Prsentation PowerPoint

1
DNA in chromatin Modelling the influence of the
sequence on the structure and dynamics of
chromatin
Benjamin Audit Laboratoire Joliot-Curie /
Laboratoire de Physique, Ecole Normale Supérieure
de Lyon Benjamin.Audit_at_ens-lyon.fr
Alain Arneodo ENS de Lyon, France Julien
Moukhtar Philippe StJean Cédric Vaillant
Emeline Fontaine Fabien Montel Cendrine
Moskalenko Yves
dAubenton-Carafa CGM, Gif-sur-Yvette, France
Claude Thermes
2
To which extent does the DNA sequence code for
the 3D structure of chromatin?
3
Long-range correlations in genomic DNAA
signature of the nucleosomal structure
4
Sequence induces local structural disorder along
the DNA polymer
frozen structure _at_ T 0
5
Trinucleotidic structural tables based on
experiments
trajectory of the double helix axis
6
Heterogeneous elastic model for DNA
7
Genomic factors to nucleosome positioning
Small scale (a) positioning sequence motifs
periodicity 10 of bending sites (AA/TT,
AAA/TTT) (b) repulsive sequence motifs
Poly(dA-dT) (c) nucleosome free energy of
formation v.s. sequence ? Large
scale sequence-induced statistical principle ?
What is the influence of long-range correlated
genomic disorder ?
8
2D DNA loops

• Nucleosome 2 coplanar loops of DNA

• Inhomogeneous 2D elastic model

(a) Winding constraint
(b) Cyclisation constraint
9
Intrinsic curvature
long-range correlations ? polymer intrinsically
more compact
10
Loop energy profiles

• Energy

• Free energy

Independently of position s

• Weak disorder Gaussian statistics

• Strong disorder, H exponential
  distribution of energetic barriers

long-range correlations ? larger fluctuations of
energy profiles
11
Correlations in loop energy profiles
long-range correlations ? persistent fluctuations
of energy profiles
12
Formation of DNA loops
Free-energy for the formation of a loop of length
l in a semi-flexible chain of length L gtgt l (2D
elastic model)
H0.5 H0.6 H0.8
long-range correlations ? facilitates the
formation of small loops
13
Mobility of DNA loops
Mean First Passage Time ? at distance N during
the diffusion of a small loop (l200) in a
semi-flexible chain of length L gtgt l (2D elastic
model)
long-range correlations ? favour local mobility
of small loops
14
If DNA loops were nucleosomes
Nucleosomes diffuse in a potential defined by the
sequence
The formation and dynamics of the chromatin
fibre is favoured by LRC
15
Genome-Scale Identification of Nucleosome
Positions in S. cerevisiae (Yuan et al., Science
309)
16
Statistical characterization
Large-scale disorder of nucleosome positioning
Long-range correlations between 200 and 1000bp
with H0.8
Small-scale periodic component at 180bp
Weak positioning fat tail (exponential)
statistics
17
Interpretation in term of the DNA loop model
1 nucleosome
In fact many nucleosomes of size l146bp with
steric repulsion? Periodic modulation of energy
profile
Experimental evidence that long-range
correlations influence nucleosomal organisation
18
Spontaneous emergence of rosettes on the
chromatin fibre
Open chromatin at replication origins ?
19
Cell nuclei contain a dense solution of
macromolecules
Typical density 20-40
Cristallisation 55
 Close packing  74
20
where depletion forces can occur (S. Asakura
F. Oosawa, J Chem. Phys. 1954)

• Entropic clustering
• depletion force

21
Depletion force on a flexible tube(Snir
Kamien, Science 2005)
L 5-6lp Compact helix
30lp plectonemes
gtgt 30lp ??? (tore, super-helices, beta-sheet)
22
Chromatin as a tube with fixed defects
23
Possible defect structures at replication origins
24
Formation of rosettes around clustered
replication origins
Understanding how chromatin structure can
contribute to genome-scale functional regulation ?
25
Atomic Force Microscopy
with deflection
Quadrant photodiode
no deflection
LASER
Cantilever
Pointe
sample
z
y
x
Piezo-electric scanner
26
Binding of DNA on mica surface
DNA

• using divalent ions Mg2, Ni2

Mg2
-
-
-
-
-
-
-

-
-
-
-
-
-

-
-

Mica
-

-
- - - - - - - - - - - - - - -

• protocol
• 5 µL of DNA solution ( lt 1 ng/µl nM) containing
  5 mM Mg2 ions and a buffer solution (TRIS 10mM,
  pH 7.4).
• incubation 2 min.
• Rinse with 1 mL ultrapure water and nitrogen blow
  to dry.

