Transcriptional Precision in Drosophila Embryos: Enhancers and Repressors

Part 1: Transcriptional Precision: Enhancers

00:00:00.00 Hi, I'm Mike Levine. I'm at the University of California at Berkeley.
00:00:04.10 And I'm going to discuss gene regulation in the early Drosophila embryo
00:00:09.05 and it's long been recognized that the precise control of gene activity
00:00:14.21 is very important in animal development.
00:00:16.21 And a very striking example of this can be seen on the first slide.
00:00:20.13 Here you're looking at the heads of adult fruit flies, Drosophila melanogaster.
00:00:27.12 On the left is a fly that's normal. It has a pair of eyes, small antennae and mouth parts.
00:00:34.27 On the right is the head of an Antennapedia mutant.
00:00:38.19 And here you see that in place of the normal antennae,
00:00:42.17 there are perfectly sculpted middle legs.
00:00:45.21 And this results from the misexpression of the Antennapedia gene
00:00:50.10 in the wrong tissue.
00:00:51.21 So normally Antennapedia is active in the middle thoracic regions of developing embryo
00:00:57.11 where it controls the development of middle legs
00:01:00.07 but in this Antennapedia mutant, the normal Antennapedia transcription unit
00:01:05.04 is misexpressed in head tissues
00:01:07.05 and this causes a transformation of antennae into legs.
00:01:10.25 And so I'm going to discuss specifically
00:01:14.12 gene regulation during early Drosophila embryogenesis.
00:01:18.08 And this is summarized on the next slide.
00:01:20.24 So here you're looking at a series of progressively older embryos
00:01:26.03 beginning with the fertilization of the egg at stage 1.
00:01:31.02 The two pronuclei fuse to form a zygotic diploid nucleus.
00:01:38.17 And then this nucleus undergoes a series of mitoses indicated by the numbers 2, 3, and 4 and so forth.
00:01:46.20 These are among the fastest mitoses known for any animal embryo system
00:01:52.05 with a division every 6 to 8 minutes.
00:01:54.16 After 13 rounds of division you wind up with a nuclear cleavage cycle 14 embryo
00:02:00.23 which contains about 6000 nuclei forming a monolayer at the periphery of the egg;
00:02:07.20 the cortex of the egg enclosing the central yolk.
00:02:11.14 And the next slide shows some more details on the last two rounds of mitoses.
00:02:17.20 You can see that the nuclei are dividing in a syncytium.
00:02:23.00 This means that there are no cell membranes.
00:02:26.22 The nuclei are in a common cytoplasm; the ooplasm of the egg.
00:02:31.13 They're at the periphery of the egg
00:02:33.21 and you see the last two division cycles
00:02:36.12 leading to that nuclear cleavage cycle 14 embryo I discussed before.
00:02:42.06 So at the nuclear cleavage cycle 14 stage, once again, you have 6000 nuclei
00:02:48.14 forming a monolayer at the periphery of the egg and there are no further mitoses.
00:02:54.00 Now, cell membranes are laid down between the individual nuclei
00:02:57.02 over a one hour period of development.
00:03:00.01 And this is the period I'm going to discuss.
00:03:02.00 This is when localized patterns of gene expression
00:03:05.02 establish the basic blaue plan or blueprint of the adult fly.
00:03:09.29 Now this system is ideally suited for visualizing gene expression in development.
00:03:16.00 First, the nuclei are at the periphery.
00:03:19.18 There's no need to focus the objective of the microscope into the interior of the egg and embryo
00:03:25.25 because all the action is at the surface.
00:03:28.23 Secondly, you can see that the system is naturally synchronous.
00:03:33.07 All these mitoses are occurring synchronously.
00:03:36.13 All the cells, all the nuclei are at the same stage of the cell cycle.
00:03:40.21 And this is a big advantage, as you'll see, for studying gene regulation.
00:03:45.29 So, in the first part of my talk, I want to emphasize
00:03:49.16 the importance of enhancers in controlling basic on-off patterns of gene expression.
00:03:56.19 An example is shown here for a segmentation gene called Even-Skipped or Eve.
00:04:02.29 Now, eve is expressed in a series of 7 stripes along the length of the embryo
00:04:09.18 and these 7 stripes foreshadow the subdivision of an embryo into a repeating series of body segments.
00:04:16.20 Although, segmentation isn't observed until several hours later in development.
00:04:22.09 The eve gene is controlled by a series of separate enhancers
00:04:28.07 located both upstream of the transcription start site (indicated by the arrow in the digram below)
00:04:34.29 and three enhancers located downstream of the transcription start site.
00:04:40.01 So, for example, a 500 base pair enhancer located about 1kb upstream
00:04:46.25 of the eve transcription start site controls the expression of stripe number 2.
00:04:51.28 A separate enhancer located about 3 kilobases upstream of the transcription start site
00:04:59.03 controls the expressions of stripes 3 and 7.
00:05:02.23 All together, the 5 enhancers located both upstream and downstream of the eve gene
00:05:09.28 control the 7 stripes of eve expression.
00:05:13.07 And the next slide shows more details for the regulation of one of these enhancers,
00:05:17.29 the Eve-Striped 2 enhancer, which has been studied in the greatest detail.
00:05:22.11 The Stripe 2 enhancer is diagrammed here.
00:05:26.10 It is about 500 base pairs in length,
00:05:28.26 located between 1 and 1.5kb upstream of the transcription start site,
00:05:35.04 and you can see it contains a cluster of binding sites
00:05:38.02 for different sequence specific transcription factors; both activators and repressors.
00:05:43.27 And the distribution of these transcription factors relative to the Stripe 2 pattern
00:05:49.01 is shown in the middle diagram.
00:05:51.15 As I'll discuss a little later, the broad, maternal Bicoid gradient
00:05:57.11 works together with Hunchback to define a broad domain in the entire anterior half of the embryo
00:06:06.05 where the Stripe 2 enhancer could be switched on.
00:06:08.16 Bicoid and Hunchback work as activators of the Stripe 2 pattern.
00:06:12.13 But, the Stripe 2 enhancer also contains binding sites
00:06:16.15 for sequence specific transcriptional repressors
00:06:18.27 including Kruppel (the binding sites are indicated by Kr) and Giant (with binding sites indicated by Gnt).
00:06:27.02 And these work as transcriptional repressors to delineate the Stripe 2 pattern,
00:06:31.27 to form both the anterior and posterior orders.
00:06:35.05 So the only place in the embryo where you have unfettered Bicoid and Hunchback activators
00:06:41.05 that can bind to the activator binding sites within the Stripe 2 enhancer
00:06:45.19 without impedance from the Giant and Kruppel repressors
00:06:49.10 is right here, right within the limits of the Stripe 2 pattern.
00:06:53.10 So all together the 500 base pair Stripe 2 enhancer integrates
00:06:59.02 both positive and negative regulatory information
00:07:02.04 to produce this sharp on-off stripe of transcription.
00:07:06.04 And the picture below shows the relative expression pattern of the Eve stripes
00:07:12.15 So you see the 7 Eve stripes in red
00:07:15.07 and the Giant expression pattern in green.
00:07:18.21 And you see that the anterior Giant pattern tightly straddles the anterior border of Eve Stripe 2.
00:07:25.03 Giant binds to the Stripe 2 enhancer to form that stripe.
00:07:29.12 Then the last point I want to make about transcriptional enhancers in development
00:07:34.07 is that multiple enhancers can work independently of one another
00:07:38.15 to produce an additive pattern of expression.
00:07:41.03 So above you see the expression pattern produced by an enhancer called the NEE.
00:07:47.12 And the NEE produces a lateral stripe of transcription.
00:07:52.16 With expression off in the bottom of the embryo due to a localized repressor called Snail.
00:07:58.24 Now, the embryo in the bottom panel shows the expression pattern of
00:08:06.11 a synthetic transgene that contains the NEE enhancer placed next to the Eve Stripe 3 enhancer.
00:08:14.00 And what you observe is an additive pattern of expression;
00:08:17.26 a criss-cross of the NEE lateral stripe and Eve Stripe 3.
00:08:22.19 And the reason why the two enhancers are working additively
00:08:26.04 is because of short range transcriptional repression.
00:08:29.19 The Snail repressor binds to the NEE as I mentioned
00:08:33.19 to keep it off in the bottom regions of the embryo.
00:08:36.19 But the binding of Snail to the NEE does not interfere with the neighboring Eve Stripe 3 enhancer
00:08:43.10 which you can see is expressed just fine in the bottom regions of the embryo
00:08:48.10 where there are high levels of the Snail repressor.
00:08:50.23 So this property of short range repression...
00:08:53.08 repressors only inhibit the activators to which they are bound.
00:08:57.26 This permits the different enhancers to work independently of one another
00:09:01.25 and give an additive pattern of expression;
00:09:04.00 in this case a criss-cross of the NEE and Eve Stripe 3.
00:09:08.21 So to summarize this introductory part of my talk:
00:09:12.13 Enhancers are typically 500 base pairs in length.
00:09:15.26 They contain binding sites for both activators and repressors
00:09:19.09 and thereby produce sharp on-off patterns of gene expression.
00:09:23.01 Multiple enhancers work in an additive pattern to produce complex patterns of expression
00:09:29.00 due to short range transcriptional repression.
00:09:31.10

