Gene expression profiles of Chlamydophila pneumoniae during the acute and iron mediated persistent infection

The obligate intracellular, gram-negative bacterium, Chlamydophila pneumoniae (Cpn), has impact as an acute and chronic disease-causing pathogen. Little is known about changes in the Cpn transcriptome during its biphasic developmental cycle (the acute infection) and persistence stages that are linked to chronic diseases. To analyze Cpn CWL029 gene expression, we designed a pathogen-specific oligo microarray and optimized the extraction method for pathogen RNA. Throughout the acute infection, ratio expression profiles for each gene were generated using 48 hours post infection as a reference. Based on these profiles, significantly expressed genes were separated into 12 expression clusters using SOM clustering and manually into the ‘early’, ‘mid’, ‘late’ and ‘tardy’ cluster classes. The latter two were differentiated because the ‘tardy’ class showed steadily increasing expression at the end of the cycle. The transcriptome of the Cpn elementary body (EB) and published EB proteomics data were compared to the cluster profile of the acute infection. We found an intriguing association between ‘late’ genes and genes coding for EB proteins, whereas ‘tardy’ genes were mainly associated with genes coding for EB mRNA. It has been published that iron depletion leads to Cpn persistence. We compared the gene expression profiles during iron depletion-mediated persistence with the expression clusters of the acute infection. This led to the finding that establishment of iron depletion-mediated persistence is more likely a mid-cycle arrest in development rather than a completely distinct gene expression pattern. Taken together, here we describe the Cpn transcriptome during the acute infection with ‘late’ genes correlating to EB proteins and ‘tardy’ genes to EB mRNA. Expression profiles during iron mediated-persistence led us to propose the hypothesis that the transcriptomic ‘clock’ is arrested during acute mid-cycle. Keywords: time course, iron, persistence, Chlamydia, Chlamydophila, pneumoniae, recovery, pathogen, life cycle, developmental cycle A 50mer oligo array containing all open reading frames of the genome projects of CWL029 (http://www.stdgen.lanl.gov), AR39 (garnet.Berkeley.edu) and J138 (http://kantaro.grt.kyushu-u.ac.jp/J138/ident/index.html) was designed. To ensure maximum specificity, oligo sequences were compared for homology against the human genome using the BLAST algorithm. Design and spotting was done by MWG-Biotech AG (Germany). Cross-hybridization of eukaryotic RNA was confirmed to be marginal by labeling and hybridizing eukaryotic total RNA with the described procedure. Arrays were blocked prior to hybridization by incubation for 45 min at 42°C in blocking solution (4x SSC, 0.5% SDS, 1% BSA), washed five times with H20 and dried by centrifugation (1000 rpm, RT, 2 min). Microarray experiments were done as two-color hybridizations. In order to compensate for dye-specific effects and to ensure statistically relevant data, a color swap was performed. Reverse transcribed cDNA pools were mixed with hybridization buffer (Ambion), denatured, snap-cooled, pipetted onto the array using Lifter Slips (Erie, USA) and placed in a sealed humidified hybridization chamber (Scienion, Germany) at 42° C for 48 hours without shaking. After hybridization the array was washed and dried according to the Ambion protocol. Scanning of microarrays was performed with 5 μm resolution using a microarray laser scanner (Agilent). Features were extracted using the Agilent Technologies image analysis software (Version A7.5) using default settings for non-Agilent microarrays. A local background subtraction method was used and the background was adjusted globally to zero. Dye normalization was done using the rank consistency method and applying a local weighted regression normalization (LOWESS). The ratio between both channels, the log ratio error and the p-values were calculated using default settings. Data analysis was carried out on the Rosetta Inpharmatics platform Resolver Buildt 4.2. Transcriptome analysis was carried out with two biological replicates for 36 hpi and at least three biological replicates for all other time points. Samples were derived from independent infections, RNA preparations, labeling reactions and hybridizations. We selected genes using a 1.8-fold expression cut-off. All data of the individual sets comprising different time points were combined as Ratio Experiments using negative polarity for the color swap dye-reversal hybridizations. This was achieved by combining the Ratio Profiles, representing single hybridizations, and using an error weighted average based on p-values.