Transcript Aspergillus nidulans - Center for Biological Sequence Analysis

Transcription analysis of glucose derepressed Aspergillus nidulans mutant
using Febit Geniom® One system
By Ph.D student Jesper Mogensen
Center for Microbial Biotechnology
&
H. Bjørn Nielsen
Center for Biological Sequence
Analysis

- science of green technology
Why this project?



We would like to be the first to publish results using
”full-genome” oligonucleotide array for Aspergillus
nidulans
It was decided to compare the wild-type A. nidulans
against a glucose de-repressed creA mutant
because of the severe effects observed in this
mutant (change in morphology)
It was decided to use the two most different growth
conditions (repressing and non-repressing) in shakeflask experiments to avoid time consuming
fermentations
CM
B/
H. Bjørn Nielsen, Microarray workshop, August 2004, Thailand
Slide 2
Why is Aspergillus nidulans interesting?

Aspergillus nidulans is interesting because:





More than 55 years of genetic experience
Can grow on complex substrates
It is ”easy” to work with in the laboratory
Is used as model organism
Sequence available for A. nidulans (Broad Institute)
CM
B/
H. Bjørn Nielsen, Microarray workshop, August 2004, Thailand
Slide 3
CreA: The history




CreA = carbon (catabolite) repressor
Discovered in a study about nitrogen metabolite
repression in A. nidulans in 1973 (Arst et al. 1973)
Metabolically favourable carbon sources such as Dglucose are preferred, and when these are available,
enzymes and permeases that allow the utilization of
alternative carbon sources are not produced
The corresponding repressor in S. cerevisiae is
called Mig1 (which is more intensively studied than
CreA)
CM
B/
H. Bjørn Nielsen, Microarray workshop, August 2004, Thailand
Slide 4
The creA mutant, characteristics



CM
B/
With no CreA protein present in the cells, the genes for less
favourable carbon sources will not be repressed
On a mixture of glucose and e.g ethanol both carbon sources
will be utilized at the same time
Fails to grow on complete-medium with allyl-alcohol (ethanol
agonist) because it is converted to acrolein by ADH (alcohol
dehydrogenase)
Complete medium
Complete medium + 5 mM allyl-alcohol
H. Bjørn Nielsen, Microarray workshop, August 2004, Thailand
Slide 5
CreA…what is known in Aspergillus?

CreA inhibits transcription of many genes by binding to specific
sequences in the promoter of these genes:
1.
2.
3.
CM
B/
Catabolism of less preferred carbon sources
Gluconeogenic and glyoxalate cycle enzymes
Genes related to secondary metabolism (e.g. penicillin)

Strong repressing carbon sources: glucose, xylose,
sucrose and acetate

Intermediate repressing carbon sources: mannose,
maltose, fructose, mannitol and galactose

Non repressing carbon sources: glycerol, lactose,
arabinose and ethanol

Northern blot analysis of creA mRNA reveals a complex
expression profile dependent on carbon source
H. Bjørn Nielsen, Microarray workshop, August 2004, Thailand
Slide 6
CreA…what is missing?




General overview of affected genes in creA deleted
mutant using transcription analysis
Understanding the interaction with CreB, CreC
and/or other regulating proteins
Understanding the mechanism behind the CreA
induction and repression
Understanding the pleiotropic effects

Pleiotropism: The control by a single gene of several distinct
and seemingly unrelated phenotypic effects (e.g.
morphology)
CM
B/
H. Bjørn Nielsen, Microarray workshop, August 2004, Thailand
Slide 7
Why are microarrays interesting?