Part 2: Transcriptional Precision: Shadow Enhancers

00:00:01.01 So I've discussed how enhancers produce basic on-off patterns of gene expression
00:00:06.22 in the early Drosophila embryo
00:00:09.01 and now I'd like to address the question of transcriptional precision.
00:00:13.03 How, for example, you get the even-skipped stripes turned on in the right place,
00:00:19.28 at the right time in embryo after embryo in large populations of embryos
00:00:26.03 And to introduce the underlying mechanisms for precision and robustness,
00:00:32.17 getting that pattern right every time and in every embryo,
00:00:36.05 I'd like to introduce dorsal-ventral patterning.
00:00:39.10 That is the patterning of the back-belly axis of the developing embryo
00:00:45.08 and it is controlled by a sequence-specific transcription factor called Dorsal.
00:00:50.15 Here you see a side view of a 2 hour Drosophila embryo
00:00:56.23 stained with an antibody against the Dorsal protein.
00:01:00.03 And you can see that the protein is distributed in a broad nuclear gradient
00:01:05.17 with peak levels present in the bottom or ventral regions of the embryo, the future belly of the fly
00:01:12.03 and lower levels in more lateral and dorsal regions, the future back of the fly.
00:01:18.05 Now, this Dorsal nuclear gradient controls dorsal-ventral patterning
00:01:23.18 by regulating something like 50 target genes in a concentration dependent fashion.
00:01:29.18 And some examples of these targets are shown on the next slide.
00:01:33.22 So here's a side view of an embryo that's just 1 hour older than the one I showed you on the previous slide.
00:01:40.28 This is a 3 hours embryo and it was stained to show the expression of 4 different target genes
00:01:47.16 that are directly regulated by the Dorsal gradient.
00:01:50.22 The gene visualized in yellow is activated by peak levels of the Dorsal gradient.
00:01:56.06 Those visualized in blue, green, and red, by progressively lower levels of the gradient.
00:02:03.04 And one point I want to make concerns that gene expression pattern seen in yellow.
00:02:08.29 This corresponds to a gene called Snail which I mentioned in the previous slide.
00:02:14.10 Snail works as a localized repressor to establish the boundaries of the developing mesoderm.
00:02:22.10 So those yellow expressing cells will invaginate during gastrulation to form the internal mesoderm.
00:02:28.28 Snail works as a repressor to delineate a boundary between the mesoderm in yellow
00:02:34.29 and the lateral ectoderm in blue and green.
00:02:38.04 And I'll come back to that point of Snail as a boundary repressor a little later.
00:02:43.11 The next slide summarizes the basic mechanisms by which the Dorsal gradient
00:02:50.12 produces different thresholds of gene expression.
00:02:53.19 Now, if I can direct your attention to the bottom of the slide,
00:02:57.16 you'll see that there's a high magnification view of the Dorsal gradient
00:03:02.12 this time looking directly at the ventral midline of the embryo.
00:03:07.12 That broad Dorsal nuclear gradient directly activates an immediate early target gene called Twist
00:03:14.11 which is distributed in a steeper gradient. It's sharper than the input Dorsal gradient.
00:03:20.15 And then Dorsal and Twist work together to activate this target gene Snail
00:03:26.14 which I just mentioned and which is sharply expressed within the ventral mesoderm
00:03:31.10 with sharp, lateral borders and that Snail expression pattern delineates the boundary
00:03:36.05 between the future mesoderm and the lateral, neurogenic ectoderm.
00:03:40.11 So, the Dorsal gradient generates 3 basic patterns of gene expression
00:03:47.11 and these are produced in collaboration with Twist and Snail
00:03:52.14 which together, with Dorsal, work as sequence specific transcription factors.
00:03:56.28 The three basic patterns are called Type 1, Type 2, and Type 3.
00:04:02.05 The Type 1 expression pattern is represented by Snail, which I've mentioned.
00:04:07.10 Snail expression is restricted to the ventral most regions of the embryo
00:04:11.28 where the highest levels of the Dorsal and Twist activators are.
00:04:15.23 Type 2 target genes, and an example is shown in panel E,
00:04:20.02 shows a narrow lateral stripe of gene expression
00:04:24.11 where there are intermediate to low levels of the Dorsal and Twist gradient.
00:04:29.20 And then finally, the Type 3 pattern seen in panel F shows
00:04:34.23 a broad lateral stripe which is activated in response to the lowest levels of the Dorsal gradient.
00:04:40.17 So these are the 3 basic thresholds of gene expression produced by the Dorsal gradient.
00:04:45.27 Now, to the right of each of these 3 expression patterns, you see
00:04:51.04 summaries of the enhancers which control the expression patterns.
00:04:55.12 So the Type 1 expression pattern is regulated by an enhancer
00:05:00.09 that contains unlinked Dorsal and Twist activator sites.
00:05:04.23 They are uncolored indicating that they are low affinity, suboptimal binding sites.
00:05:11.23 So the only way you can turn on Snail and occupy those binding sites
00:05:16.20 is with the highest levels of the Dorsal and Twist activator.
00:05:19.17 In contrast, the Type 2 enhancer contains at least 1 pair of optimal Dorsal and Twist binding sites.
00:05:29.01 The 2 binding sites are closely linked and have a specific arrangement.
00:05:34.23 The asymmetric Twist site points towards the Dorsal site.
00:05:38.21 And these features of arrangement and high affinity sites
00:05:43.04 allows intermediate to low levels of the Dorsal gradient to activate the Type 2 target gene.
00:05:49.05 And then finally, Type 3 target genes which respond to the lowest levels of the Dorsal gradient,
00:05:54.21 have multiple, high affinity Dorsal binding sites (once again indicated by the green ball)
00:06:02.00 as well as closely linked binding sites for ubiquitous transcriptional activators such as Zelda.
00:06:08.10 And this linkage between Zelda and optimal Dorsal binding sites
00:06:12.08 ensures that the lowest levels of the Dorsal gradient are able to activate the Type 3 genes.
00:06:17.24 You'll notice that both the Type 2 and the Type 3 enhancers contain binding sites
00:06:23.17 for the Snail repressor and that keeps these enhancers off in the bottom of the embryo
00:06:30.04 and the future mesoderm. Off and off due to the Snail repressor.
00:06:34.01 So all together these three different thresholds of gene expression
00:06:38.24 in response to the Dorsal gradient depend on enhancers with
00:06:43.17 different binding affinities of Dorsal and Twist activators,
00:06:46.12 with the presence or absence of Snail repressor binding sites,
00:06:50.24 and depending on the so-called grammar, the local arrangements of binding sites,
00:06:56.00 such as the close linkage of Dorsal and Twist in the Type 2 enhancers,
00:07:00.01 and the linkage of Dorsal and Zelda seen in Type 3 enhancers.
00:07:03.01 All together, these features of the enhancers determine
00:07:07.03 the kind of output expression patterns that are attained in response to the Dorsal gradient.
00:07:12.23 So now I want to get, finally, to the question that I promised earlier
00:07:18.22 and that's transcriptional precision.
00:07:21.05 How do you get these expression patterns on not just in 1 lucky embryo,
00:07:26.09 but in all embryos, in natural populations
00:07:30.03 which are subject to all sorts of variation in genetic background and in the environment.
00:07:35.25 And the first mechanism I'm going to discuss is called the shadow enhancer.
00:07:41.03 And these were identified using whole genome binding assays,
00:07:45.20 looking at the genome wide distribution of Dorsal, Twist, and Snail binding
00:07:50.20 throughout the Drosophila genome in the early Drosophila embryo.
00:07:54.13 An example of these whole genome ChIP binding assays is shown here.
00:07:59.11 So this is a Dorsal target gene called Sog. It's a Type 3 target gene.
00:08:05.07 It's activated by the lowest levels of the Dorsal gradient.
00:08:09.05 At the bottom here you're seeing an embryo
00:08:12.01 which is oriented so that the mesoderm is facing you.
00:08:15.25 The Sog expression pattern is in the lateral, neurogenic ectoderm
00:08:20.21 in response to low levels of the Dorsal gradient
00:08:23.06 and you see that the gene is off in the ventral mesoderm due to the Snail repressor.
00:08:28.01 Now, the whole genome ChIP binding assay identified a cluster
00:08:33.27 of Dorsal, Twist, and Snail binding sites within the first intron of the Sog transcript
00:08:39.28 and this corresponds to the known enhancer which regulates Sog expression.
00:08:45.18 But, to our surprise, these whole genome binding assays identified
00:08:49.27 a second cluster of binding sites located in a remote position
00:08:54.11 over twenty kilobases upstream of the Sog transcription unit on the other side of this linc gene.
00:09:01.15 And despite this unexpected location, this genomic DNA fragment
00:09:07.01 containing this cluster of binding sites works as an enhancer to regualte Sog expression.
00:09:12.26 It produces a pattern of expression very similar to the one
00:09:16.08 produced by the intronic enhancer and seen for the endogenous gene.
00:09:20.28 So we have 2 separate enhancers separated by 25 kilobases
00:09:26.17 that produce basically the same pattern of expression.
00:09:29.24 Now this is different from what I discussed in the introduction regarding the Eve stripes.
00:09:35.00 There you have separate enhancers for different stripes.
00:09:38.00 In this cases we have 2 separate enhancers for the same pattern of expression.
00:09:42.05 And we refer to the remote enhancer, the more distal enhancer, as the shadow enhancer.
00:09:47.15 Shadow enhancers are sometimes found in really weird locations
00:09:52.19 like the introns of neighboring genes and an example of that is shown on the next slide.
00:09:57.09 So here you see another Dorsal target gene called Brinker,
00:10:01.14 which is expressed in the lateral neurogenic ectoderm
00:10:05.22 in response to low levels of the Dorsal gradient.
00:10:08.16 It's kept off in the ventral mesoderm by the Snail repressor.
00:10:12.14 The whole genome ChIP assays identifies a cluster of Dorsal, Twist and Snail binding sites
00:10:19.06 at the 5' end of the gene about 10 kilobases upstream of the Brinker transcription start site.
