1. The Evolution of a Heterochromatic Domain in Drosophila:
Investigating the Strange
Dot Chromosome
Sarah C R Elgin
January 2012

2. Goal: to understand the organization and functioning of the dot
A collaborative investigation involving:
- former members of the Elgin Lab: Lee Silver, Carl Wu, TC James, Joel Eissenberg, Lori Wallrath, Fang Lin Sun, Karmella Haynes
- current members of the Elgin Lab: Nicole Riddle,
Tingting Gu, Chris Shaffer, Wilson Leung
- modENCODE: Gary Karpen, Mitzi Kuroda, Vincenzo Pirrotta,
Peter Park, and their colleagues
- Faculty and students of the Genomics Education Partnership
Goal: to understand the organization and functioning of the dot
chromosome in Drosophila, an unusual heterochromatic domain.
Funding: HHMI Professors Program
NIH General Medical Sciences, National Human Genome Research Institute

3. Minimum Haploid DNA Content - the C Value Paradox
Genomics Education Partnership
4/23/2017
Minimum Haploid DNA Content - the C Value Paradox
Instructor: Professor Sarah CR Elgin
Slide: Gabriella Farkas
Sources: Gene Regulation for Higher Cells: A Theory. R Britten & E Davidson, Science, 1969 Jul 25;165(891):
Notes: The genomes of eukaryotes have a surprisingly large amount of DNA. It has been suggested that this is related to the “complexity” of the organism- but complexity is hard to define! (Note the lack of a Y axis in this representation, published in Science!)
Britten and Davidson, 1969 Science 165:349
Copyright © 2011, Washington University in St. Louis

4. Genomics Education Partnership
4/23/2017
Larger genomes reflect high levels of repeats - retroviral and DNA transposon remnants (TEs)
Source: CD Allis, T Jenuwein, D Reinberg, Overview and Concepts, in “Epigenetics” (2007)ed Allis, Jenuwein, Reinberg, Caparros, Cold Spring Harbor Laboratory Press, NY.
Notes: Simple eukaryotes such as yeast often have a small genome in which the protein-coding genes are the major fraction. However, genomes of higher organisms are primarily made up of repetitious DNA and non-coding sequences, probably derived from repeats. Thus the increase in genome size appears at least in part to be the consequence of retroviral and transposon invasion of the genome, as well as local duplications etc.
Allis et al: Epigenetics 2007
Copyright © 2011, Washington University in St. Louis

5. Considerations for Genome Sequencing
Genomics Education Partnership
4/23/2017
Considerations for Genome Sequencing
Satellite DNA, a sequence of tandem repeats, is very difficult to sequence, as there are few markers to help order subclones; hence centromeric regions of the chromosomes are usually left unsequenced.
Other repetitious DNA, derived from transposable elements, also causes difficulties; because one finds nearly identical sequences located in different regions of the genome, mistakes can be made in assembling sequence data. High quality discrepancies can identify these.
Much of the repetitous DNA is packaged into heterochromatin, which maintains these regions in a compact and transcriptionally silent form.
However, in many higher organisms, protein-coding genes are found embedded in repetitious DNA. Check out your favorite human gene on the UCSC Browser by taking off RepeatMasker!
Instructor: Professor Sarah CR Elgin
Source: Summary by SCR Elgin
SCR Elgin
Copyright © 2011, Washington University in St. Louis

6. Conserved noncoding - regulatory? 3.5%
Eukaryotic genomes are very large – and most of that DNA is non-coding!
Human Genome 3 Gb
~2 m/cell !
Coding exons
1.5%
TEs –
retroviruses,
DNA transposons
Conserved noncoding - regulatory? 3.5%
Slide: Xiaohui Xie, MIT (2007).
Notes: Is all of this repetitious DNA garbage – would we eliminate it from our genome if we could? Or is it junk? Something that might prove useful in the future? Repetitious sequences are used in chromosome structures such as the centromere, and there is evidence that at least some regulatory sequences (perhaps many) derived from repetitious sequences. But we must insure that transposable elements remain quiescent – not transposing, which would result in mutations. The DNA must be packaged to fit into the nucleus, and that same packaging into chromatin may be the key to silencing.
Key Questions:
Is it junk or garbage?
How is DNA packaged into a nucleus?
How is silencing maintained – while
allowing appropriate transcription ?

7. What determines phenotypes? It’s not just your DNA….
Environment (diet)
(grey bars = folate)
Phenotype
Epigenetics ?
Genotype
Slide: Nicole Riddle, Washington University in St Louis.
Source: RA Waterland, RL Jirtle (2003) Transposable elements: targets for early nutritional effects on epigenetic gene regulation. Mol Cell Biol 23: 5293 – 300.
Notes: Environment can have a big impact on phenotype. Example: phenylketonurea - metabolic, genetic disorder leads to severe developmental phenotypes, retardation etc, but if the environment, ie diet is modified, no symptoms are observed. Data is shown here for the agouti mouse – the yellow coat color is due to an active transposable element (TE) driving a transcription factor in an incorrect tissue; if you feed a methyl donor (ie folate) to the mouse and you can shift the phenotype back to wildtype (brown).
Development
(Waterland and Jirtle 2003)

8. Chromatin structure = epigenetics !
What sets and maintains tissue-specific gene expression patterns? Differences are heritable through mitosis, but independent of DNA sequence.
DNA modification (mC)
Chromatin structure
Nuclear localization
It’s all about silencing!
How is chromatin assembled?
When, where and how does
gene silencing occur?
Incorrect silencing can lead
To genetic disability, as seen
In Fragile X syndrome
Zoghbi and Beaudet 2007
Slide: Illustrations from Nicole Riddle
Source: Fragile X Foundation; Zoghbi and Beaudet, 2007.
Notes: The basic question of interest is how is chromatin assembled and modified to maintain appropriate levels of gene expression. In particular, heterochromatin formation biases to gene silencing. And of course it must be done correctly. FRX is caused by a trinucleotide repeat expansion in the 5’ regulatory region of FMR1, that then causes hypermethylation and silencing of the gene. Kids with FRX have big ears and high foreheads and elongated facial features, and have some level of mental disability.
Fragile X Foundation 8

