Microarray
analysis of
Mycobacterium
avium subspecies paratuberculosis (MAP) intracellular and
extracellular transcriptomes

Contents
Introduction
Experimental Aims
Microarray
construction and validation
cDNA
generation and hybridization to microarrays
Differential
expression analysis results
Conclusions
and further work
Introduction
These experiments form part of
the Department of Surgery, SGHMS research programme investigating the molecular
biology of MAP and its involvement as a causative agent of Crohn's disease in
humans. Contributors to this work include Dr Tim Bull, Professor John
Hermon-Taylor, Dr Joe Sheridan, Dr Jason Hinds *, Dr Phil Butcher *.
These results are
preliminary ONLY, have not as yet been submitted for publication and must not
be published, referenced or copied without permission from Dr. Tim Bull
(tim.bull@sghms.ac.uk).
(* Microarray Facility,
Department of Medical Microbiology, SGHMS)
Back to top
Experimental
Aims
- To
extract mRNA populations representational of the complete transcriptome
from MAP cultured in extracellular and intracellular environments in
sufficient quantity and quality for microarray analysis.
- To
evaluate the MAP sub-microarray.
- To determine MAP specific genes that are
differentially expressed in an intracellular environment.
Back to top
Microarray
construction and validation
- Generation
of PCR products
Primers (21mers) designed with Primer3 [Primer3 Input (primer3.cgi v
0.2c)] as specific for MAP or MTB were used to amplify products from 250ng
MAP total genome DNA per sample extracted from MAP (strain 989) or MTB
(strain H37Rv) respectively. MAP products were amplified using the
following conditions: 1 cycle (96°C: 3min); 40 cycles (94°C : 30sec; 58°C
: 30sec; 72°C : 90sec) ; 1 cycle (72 °C : 5 min). All MAP amplification
products were cleaned once with Qiagen PCR Cleanup kit (Qiagen, UK) and
purity checked (see Figure 1).
MTB products were amplified and purified as described (http://www.sghms.ac.uk/depts/medmicro/bugs/Mtuberculosis/mtuberculosis.htm).
- Microarray
gridding
25 Arrays were printed onto Poly-L-lysine coated slides using the BioRobotics
MicroGrid II split pin technology generating a 4 x 4 meta grid containing
256 spots delivered using a 16 pin tool. To eliminate background bias and
reduce pipetting fluctuations, gene products from MAP genes were spotted
four times onto the same array. The array contained 46 MAP gene products
including GS cassette genes, 16SrDNA, sodA, katG, 65Kd heat
shock Ag, IS900 transposase p43 and ORF’s immediately
adjacent to fourteen IS900 loci. Also included were a random
selection of 100 MTB gene products also spotted singularly or in
duplicate. Each 4 x 4 grid also contained a dilution series of 16SrDNA,
Cy5, Cy3 dCTP fluorescent spot controls.
- Validation
To validate the quality of the gridding and to identify MTB products with
sufficient homology to generate a significant signal, two array slides
were hybridised at 65°C for 16hr with 2.5mg
MAP 989 DNA labelled with Cy3 dCTP (see Figure 2). This showed
that all 46 MAP specific genes and 40% of MTB genes (18 of 45 genes) were
significantly reactive and gave acceptably consistent signals over each of
the spots.
Back to top
Amoebic culture and mRNA extraction
- Growth
of MAP in amoebae
Acanthamoebae polyphaga cultures were grown to 80% confluency in a
750cm3 tissue culture flask at room temperature using Proteose
Peptone Glucose (PPG) medium. MAP cultures grown on Middlebrooks 7H11 agar
supplemented with glycerol and mycobactin J (MAP medium) were harvested
and diluted to 1 x 107 cfu/ml (as estimated by nephelometry).
Cultures were washed once in sterile PBS, passed slowly through a small
gauge needle (25G) in PBS to
separate aggregates and then inoculated into a single flask of A.
polyphaga and incubated at RT for 48 hrs. Samples were then treated
with 100mg/ml Amikacin for 2 hours and the PPG media carefully poured off
and replaced. MAP extracellular control cultures contained PPG only and
were not treated with Amikacin. Cultures were then incubated at RT for
8-10 weeks without changing the PPG media.
- Viability
of MAP in amoebic cultures
MAP cultures were stained for acid-fast bacilli (Auramine/Rhodamine) and
sub-cultured on MAP medium for viability. MAP could not be recovered from
amoebic flasks on MAP medium. However further experiments using long term
cultures of MAP in amoebae (not shown here) suggest that MAP, although now
unable to grow on solid media, are still capable of replication within an
intracellular environment and can survive in amoebic cultures for at least
12 months. Microscopy estimated an MOI of 3-5 bacilli per cell with over
70% of amoebae infected after 8 weeks. Each flask contained approximately
1 x 106 amoebae after 8 weeks incubation.
- mRNA
extraction
Five flasks of MAP infected amoebae and five of MAP in PPG controls were
harvested and combined into test and control pools. Amoebae were then
differentially lysed with 1% Triton-X100 for 20 min and the intact MAP
pelleted and washed in Tris HCl (pH8.0). Pellets were then mixed with
0.6ml TRIZOL and ribolysed @ 6.5 for 45secs (Hybaid Ribolyser). mRNA
populations were cold extracted with chloroform, precipitated with
isopropanol and washed in 70% ethanol. Samples were further treated with
DNAseI for 15mins, purified through an mRNA extraction column (Qiagen) and concentrations estimated
against RNA standards on a 1.5% agarose gel (see Figure 3).