00:10:25.14 There's the Brinker gene itself.
00:10:27.04 And this 5' cluster corresponds to the known enhancer which regulates Brinker expression.
00:10:34.06 But, again, unexpectedly, the whole genome binding assays identified
00:10:39.16 a second cluster of binding sites located within the intron of this neighboring gene called Atg5
00:10:46.02 which is unrelated to Brinker and is not even turned on in the early embryo
00:10:50.27 and despite this odd location, this cluster of binding sites works as an enhancer
00:10:56.10 to regulate Brinker expression and produce a lateral stripe pattern
00:11:01.18 very similar to that seen for the endogenous gene and that generated by the normal 5' enhancer.
00:11:07.28 So here we have a situation similar to what I described for Sog--2 enhancers
00:11:13.28 separated by about 25 kilobases; 1 located 5', 1 located 3' within the neighboring gene.
00:11:20.29 These two enhancers produce very similar patterns of expression.
00:11:24.25 So what's the deal? I mean, something on the order of 1 half of the 50 Dorsal target genes
00:11:32.17 are thought to contain shadow enhancers. What are they doing?
00:11:35.28 We certainly didn't expect them
00:11:38.03 and they certainly weren't identified using conventional gene fusion assays
00:11:42.10 because of their very remote locations.
00:11:44.15 And so to get at the question of what is the shadow enhancer do for development
00:11:51.01 a graduate student in the lab, Mike Perry, analyzed the regulation of the Snail gene.
00:11:56.19 As I mentioned before, Snail is activated in response to high levels of the Dorsal gradient
00:12:03.01 within the ventral regions of the embryo which will form the mesoderm.
00:12:07.24 Snail is regulated by 2 separate enhancers; a primary enhancer and a shadow enhancer
00:12:14.17 separated by about 5 kilobases and you see that the shadow enhancer is located
00:12:19.20 within the first intron of this neighboring locus Tim17
00:12:24.07 which again is not expressed in the early embryo and is unrelated to Snail.
00:12:29.15 Despite this location within the Tim17 locus, the shadow enhancer produces
00:12:35.27 a pattern of expression very similar to that observed for the endogenous gene
00:12:40.13 including very sharp lateral limits of expression at
00:12:43.21 the boundary between the future mesoderm and the lateral neurogenic ectoderm.
00:12:47.28 So Mike Perry made a series of BAC transgenes which mutagenize
00:12:54.07 either the proximal primary enhancer or the distal shadow enhancer
00:13:00.20 or kept both enhancers intact.
00:13:03.05 And these transgenes encompass about 25 kilobases of genomic DNA,
00:13:08.28 include the entire Snail and Tim17 loci and associated regulatory DNAs.
00:13:15.26 In some of the experiments Mike replaced the Snail transcription unit with the yellow reporter gene.
00:13:24.05 This contains a large intron which permits us to detect nascent transcripts and,
00:13:29.20 as you'll see, this is the most quantitative means for assessing gene activation in the embryo;
00:13:36.10 to do in situ hybridization detecting nascent transcripts using intronic products.
00:13:41.04 Alright. So, Mike made the set of 3 BAC transgenes;
00:13:46.14 both enhancers intact, mutagenize either the shadow or the proximal primary enhancer.
00:13:53.01 An example of this analysis is shown in the next couple of slides.
00:13:56.29 So here you're looking at the early activation of the yellow marker gene
00:14:03.28 from these BAC transgenes
00:14:06.07 and we're looking at the optimal culturing condition of 22 degrees.
00:14:10.00 This is the normal temperature, room temperature,
00:14:13.02 that we normally raise our strains of Drosophila melanogaster.
00:14:18.19 And you can see that whether you have 1 enhancer or 2 enhancers
00:14:23.16 the initial activation pattern is fine.
00:14:27.23 You see homogeneous expression, sharp lateral limits within the central mesoderm.
00:14:32.19 So in this case removal of the proximal enhancer does not alter the early activation in gene expression.
00:14:42.14 But, this is an optimal temperature, 22 degrees.
00:14:46.20 At high temperatures, 30 degrees, which are stressful for the flies
00:14:51.15 but natural populations deal with it all the time,
00:14:54.05 the transgenes containing a single enhancer begin to break down
00:14:58.06 and that can be seen on the next slide.
00:15:00.11 So now, at 30 degrees, the BAC transgene containing just the shadow enhancer
00:15:06.28 with the proximal enhancer disrupted: this expression pattern is abnormal.
00:15:13.22 You see gaps in expression and the lateral limits start to get jagged.
00:15:18.16 In contrast, the transgene that contains both enhancers intact continues to produce
00:15:26.13 a normal pattern of expression; homogeneous expression with sharp lateral limits.
00:15:31.17 And so these types of experiment, and we've done many more like this,
00:15:35.21 suggest that the shadow enhancer helps ensure that critical patterning genes like Snail
00:15:42.01 are turned on in the right place at the right time in embryo after embryo
00:15:46.19 in large populations of embryos subject to both environmental variations,
00:15:51.14 such as increases in temperature, and also genetic variation
00:15:55.04 which is a common feature of natural populations.
00:15:59.11 So the shadow enhancer helps confer robustness.
00:16:03.28 In addition, shadow enhancers help improve the precision, the clarity,
00:16:10.01 the sharpness of gene expression and I'll show one example regarding segmentation.
00:16:16.17 Now I mentioned earlier the bicoid gradient. It is one of the activators of eve stripe 2.
00:16:24.20 Bicoid is already present in the unfertilized egg.
00:16:27.22 The protein is distributed in a broad gradient
00:16:30.19 with peak levels in anterior regions and lower levels in more posterior regions.
00:16:35.17 This bicoid gradient controls segmentation by leading to sequential patterns
00:16:41.14 of gap gene expression and as I showed you, directly regulating pair-rule stripes
00:16:46.29 of segmentation gene expression such as eve stripe 2.
00:16:50.19 The next slide shows the gap genes that are regulated by bicoid.
00:16:57.03 This just shows three of the gap genes. There are several others not shown here.
00:17:00.17 These are hunchback, Kruppel and knirps which are expressed sequentially
00:17:06.00 along the anterior to posterior axis.
00:17:09.15 All of these gap genes contain shadow enhancers
00:17:12.24 but I want to just focus yoru attention to the top panel
00:17:15.18 of the hunchback expression pattern.
00:17:17.24 Hunchback shows a simple on/off pattern of gene expression.
00:17:23.26 It is on in the anterior half of the embryo and off in the posterior.
00:17:29.02 It is switched on in response to high and intermediate levels of the bicoid gradient.
00:17:34.18 But low levels of bicoid in posterior regions of the embryo
00:17:38.11 are insufficient to activate hunchback transcription.
00:17:42.04 This on/off switch in transcription is controlled by a classical enhancer
00:17:48.10 that was identified over 20 years ago in the labs of Nusslein-Volhard and Gary Strulh.
00:17:54.05 The enhancer is located right at the proximal p2 promoter shown in blue here.
00:18:00.21 And it contains a series of optimal bicoid binding sites
00:18:05.13 that respond to high and intermediate levels of the bicoid gradient to give that on/off switch.
00:18:11.13 About a year ago, a graduate student in my lab Mike Perry identifed a shadow enhancer
00:18:18.17 unexpectedly that's located 5 kilobases upstream of the classical enhancer.
00:18:24.23 within the extended hunchback regulatory region and it's indicated in red there.
00:18:31.24 So this is a similar situation to what I described for snail.
00:18:35.24 We have two separate enhancers separated by 5 kilobases
00:18:40.16 that produce very similar patterns of expression.
00:18:43.01 You see that the shadow enhancer gives and on/off switch in transcription
00:18:46.18 very similar to that obtained with the classical, proximal enhancer.
00:18:51.04 So again, why two enhancers?
00:18:53.18 In this case it doesn't matter what temperature you look at.
00:18:57.18 Two enhancers give a sharper pattern of hunchback expression than either one of them.
00:19:03.05 And some evidence for that is shown here.
00:19:05.10 Once again, we are looking at a series of BAC transgenes
00:19:10.01 that are about 20 to 25 kilobases in length and einclude the extended 5' hunchback regulatory sequence.
00:19:18.14 The hunchback transcription unit was replaced with the yellow reporter gene,
00:19:23.14 once again permitting detection of nascent transcripts using intronic probes.
00:19:28.23 We have a BAC transgene with both enhancers intact on the right.
00:19:33.07 The classical enhancer is disrupted in the middle panel
00:19:37.27 and the newly identified shadow enhancer is disrupted in the left most panel.
00:19:43.18 And you can see that the BAC transgenes containing just a single enhancer
00:19:48.12 give rise to kind of erratic patterns of activation.
00:19:52.24 And we're looking at transcriptional activation literally within 20 minutes
00:19:56.24 after the first formation of the bicoid gradient.
00:19:59.22 So this is very early activation in response to the gradient.
00:20:03.20 And you can see in the first two panels (the left and the middle panel)
00:20:08.00 the patterns are erratic and the borders is indistinct,
00:20:12.21 and there's even ectopic expression outside the normal limits
00:20:17.09 of the hunchback transcription pattern.
00:20:19.08 However, with two enhancers, as seen on the right, you get a much more
00:20:23.17 faithful pattern of expression, homogenous activation, and a fairly sharp posterior border.
00:20:29.22 And those numbers on top refer to the discordance between the expression pattern
00:20:36.09 produced by the transgene and the endogenous hunchback expression pattern.
00:20:41.05 So for example, in the middle panel, 24% of the nuclei
00:20:45.18 which express the endogenous hunchback gene fail to show expression of the BAC transgene.
00:20:53.26 So that's a very high rate of discordance.
00:20:55.27 Whereas, when you have both enhancers intact you only have a 3% discordance.
00:21:00.26 That pattern the with two enhancers is much more faithful than
00:21:04.10 the patterns produced with either the shadow enhancer or the primary proximal enhancer alone.
00:21:10.15