9. Chromatin formation: First step - packaging in a nucleosome array
Second - differential packaging into heterochromatin & euchromatin
DNA
Chromatin
Histone protein core
Source: Left, Lodish et al, Molecular Cell Biology, 4th edition; right, Felsenfeld et al, (2003) Nature 421:448.
Notes: Left, a generic illustration of a eukaryotic cell. Note the clumps of denser material in the nucleus, referred to as heterochromatin; these contrast with the less dense nuclear material, the euchromatin. Right: a condensed metaphase chromosome being gradually unraveled to reveal a chromatin fiber. DNA does not exist as a free molecule in the nucleus. It is packaged into chromatin which is defined as DNA plus its associated proteins. The basic subunit of chromatin is the nucleosome, which is 146 bp of DNA wrapped around a histone octamer core. Note that while we have good data on nucleosome structure, the illustrations of higher order packaging, while consistent with current data, are not well established.
Chromosome
(metaphase)
Lodish et.al., Molecular Cell Biology, 4th Edition
Felsenfeld et al. Nature 2003, 421: 448

10. Electron Micrograph of Chromatin Fibers (rat thymus nucleus)
Genomics Education Partnership
Electron Micrograph of Chromatin Fibers (rat thymus nucleus)
4/23/2017
Instructor: Professor Sarah CR Elgin
Slide: Gabriella Farkas
Sources: based on Figure 1. from Olins AL,Carlson RD, Olins DE.. J Cell Biol Mar;64(3):
Visualization of chromatin substructure: upsilon bodies.
Notes: Prior to 1970, most biologists thought that the DNA in the eukaryotic nucleus might exist in two forms- relatively accessible (naked), if transcriptionally active, but inaccessible and coated with histones if inactive. This model of a chromatin fiber with DNA on the inside, histones on the outside, was turned inside-out by a series of experiments carried out in the early 1970’s. Some of the first evidence leading to the new model was electron micrographs such as that shown above, showing chromatin fibers released from a nucleus, with the appearance of beads on a string. We now know that the DNA is initially packaged by association with the core histones to form nucleosomes, with the DNA wrapped around the outside of the histone “bead.” This creates a fiber of ca. 10 nm diameter, referred to as a “string of beads,” or more properly, “string wrapped around beads.”
0.1 mm
Olins et. al., 1975 J. Cell Biol, 64:528
Copyright © 2011, Washington University in St. Louis

11. Genomics Education Partnership
4/23/2017
“A eukaryotic chromosome made out of self-assembling 70A units, which could perhaps be made to crystallize, would necessitate rewriting our basic textbooks on cytology and genetics! I have never read such a naïve paper purporting to be of such fundamental significance. Definitely it should not be published anywhere!”
Instructor: Professor Sarah CR Elgin
Sources: K D van Holde, “Chromatin”, 1989, ISBN , Springer-Verlag, New York.
Notes: Such pictures were obtained both by Olins & Olins (Oak Ridge Natl. Lab.) and by Chris Woodcock (U. Mass. Amherst). However, publication was slowed by reviews such as that above. The reviewer correctly recognized that the pictures suggest a paradigm shift in our thinking about the chromatin fiber, and is concerned that the electron micrographs may be showing an artifact. However, on-going experiments by others showed that the subunits (nucleosomes) could be isolated and characterized, and indeed these subunits have been crystallized – although it took many more years effort to do so.
Anonymous review of paper submitted by C.F.L. Woodcock, 1973, showing EM pictures of nucleosome arrays.
Quoted in “Chromatin” by K.D. van Holde, 1989
Copyright © 2011, Washington University in St. Louis

12. The Structure of the Nucleosome Core
Genomics Education Partnership
4/23/2017
The Structure of the Nucleosome Core
Resolution: 2.8 Å
Half of the nucleosome structure is shown
One turn of the DNA helix is visible (73 bp)
View is down the superhelix axis
Protein - DNA contact: white hooks
Instructor: Professor Sarah CR Elgin
Slide: Gabriella Farkas
Sources: Figure 1 from: Chromatin structure: The nucleosome core all wrapped up. Rhodes, D., Nature 389, (18 September 1997); after Luger et al, 1997, Nature: 389: (18 Sept 1997)
Notes: The illustration shows how DNA is bound and organized by the histone core. After 7 yrs effort, Tim Richmond and his colleagues crystallized the intact nucleosome core with DNA, giving a high resolution structure (2.8 A). To get good crystals he had to use an inverted repeat DNA (symmetrical around the dyad axis) and histones prepared in E. coli, to avoid any post-translational modifications (see below). Note the numerous DNA-histone contacts (white hooks) – this is a very stable structure, maintaining dsDNA, blocking transcription.
Rhodes, 1997 Nature 389:231, after
Luger et. al., 1997 Nature 389:251
Copyright © 2011, Washington University in St. Louis

13. DNA packaging domains Euchromatin Heterochromatin Less condensed
Chromosome arms
Unique sequences;
gene rich
Replicated throughout S
Recombination during meiosis
Heterochromatin
Highly condensed
Centromeres and telomeres
Repetitious sequences;
gene poor
Replicated in late S
No meiotic recombination
Slide: Gabriella Farkas.
Notes: While all DNA is packaged into nucleosome arrays, the nucleosomes vary in the post-translational modifications of the histones and in the associated proteins. So while the distinction between euchromatin and heterochromatin was first recognized by the cytology (less dense vs. more dense regions in the nucleus when staining the DNA), and expanded to the collection of properties given on the slide, we are now able to define heterochromatin in biochemical terms. Specific residues in the histones’ amino-terminal tails are enzymatically modified, most often by acetylation, methylation, and phosphorylation. These modifications influence chromatin packaging and the level of gene activity at different regions of the genome.
Transcriptional activators
Hyper-acetylated histone tail
Heterochromatin Protein 1 complex
Hypo-acetylated histone tail; methylated H3/K9

14. 1 2 Heterochromatin formation – silencing counts!
How is heterochromatin organized and packaged to promote silencing? 1 2
The fourth chromosome appears heterochromatic
but has ~80 genes:
- do these genes have unusual characteristics?
- how has the chromosome evolved?
-- how do these genes function?