Possible to get an overview of (all) metabolic
pathways  easier to direct the gene-manipulation
to the right ”target” in the cell to get e.g. higher yields
Possible to get an overview of the pleitropic effects
that – at first sight - are not easy to predict
CM
B/
H. Bjørn Nielsen, Microarray workshop, August 2004, Thailand
Slide 8
Febit geniom one array system, I



CM
B/
All in one micro-array system
Enables micro-array research on any organism with known
sequence data
Micro-array synthesis and hybridisation is performed inside a
three dimensional micro-channel structure - the DNA
processor®
Febit geniom one
The DNA processor/chip
H. Bjørn Nielsen, Microarray workshop, August 2004, Thailand
Slide 9
Febit geniom one array system, II




CM
B/
6500 ”spots” are available in each of the 8 channels of
the Febit DNA processor
Each ”spot” corresponds to one oligonucleotide
5-8 oligonucleotides are used per gene  800-1300
genes can be analysed in one channel!
In total app. 10,000 genes can be analysed on one chip
or the chip can be divided into eight different arrays which
can be analysed at the same time
Blank micro-channel structure
Processed micro-channel structure
H. Bjørn Nielsen, Microarray workshop, August 2004, Thailand
Slide 10
Micro-array manufacturing
Digital projector system
UV light
Phot olabile
prot ect ion group
T
+
Microst ruct ure
1. Immobilized surface

2. Deprotection
T
T
T
3. Add amidite
4. Elongation
In this experiment the oligo length was set to 24-26
bp
CM
B/
H. Bjørn Nielsen, Microarray workshop, August 2004, Thailand
Slide 11
Design of oligonucleotides


Desired oligo length between 24 and 26 bp.
From 4 to 8 probes pr. gene
CM
B/
H. Bjørn Nielsen, Microarray workshop, August 2004, Thailand
Slide 12
Febit experiment overview, I





CM
B/
Sequence data for oligo design was obtained from
MIT/Whitehead (now Broad institute) Aspergillus
nidulans database
3278 annotated genes were selected (from 9540
putative genes). Genes with pfam-entry were
preferred!
Wildtype A. nidulans (A187) will be compared with
creA mutant
Growth conditions: minimal-medium + glucose or
ethanol as carbon source
2 strains, 2 growth conditions, 5 biological replicates
 20 individual shake flasks
H. Bjørn Nielsen, Microarray workshop, August 2004, Thailand
Slide 13
Febit experiment overview, II
5 x glucose
5 x ethanol
5 x glucose
5 x ethanol
CM
B/
H. Bjørn Nielsen, Microarray workshop, August 2004, Thailand
Slide 14
Pictures of wild type and creA mutant
CM
B/
A187, glucose
A187, ethanol
creA, glucose
creA, ethanol
H. Bjørn Nielsen, Microarray workshop, August 2004, Thailand
Slide 15
Selection of replicates

The three most similar replicates (out of five) from
each condition were selected based on:



Dry-weight measurements
pH measurements
HPLC measurements (glucose/ethanol, glycerol)
A187, glucose
12,0000
10,0000
8,0000
Glycerol g/L x 50
Glucose g/L
pH
Dry-weight g/L * 5
6,0000
4,0000
2,0000
Dry-weight g/L * 5
pH
0,0000
Glucose g/L
A187, G1
CM
B/
A187, G2
A187, G3
Glycerol g/L x 50
A187, G4
A187, G5
H. Bjørn Nielsen, Microarray workshop, August 2004, Thailand
Slide 16
Experimental Design

Everything was don in batch to capture the systematic noise
A187
CreA
Glucose Ethanol Glucose Ethanol
3x
CM
B/
H. Bjørn Nielsen, Microarray workshop, August 2004, Thailand
Slide 17
RNA purification




Frozen mycelium was crushed using steel balls (2 x 2
mm + 1 x 5 mm) in a 2 mL Eppendorf tube fitted in a high
velocity shaker
Total-RNA was extracted using RNeasy mini-kit from
Qiagen
Quality of RNA was checked on spectrophotometer,
1% agarose gel and Agilent Bioanalyzer
Quantity measured on spectrophotometer
23S
16S
5S + tRNA
CM
B/
H. Bjørn Nielsen, Microarray workshop, August 2004, Thailand
Slide 18
From total-RNA to amplified antisense RNA
Biotin-11-CTP
Biotin-16-UTP
CM
B/
Fragmentation of aRNA
Loading of sample on Febit array
Hybridization
Labeling with streptavidin::fluorescein dye
H. Bjørn Nielsen, Microarray workshop, August 2004, Thailand
Slide 19
Data Analysis - Normalization
CM
B/
Q-spline. Workman et al. Genome Biol. (2002)
H. Bjørn Nielsen, Microarray workshop, August 2004, Thailand
Slide 20
Data analysis - Batch to batch variation
Within batch variation is lower than
the between batch variation