Part 3: Transcriptional Precision: Paused Polymerase II

00:00:00.20 So I've discussed the shadow enhancers as a mechanism of transcriptional precision
00:00:07.05 ensuring robust expression in natural populations
00:00:11.17 and more precise, sharper patterns of gene activation.
00:00:16.28 A second mechanism of transcriptional precision that I'd like to discuss today
00:00:22.08 is paused RNA polymerase.
00:00:24.17 And paused PolII was identified initially performing whole-genome ChIP binding assays
00:00:33.00 using antibodies against RNA polymerase II.
00:00:36.16 And of course this was done on a whole genome scale
00:00:38.23 but what I'm showing you here is just one example of paused polymerase.
00:00:43.10 This is for our old friend the Sog gene which is activated by
00:00:47.17 low levels of the dorsal gradient in the lateral neurogenic ectoderm.
00:00:52.14 Again, Sog is kept off the ventral mesoderm by the snail repressor.
00:00:56.06 What you see here is the binding of RNA polymerase II to the promoter region of this Sog gene
00:01:02.22 before it's switched on in the early embryo.
00:01:05.13 So I refer to this as paused polymerase.
00:01:07.19 We know that this polymerase is fully activated
00:01:11.23 and that is indicated by as assay called a nuclear run-on assay
00:01:15.16 which I'm not going to discuss in any particular detail.
00:01:18.05 But the assay not only identified where polymerase is bound,
00:01:22.10 but where, specifically, activated polymerase is found in the genome.
00:01:26.26 So we know that this polymerase: It has begun transcription.
00:01:31.16 It has escaped from the promoter,
00:01:33.15 melted the DNA, made a short nascent transcript of about 30-50 nucleotides,
00:01:38.03 and then arrests 30-50 nucleotides downstream of the Sog transcription start site.
00:01:44.18 So prior to Sog activation, this gene is ready to go.
00:01:49.03 Polymerase is bound, and engaged, and activated.
00:01:53.11 And this situation is not just true for Sog,
00:01:56.23 but most of the genes that control embryonic development
00:02:00.17 show paused polymerase prior to the activation.
00:02:03.29 So these are the Hox genes,
00:02:05.19 genes encoding different signaling pathways
00:02:09.05 such as Notch signaling, Wnt and Hedgehog signaling.
00:02:12.28 These types of genes tend to contain paused polymerase in the Drosophila embryo.
00:02:18.00 And the question is why. What does paused polymerase do for gene activation?
00:02:23.08 And a hint came from quantitative in situ hybridization assays
00:02:30.07 done by a graduate student in my lab Allister Boettiger.
00:02:34.10 And what Allister did was to look very carefully at the first nascent transcripts,
00:02:40.16 at the moment of activation, produced by genes that contained paused polymerase
00:02:47.05 or lacked paused polymerase.
00:02:48.13 And here you're seeing the early activation of two different genes
00:02:53.22 activated by low levels of the Dorsal gradient; Sog in green and Thisbe in red.
00:03:00.09 You can see they're both turned on in the lateral regions that will form the neurogenic ectoderm
00:03:06.20 however Sog contains paused polymerase whereas Thisbe does not.
00:03:15.29 Thisbe lacks paused polymerase at its promoter prior to activation.
00:03:20.22 And so what you see in the case of Sog which contains the paused polymerase,
00:03:24.27 is that, within a few minutes of the first detection of the first nascent transcripts,
00:03:32.12 over 90% of the nuclei, which will eventually express the gene, show transcription.
00:03:37.20 And so a very synchronous pattern of gene activation.
00:03:42.00 In contrast, in the case of Thisbe, which lacks the paused polymerase,
00:03:45.19 you can see that only about 50% of the nuclei show nascent transcripts
00:03:51.04 within the first several minutes of gene activation.
00:03:53.15 And it will take another 20-30 minutes before this pattern fills out
00:03:58.21 and becomes uniform as seen for Sog.
00:04:01.27 So paused polymerase fosters rapid and synchronous patterns of gene activation.
00:04:08.18 And this synchrony is a property of the core promoter.
00:04:14.20 And some evidence for that is shown here.
00:04:17.05 So if you look at the bottom panel, you see the early activation of a gene called Pannier.
00:04:24.16 Pannier is switched on in the dorsal ectoderm of the early embryo.
00:04:29.18 Its initial transcription is highly erratic or stochastic.
00:04:34.27 Only about 30% of the nuclei which will eventually express the gene
00:04:39.16 show nascent transcripts during the early phase of its activation.
00:04:44.05 Now what you are looking at in these top panels are the same Pannier transgenes
00:04:50.11 except that the core promoter of Pannier, which lacks paused polymerase,
00:04:55.18 has been replaced by two paused promoters; either Snail in the middle or Tail-up.
00:05:03.00 Ok? So, we're switching the non-paused Pannier promoter on the bottom
00:05:06.12 to the paused paused Snail or Tail-up promoters but otherwise the transgene is kept the same.
00:05:11.25 And what you observe is much more synchronous patterns of gene activation
00:05:17.00 with the paused Snail promoter or the paused Tail-up promoter
00:05:21.20 as compared to the natural Pannier promoter which comes on in a highly erratic fashion.
00:05:26.23 So it is the minimal promoter, in this case just 100 nucleotides
00:05:31.05 centered around the transcription start site that is sufficient
00:05:35.05 for pausing of the polymerase and producing synchronous patterns of gene activation.
00:05:40.29 Now, it turns out that many of the most critical embryonic patterning genes
00:05:48.23 contain both paused polymerase and shadow enhancers.
00:05:53.20 And I want to show you evidence that this combination
00:05:56.14 of paused polymerase and shadow enhancer is a potent one-two punch
00:06:01.22 for rapid, synchronous, robust, precise patterns of gene expression
00:06:07.16 that also produce very high rates of RNA synthesis.
00:06:11.24 And I'm going to come back to our old friend the Snail gene to illustrate this point.
00:06:18.03 So Snail, like a number of key patterning genes in the early embryo,
00:06:23.08 contains both paused polymerase and a shadow enhancer;
00:06:28.19 two enhancers for the early pattern of gene activation.
00:06:32.22 And what you see in panel A is a high magnification view of the Snail expression pattern.
00:06:40.16 I've mentioned how Snail is activated by high levels of the dorsal gradient.
00:06:44.17 It shows very sharp limits of gene expression that delineate the boundary between
00:06:50.16 the mesoderm in red and the lateral ectoderm, the neurogenic ectoderm
00:06:55.12 which will produce the nervous system of the fly.
00:06:57.29 And I want to be very clear about this boundary. It is a critical boundary in the early embryo.
00:07:03.10 The mesoderm cells (the red expressing cells) will invaginate during gastrulation
00:07:08.16 to form internal tissues and muscles.
00:07:11.29 In contrast, just on the other side of the Snail border,
00:07:15.24 these lateral cells will remain initially on the surface of the embryo,
00:07:20.25 forming surface ectoderm which will eventually give rise to the nervous system
00:07:25.02 and to the nerve core of the adult fly.
00:07:27.24 So very different fates across that border.
00:07:30.10 Sharp borders are often the sites of active cell signaling
00:07:35.10 and in this case, there's Notch signaling across the border.
00:07:38.09 The Snail expressing cells send a Notch signal
00:07:41.25 to the row of cells just on the other side of the border.
00:07:45.06 That row of cells receives the Notch signal
00:07:48.16 and activates a regulatory gene called Sim which is important for
00:07:53.11 establishing the ventral midline of the nervous system
00:07:57.03 and the midline is important for the patterning of the entire nervous system
