Streptococcus pneumoniae early response genes to human lung epithelial cells

Background Streptococcus pneumoniae infection starts from colonization of the host respiratory tract where interaction with host respiratory tract epithelial cells occurs. To investigate pneumococcal genes that are involved in the early stage of interaction with host epithelial cells, transcriptional responses of an encapsulated pathogenic pneumococcal strain TIGR4 upon exposure to human lung epithelial cells A549 for 0.5 h and 1 h time periods were investigated by using TIGR (JCVI) microarray technology. Gene expression changes were validated by quantitative real-time PCR (qRT-PCR) analysis. Findings We observed different transcriptional profiles at two incubation time periods in which most gene expressions were down-regulated at 0.5 h but up-regulated at 1 h. Many genes associated with ribonucleotide biosynthesis were down-regulated at both time points, whereas the genes associated with cell envelope, energy metabolism, transport and protein synthesis were mostly up-regulated at 1 h. Furthermore, these profiles were compared to the transcriptomes of a TIGR4-derived strain in response to human macrophages for the same time periods. We found one set of genes that exhibited similar expression changes upon exposure to both types of host cells, including cell envelope-associated bgaA (SP0648) and nanA (SP1693), and uncharacterized gene clusters such as SP1677–SP1680 and SP1688–SP1690. Conclusion These data indicate that at the early stage of interaction with host epithelial cells, a complex gene regulation and expression change occur in bacteria. Some of them might play an essential role during pathogen-host interactions and for the establishment of infection.

an initial step for infection. Many factors that contribute to the colonization and/or invasion of host epithelial cells have been characterized in S. pneumoniae (recently reviewed by: [1][2][3]). However, it is becoming obvious that multiple factors are involved in this complex process [4].
Microarray-based transcriptome studies have been used in many pathogens, investigating their transcriptional responses to host cells [5]. However, they were rarely performed at an early stage of interaction time period, a stage that might be critical for microbes to establish an infection. This is likely due to the difficulty of obtaining sufficient bacterial RNA from a mixture of bacteria and host cells. In S. pneumoniae, transcriptome studies were initiated by Orihuela et al. [6] in which an unencapsulated derivative of TIGR4 was investigated following exposure to human pharyngeal epithelial cells (Detroit 562) for 3 h. By using self-spotted pneumococcal oligonucleotide (oligo) microarrays we have also examined gene expression changes of an encapsulated serotype 3 clinical isolate and one unencapsulated avirulent laboratory strain following incubation with human lung epithelial cells (A549) for 1 h and 3 h, respectively [7]. Nevertheless, a lack of information exists regarding pneumococcal gene expression at an early stage of interaction with host cells. The strain-specific gene regulation features of S. pneumoniae [8] also prompted our research interests on other serotype strains.
In this study, we have developed a system which can be used to isolate enough bacterial RNA for microarray analysis from encapsulated pathogenic strains following incubation with A549 cells for a short time period. By using TIGR microarrays, we performed transcriptome studies on an encapsulated wild-type strain TIGR4. This study highlighted the gene transcriptional profiles in S. pneumoniae and revealed the potential roles of some target genes during pathogen-host interactions.

Incubation of bacteria and host cells
Culturing and incubation of pneumococcal strain TIGR4 (provided by Dr. Caroline A. Obert, St. Jude Children's Research Hospital) and human lung epithelial cells A549 were performed as previously described [7] with minor modifications. Briefly, bacteria grown to early logarithmic-phase at OD 600 0.3 were collected by centrifugation, re-suspended in antibiotic-free MEM complete medium supplemented with 1% fetal bovine serum (FBS), and incubated with host cells in T75 flasks at a multiplicity of infection 120:1. After incubation, non-adherent bacteria were removed by washing 3 times with 5 ml of antibioticfree cell culture medium. Host cells were removed by incubation with a host cell lysis buffer containing guanidine thiocyanate (Sigma), β-mercaptoethanol, phenol and ethanol at room temperature for 10 min. Bacterial samples were collected by centrifugation for RNA isolation. Bacteria incubated with cell culture medium for different time points, treated with RNALater (Ambion), were collected as medium control samples.