Back to top
cDNA generation and
hybridization to microarrays
- cDNA
10mg
mRNA samples were mixed with 3mg of a Random Primer Set (Invitrogen)
heated to 95°C
for 5 min then snap cooled on ice. cDNA was then generated by adding First
Strand Buffer, DTT, 450U Superscript Reverse Transcriptase (Invitrogen)
and dNTP’s to the following final
concentrations 0.5mM dATP, dTTP, dGTP, 0.2mM dCTP and Cy3 (test sample) or
Cy5 (control sample) (Amersham
Pharmacia) to 0.3mM in a final volume of 50ml
and incubated in the dark @ 25°C for 10min followed by 42°C
for 90min. Cy3 and Cy5 samples were then combined and purified through a
MinElute column (Qiagen), eluted into
10.5ml
H2O then mixed with filtered Stock Hyb Buffer (Final conc 4x
SSC, 0.3% SDS) to a final volume of 16ml.
- Hybridization
Microarrays were pre-hybridised in buffer (3.5x SSC, 0.1% SDS, 10mg/ml
BSA) @ 65°C
for 20min followed by thorough rinsing in dH2O and then
propan-2-ol with subsequent drying by centrifugation @ 1500rpm for 5 min.
16ml
purified Cy3/Cy5 labelled sample was heated to 95°C
for 2min then added directly to the microarray slide. A coverslip was
applied and the microarray hybridised in a hybridization cassette (Telechem International) @ 65°C
in the dark for 18hr. Microarrays were then vigorously washed twice in
0.06x SSC, dried by centrifugation @1500 rpm for 5 mins and immediately
read with a dual laser 428 Scanner (Affymetrix).
Back to top
Differential
expression analysis results
- Analysis
calculations
Imagene V4.2 software (Biodiscovery, Inc) was
used for array imaging, normalization and signal enhancement. Genespring (Silicon
Genetics) software was used for preliminary cluster analysis and
signal output manipulation (see Figure 4 Microarray
Image.htm). Intracellular/extracellular signal ratios (Cy3/Cy5) were
calculated by (Data_Mean_signal –
Data_Mean.background) / (Reference_Mean_signal –
Reference_Mean.background). Replicate ratios for each gene spot were averaged
and are displayed with mean and standard deviation
(see Table 1).
- Controls
and signal cutoffs
16SrDNA (from MAP and MTB products) control dilution spots gave
differential signal ratios of 1.42 ±
0.08 indicating that the proportion of RNA hybridized to the microarray
from intracellular and extracellular samples was relatively equal. All
spot blanks gave consistently low total signal values, which were
significantly lower (5-500x lower) than spots with products. The cut-off
total signal value was defined as Mean + 2SD of the total signal values
(Ref signal value + Data signal value) from a set (x16) of 16SrDNA
(diluted 1:125) spots distributed across the whole microarray. Signal data
from spots giving total signal values below this figure (in this case 7700)
were discarded from the analysis. The total signal values of 90% spots
with MTB DNA products were significantly lower than in those with MAP DNA
spots.
- Transcriptome
analysis
Gene clusters were arbitrarily assigned to 6 groups with mean ratios x0.5,
x1, x2, x3, x4 and > x5 of the 16SrDNA control mean ratio.
x1 group
The majority of MTB homologues (117 gene products) and MAP gene products
specific for IS900, p43 from IS900 and the 65Kd MAP
heat-shock antigen showed no significant differential expression between
intracellular and extracellular signal expression.
x2 and x0.5 groups
17 MTB homologues: 0 MAP genes in the extracellular culture and 2 MTB
homologues: 24 MAP genes in the intracellular culture showed more than
double the expression signal relative to16SrDNA.
x3, x4 and >x5 groups
12 MAP genes showed more than three fold increase in expression signal in
the intracellular samples. Maximum increase in expression signal was seen
in gsc and gsd genes from the GS cassette. The highest
increase in expression signal from IS900 associated genes was associated
with the des gene from Locus 5 and a tetR regulation gene
from Locus 6.
Differential expression signal enhancement associated with IS900
loci
All signal ratios associated with IS900 Loci show at least some
increase in intracellular expression signal. In 7 of 11 Loci with ORF
interrupted by IS900, an increase (column Left-Right in Table 2) in expression signal of the
interrupted region (Left side of IS900) was observed relative to the
region immediately before the IS900 (Right side of IS900).
Back to top
Conclusions and
further work
a)
The MAP microarray functioned well, especially with MAP
specific PCR products but also with MAP cDNA hybridisation to MTB derived
product targets.
b)
We were able to extract sufficient mRNA from MAP grown
intracellularly (within amoebae for at least 8 weeks) and extracelluarly (in
proteose peptone media) to generate cDNA populations suitable for microarray
analysis.
c)
The following genes were demonstrated to have a significantly
increased expression signal ( >2 fold) in MAP from intracellular samples
All GS cassette genes including gsa, gsbA, gsbB, gsc,
gsd and mpa.
MAP specific genes sodA, hupB, drrC
MAP genes specific for both adjacent genomic regions of IS900 Loci 1, 2,
4-6, 8-11, 13, 14.
MAP genes specific for one adjacent genomic region of IS900 loci 3, 7,
12.
MAP homologues to MTB genes (especially homologues to Rv0046c and Rv0057)
The following genes were demonstrated to have a highly increased expression
signal (>4 fold) in intracellular samples
MAP specific genes from the GS cassette: gsc and gsd.
MAP specific genes homologous to desA1 from IS900 locus 5 and tetR
regulation genes from IS900 locus 6
d)
MAP signals for all MTB spots were relatively weak. The
production of a microarray containing the whole MAP genome in individual PCR
products would be most desirable.
Back to top
Last revised: 21.01.02