00:08:01.06 similar to the role of the floor plate in the vertebrate neural tube.
00:08:05.27 So I want to discuss the formation of that critical, sharp Snail border,
00:08:10.19 that site of signaling and activation of Sim
00:08:15.16 within the ventral most regions of the neurogenic ectoderm.
00:08:18.23 Now, the sharp border depends on mutual repression
00:08:24.16 between neurogenic repressors present in lateral regions of the embryo
00:08:29.19 and Snail itself which functions, as I mentioned, as a transcriptional repressor.
00:08:34.09 So you have repressors fighting each other to sharpen that border
00:08:38.12 and you'll see this in more detail in a moment.
00:08:40.27 On the right you see the consequences of removing the Snail repressor.
00:08:45.28 So this is a mutant embryo that fails to make the Snail repressor.
00:08:51.09 The Snail RNA you can see is unevenly distributed. There's no longer a sharp border.
00:08:58.29 Snail expression is being rapidly shut down
00:09:02.07 and neurogenic repressors that are normally restricted to lateral regions
00:09:06.15 are now de-repressed throughout the embryo,
00:09:08.24 converting what should become the mesoderm into extra lateral neurogenic ectoderm.
00:09:15.12 This process can be seen in a little more detail in this slide.
00:09:20.11 So here you see the refinement of the lateral limits of the Snail expression pattern.
00:09:26.08 When Snail first comes on the borders are rather indistinct
00:09:31.12 but in a period of just 30 to 40 minutes,
00:09:35.16 they are quickly sharpened to form that razor sharp, not straight but sharp boundary
00:09:41.28 between the future mesoderm and the neurogenic ectoderm.
00:09:45.03 The sharpening of the Snail expression pattern, as I mentioned,
00:09:48.15 depends on mutual repression between Snail and lateral neurogenic repressors such as Vnd.
00:09:55.22 Vnd is not the only repressor involved in this process
00:09:59.16 of sharpening the Snail border but it's one of them.
00:10:01.20 So the idea is that Snail gets turned on.
00:10:04.14 It represses Vnd, restricts Vnd to the lateral neurogenic ectoderm
00:10:09.24 and then Vnd returns the favor by feeding back on Snail, repressing it to sharpen that border.
00:10:16.05 Now this is a classical bistable threshold switch
00:10:20.25 with Snail and Vnd duking it out to form that sharp border.
00:10:25.20 It gives a sharp border but it's a dangerous mechanism
00:10:28.26 because it depends on Snail being very quickly expressed
00:10:33.23 in order to achieve critical, threshold levels in the early embryo
00:10:37.23 to keep these neurogenic repressors out of the mesoderm
00:10:41.01 and restricted to the lateral neurogenic ectoderm.
00:10:44.03 If you don't make enough Snail at the right place and the right time,
00:10:47.10 then the mesoderm is not formed and it's transformed into neurogenic ectoderm.
00:10:54.01 So to get a feeling for just how fast is the rate of Snail synthesis
00:11:00.24 in this bistable threshold switch and the sharpening of the Snail border,
00:11:06.20 Allister developed a quantitative in situ hybridization assay.
00:11:10.22 And I'm not going to take you through all the hideous details here
00:11:14.18 but basically he takes a fluorescent probe, hybridizes to staged embryos,
00:11:20.16 uses a variety of deconvolution methods to identify individual hybridization dots
00:11:28.01 and we have evidence that each dot represents a single Snail mRNA.
00:11:34.00 And then he counts the dots to get a feeling for the number of Snail products
00:11:38.16 present in a given cell at a given stage in development.
00:11:41.20 On the right here you see a massaged image containing
00:11:47.06 4 consecutive nuclei across the Snail border.
00:11:52.05 Now, the embryo is still a syncytium.
00:11:56.09 The cell membranes have not completely formed between the nuclei
00:11:59.09 so what I'm calling a cell is the nucleus plus the most closely associated cytoplasm.
00:12:05.11 And you can see that the two cells to the left have
00:12:10.00 high levels of the Snail hybridization dots (of Snail mRNAs).
00:12:14.11 The nuclei or cells on the right basically lack Snail transcripts
00:12:19.04 and there's the sharp border.
00:12:20.06 You'll also see that the hybridization dots tend to be asymmetrically distributed
00:12:26.15 within the cell with more dots on the apical side of the cell than in the basal cytoplasm.
00:12:32.10 And this is true for many of the mRNAs encoded by developmental patterning genes
00:12:38.04 in the early fly embryo. They tend to show this apical localization.
00:12:42.24 So the next slide shows some of the actual numbers.
00:12:46.15 So what you're looking at here are
00:12:49.23 each dot is counting the Snail mRNAs in a given cell at a given stage
00:12:56.19 and so at the beginning of nuclear cleavage cycle 13,
00:13:00.12 there is an average of 50 Snail mRNAs per cell.
00:13:04.21 And then this number rises to 120 mRNAs per cell
00:13:10.08 during the 15 minute window that is the interphase period of nuclear cleavage cycle 13.
00:13:17.03 Then the cell divides and of course you go from an average of 120 mRNAs per cell
00:13:22.09 down to 60 mRNAs per cell at the beginning of nuclear cleavage cycle 14
00:13:28.00 and then during the next 25 minutes this number rises to its steady state level
00:13:34.24 from 60 to an average of 180 mRNAs per cell over a period of about 25 minutes.
00:13:42.19 This is a very fast rate of synthesis and that's summarized on the next slide.
00:13:48.27 So the steady state, as I mentioned, is 180 Snail mRNAs per cell
00:13:56.06 and this was reached in a period of about 25 minutes
00:13:59.17 during the first half of nuclear cleavage cycle 14.
00:14:03.00 That is a rate of synthesis of 20 mRNAs per minute
00:14:06.24 assuming a half-life of 10 minutes and we think that's an over-estimate.
00:14:12.04 The actual half-life for the Snail mRNA may be more like 7 to 8 minutes.
00:14:16.01 But even assuming a half-life of 10 minutes, that means that one Snail mRNA
00:14:22.14 is synthesized every 5 to 10 seconds per allele.
00:14:27.09 and that is a very fast rate of synthesis. That's pushing the theoretical limit.
00:14:32.12 The limit of transcription rates in the early Drosophila embryo is 1 RNA per 4 to 5 seconds.
00:14:39.12 That number is based on the rate of polymerase elongation (about 20 nucleotides per second).
00:14:45.18 That was measured by John Liss some years ago
00:14:48.04 and we have evidence that it's completely consistent with the rate of 20 nucleotides per second
00:14:53.24 and it's only possible to load one polymerase at the promoter
00:14:58.07 every 4 to 5 seconds because the polymerase is very large and bulky
00:15:03.23 and you can only load one every 80 to 100 nucleotides.
00:15:07.20 So that's your rate of synthesis.
00:15:09.27 One transcript every 4 to 5 seconds due to the rate of polymerase elongation
00:15:14.24 and the size of the polymerase and the ability to load polymerase onto a crowded promoter.
00:15:20.00 You've got to think about bumper to bumper traffic on the freeway
00:15:23.13 trying to all go at the speed limit.
00:15:25.29 And so we were within a factor of 2 of the theoretical limit
00:15:30.15 of the rate of synthesis of Snail mRNA
00:15:33.08 and as I said, that's important to ensure that Snail represses the neurogenic repressors
00:15:38.15 during this mutual fight that sharpens the Snail border.
00:15:44.00 So you always get that right by getting a very fast rate of synthesis
00:15:48.04 and this is fostered by both shadow enhancer and paused polymerase.
00:15:52.23 So I just want to finish this part of my talk by emphasizing the importance
00:15:59.16 of paused polymerase in giving rapid and synchronous patterns of gene activation.
00:16:07.00 Paused polymerase is also associated with genes that are
00:16:10.11 subject to very rapid rates of synthesis as seen for Snail.
00:16:14.02