27
AFM image of 2200 pb virus DNA fragments
28
Image analysis
29
Image analysis
30
Introducing sequence effects in polymer
elasticity modeling

• semi-flexible polymer, WLC model

persistence length
Elastic modulus

• taking into account sequence induced structural
  disorder in the WLC model

lp Aeff
31
Influence of LRC on persistence length measurement
Uncorrelated structural disorder WLC is still
valid but with an effective elastic modulus
(Trifonov 1988)
Long range correlated structural disorder WLC
no more valid
double integral single integral
No analytical expression numerical calculation
32
Intrinsically straight DNA fragment (200 bp)
Construction of a straight artificial DNA
Mica ADN D1 (7ng/µl) MgCl2 5mM, Tris 10mM
(-EDTA) pH7.4, 2, rin. 1 ml, Argon
M. Vologodskaia and A. Vologodskii, J. Mol. Biol.
(2002)
33
Highly curved DNA fragment (200 bp)
Construction of an artificial DNA with very high
curvature
Mica ADN C1 (0.3ng/µl) MgCl2 10mM, Tris 10mM
(-EDTA) pH7.4, 2, Argon
50 nm
34
Image analysis of 200bp straight and highly
curved DNA fragments
we are able to distinguish the two fragments by
measuring the persistence length
BUT
200 bp is too short to accurately estimate A
need for longer fragments
35
Highly curved DNA fragment inserted in a random
plasmid
DNA Plasmid pBSKII (3161 bp)
ScaI
XhoI
EcoRI
insert 200 bp of the highly curved fragment
cut with XhoI (restriction enzyme)
cut with EcoRI XhoI cut with
ScaI
A B C
36
Blind test with the AFM
A B C
37
Blind test with the AFM
A B C
38
A
s 100 bp
lp(s) (bp)
s (bp)
B
lp(s) (bp)
100 nm
s (bp)
C
lp(s) (bp)
s (bp)
39
Blind test with sequence analysis
window s 100 bp
ltR2gt/2s (bp)
n along DNA sequence (bp)
40
DNA selection and preparation

• DNA with no structural disorder (s 0) 1
  intrinsically straight DNA in the absence of
  thermal fluctuations (length 800 bp).
• DNA with uncorrelated structural disorder (H
  0.5) 1 DNA fragment extracted from
  Hepatitis C RNA virus genome (length 2200
  bp).
• DNA having long range correlated structural
  disorder (0.5 lt H lt 1) 2 DNA fragments
  extracted from the human genome chromosome 21
  and chromosome 8
  (length 2200 bp).