15. Fruit Flies! Short life cycle, easily maintained: good genetic tools
Polytene chromosomes: excellent cytology
Biochemical approaches
Simple genome, good
reference sequence
PEV – reporter for gene silencing, heterochromatin formation
Metazoan useful for behavioral, developmental and human disease research
Mary Lou Pardue, MIT
Slide: Polytene chromosome showing in situ hybridization from Mary Lou Pardue, MIT.
Notes: The fruit fly is a great model organism for epigenetic studies: It’s short life span and simple genome (4 chromosomes) facilitates genetic studies; phenotypes easy to score; biochemical approaches are possible; the salivary gland polytene chromosomes provide a pretty high resolution visualization of the chromosomes, allowing studies of the distribution of DNA sequences (in situ hybridization) and proteins (immunofluorescent staining). We can use a reporter gene (white) to monitor the chromatin environment; when this gene is in its normal euchromatic environment, it is fully expressed giving a red eye, but when moved to heterochromatin (by rearrangement or transposition) it is silenced in some of the cells in which it should be active, resulting in variegation. Many of the deleterious mutations that cause health problems in humans can be modeled in the fruit fly, including Fragile X. (Flies are relevant!)
euchromatin heterochromatin
expressed silenced

16. Using a white transgene to sample chromatin environments
Genomics Education Partnership
4/23/2017
Using a white transgene to sample chromatin environments
inject
transposon
carrying
white gene
P[white+]
white67c23
Instructor: Professor Sarah CR Elgin
Slide: Gabriella Farkas
Sources: Elgin lab
Notes: In flies we can use a P transposable element to put fragments of DNA back into the genome. A P element carrying the white gene can restore eye color; if we mobilize the P element, it will transpose at random. In most cases a red-eyed fly is recovered, indicating full expression of the gene, but in 1% of the cases we recovered flies with a variegating eye phenotype, suggesting insertion into heterochromatin.
mobilize P element
by crossing to stock
with transposase
insertion into
euchromatin
(99%)
insertion into
heterochromatin
( 1%)
Elgin Lab
Copyright © 2011, Washington University in St. Louis

17. Transposition of a P element reporter allows sampling of euchromatic and heterochromatin domainsX 2L 3L 2R 3R 4
Silenced 1%
Source: L Wallrath & SCR Elgin (1995) Genes Dev 9: 1263 – 1277.
Notes: The P element construct used is diagrammed above. The hsp70-driven white gene gives us the eye phenotypic marker; the hsp26 gene marked with plant DNA has been used for the chromatin structure studies that follow. In situ hybridization showed that the variegating P element inserts all map to the pericentric heterochromatin, telomeres, Y, or fourth chromosome, all previously known heterochromatic domains (not to scale).
Active
99%
And the Y chromosome
Wallrath and Elgin, 1995

18. Assessing chromatin structure- same gene, different environments Analysis based on nuclease digestion of chromatin
Slide: Gabriella Farkas
Source: Based on experiments in FL Sun, M Cuaycong, SCR Elgin (2001) Mol Cell Biol 21: 2867 – 2879, and D Cryderman, H Tang, C Bell, DS Gilmour, LL Wallrath (1999) Nucleic Acids Res 27:
Notes: Now we can ask what happens to a gene that is normally in euchromatin when it is placed in a heterochromatic environment. It is being silenced inappropriately – is this due to a change in chromatin packaging? We can carry out chromatin mapping studies using nucleases to examine the chromatin structure of the native gene. Digesting with micrococcal nuclease will reveal the nucleosome pattern, while digesting with DNase I will show where DNase hypersensitive sites – nucleosome free regions – are present. We find that when a gene is moved from a euchromatic to a heterochromatic domain, the nucleosome array becomes much more regular, with the loss of DH sites. The absence of DH sites most likely contributes to the observed silencing.
The euchromatic hsp26 transgene: - DH sites: accessibility at the TSS, upstream regulatory region irregular nucleosome array
The heterochromatic hsp26 transgene:
- loss of DH sites
- regular nucleosome array

19. Looking for heterochromatic proteins by immunofluorescent staining of the polytene chromosomes: discovery of HP1a
James & Elgin,1986; James et al 1989
Source: TC James & SCR Elgin (1986) Mol Cell Biol 6: 3862 – 3872; and TC James et al (1989) Eur J Cell Biol 50: 170 – 180.
Notes: To look for proteins preferentially associated with heterochromatin, we prepared monoclonal antibodies against proteins that bind tightly in the nucleus, and screened by using the antibodies for immunofluorescent staining of the polytene chromosomes. Heterochromatin Protein I (HP1) is preferentially associated with the chromocenter, small fourth chromosome, telomeres, and a few sites in the chromosome arms. (Note that in polytenization, the pericentric heterochromatin is under-replicated, and all of the euchromatic chromosome arms fuse in a common heterochromatic chromocenter, C above.) C C
HP1
Phase

20. Heterochromatin-associated gene silencing is dependent on HP1
Mutations in
Source: Eissenberg et al (1990) Proc Natl Acad Sci USA 87: 9923 – 9927.
Notes: This experiment provides key genetic evidence that HP1 plays a role in the mechanisms of silencing; loss of HP1 (either a truncation mutant or a mutation in the chromo domain) has this dominant phenotype, a loss of silencing in a reporter showing a variegating phenotype. The original mutation was recovered by T Grigliatti in a screen for mutations with this phenotype – suppression of variegation. Eissenberg et al then sequenced the DNA from these flies, proving that the mutation is in a gene coding for HP1a. The homozygous mutation is lethal, showing that HP1 is an essential protein.
gene for HP1a
Mutations recovered by T Grigliatti as suppressors of PEV.
Dosage dependent response.
Eissenberg et al, 1990, PNAS 87: 9923

