Download Slide 1

Evidence That G-Quadruplexes Regulate Transcription in S. Cerevisiae
Steven Hershman, Qijun Chen, Julia Lee, Marina Kozak, Jasmine Smith,
Alex Chavez, Li-San Wang and F. Brad Johnson
Department of Pathology and Laboratory Medicine, University of Pennsylvania, Philadelphia, PA 19104
Model: Rap1p-G4 interactions
Results
Background
Structural biology and biochemistry of G-quadruplexes (G4s)
In silico prediction shows yeast genome has high G4 forming potential
 G-quadruplexes (G4s) are four
stranded DNA structures that form
when guanine residues participate in
Hoogsteen hydrogen bonding (Left)
forming boxes known as G-quartets.
These G-quartets stack to from a
quadruplex (Right).
 G4s are particularly stable under
physiological pH and salt conditions.
 Top: The enrichment ratio of
quadruplexes of different
lengths in ORFs and
promoters. Control
sequences were simulated by
conserving position-wise
A/T/G/C frequencies. G4
sequences show a greater
preference for promoters
compared with open reading
frames, especially at lower,
more trustworthy, G4
lengths.
 Bottom: A sliding window
was used to measure the
position of G4 sequences of
length 50 or less relative to
the translation start site
across all genes. The greatest
peak was found to occur at
about 425 base pairs
upstream of the promoter. In
yellow is an example control
sequence, which has many
fewer G4 forming sequences.
 In humans, sequences with G4 forming potential are overrepresented, particularly in
promoters and nucleosome-free regions (Huppert, 2005).
 Additionally, G4 forming sequences are overrepresented in proto-oncogenes and
underrepresented in tumor suppressor genes (Eddy, 2006).
Role of G-quads in telomeres; its connection with aging
 Telomeres are DNA at the ends of chromosomes.
 Across all organisms, telomeres have guanine-rich sequences. Many of these sequences
have been shown to form G4 in vitro (some in vivo).
 As human cells age telomeres shorten.
 When telomeres shorten, proteins that are associated with the ends of telomeres are
released.
 Rap1p is a transcription regulator that is known to bind to double-stranded DNA
(including telomere DNA) and to G4s.
 Individuals with the premature aging disease Werner syndrome lack a DNA helicase,
WRN, that is preferentially unwinds G4s and helps maintain telomeres.
Yeast as a model organism to study aging
 Yeast naturally maintain telomere length using the enzyme telomerase.
 Yeast tlc1 mutants lack telomerase activity and shorten their telomeres with cell
division. Eventually this telomere shortening causes permanent cell-cycle arrest
(senescence).
 Yeast tcl1 mutants are model for studying human cell senescence.
Rap1p
Fold Enrichment G4
20
18
 Rap1p leaves its telomere binding sites as
telomeres become short at senescence.
ORF
16
Promoter
14
ON
12
 Top: When Rap1p interacts with
quadruplexes in the absence of a double
stranded binding site, it activates
transcription.
10
8
6
4
2
0
25
35
50
75
100
125
150
200
250
500
1000
Maximum G4 Length
 Bottom: If there is a double stranded
binding site present, then Rap1p inhibits
transcription.
Rap1p
OFF
200
Number of G4 within 200bp
Evidence of non-telomeric G4s in the human genome
 To explain the Rap1p/tlc1/G4 overlap
results, we propose the model to the right.
180
Genome
160
Conclusions
Control
140
120
 In the yeast genome, the increased frequency and location preference to promoters of Gquadruplex sequences suggests a role in gene regulation.
 G-quadruplexes may help coordinate transcriptional responses to telomere shortening.
Rap1p might be involved in the mechanism.
100
80
60
40
20
0
-1500
-1250
-1000
-750
-500
-250
0
Promoter
250
500
750
1000
1250
1500
Future Projects
ORF
G4 potential is correlated with senescence
Genes with G4 sequences of length 50 or less in
their promoters are correlated with genes that
are upregulated 2-fold in cells lacking tlc1 that
have reached senescence.