Part 4: Transcriptional Precision: Repressors

00:00:00.08 I've discussed two mechanisms of transcriptional precision in the early Drosophila embryo;
00:00:06.16 shadow enhancers and paused polymerase.
00:00:09.25 These two mechanisms ensure robust, rapid and uniform patterns of gene activation,
00:00:19.00 and, when used together, very rapid rates of RNA synthesis.
00:00:24.04 I want to finish up be discussing mechanisms of transcriptional repression.
00:00:30.12 And I've already alluded to the importance of repression in the early Drosophila embryo.
00:00:35.26 For example, repressors form the two borders of the Eve-Stripe 2 pattern.
00:00:40.28 The Snail repressor establishes the boundary between the mesoderm and neurogenic ectoderm.
00:00:46.25 And now I want to address how does this repression work.
00:00:51.14 And I will turn specifically to Snail repressor.
00:00:55.09 So here, once again, we're looking at the expression of the Sog gene in the early embryo.
00:01:02.10 As I've mentioned a few times, Sog is activated by low levels of the Dorsal gradient
00:01:06.28 in the lateral neurogenic ectoderm.
00:01:09.19 It's kept off in the ventral mesoderm by the Snail repressor
00:01:14.26 and of course Sog is one of those magic genes
00:01:18.07 that has both paused polymerase and a shadow enhancer.
00:01:21.29 The enhancers aren't shown here, but earlier I showed you that
00:01:24.26 Sog has both a primary and intronic enhancer and a remote, 5' shadow enhancer.
00:01:30.22 So Sog, like Snail, comes on quite quickly and uniformly in the early embryo.
00:01:37.07 Now, Sog, in fact, beats Snail to the punch
00:01:41.17 and is turned on a little bit before there are critical levels of the Snail repressor.
00:01:47.07 So, initially, Sog is turned on in both the ventral regions that will form the mesoderm as well as
00:01:54.02 the lateral regions that will form the neurogenic ectoderm where Sog becomes stably expressed.
00:01:59.13 And a graduate student in the lab, Jacques Bothma has done high-resolution imaging
00:02:04.27 of this loss of Sog expression. (although it's only transiently expressed).
00:02:10.22 This loss of transient expression of Sog in the ventral mesoderm due to the Snail repressor.
00:02:16.20 So that's what I'm going to show you.
00:02:18.22 The experimental design is summarized here.
00:02:22.11 So this summarizes the Sog transcription unit.
00:02:26.27 It's a fairly large gene (20 kilobases) with a larger 5' intron and 3' intron.
00:02:34.17 Jacques made separate fluorescent probes for the 5' intron and the 3' intron.
00:02:40.17 The two probes are separated by 10 kilobases.
00:02:43.29 And then he mixed these together and hybridized them to staged embryos.
00:02:47.24 When the Sog gene first comes on, you only see green hybridization signals
00:02:54.03 and that's because the polymerase is only located within the first intron
00:03:00.02 so the only nascent transcripts present at this early stage of activation of Sog
00:03:05.22 contain only 5' intron sequence.
00:03:08.28 But after a delay, the polymerase enters the 3' intron
00:03:13.17 and as it does so, the 5' intron is spliced out.
00:03:18.14 That's indicated by this little lariat.
00:03:20.07 Then, the polymerase extends into the 3' intron and
00:03:24.27 now the 3' intronic hybridization probe cross reacts with these nascent transcripts.
00:03:31.26 So after a delay, the Sog gene is stably expressed.
00:03:35.22 You see yellow hybridization signals because the nuclei contain a mixture
00:03:40.21 of nascent transcripts containing either 5' intron sequences or 3' intronic sequences.
00:03:47.06 So now the question is: How does Snail work as a repressor?
00:03:51.11 The Snail repressor kicks in and will switch the gene off in the ventral mesoderm as I mentioned.
00:03:57.17 And there are at least two possible mechanisms.
00:04:01.11 One possibility is that Snail disrupts polymerase elongation.
00:04:06.13 The polymerase is stopped right in its tracks
00:04:08.25 so the polymerase falls off and the gene is immediately shut off.
00:04:13.06 And so according to this view,
00:04:15.07 we should go from a yellow hybridization signal when the gene is on
00:04:18.16 to no hybridization signal at all, just a blue, unlabeled nuclei.
00:04:23.29 So that's if Snail works at the level of polymerase elongation.
00:04:28.21 A second mechanism is if the polymerase works at the level of initiation
00:04:33.25 by blocking the release of the paused polymerase from the promoter.
00:04:38.10 According to this mechanism, once Snail kicks in, there's no new synthesis.
00:04:45.02 There's no new release of polymerase
00:04:47.03 but the elongating polymerases that were released
00:04:50.18 prior to the onset of Snail repression are able to complete transcription.
00:04:54.18 And if so, we should see an intermediate stage of hybridization
00:04:59.17 which we call the red flash.
00:05:01.24 These are the elongating polymerases completing transcription of the gene
00:05:07.10 that had been released from the promoter prior to the onset of Snail repression.
00:05:13.01 So according to the first mechanism (inhibition of elongation), we should go from yellow to off.
00:05:19.10 According to the second mechanism of blocking the release of polymerase
00:05:23.24 or transcription initiation, we should go from yellow to red to off.
00:05:28.21 And this is what we see.
00:05:30.26 So, we obtained evidence that Snail works at the level of transcriptional initiation,
00:05:36.05 blocking the release of polymerase, but the polymerase that had already been released,
00:05:40.25 completes transcription even after the onset of Snail repression.
00:05:46.20 And some evidence for that is shown here.
00:05:49.08 So here you're looking at progressively older nuclear cleavage cycle 14 embryos.
00:05:55.04 These embryos were double-labeled
00:05:57.19 with the green 5' intronic probe and the red 3' intronic probe
00:06:03.06 looking at Sog expression and repression.
00:06:06.26 When Sog is first switched on, you see only green hybridization signals.
00:06:12.23 Activation's fairly synchronous as we would expect for a gene containing paused polymerase.
00:06:18.00 And it's very rapid as we would expect for a gene that
00:06:20.22 has both paused polymerase and a shadow enhancer.
00:06:23.17 And you see expression of both lateral regions where the gene will be stably expressed
00:06:27.11 and ventral regions where it's going to get shut off.
00:06:29.21 At this early stage, the only nascent transcripts contain hybridization with the 5' intronic probe.
00:06:37.20 The polymerase has not yet entered the 3' intron.
00:06:40.16 After a ten minute delay, the nuclei stain yellow
00:06:44.28 and that's because now the polymerases have entered the 3' intron
00:06:48.13 and we get hybridization with both the green 5' intronic probe and the red 3' intronic probe.
00:06:55.14 And this delay of ten minutes, going from green to yellow,
00:06:58.04 is consistent with the rate of polymerase elongation of 1kb a minute
00:07:03.05 as I mentioned before and was shown previously in John Lis's lab.
00:07:07.07 OK, so now Sog is stably on. The Snail repressor kicks in.
00:07:12.17 And you see that Sog is getting gradually repressed in the ventral mesoderm over time.
00:07:19.05 And it takes a long time to switch off Sog expression; about 20 minutes between
00:07:25.06 the onset of the Snail repressor and the elimination of Sog expression in the ventral mesoderm.
00:07:32.14 And you see an intermediate stage of the red flash right here
00:07:36.23 and that's the elongating polymerases completing the transcription of the 3' intron
00:07:45.04 that had been released prior to the onset of Snail repression.
00:07:48.17 so this red flash is very good evidence that Snail works
00:07:52.05 at the level of initiation not elongation.
00:07:55.13 Elongating polymerases complete transcription even after the Snail repressor has kicked in.
00:08:01.09 But this mechanism of repression, blocking the release of polymerase from the promoter,
00:08:07.03 blocking initiation, means that there's going ot be a repression lag.
00:08:10.22 And that lag is commensurate with the size of the gene.
00:08:14.15 Sog is 20 kilobases in length and it takes 20 minutes to fully repress Sog expression,
00:08:20.07 to completely eliminate its expression in the ventral mesoderm.
00:08:23.23 And that's due to the elongating polymerases completing transcription
00:08:27.25 even after the onset of Snail repression.
00:08:30.26 We estimate that during the repression lag of 20 minutes
00:08:34.24 as many as 50 to 100 unwanted Sog mRNAs could be made in the mesoderm
00:08:41.24 where they're not needed.
00:08:44.08 Moreover, those ectopic Sog mRNAs may cause developmental problems.
