is lethal) no NON balanced (CyO+) progeny will survive.


Which brings me back to my first point. Does essential mean important? DMAP1 – whatever it does – I am betting is not as ubiquitous a function as RNA polymerase III. And yet, loss of function in DMAP1 is probably lethal, while the cell can probably survive losing CG33785/6. Perhaps a better way to think of this is to take Mother Nature’s point of view. If something is important, REALLY important, required to get to the next generation, are you going to have only ONE copy of it?


DMAP1 Fertility Tests

So my colleague Kathleen and her inexhaustable supply of undergrads at Simon Fraser University in B.C., Canada, have generated seven potential DMAP1 lesions, using the male recombination screen described in the last post. Recall that in this screen, a P transposable element very close to the upstream gene CG33785 was induced to transpose in the male germline, hopefully generating a deletion in the direction of DMAP1, and thereby recombining a recessive brown eye colour mutation onto the same chromosome. Note here that four out of seven potential lesions produce viable brown-eyed (and therefore homozygous) flies. They look normal but are they really? In Figure 1, each potential mutant that produces homozygous progeny is tested for fertility. For each strain, four crosses, with virgin male and female parents, 3-4 of each sex (i.e, as close to identical conditions as possible). First, balanced males and females are crossed to each other (grey and black representing Cy males and females respectively – remember Cy is the dominant curly wing mutation that marks the SM6 balancer). Then, balanced males are crossed to homozygous females (which are dark blue in my colour scheme) – this tests the homozygous female fertility. Then the other way around: homozygous males (light blue) crossed to balanced females – this tests the homozygous males’ fertility. Finally, homozygous males and females to each other.

Fertility tests

The total number of flies for each cross exceeded 100, but I expressed the proportions of each genotype as a percentage of 100 in this graph. For balanced parents crossed to each other, the ratio of balanced to non balanced progeny (or blue shades to grey shades) should be 0.5.  (Think A/a X A/a, where A is Cy, and Cy/Cy is lethal. Half as many a/a to A/a, or 1/3 : 2/3). Note that doesn’t really happen except for 12-14 (and that ratio turns out to be 0.48).  For the two middle crosses, heterozygous males to homozygous females and the other way around (think A/a X a/a) half the progeny should be homozygous, and half should be balanced, or 1:1.  So you should see as much grey/black and light/dark blue for those crosses. Do you? Of course, homozygotes crossed to each other are depicted solely in the blue shades.

Clearly two of the potential mutants show sterility defects: 20-13 is male AND female sterile (no progeny at all for the second, third and fourth crosses), and 12-14 looks male sterile. Incidentally, in the very first post about this project, I yanked the ovaries out of 20-13/20-13 females, (look here), and asserted the testes didn’t look quite right either, so that’s consistent so far. (Could still be something in the background, but relish it for now!)

As for 20-11 and K2-13, the second and third crosses don’t produce quite the right ratios.  In addition, 20-11/20-11 parents produced approximately 70% fewer progeny than 20-11/SM6 parents, and for K2-13 that reduction was around 50%. So there might be viability and fertility issues for both lines.

In summary, 20-13 is homozygous male and female sterile, 12-14 is male sterile, 20-11 and K2-13 homozygotes may be semi fertile and perhaps even semi viable (particularly 20-11). But are these all mutations in the same gene? Stay tuned.