21. HP1 interacts with both the modified histone H3K9me2/3 and the modifying enzyme
Chromo
Shadow
SU(VAR)3-9
Sources: Tschiersch et al (1994) EMBO J 13: 3822 – 3831; Lachner et al (2001) Nature 410: 116 – 120; Bannister et al (2001) Nature 410: 120 – 124.
Notes: How might heterochromatin be assembled? Work in mammalian cells showed that the chromo domain binds H3 iff it is methylated at K9 (Jenuwein, Kouzarides). The shadow domain generates HP1 dimers, forming a platform that binds several other chromosomal proteins. Among these is SUV3-9; Jenuwein showed that the human homolg has an enzymatic activity that methylates histone tails at lysine 9. Thus HP1 can both recognize a histone modification (H3K9 methylation), and promote that modification by interacting with the enzyme. This suggests a mechanism of recruit/methyl/recruit that could be used both in maintenance and propagation of heterochromatin. Note that because of its dimerization through the CSD, HP1 could work as a bifunctional crosslinker.
Histone 3
methyl-Lys9
H3 K9 methyl
transferase
[(SU(VAR)3-9 identified in screen by Reuter;
H3 interaction first shown from work in mammals – Jenuwein, Kouzarides;
demonstrated in flies by Imhof.]

22. Model for spreading of heterochromatin
Slide: Gabriella Farkas, SCR Elgin lab.
Notes: Animation of spreading of heterochromatin: HP1 recognizes H3K9me3 and binds; it then recruits SU(VAR)3-9, which can modify the next nucleosome accordingly.

23. Establishing silencing: a multi-step process
Loss of euchromatin marks
Gain of heterochromatin marks
Source: Elgin, SCR & Reuter, G (2007) in “Epigenetics,” ed Allis, Jenuwein, Reinberg, Caparros; Cold Spring Harbor Laboratory Press, New York.
Notes: Of course it is not that simple! Over 100 Su(var) mutations have been reported – over 100 proteins needed to generate / maintain silent chromatin! Some of these are needed to remove the “active marks” – for example, dLSD1, the enzyme that demethylates H3K4. And some are required to add the “silencing marks”, as well as being a structural component of the heterochromatin. Thus a cascade of events is required to shift a locus from a euchromatic, activatable status to a heterochromatic, silenced status. Remember – these processes are reversible – but seem to be fairly stable changes during differentiation, as a specific cell type acquires a limited set of active genes.
wm4 reporter (screens by Reuter, Grigliatti, others)

24. 2 Heterochromatin formation on the dot chromosome…
The fourth chromosome appears heterochromatic
but has ~80 genes:
- do these genes have unusual characteristics?
- how has the chromosome evolved?
- how do these genes function? 2

25. The Drosophila melanogaster fourth chromosome exhibits an amalgam of euchromatic and heterochromatic properties (HP1a association)
Heterochromatic properties:
late replication, lack of recombination
high repeat density (30%)
antibody staining of HP1, H3K9me2/3
But…
the fourth has ~ 80 genes in distal 1.2 Mb
these genes are transcriptionally active!
James & Elgin,1986; James et al 1989
Slide: Wilson Leung
Source: James & Elgin (1986); James et al (1989); see previous slide.
Notes: The fourth chromsome of D melanogaster (sometimes called the “dot” chromosome from its metaphase appearance, but more properly known as the Muller F element) appears entirely heterochromatic by classical criteria, but the distal 1.2 Mb has 80 genes, a gene density similar to that found in the euchromatic arms. C C
HP1
Phase

26. Most hsp70-white reporters exhibit variegation
on insertion into the fourth chromosome
2-M1021
39C-12
2-M390
39C-52
Source: Sun et al (2004) Mol Cell Biol 24: 8210 – 8220, and Riddle et al (2007) Genetics 178:1177 – 1191.
Notes: A map of the fourth chromosome distal portion (centromere to left) showing the presence of TEs and the genes (bars below). Each triangle indicates a line carrying a single P element reporter inserted at that site, with the eye phenotype shown, red or variegating, indicating a euchromatin or heterochromatin environment. Most fourth chromosome genes lie in a domain that appears to be heterochromatic, at least in the eye primordia. Note that the heat shock promoter used here is active/activatable in almost all tissues.
Sun et al 2004; Riddle et al 2007

27. Our GEP Research Goal: Use comparative genomics to learn more
about heterochromatic domains, analyzing the
dot chromosomes and a control euchromatic
region of Drosophila genomes
Status
Reference
Completed
Source: Evolutionary tree from FlyBase.
Slide: Wilson Leung
Notes: The research accomplished by GEP students. What can we learn from the comparative genomics of the dot chromosome? Initially we have looked at gene and chromosome evolution in several distant species (ca. 40 mil yr); this year we are also starting work on D. ananassae, where the dot chromosome has undergone an unusual expansion (10-fold larger – all repeats!).
Annotation
Sequence Improvement
New Project
FlyBase:

28. Genomics Education Partnership (GEP)
Source: GEP website,
Slide: Wilson Leung
Partners are generally PUI schools; faculty join by
attending one-week workshop at WU. Shared
work organized on GEP website.

29. The D.melanogaster & D. virilis dot chromosomes
are 25% - 30% repetitious DNA
(typical – but up to 80% in D. ananassae)
Source: Leung et al (2010) Genetics 185: 1519 – 1534.
Slide: Wilson Leung
Notes: Our first published analysis is a comparison with D. virilis. We find that both D melanogaster and D virilis dots have similar levels of repetitious DNA (27-30%), intermediate between that found in the euchromatic arms (~10%) and that found in the pericentric heterochromatin (~60-70%). Note that our ability to find repeats depends on having good libraries of repeat sequences for a given species. The commonly available libraries are based on D melanogaster; since other species have been invaded by different retroviruses and transposons, their repeats may be underrepresented unless a special effort is made to find these repeats. Looking at the types of repeats, we see a higher relative concentration of DNA transposons (red) on the dot chromosomes and a higher concentration of retroviral sequences (green) in the heterochromatic regions. Note that there are two D virilis dot chromosome sequences available for study, one assembled as part of the 12 Genomes Project and a different strain, finished and analyzed at higher accuracy by our GEP/Bio 4342 students.
D mel D vir D mel D vir D vir D mel D vir D vir D vir
Leung et al 2010

30. Dot chromosome genes: introns are larger, exons show less codon bias
Euchromatic
Leung et al Genetics 185:
Codon Bias
Dot
Heterochromatic
Source: Leung et al (2010) Genetics 185: 1519 – 1534.
Slide: Wilson Leung
Notes: We find that the dot chromosome genes on average are larger than genes in the euchromatic arms (8 kb vs 2 kb) and have more introns (8 vs 3). Most of the size difference can be explained by more and larger intron sizes due primarily to repetitious DNA. (The graph shows the percentage of introns that are a given size (x axis) or smaller.) We also find that codon bias is reduced on the dot chromosomes, presumably because the lack of crossing over. (Codon bias refers to the use of codons – in theory an organism could use all 64 codons, but in practice most organisms use fewer.)
Intron Size
D. melanogaster Dot
D. melanogaster Euch.
D. melanogaster Het.
D. virilis Dot
D. virilis Euch.