00:08:49.15 And so if you want to get rid of transcripts that are produced after the onset of repression,
00:08:55.09 that arise due to this repression lag,
00:08:57.24 you might have to rely on post-transcriptional mechanisms such as microRNAs.
00:09:02.21 And Sog is indeed regulated by a specifically expressed microRNA
00:09:08.03 possibly getting rid of, cleaning out, those transcripts made after the onset of Snail repression
00:09:15.06 due to this repression lag.
00:09:17.06 Large genes, really large genes, have a more severe problem with repression lag.
00:09:22.28 There are Hox genes in the Drosophila genome, like Antennapedia and Ubx,
00:09:26.24 which are on the order of 100 kilobases in length.
00:09:29.14 It literally takes an hour and a half to repress these genes.
00:09:33.12 Both Antennapedia and Ubx are subject to extensive regulation by microRNAs
00:09:38.11 possibly due to this repression lag issue.
00:09:41.22 So now, once Snail represses Sog in the ventral mesoderm,
00:09:47.25 paused polymerase is once again established at the Sog promoter.
00:09:52.06 Although, Sog is never going to get turned on in the ventral mesoderm.
00:09:56.17 It's turned off in the ventral mesoderm by 3 hours after fertilization.
00:09:59.17 No more Sog transcription in the mesoderm.
00:10:01.28 And it will never come back on in the mesoderm throughout the lifecycle of the fly.
00:10:06.21 Never again.
00:10:07.13 Over the next several hours of development, the paused polymerase
00:10:12.02 at the promoter regions of Sog in the mesoderm, is lost.
00:10:17.20 And so you go from a gene that's repressed but poised
00:10:22.10 it's poised by virtue of containing the paused polymerase
00:10:24.27 to repressed and silent which lacks a paused polymerase.
00:10:29.18 So this is a more stable form of gene silencing.
00:10:32.25 I want to discuss this transition between repression achieved by Snail
00:10:37.29 but you still have the paused polymerase
00:10:40.14 to the silencing of this gene in the mesoderm
00:10:43.09 because you've now removed the paused polymerase
00:10:46.00 and the gene will never again come back on in the mesoderm.
00:10:50.04 This transition from repressed but poised or paused to silenced
00:10:57.01 is mediated by a special group of repressors called the Polycomb repressors.
00:11:02.19 And an example is shown on the next slide.
00:11:04.14 This shows a Polycomb mutant called Esc or Extra Sex Combs.
00:11:11.28 And what you're looking at is the cuticle
00:11:15.01 or the secretion of the ventral skin in late staged embryos.
00:11:20.20 In Polycomb mutants such as Esc, all the body segments
00:11:26.13 acquire the phenotype of the posterior most eighth abdominal segment or A8.
00:11:32.17 So in Polycomb mutants the anterior abdomen, the thorax and even some of the head segments
00:11:39.01 are transformed into A8.
00:11:42.01 That's what you see in the Esc mutant.
00:11:44.15 This transformation is due in part to the misexpression of Hox gene called Abdominal B.
00:11:52.09 Abdominal B is normally transcribed at the highest levels in the developing A8 segment
00:11:59.13 but in Polycomb mutants, Abdominal B is transcribed at high levels throughout the embryo.
00:12:05.12 And that transforms all the body segments to an A8 identity.
00:12:09.28 Now, I mentioned in the very beginning an anterior Hox gene called Antennapedia
00:12:16.28 and there are others like Ultrabithorax and Sex Combs Reduced.
00:12:23.02 Antennapedia is transcribed normally in the middle thorax in T2.
00:12:29.11 But in the Esc mutants, AbdB is transcribed everywhere
00:12:34.09 and AbdB is a repressor and it switches Antennapedia off.
00:12:38.07 And that leads to this transformation of the thorax into extra A8 posterior segment identity.
00:12:46.08 So a postdoc in the lab, Vivek Chopra, did whole genome Pol2 ChIP binding assays
00:12:54.07 to look genome-wide at the distribution of RNA polymerase
00:12:57.11 in advanced staged wild type embryos and Esc mutants.
00:13:02.11 And he found a significant increase in the amount of paused polymerase
00:13:08.16 in Esc mutants at a couple of thousand different genes.
00:13:12.10 So an example is shown here.
00:13:14.08 This is the Antennapedia gene itself.
00:13:17.01 Antennapedia has several promoters. This is proximal most promoter.
00:13:21.09 In wild type embryos, you see a little bit of polymerase binding at the promoter,
00:13:26.13 a little bit of paused polymerase.
00:13:27.16 But in the Esc mutant, you see a significant increase
00:13:30.29 in the amount of paused polymerase at the promoter.
00:13:34.00 And this was surprising to us because, as I said,
00:13:36.28 in Esc mutants AdbB is transcribed everywhere and Antennapedia is shut off.
00:13:41.21 So you see more polymerase binding in the Esc mutant
00:13:46.07 at the Antennapedia promoter even though the gene is off in these mutants.
00:13:50.29 And this trend of increased polymerase binding, increased paused polymerase,
00:13:56.19 in Esc mutants, as I said, is seen for couple of thousand genes.
00:14:00.08 A few more examples are shown in the next slide.
00:14:03.00 So here are two patterning genes; Poxmeso, Twin of Eyegone.
00:14:08.09 In advanced staged wild type embryos these genes contain little or no paused polymerase
00:14:14.17 but in Esc mutants you start seeing significant paused polymerase at the promoters.
00:14:19.17 So, this is the trend: In Polycomb mutants, there is an increase in paused polymerase.
00:14:26.08 And our model for what's going on
00:14:30.00 (we have some additional evidence for this which I'm not going to show right now)
00:14:33.22 is summarized in this last model slide.
00:14:37.07 So, here's a gene that's repressed but poised because is contains paused polymerase.
00:14:44.27 Again, think about the Sog gene in the mesoderm of a 3 hour embryo
00:14:49.19 after Snail has finally shut it down after the repression lag.
00:14:53.29 The gene is off but you have the paused polymerase.
00:14:56.22 The Sog gene will never come back on in the mesoderm
00:15:00.06 and so eventually the Polycomb repressors take off the paused polymerase at the promoter
00:15:06.08 and replace it with a positioned nucleosome.
00:15:09.16 So what we believe the Polycomb repressors are doing
00:15:12.15 is to convert repressed genes that are poised into repressed genes that are silenced
00:15:18.07 that now have a nucleosome at the promoter instead of paused polymerase.
00:15:23.14 And that's a more permanent form of repression.
00:15:27.00 It's harder to turn this gene back on than this type of gene.
00:15:30.12 So then if we take away Polycomb, as seen in the Esc mutants,
00:15:34.00 of course there is an increase in the paused polymerase
00:15:36.07 because we haven't achieved this reaction of replacing
00:15:39.19 the paused polymerase at repressed genes with the positioned nucleosome.
00:15:43.21 So that's our model.
00:15:45.11 So I'm going to just end by acknowledging the associates in the lab who did the work.
00:15:52.26 The first part on shadow enhancers and robustness in response to changes in temperature
00:16:00.14 that was done by a graduate student, Mike Perry.
00:16:04.11 The studies on paused polymerase
00:16:07.19 promoting rapid and synchronous patterns of gene activation
00:16:11.26 and paused polymerase working with shadow enhancers
00:16:14.14 to give very high rates of synthesis such as the rapid synthesis of Snail transcription
00:16:19.15 this was the work of a graduate student named Alistair Boettiger.
00:16:22.18 The visualization of repression and the repression lag I discussed
00:16:28.21 regarding the inhibition of Sog expression in the ventral mesoderm by Snail
00:16:34.14 that was the work of a graduate student named Jacques Bothma.
00:16:37.22 And finally, this last part that the Polycomb mutants such as Esc
00:16:43.06 replace the paused polymerase at repressed genes with a positioned nucleosome
00:16:48.05 to convert them to a silent state
00:16:51.04 that was the work of a postdoc named Vivek Chopra.
00:16:54.06 The studies on paused polymerase were launched in collaborative studies
00:16:58.29 done initially with Julia Zeitlinger and Rick Young at MIT.
00:17:03.20 Julia was a postdoc at that time
00:17:06.03 and she now has her own lab at the Stowers Institute at Kansas City.
00:17:09.20 And more recently, we've collaborated with a postdoc named Leighton Core
00:17:14.24 in John Lis's laboratory at Cornell
00:17:16.28 on the paused polymerase and that fact that the bound polymerase
00:17:22.10 is in an active state based on whole genome nuclear run-on assays.
00:17:26.16 And with that I will thank all of you out there in the ether for your time.
00:17:32.17