CM
B/
H. Bjørn Nielsen, Microarray workshop, August 2004, Thailand
Slide 21
Data analysis - Blocking

We can capture the batch variation by blocking
Two-way ANOVA
Effect 1
Ethanol
Glucose
Effect 2
A187
CM
B/
CreA
Batch A B C
H. Bjørn Nielsen, Microarray workshop, August 2004, Thailand
Slide 22
Data analysis - Result of the Statistic

We get 3 p-values from a two way ANOVA
A genotype
p-value
A genotype
p-value
A growth
media
p-value
A growth
media
p-value
An interaction
p-value
An interaction
p-value
Two-way ANOVA
Media effect
Ethanol
Glucose
CM
B/
A187
CreA
Genotype
effect
Batch A B C
H. Bjørn Nielsen, Microarray workshop, August 2004, Thailand
Slide 23
Data analysis - Threshold

A threshold of p < 0.01% for the min p-val was used

I.e. if a gene is significant in one of the 3 tests it is used

This resulted in 200 significant genes

With 3278 genes on the chip, we estimates
33 false positives.
CM
B/
H. Bjørn Nielsen, Microarray workshop, August 2004, Thailand
Slide 24
Clustering of genes into 12 clusters
Clustering of genes according to expression level in each
condition (k-means)

CM
B/
H. Bjørn Nielsen, Microarray workshop, August 2004, Thailand
Slide 25
The influence of CreA, I

Theory: CreA is always present at a certain level and represses genes
under both repressing and de-repressing conditions
(11 out of 200 most significant genes)
BUT repression is for some genes less under
de-repressing conditions (cluster C)
(24 out of 200 most significant genes)
CM
B/
H. Bjørn Nielsen, Microarray workshop, August 2004, Thailand
Slide 26
The influence of CreA, II

In this cluster, the CreA doesn’t seem to have any effect! This
cluster of genes seems to be induced by glucose and
repressed by ethanol independently of CreA
(23 genes out of 200 most significant genes)
CM
B/
H. Bjørn Nielsen, Microarray workshop, August 2004, Thailand
Slide 27
The influence of CreA, III
CreA

Repres.
Gene X
Theory: CreA acts as a repressor of a repressor of gene in
cluster L
(33 genes out of 200 most significant genes)
CM
B/
H. Bjørn Nielsen, Microarray workshop, August 2004, Thailand
Slide 28
The influence of CreA, IV
CreA
Gene X
Inducer
EtOH

Theory: CreA acts as a repressor of an ethanol activated
inducer of genes in cluster I
(35 of the 200 most significant genes)
CM
B/
H. Bjørn Nielsen, Microarray workshop, August 2004, Thailand
Slide 29
Induction or de-repression…or both?
EtOH
CreA

CM
B/
alcR
alcA
In cluster A, B and H (including ADH1) the genes seems to be
both de-repressed and induced at de-repressing conditions!
(38 of the 200 most significant genes)
H. Bjørn Nielsen, Microarray workshop, August 2004, Thailand
Slide 30
Conclusions


With the expression analysis it has been possible to
cluster the 200 most significant genes in 12 different
clusters, where the genes seem to be either directly,
indirectly or not regulated by CreA
With the use of micro-array analysis it is possible to
get an overview of the pleiotropic effects (indirect
regulation by CreA)
CM
B/
H. Bjørn Nielsen, Microarray workshop, August 2004, Thailand
Slide 31
Future plans





Search for CreA binding sites in promoters
Better annotation of most significant genes
Run Northern blot (ugpA, gfaA, pcmA, ungA, creA)
(Gerald Hofmann)
Compare the expression profile with flux analysis
model (Helga David)
Biological interpretation of result
CM
B/
H. Bjørn Nielsen, Microarray workshop, August 2004, Thailand
Slide 32
Acknowledgment





Jesper Mogensen
Professor Jens Nielsen
Gerald Hofmann
Michael Lynge Nielsen
People at CMB and CBS
CM
B/
H. Bjørn Nielsen, Microarray workshop, August 2004, Thailand
Slide 33