31. Initial analysis of Drosophila virilis dot chromosome fosmids
Genomics Education Partnership
4/23/2017
Initial analysis of Drosophila virilis dot chromosome fosmids
Sources: Slide made from Figure 4. in publication: Comparison of dot chromosome sequences from D. melanogaster and D. virilis reveals an enrichment of DNA transposon sequences in heterochromatic domains
Slawson et. al., Genome Biol. 2006; 7(2): R15.
Notes: While most of the same genes are found on the dot chromosomes in D melanogaster and D virilis, there have been extensive rearrangements, presumably by inversions. For example, pan and Caps are on opposite ends of the dot in D melanogaster, but are found on the same fosmid (#30, dark blue) in D. virilis. In contrast, the gene cluster of fosmid #103 (turquoise) has maintained the same order and orientation – perfect synteny.
Almost all of the same genes are present (27/28), but rearrangements within the chromosome are common!
Slawson et. al., 2006 Genome Biology, 7(2):R15.
Copyright © 2011, Washington University in St. Louis

32. Comparison of gene order and orientation
D. melanogaster
Source: Leung et al (2010) Genetics 185: 1519 – 1534.
Slide: Wilson Leung
Notes: Looking at the chromosome as a whole (from the most proximal to the most distal gene), considering only those genes that are present on the dot in both species, and treating this as an abstract problem, we find that a minimum of 33 inversions are required to convert one chromosome to the other. These range from inversion of large fragments to inversion of single genes. (Note this is an abstract solution to the problem – we do not know how these changes actually occurred over evolutionary time.)
D. virilis
72 genes on both the D. virilis and D. melanogaster dots.
A minimum of 33 inversions are needed to convert order and orientation!
Leung et al 2010

33. “Wanderer” genes move between the dot chromosome and a
euchromatic site in the long arms; they adopt the
properties (gene size, codon bias) of their local environment
CG5262
CG9935
CG5367
rho-5
CG1732
dot
Source: Leung et al (2010) Genetics 185: 1519 – 1534.
Slide: Wilson Leung
Notes: Careful annotation has convinced us that there are eight genes that are in a euchromatic domain (in the other chromosome arms) in one species, and in a heterochromatic domain (on the fourth) in the other species. (Some other candidates were discarded as being transposable elements, not true genes.) These genes adopt the properties of the surrounding domain – i.e., they are larger and show less codon bias when they are on the fourth chromosome, indicating that these properties are a consequence of being in that domain, not a property of the gene per se. We hope to identify and analyze more such genes in our comparative studies. Because we have restriction maps that confirm the fosmid assemblies, we know that these genes genuinely moved – but how is a mystery!
CG11077
CG4038
CG11076
dot: D. virilis
dot: D. melanogaster
Leung et al 2010

34. Some things to look for while annotating dot chromosome genes….
Is there a homologous gene in D. melanogaster?
Is it on the dot chromosome?
Are all of the isoforms found in D melanogaster present?
How many exons?
Any unusual splice sites?
What is the order and orientation of genes compared to D. melanogaster?
Are there repetitious elements nearby?
Check out your gene on FlyBase – what is the pattern of expression in
D. melanogaster? Has a function been described?
Many dot chromosome genes are expressed at a high level - how can
this occur in a heterochromatic domain?
Slide: SCR Elgin.

35. Chromatin Immuno-precipitation - ChIP (cells or nuclei)
1. Crosslink proteins to DNA
2. Isolate chromatin and sonicate
qPCR
3. Incubate with antibody
ChIP-chip*
Slide: Nicole Riddle
Notes: More recently we have looked at genome-wide patterns of histone modification using ChIP (chromatin immunoprecipitation) experiments. The chromatin is cross-linked (generally using formaldehyde), and then sheared into small fragments of a few hundred base pairs. The fragments are incubated with an antibody specific for a particular histone modification (for example, H3K9me3) or chromosomal protein; all of the chromatin fragments that bind the antibody are then collected. The cross-linking is reversed, and the selected population of DNA identified, either by qPCR (to look at a particular locus), or by using an oligonucleotide array to look at the whole unique genome, or by re-sequencing the DNA, and the identified population of DNA fragments is mapped back onto the sequenced genome. The resulting map shows the distribution of the particular histone modification or chromosomal protein. The modENCODE project ( has mapped numerous histone modifications and chromosomal proteins across the genome of D. melanogaster.
4. Isolate AB/chromatin complexes
ChIP-seq
5. Isolate DNA from complexes

36. Mapping chromatin marks by ChIP-chip: Chromosome arm 3L shows a distinct shift between heterochromatin and euchromatin
Centromere
S2 cells
Euchromatin
Heterochromatin
HP1a
Su(var)3-9
H3K9me2
H3K9me3
genes
Enrichment (log intensity ratio values)
Slide: Nicole Riddle
Source: Data discussed in Riddle et al (2011) Genome Res 21: 147 – 63
Notes: The ChIP data is presented here in a “wiggle” graph indicating enrichment; significant enrichment is indicated by the solid pink bar. This can be related to the gene map on the bottom line (genes in green). Here we are looking at the region at the base of chromosome 3L, where we see a transition from the euchromatic domain (left) to the pericentric heterochromatin (right). One sees a consistent shift to enrichment for the histone modifications (H3K9me2/3) and chromosomal proteins (HP1a and Su(var)3-9, an H3K9 histone methyl transferase) that we have found associated with heterochromatin formation and silencing.
Euchromatin / heterochromatin transition point from Flybase
Pink boxes show significant enrichment (0.1% false discovery rate)