Tlc1
up
G4-DNA
354
33
244
Fisher Exact P=0.00036
 GGG repeats in quadruplex forming sequences can be mutated so that they should no
longer form quadruplexes. We would like to test how this affects regulation of yeast
genes that are bind Rap1, but not in a double stranded DNA-dependent manner.
 If quadruplexes play a role in gene regulation, then by adding drugs that stabilize
quadruplex formation, we should be able to find genes who expression levels change.
We would like to explore this with various quadruplex binding drugs in microarray
experiments.
 In addition to intrastrand quadruplexes, it is also possible that multiple strands can
come together to form a quadruplex. We would like to explore this possibility
computationally by allowing the four GGG repeats to occur on either the sense or
anti-sense strand.
G4 potential is correlated with non-double strand Rap1p DNA targets
Scientific problem
Are there in vivo non-telomeric G4s in yeast?
If so, what role do they play?
Computational Prediction of G4 DNA
Regular Expression (Huppert, 2005)
The most conservative G4 prediction algorithms use a regular expression where repeats
of at least 3 guanine residues are separated by loops of 1-7 other base pairs.
G3+N1-7G3+N1-7G3+N1-7G3+
Sliding Window (Eddy, 2006)
An alternative method for finding G4s
relies on looking at the distance between
the first and fourth GGG repeat. This
allows for more flexible lengths in the loop
regions.
Non-dsDNA
G4-DNA
 Top: A list of genes that associate with Rap1p targets
Rap1 but not through double stranded
DNA binding was developed by
390
55
222
creating a list of genes that associate
with Rap1p through chip-chip assays
and removing any gene whose double
stranded DNA associated to Rap1 in
Fisher Exact P=1.4 x 10-13
vitro. This list of genes was enriched
with G4 potential.
Class
Go Term
P Value
telomerase-independent
 Genes that interact with Rap1 are
telomere maintenance
BP
correlated with genes that are
8.11E-08
downregulated at senescence (Fisher
mitotic recombination
BP
3.00E-06
Exact P=0.00087). G4 cannot explain
DNA helicase activity
MF
9.87E-06
this connection because the tlc1-G4
helicase activity
MF
0.000121
connection was based on genes that
energy derivation by
were upregulated.
oxidation of organic
compounds
0.001313
 Bottom: A GO analysis was completed BP
on the genes in this double overlap.
transcription factor activity
MF
0.001485
Classes that involved metabolism and
DNA recombination
BP
0.001967
transcription were overrepresented.
generation of precursor
This echo similar results found in E.
metabolites and energy
BP
0.00367
Coli by Rawal (2006).
Literature Cited:
 Eddy J and Maizels N (2006). Gene Function Correlates with Potential for G4 DNA
Formation in the Human Genome. Nucleic Acids Research. 34(14):3887-96.
 Huppert JL and Balasubramanian S (2005). Prevalence of quadurpelxes in the human
genome. Nucleic Acids Research. 33, 2908-2916.
 Huber MD, Lee DC, and Maizels N (2002). G4 DNA unwinding by BLM and Sgs1p:
substrate specificity and substrate-specific inhibition. Nucleic Acids Research. 30, 39543961.
 Williamson JR (1994). G-Quartet Structures in Telomeric DNA. Annu. Rev Biophys.
Biomol. Struct. 23, 703-730. OR Williamson, J.R., Raghuraman MK, and Cech TR (1989).
Monovalent Cation-Induced Structure of Telomeric DNA: The G-Quartet Model. Cell.
59, 871-880.
 Rawal P, Kummarasetti VB, Ravindran J, Kumar N, Halder K, Sharma R, Mukerji M, Das
SK, Chowdhury S (2006). Genome-wide prediction of G4 DNA as regulatory motifs: role
in Escherichia coli global regulation. Genome Res., 16, 644-655.
Acknowledgements:
 Biostatistics Core, School of Medicine, U. of Pennsylvania for early advice as to which
statistical analyses to use
 Cara Winter, Wager Lab, Department of Biology, U. of Pennsylvania for help with
microarray analysis
 Supported by the Vagelos Program, Roy and Diana Vagelos Science Challenge Award
and National Institute on Aging Grant 1R01AG021521