Preparation of bacterial RNA
Isolation of bacterial RNA was performed with RiboP-ure™-Bacteria Kit (Ambion) or a modified method using RNeasy MiniKit (Qiagen) as previously described [7]. From each flask of cell infection, about 2~4 μg bacterial total RNA with less than 10% of eukaryotic RNA contamination could be generated. Medium control RNA samples at each incubation time point were generated by pooling RNAs isolated from 3 separate assays. Genomic DNA contamination was removed by the treatment with RNasefree of DNase I (Ambion).

Microarray experiment and analysis
TIGR (J. Craig Venter Institute) S. pneumoniae 70-mer oligo microarray (version 6), provided by the Pathogen Functional Genomics Resource Center (PFGRC), was used in this study. The cDNA synthesis, Cy-dye labelling, and microarray hybridization were carried out according to TIGR's standard operating procedures (SOPs) http:// www.tigr.org. Hybridization signals were captured with a GenePix 4200A scanner (Axon Instruments) and the data were processed and analyzed through ArrayPipe http:// www.pathogenomics.ca/arraypipe [9]. This includes flagging of marker spots, background correction, printTip Loess normalization with Limma, and statistical analysis with Limma's eBayes moderated t-test [10]. Gene expressions of fold change ≥ 2.0 (bacteria incubated with host cells vs. bacteria incubated with media) with statistical significance (p ≤ 0.05) were classified as being significantly changed. In this study, eight independent hybridizations, including four labelled in dye flips, using RNA samples isolated from eight separate assays were performed for each incubation time point.

Quantitative real-time PCR (qRT-PCR) analysis
The oligo primers used for qRT-PCR analysis (Table 1) were designed from S. pneumoniae TIGR4 genome sequences by using Clone Manager Suite 7 (Scientific & Educational Software) and synthesized by Invitrogen. The qRT-PCR reaction and analysis were performed as previously described [7]. For each gene, duplicate reactions were performed on the RNA samples isolated from at least two separate assays for each incubation time point.

Results and discussion
Transcriptional responses of S. pneumoniae to host epithelial cells Microarray analysis revealed many gene expression changes following exposure to A549 cells (Table 2). At 0.5 h, most gene expressions were down-regulated (35 vs. 16) and a smaller number of genes changed ( Fig. 1). At 1 h, more genes were changed and most of them were up-regulated (50 vs. 25) (Fig. 2). Furthermore, most of those changed genes were only defined at a certain incubation time period (Fig. 3). These data indicate divergent transcriptional profiles between 0.5 h and 1 h incubation time periods. Repressed transcriptional profiles at 0.5 h (Fig. 1) suggest that the interaction with human respiratory tract epithelial cells, a natural reservoir for S. pneumoniae, might be a favourable situation for pneumococci. This is in contrast to the S. pneumoniae transcriptomes to macrophages, where most genes that showed transcriptional changes at the early stage of interactions were up-regulated (Song XM, Connor W, Hokamp K, Babiuk LA, Potter AA: Transcriptome studies on Streptococcus pneumoniae, illustration of early response genes to THP-1 human macrophages, submitted). When incubated for 1 h, bacterial survival, growth and virulence mechanisms appear to be activated, apparent from an induced expression of genes in cell envelope, energy metabolism, transport, protein synthesis, and hypothetical proteins (Fig. 2).
We also observed a common change between two incubation time points, that more than 10 purine and pyrimidine ribonucleotide biosynthesis genes, including purine and pyrimidine regulatory genes purR and pyrR, were consistently down-regulated (Table 2; Figs. 1, 2). The roles of ribonucleotide biosynthesis and their gene regulation mechanism in S. pneumoniae are largely unknown. However, down-regulation of these genes in pneumococci appears to occur only at an early stage of interaction with host epithelial cells, but not at 3 h [6,7]. It also might be specific to the pneumococcal strains and the types of host cells because most of those ribonucleotide biosynthesis genes were unchanged in a serotype 3 strain [7] or when the TIGR4-derived strain was exposed to the host macrophages (Song XM, Connor W, Hokamp K, Babiuk LA, Potter AA: Transcriptome studies on Streptococcus pneumoniae, illustration of early response genes to THP-1 human macrophages, submitted). Perhaps this is the shift of bacteria to parasitism enabling the uptake of substrates from the host cells [11], or the indication of metabolic changes in different pneumococcal strains in different host environment.