37. Chromosome 4 is largely heterochromatic, but shows distinct peaks of H3K4me2/3, indicating transcription start sites
Centromere
Telomere
HP1a
H3K9me2
Enrichment (log intensity ratio values)
H3K9me3
Slide: Nicole Riddle
Source: Data discussed in Riddle et al (2011) Genome Res 21: 147 – 63
Notes: What about the genes on the fourth chromosome? ChIP mapping (as well as our earlier analysis) shows that that region is largely heterochromatic. [Pink bars above show a significant enrichment in association of that mark, mapped against the chromosome sequence (bottom line; genes shown in green).] However, we also see evidence of a mark associated with gene activation – H3K4me2. H3K4me2 is associated with active genes, generally enriched at the 5’ end, and a set of significant peaks is seen here.
H3K4me2
genes

38. A model of 9 chromatin states, based on clustering of histone modification marks, identifies large-scale genomic domains
Slide: Nicole Riddle
Source: Figure 1a and 1b from Kharchenko et al (2011) Nature 471:
Notes: We have repeated the ChIP experiment with 18 histone modification marks, and find that we can sort the chromatin fragments into 9 types. State 1 is associated with Transcription Start Sites, 2-4 with enhancers and regulatory regions; 5 is specific to the male X chromosome, required for dosage compensation; 6 is the Polycomb state, maintaining silencing at certain developmentally regulated genes; 7 is constitutive heterochromatin, and 8 may be facultative heterochromatin; and 9 has no distinguishing features. Note green labels are used for histone marks are associated with activation, red for silencing marks. Mapping the chromatin states onto the Drosophila melanogaster genome is shown above. Note the heterochromatic regions around the centromeres (right) and entire fourth (state 7); heterochromatin-like state 8 is distributed within the chromosome arms in a cell-type specific pattern. State 6 is almost entirely on the X chromosome.
 TSS (red; H3K4me3 rich)
 Polycomb (grey; H3K27me3 rich)
Heterochromatin (dark blue; H3K9me3)

39. An expanded view of the fourth chromosome reveals
TSS (state 1, red) and Pc (state 6, grey) domains
interspersed within heterochromatin (states 7 & 8, blue).
Pericentric heterochromatin
10 Mb
chr3L
chr4
500 kb
Red
Variegating
Slide: Wilson Leung
Source: mapped according to the 9 state model of Kharchenko et al (2011).
Notes: We can now map the different chromatin states across an expanded fourth chromosome. We find that while much of the chromosome is in states 7 and 8 (blue and light blue, associated with H3K9me2/3), there are domains in state 1 (red, associated with H3K4me2/3, transcription start sites) and state 6 (grey, associated with H3K27me3, regulated by the Polycomb system). Our variegating reporters lie in domains that are in state 1 or state 7/8 in Bg3 cells; note that the reporters showing full expression lie in domains that are found in state 6 in at least one cell type, indicating that this state is permissive for gene expression.
BG3 cells, chromatin states: 1 2 3 4 5 6 7 8 9

40. expression patterns, including expression in the adult
Might fourth chromosome genes function early, and be silenced later? No, fourth chromosome genes show a variety of
expression patterns, including expression in the adult
Slide: Susan Celniker, Lawrence Berkeley National Lab, CA,
Source: modENCODE data produced by the Celniker group.
Notes: Fourth chromosome genes are expressed at similar levels as euchromatic genes – some at consistently high levels, others at low levels, and some in a developmentally regulated pattern. There is no obvious bias in the type of gene on the fourth.
S Celniker, modENCODE

41. Most 4th chromosome genes lie in heterochromatic space (blue),
but active genes achieve state 1 (red) at the TSS
Source: modENCODE data assembled in a local UCSC-style browser by Wilson Leung.
Notes: Most of the fourth chromosome genes lie in heterochromatic domains (chromatin state 7), but active genes exhibit state 1 (typical of transcription start sites) at the 5’ end of the gene, with DNase I hypersensitive sites as observed for euchromatic genes. Note the flanking repetitious element (1360).
1360

42. Active fourth chromosome genes show depletion of HP1a and H3K9me3 at the TSS, but enrichment across the body of the gene
chromosome 4
Average enrichment
Slide: Nicole Riddle
Source: Riddle et al (2011) Genome Res 21: 147 – 63. Figure 6B and supplemental figure.
Notes: Looking both at Bg3 cells and S2 cells (D melanogaster cell lines), and grouping the data around the TSS of all active genes, one sees a depletion of H3K9me2/3 (silencing marks) and an increase in H3K4me3 (an activation mark) just downstream of the binding site of the RNA polymerase. HP1a appears to show less depletion, and in fact might have a peak at -800.
TSS-relative position
RNA pol II H3K4me3
HP1 H3K9me2 H3K9me3
Transcription levels are similar to euchromatic genes!

43. Future: try to determine what feature drives 4th chromosome
The fourth chromosome: a repeat rich domain with “heterochromatic” genes
Source: modENCODE data assembled in a local UCSC-style browser by Wilson Leung.
Notes: We conclude that fourth chromosome genes have a special mechanism to block heterochromatin spreading across the promoters (TSS) of the active genes, or to displace such marks as part of the activation process. Note that a typical TSS chromatin has been assembled here, even though a 1360 repetitious element, a likely target for silencing, is nearby It will be interesting to see whether these genes have a special TSS motif not found in genes normally resident in euchromatin.
1360
Future: try to determine what feature drives 4th chromosome
gene expression that is absent from euchromatic genes (hsp70).

44. Eight new genomes Expanded dot chromosomes?
Source: Thomas Kaufman, Indiana University.
Notes: Eight additional Drosophila species have been sequenced as part of the modENCODE project. These species were selected to be an appropriate evolutionary distance from D. melanogaster to facilitate motif hunting. Interestingly, some of these species appear to have an expanded dot chromosome, and some a rather small dot chromosome. The GEP plans to annotate the dot chromosomes from 2-3 of these species in order to search for TSS motifs specific to fourth chromosome genes.
Expanded dot
chromosomes?

