Microarray Analysis Reveals Previously Unknown Changes to Toxoplasma Infected Host Cells


Cells, Parasites, And Reagents

Human foreskin fibroblasts (HFFs) were used from passage 10-16 for all assays described. The PDS strain of T. gondii, cloned from ME49, was routinely passaged in HFFs using standard T. gondii culture conditions with DMEM supplemented with 10% FBS (Life Technologies; Rockville, MD). All fine chemicals were from Sigma (St. Louis, MO) unless otherwise specified.

Infection Assays

All experiments were performed using PDS tachyzoites. Infected monolayers were scraped and syringe-lysed with a 27-gauge needle. Parasites were pelleted at 600 x g and washed 2 times in serum-free DMEM. Parasites were counted and added to HFF monolayers (4-7 d old) at an approximate MOI of 5. Infection assays were routinely performed in T75 flasks grown in 15 ml media (or a proportionate volume for different size flasks). Flasks were grown in 5 % CO2 for the indicated times after which the medium was aspirated and total RNA prepared using the RNAeasy midi kit (Qiagen; Valencia, CA) according to the manufacturer's instructions. For the transwell experiments, HFFs were plated in 10 ml media in the lower chamber of a 10 cm transwell tissue culture insert (Corning-Costar; Corning, NY) and 5 ml media in the upper chamber for 4-7 d prior to infection. 1.4 x108 tachyzoites in serum free media were added to the upper chamber and the plates were incubated for 4 h at 37C, 5% CO2.

Microarray Hybridization

cDNA microarrays were synthesized at Stanford University using standard protocols (Eisen, 1999). The microarrays were spotted with 18,000 - 27,000 clones that are available from Research Genetics (Huntsville, AL). cDNA probe preparation was performed essentially as described (Eisen, 1999). Briefly, 20-25 mg of total RNA was annealed to 4 mg oligo-dT18 (New England Biolabs; Bedford, MA) and converted to first-strand cDNA using Superscript II (Life Technologies; Rockville, MD). The reaction was stopped by the addition of 1mM EDTA and RNA hydrolyzed with NaOH at 65C. The reaction was washed with TE and concentrated in a microconcentrator (Millipore; Waltham, MA) to a volume < 20ul. Second-stand cDNA was synthesized and labeled with Cy3-dUTP (infected) or Cy5-dUTP (uninfected) (Amersham-Pharmacia; Uppsala, Sweden) via a random prime reaction using random nonamers. Comparison of labeling cDNA using random nonamers with direct incorporation of the dyes during the first-stand cDNA synthesis reaction, was performed as described (Eisen, 1999), and did not indicate any bias due to the random oligonucleotides. Following labeling, the reactions were washed with TE and human C0t1 DNA (20 mg) (Life Technologies; Rockville, MD), yeast tRNA (20 mg) (Life Technologies; Rockville, MD), and oligo-dA18 (20 mg) was added. Microarrays were hybridized overnight with the labeled cDNA in 3.4 X SSC, 0.3% SDS at 65C and washed sequentially in 2X SSC/0.1% SDS, 1X SSC, and finally 0.2X SSC. The microarrays were dried and scanned using a GenePix 4000 microarray scanner (Axon Instruments; South San Francisco, CA). Each timepoint and condition was repeated at least 2 times.

Data Analysis

After the microarrays were scanned, the images were analyzed with the Scanalyze software (written by Mike Eisen and available at http://rana.Stanford.EDU/software/) in order to determine the fold change for each spot . In addition, spots of poor quality were annotated and discarded from subsequent analysis. The fluorescence intensities of the remaining spots were normalized using previously published methods (Alizadeh, 2000). Briefly, the normalization factor was calibrated independently for each microarray by applying a scaling factor so that the median fluorescence ratio of spots above background on each microarray was 1.0 (Alizadeh, 2000). The data were then filtered such that only spots with intensities that were 3 times greater than background in either channel were used in the analysis. Clusters of spots were generated by applying further filters that retrieved spots that displayed a two-fold or greater change in an experiment. For all such clusters, each spot was manually examined to assess its quality and those that exhibited poor quality throughout the analysis were removed. Spots of poor quality in individual microarrays were discarded from the dataset and are represented as gray bars. Data clustering was performed using the Cluster program and figures generated using Tree View (both available at http://rana.Stanford.EDU/software/) (Eisen, 1998).


Questions, comments, or feedback??? Ira Blader