Videos in this Talk

Part 1: Transcriptional Precision: Enhancers
Audience:
- Student
- Researcher
- Educators of H. School / Intro Undergrad
- Educators of Adv. Undergrad / Grad
Part 2: Transcriptional Precision: Shadow Enhancers
Audience:
- Researcher
- Educators of Adv. Undergrad / Grad
Part 3: Transcriptional Precision: Paused Polymerase II
Audience:
- Researcher
- Educators of Adv. Undergrad / Grad
Part 4: Transcriptional Precision: Repressors
Audience:
- Researcher
- Educators of Adv. Undergrad / Grad

Speaker: Michael Levine

Total Duration: 1:05:38

Recorded: January 2012

All Talks in Development and Stem Cells

Talk Overview

Levine discusses the important role of precisely regulating gene expression during animal development. The Drosophila embryo provides a model system because during early embryogenesis a syncytium with about 6000 synchronously dividing nuclei is formed before membranes are laid down to separate the nuclei and areas of regulated gene expression are created. These regions of specific gene expression ultimately determine the body plan of the organism.

In Part 1 of his lecture, Levine emphasizes the importance of enhancers in regulating localized gene expression. Enhancers are elements located up or downstream from the transcription start site of a gene and they bind activators and repressors of gene expression. By integrating the effect of activators and repressors, enhancers produce sharp on/off boundaries of gene expression.

In Parts 2 and 3, Levine describes mechanisms used in the embryos to maintain the precision and robustness of gene expression. One mechanism is “shadow enhancers”. These are secondary enhancers found at a significant distance from the transcription start site that act in conjunction with primary enhancers to ensure the sharp patterns and reproducibility of gene expression. A second mechanism of ensuring precise activation of gene expression is “Paused Polymerase”. Active RNA polymerase II binds to the transcription start site, makes a very short nascent transcript and then pauses. When activators bind to the promoter, pol II is ready to go and is rapidly activated across many nuclei. When used together, shadow enhancers and paused pol II provide rapid, uniform, robust and reproducible gene activation.

In the fourth part of his talk, Levine focuses on the role of repressors in regulating gene expression. He shows data from his lab that explain the mechanism of action of the Snail and Polycomb repressors.

Speaker Bio

Michael Levine

Mike Levine is a Professor in the Department of Molecular and Cellular Biology and the Center for Integrative Genomics at the University of California at Berkeley. His current position completes a circle that began when he was an undergraduate student at UCBerkeley. Levine went on to receive his PhD from Yale University and he did… Continue Reading