45. 1 2 Heterochromatin formation on the dot chromosome…
Heterochromatin formation changes chromatin at
the nucleosome level, eliminating HS sites at the TSS of euchromatic genes; silencing is dependent on HP1a 1 2
Fourth chromosome genes are larger, have more introns, and less codon bias than euchromatic genes
Fourth chromosome genes show high levels of HP1a and H3K9 methylation over the body of the gene,
but maintain access at the TSS.
Next steps: what makes fourth chromosome
genes robust? Lets look for fourth chromosome motifs!
Slide: SCR Elgin

46. Question Slides

47. A role for POF? A different configuration for HP1a?
Are there chromosomal proteins unique to the fourth chromosome? Yes – POF (Painting of Fourth) is uniquely associated with active genes on the fourth chromosome
Slide: Nicole Riddle, Washington University in St Louis
Green = HP1a Red = POF
A role for POF? A different configuration for HP1a?
(See J Larsson, PLoS Genet. 11:e209 for more on POF)

48. Drosophila melanogaster: 1/3 heterochromatin
Drosophila melanogaster: 1/3 heterochromatin. Pericentric heterochromatin is under-replicated in polytene chromosomes; the arms fuse in the chromocenter
Slide: Elgin lab, Washington University in St Louis.
Source: TS Painter (1934) J Hered. 25:
Notes: The D. melanogaster karyotype is shown; about 1/3 is heterochromatic (black domains), including all of the Y chromosome and all of the 4th chromosome. Polytene chromosomes form in the large cells of the larvae, which grow without division; endoreduplication of the euchromatic arms occurs (10 times in the salivary gland), but no mitosis – the chromatin strands remain juxtaposed in good order. The pericentric heterochromatin is not duplicated, and the chromosome arms fuse in a common chromocenter, giving the 5-arm pattern.
Drawing of polytene chromosomes modified from TS Painter, 1934, J. Hered 25:

49. HP1 from mammals can rescue mutations in flies and yeast!
HP1 sequence from Drosophila, mouse, human and mealy bug identifies chromo domain & chromo shadow domain
Clark and Elgin, 1992 Nucleic Acids Res. 20:6067
Source: R Clark and SCR Elgin (1992) Nucleic Acids Res. 20: Rescue experiments by Ma et al (2001) Chromosoma 109: 536 – 44.
Notes: Heterochromatin protein 1 is highly conserved – it is found in many fungi and in most animals checked to date. There are two conserved domains, the N terminal chromo domain (variations also found in several other chromosomal proteins) and the C-terminal chromo shadow domain. In the above comparison, gold residues are identical and orange are conserved. The chromo domain of HP1a binds to H3K9me2/3, while the chromo shadow domain dimerizes, and forms a hydrophobic binding surface, allowing interaction with several other proteins.
HP1 from mammals can rescue mutations in flies and yeast!

50. HP1 is a trans-acting modifier of PEV
Position Effect Variegation in Drosophila: an assay for heterochromatic packaging
white
Wild Type
Inversion i
Slide: Elgin lab
Source: See review by Grewal and Elgin (2002) Curr Opin Genet Develop 12: 178 – 87; demonstration of HP1a copy number impact is from J Eissenberg, St Louis University Medical School.
Notes: PEV is the variegated expression pattern that results when a gene is moved by rearrangement or insertion in or near heterochromatin,as demonstrated by this white inversion giving rise to a mottle white eye. The gene is being silenced in some of the cells in which it is normally active. Given a variegating phenotype, one can screen for second site mutations that suppress or enhance that phenotype. Such studies have lead to identification of proteins involved in chromatin structure formation.
HP1 is a trans-acting modifier of PEV
Su(var)
(1 copy HP1)
E(var)
(3 copies HP1)
Eissenberg et al

51. The heterochromatic hsp26 transgene:
The heterochromatic hsp26 transgene: - loss of accessibility at the TSS when in heterochromatin - reversed in an HP1 mutant background
Cryderman et al 1999 (Wallrath & Gilmour labs)
Source: Cryderman et al (1999) Nucleic Acids Res 27:
Notes: Digesting the chromatin with Xba!, we see loss of accessibility at the regulatory region (loss of the HS site; see loss of band at the proximal XbaI site); this is partially restored by crossing in a mutation in HP1 [Su(var)2-5].

52. Chromosome 4 short arm is a unique heterochromatic domain
Source: Haynes et al (2007) Genetics 175: 1539 – 42.
Notes: The distal arm of the fourth chromosome is a unique heterochromatic domain. Reporters in that domain, but not in the pericentric heterochromatin of the fourth, are sensitive to dosage of the distal arm, showing more silencing when only one copy is present, and less silencing when three copies are present. This titration effect implies that there are one or more proteins needed for silencing that are unique to this domain.
Haynes et al. 2007

53. Translocation away from the chromocenter results in loss of silencing – spatial organization plays a role
Source: Cryderman et al (1999) EMBO J 18: 3724 – 35.
Notes: A fly line with a reporter near the end of the fourth chromosome was subject to x-ray mutagenesis, and mutants selected for segregation of the reporter on a chromosome other than the fourth. This recovered translocations. When the reporter (with ~ I Mb of fourth chromosome DNA) is moved to the tip of the second chromosome, it looses silencing; however, when it is moved to the third chromosome but at the base – close to the chromocenter – silencing is maintained. This indicates that the 3-D spatial organization of the nucleus is important for silencing, perhaps due to local concentration of HP1a.

54. Define chromatin states by K-means clustering (using enrichment values for 1 kb chromatin fragments)
H3K4me3 
H3K9me3 
Assign each data point to closest mean.
The centroid of each cluster is the new mean.
Repeat the last two steps.
Slide: Nicole Riddle, Washington University in St Louis
Notes: The average value of enrichment for a histone modification mark for each 1 Kb fragment of chromatin is obtained, allowing that fragment to be plotted accordingly; in this example two modifications are used. A reiterative process is used to optimize the clustering of the data points. This approach, along with more sophisticated techniques, was used to derive the 9-state chromatin model described by Kharchenko et al (2011).
Select k means as starting points.