41
AFM images of 800 bp straight DNA (no structural
disorder, s 0) in 2D
42
Experimental local persistence length for a 800
pb straight DNA fragment, N 127
fitted by worm like chain model A2D
Adyn 500 bp
43
Comparison with cryo-EM measurements on straight
DNA fragments
Adyn 500 bp A2D A3D A2D /2 250 bp 85
nm !
s0 0
A3D 82 nm
44
AFM images of 2200 bp virus DNA (uncorrelated
structural disorder H 0.5) in 2D
45
Experimental local persistence length for a 2200
pb virus DNA fragment, N 69
fitted by worm like chain model Aeff_2D
425 bp
Aeff 425 bp
Aeff 445 bp
S (bp)
46
Experimental local persistence length for a 2200
pb virus DNA fragment, N 69
for uncorrelated structural disorder, Trifonov
Aeff_2D 425 bp Adyn 500 bp
s0 0.019
By AFM, we can measure the amplitude of the
sequence induced structural disorder for a 2200
bp DNA fragment (when no LRC)
47
Experimental local persistence length for a 2200
pb virus DNA fragment, N 69
for uncorrelated structural disorder, Trifonov
Aeff_2D 425 bp Adyn 500 bp
s0 0.019
By AFM, we can measure the amplitude of the
sequence induced structural disorder for a 2200
bp DNA fragment (when no LRC)
Aeff_2D 425 bp A3D A2D /2 212 bp 72
nm !
48
AFM images of 2200 bp human DNA from chromosome
21 (LRC, H0.8) in 2D
49
AFM images of 2200 bp virus DNA (uncorrelated
structural disorder H 0.5) in 2D
50
AFM images of 2200 bp human DNA from chromosome
21 (LRC, H0.8) in 2D
51
Experimental local persistence length for a 2200
pb chr. 21 DNA fragment, N 102
chr. 21 experimental results
double integral calculation fitting with worm
like chain model ?
52
Experimental local persistence length for a 2200
pb chr. 21 DNA fragment, N 102
chr. 21 experimental results
double integral calculation fitting with worm
like chain model ?
Aeff 400 bp
53
Experimental local persistence length for a 2200
pb chr. 21 DNA fragment, N 102
chr. 21 experimental results
double integral calculation fitting with worm
like chain model ?
Aeff 400 bp
Aeff 350 bp
54
Experimental local persistence length for a 2200
pb chr. 21 DNA fragment, N 102
chr. 21 experimental results
double integral calculation fitting with worm
like chain model ?
Aeff 400 bp
Aeff 350 bp
Aeff 300 bp
55
Experimental local persistence length for a 2200
pb chr. 21 DNA fragment, N 102
chr. 21 experimental results
56
Simulation of the 2D thermo fluctuations for a
2200 bp DNA fragment with LRC (H 0.83, s
0.007, Adyn 530 bp), N 110
y (bp)
x (bp)
57
Simulation of the 2D thermo fluctuations for a
2200 bp DNA fragment with LRC (H 0.83, s0
0.007, Adyn 530 bp), N 110
numerical results
58
Experimental local persistence length for a 2200
pb chr. 21 DNA fragment, N 102
chr. 21 experimental results
59
AFM images of 2200 bp human DNA from chromosome 8
(LRC, H0.8) in 2D
60
Experimental local persistence length for a 2200
pb chr. 8 DNA fragment, N 92
chr. 8 experimental results
61
Experimental local persistence length for a 2200
pb human DNA fragment, N 92
single integral calculation chr. 21 vs chr. 8

Chr. 21
Chr. 8
62
Influence of LRC on persistence length measurement
Single integral calculation for AFM trajectories
obtained in the same exp. conditions
63
Summary

• AFM experiments
• we are able to detect locally a strong curvature
  sequence effect
• when comparing lp for virus DNA (no LRC) and
  straight DNA (no disorder) we measure the
  amplitude of disorder in the sequence (s0)
• when comparing human DNA (with LRC) and virus
  DNA (no LRC), we show that there is a strong
  effect of the long-range correlations on the lp
• the WLC model cannot fit the data anymore
• the LRC model seems to account for the observed
  decrease in lp when averaging over the sequence
  induced disorder (numerical result) a
  posteriori validation of the model

64
Whats next AFM experiments

• Human DNA genomic DNA (all chromosomes cut in
  pieces of 2000 bp)
• Change the GC content of the human DNA low GC
  effect on s0?
• Change H ?
• Change Adyn change exp. conditions Mg2,
• 
  pre-treating the surface with Ni2
• focus on functional regions of human genome
  gene promoters,
• 
  replication origins,
• measurement in liquid time averaging on a
  single DNA trajectory instead of thermo averaging
  

65
Strategy to measure the sequence dependent
nucleosome dynamics
1000 bp DNA ? LRC
Long range correlated DNA
straight DNA
uncorrelated DNA
gt 1 kb
Hyperdiffusion ?
D gt D?
D ?
66
Structure and dynamics of nucleosomes histone
variants, and chromatin remodeling
DNA sequence effects on the structural and
mechanical properties of the double helix
67
Analysis of the correlations in the DNA curvature
definition of the curvature
r
l
l
n
n
2
2
n
C
n
l
C
n
n
l
68
Julien Moukhtar
Emeline Fontaine
Fabien Montel
Guillaume Lavorel
Philippe Saint-Jean
Benjamin Audit
Cédric Vaillant
Françoise Argoul
Alain Arneodo
Cendrine Moskalenko