Jeremy Logue

1
Presented by, Jeremy Logue
2
Big Question

• How does DNA sequence contribute to nucleosome
  positioning?
• And can we predict these positions by analysing
  DNA sequence?
• Segal et al. attempt to predict the positions
  of all the nucleosomes in a
• genome, based soley on DNA sequence.
• Gene activity could then be dictated by a
  nucleosome code (i.e. masking
• and exposing gene promoters).
• Two decades ago, Satchwell et al. demonstrated
  significant periodicities
• of dinucleotides and that DNA bending or
  flexibility helps in determining
• nucleosome position.

3
The Nucleosome
4
Probabilistic Nucleosome-DNA Interaction Model
Sequences aligned and reverse complements about
their centers, a dinucleotide distribution
was generated at each position. Conditional
probabilities for dinucleotides generated
according to a hidden Markov model. Thermodynamic
model accounts for steric hinderance
between nucleosomes (weighted conditional probabil
ities).
5
Dinucleotide Frequencies
6
Periodic AA/TT/TA Dinucleotides
10 bp periodic repeats of AA/TT/TA that
oscillate in phase. GC repeats oscillate out of
phase with AA/TT/TA repeats. As expected,
sequence motifs that recur periodically at
helical repeat known to facillitate DNA
bending. Confirmed by in vitro
selection experiments and by alignments
of randomly selected DNA.
7
Validation of Nucleosome-DNA Interaction Model
8
Key Dinucleotides Inferred from Alignments
Backbone inward
Backbone out
Intergenic and coding regions in yeast genome
contain many more high affinity DNA sequences
than expected by chance. Scores seperated by 10
bp are strongly correlated.
9
Predicted vs Experimentally Identified Nucleosome
Positions
GAL1-10 and CHA1 locus 54 of of predicted
stable nucleosomes within 35 bp of literature
reported positions. Predictions match
stereo- typed chromatin organization at Pol
II promoters.
Orange ovals lit reported, Black trace
probability of nucleosome starting at indicated
base pair, Blue ovals high probability, Light
blue trace average occupancy, Red and blue bars
protein-coding regions, Green ovals conserved
and bound DNA-binding sites
10
Experimentally Measured Nucleosome Occupancy
GAL1-10 and PHO5 promoters 60 of predicted
high occupancy sites confirmed in vivo. In 10
out of 11 cases, predicted regions had higher
occupancy than sites 73 bp (one-half the length
of a nucleosome) away. 50 of in vivo of
nucleosome organization can be explained
by sequence. Rest are unstable.
11
In Vitro Selected Nucleosomes
Nucleosomes Form Arrays
Individual nucleosomes are organized into higher
order arrays. Significant correlations over six
adjacent nucleosomes. Repeat length of 177 bp.
In vitro selection by salt dialysis. Yeast
genomic DNA. Intergenic regions are enriched.
12
Occupancy Across Different Chromosomal Regions
Highest predicted occupancy over centromeres
(encodes enhanced stability). Unstable
nucleosomes over highly expressed genes. Model
does not account for depleted genes, like
ribosomal proteins.
13
Functional Genes are Depleted of Nucleosomes
17 factors have significantly lower occupancy at
functional sites compared to non-functional sites.
1 factor has significantly higher
nucleosome occupancy at non-functional sites
compared with functional sites.
14
TATA Elements are Placed Outside of Stable
Nucleosomes
TATA elements lie just outside stably occupied
nucleosomes. Positions conserved among all
fungal species. May indicate that eukaryotic
genomes direct the transcriptional machinery
to functional sites by encoding
unstable nucleosomes over these elements.
15
Conclusions

• Nucleosome organization is encoded 50 by
  genome sequence.
• And this is conserved across species.
• Genomes encode the positioning and stability
  of nucleosomes in regions
• that are critical for gene regulation and for
  other specific chromosome
• functions.
• Confirms work of Satchwell et al. where
  significant periodicities of
• dinucleotides that favor DNA bending or
  flexibility where observed
• and helps in determining nucleosome position.
• May help explain how a transcription factor
  picks out relevant binding sites.
• Approach still has many limitations, new
  models should account for favorable
• nucleosome-nucleosome interactions and steric
  hinderance constraints
• implied by the three-dimensional nucleosome
  structure.
• Model does not account for competition between
  DNA binding proteins and
• nucleosomes.

16
(No Transcript)