55. Chromatin states are highly interspersed- a folded view of the chromosome (Bg3 cells)
Polycomb
Source: Figure 2 from Kharchenko et al (2011) Nature 471:
Notes: By folding the chromosome using a Hilbert space-filling curve, we can get a view of the domains that occur in this chromosome arm.
Pericentric heterochro-matin

56. Chromatin states reveal cell type specific patterns (note facultative heterochromatin – light blue)
Source: Supplemental figure from Kharchenko et al (2011) Nature 471:
Notes: Comparing the chromosome arm 3L from S2 cells (an embryonic cell) and Bg3 cells (a differentiated, neuronal cell) we see increase state 8 (light blue; potentially facultative heterochromatin) in the latter case.
BG3 S2
Chromosome 3L

57. A folded view of the chromosome reveals TSS and Pc states within chromosome 4 (Bg3 cells)
Source: Supplemental figure from Kharchenko et al (2011) Nature 471:
Notes: It is clear that the fourth chromosome has a much higher density of heterochromatin (states 7 and 8, blue and light blue) than does 3L; but note the interspersed state 1 (red; active TSSs) and state 6 (grey; polycomb domains).
4th chromosome
Pericentric heterochromatin

58. Chromosome 4 shows a distinct subset of Polycomb sites in a cell-type specific pattern
H3K9me3
(S2 cells)
(Bg3 cells)
Polycomb
genes
Slide: Nicole Riddle, Washington University in St Louis
Source: modEncode data,
Notes: The snapshot shows nearly all of chr4, with the regions differing in Pc enrichment in the different cell types clearly seen. The Pc-enriched regions show earlier replication time in S2 cells. These regions are depleted for H3K9me2/3 when Pc is present, but high H3K9me2/3 in seen when Pc is absent.

59. A subset of 4th genes is associated with Polycomb;
these domains are permissive for reporter expression
(red eyed fly).
Source: UCSC-style browser assembled by Wilson Leung, Washington University in St Louis; modENCODE data
Notes: We find that the lines showing full expression of the reporter on the fourth chromosome have the reporter inserted into a domain packaged with Polycomb in at least one cell type. Note that these domains allow assembly of 5’ DHS sites.

60. 3 Heterochromatin formation on the dot chromosome…
How are specific domains targeted for
heterochomatin formation?
Repetitious elements such as 1360 are targeted by a mechanism involving the RNAi system 3

61. Analysis of chromosome 4 identified 1360 as a target.
P[1360, hsp70-w] silencing depends on the
reporter’s position in the genome
hsp70-white
5’P
3’P
FRT
1360
Source: Haynes et al 2006 Curr Biol 16: 2222 – 27.
Notes: Our previous mapping of the fourth chromosome has suggested that the TE remnant 1360 is a target for heterochromatin formation. If this hypothesis is correct, can we construct “instant heterochromatin” by using a reporter that carries a copy of 1360 with it? No! Most insertion sites out in the euchromatic arms show a red-eye phenotype. (A few show reduced expression – solid orange – but this expression is not impacted by Su(var) mutations and must be through a different mechanism.) The one line that shows strong PEV is located close to the centromere, in a region with a high repeat density.
Haynes et al 2006 Curr Biol 16: 2222

62. FLP-mediated removal of 1360 results in loss of silencing
hsp70-white
5’P
3’P
FRT
1360
hsp70-white
5’P
3’P
FRT
+ FLP recombinase
1360 excised
FLP14
FLP16
FLP4
FLP5
Source: Haynes et al 2006 Curr Biol 16: 2222 – 27.
Notes: Is the 1360 element impacting silencing? Note that we have flanked that element with FRT sites; so if we start with the variegating line, and cross to a fly carrying FLP recombinase, about half the time the 1360 element will be eliminated by local recombination. If we look at the progeny, about half look like their parents (left) and about half show more pigmentation (right) (quantitative eye pigment levels are shown below for five progeny lines of each type). PCR was used to confirm that the lines with more pigment has lost the 1360 element, while the others had not. This demonstrates that the presence of 1360 contributes to heterochromatin formation and more effective silencing.
Haynes et al 2006 Curr Biol 16: 2222

63. Mutations in RNAi components impact PEV
Notes: The RNAi system in flies is complex, with 3 major pathways as sketched above. Mutations in the proteins marked in purple have been shown to impact silencing (as scored by PEV) and/or genome stability (scored by looking for circles of repetitious DNA).
spn-E (hls)
Dmp68/Lip
dFMR1
Results from Birchler, Elgin, Schedl, others;
note also esiRNA pathway, Siomi, Hannon & others

64. Mutations in RNAi components piwi, aubergine and homeless suppress PEV
y w; 118E-10/+ +
piwi1 +
aubQC42 +
hls∆125 +
y w; 118E-10/+
y w; 39C-12/+
y w; 39C-12/+
Source: Pal-Bhadra et al (2004) Science 303: 669 – 72.
Notes: Mutations in piwi, aubergine, and homeless result in a loss of silencing, as shown by increased pigmentation in the eyes of reporter lines. Piwi and aubergine are paz/piwi domain proteins, known to bind small RNAs. Homeless is a helicase.
piwi
aub
hls
B Leibovitch in Pal-Bhadra et al, 2004

65. A tentative model for heterochromatin targeting – HP1a – PIWI interaction (piRNA)
Cytoplasm
PIWI
transposon
HP1
Aub
piRNA
Ago3
Heterochromatin
Nucleus
Source: S Wang & SCR Elgin (2011) Proc Natl Acad Sci USA, in press.
Notes: PIWI is the only argonaute family member known to be a nuclear protein. PIWI can bind piRNAs, and can bind to HP1a. Loss of PIWI in the female germline leads to increased expression of a subset of TE’s, and a loss of HP1a associated with these sequences. This indicates that PIWI plays a role in directing heterochromatin formation to the correct sites in the genome
PIWI (binds piRNA) is an argonaute family member
and nuclear protein which interacts with HP1a