Microarray data validation
To confirm gene expression changes identified in microarray analysis, we performed qRT-PCR analysis on 16 selected genes at different incubation time point, most of them associated with cell envelope, ribonucleotide biosynthesis, SP1677-SP1680 and SP1688-SP1690 gene clusters. Except for the unchanged SP1680 at 0.5 h, all the other gene expressions changed in accordance to the microarray data, but at a greater average fold change in the qRT-PCR analysis (Figs. 4,5). Expression change of SP0057 at 1 h was only obtained from qRT-PCR assay because the strain-specific oligo probes were absent on the microarrays (Fig. 5).  macrophages, submitted). As similar experimental procedures and microarray technology were applied, we compared these two studies and revealed many common response genes at early interaction time periods, including well characterized virulence genes such as bgaA and nanA, and uncharacterized gene clusters such as SP1677-SP1680 (hypothetical) and SP1688-SP1690 (ABC transporter) ( Table 3). It indicates common features in pneumococcal gene responses to different types of host cells. Although the interactions with host epithelial cells and macrophages are mainly associated with different pathogenesis processes, reflected by the colonization of host epithelial cells and the survival from host phagocytic cells, we assume these processes are closely related and some of those genes might be assigned with multiple functions.

The exoglycosidase family genes
In S. pneumoniae, the bgaA-encoded β-galactosidase (BgaA) and the nanA-encoded neuraminidase (NanA) belong to a family of exoglycosidases exposed on the bacterial surface. Studies have demonstrated that both enzymes, especially NanA, are involved in adherence to host respiratory tract epithelial cells, possibly by clearing host cell surface structures and secreted components to enhance pathogen-host interactions [12][13][14][15]. Recently, it was demonstrated that BgaA and NanA, together with StrH (β-N-aceylglucosaminidase), act sequentially to remove sialic acid, galactose and N-acetylglucosamine [15]. These reports demonstrated the importance of S. pneumoniae to deglycosylate human targets during colonization and/or pathogenesis.
In this study, expression of bgaA (SP0648) and nanA (SP1693) was highly induced when incubated with A549 cells for 1 h in both microarray and qRT-PCR analyses ( Table 2; Fig. 4). Further qRT-PCR assay revealed an unchanged expression of strH (SP0057) (Fig. 5), correlated to the previous observation that StrH was not involved in the adherence [15]. The enhanced expression of bgaA and nanA was also observed in a TIGR4-derived strain when exposed to human macrophages for 0.5 h and 1 h time periods (Table 3). It suggests that both bgaA and nanA belong to a family of conserved early response genes. Clearing host cell surface components and accessing to the host cells are a priority for bacteria at the early stage of pathogen-host interactions.

Other genes
The cbpI (SP0069), encoding choline binding protein I, was also up-regulated in expression (Table 2; Fig. 4). The choline binding proteins (CBPs) are a family of surface proteins, many of them are involved in colonization of nasopharynx [16]. However, cbpI was the only CBP gene that was identified in this study. The function of CbpI is still unclear but its crystal structure has been solved [17].
Whether it is important in colonization, most CBPs might not be required at the early stage of interaction with host epithelial cells.
Because of strain-specific gene regulations in S. pneumoniae [7,8], different microarray technologies and experimental conditions, some potential gene targets might be missed in our transcriptome studies. For example, the pspC (SP1417) gene was reported to be up-regulated in a serotype 2 strain D39 within 1 h post-infection in mice [18]. However, expression change of pspC was not identified in our assays, despite of a degenerated PspC carried by the TIGR4 genome (TIGR). Another unchanged gene cluster was the rlrA pathogenicity islet genes (SP0461-SP0468) encoding pneumococcal pili [19,20]. All of these TIGR4-specific oligo probes were carried by the TIGR microarrays, and they were clearly identified in our studies of the regulation mechanisms for the pilus locus genes (Song XM, Connor W, Hokamp K, Babiuk LA, Potter AA: The growth phase-dependent regulation of the pilus locus genes by two-component system TCS08 in Streptococcus pneumoniae, submitted). We could therefore exclude the technical concern for these genes in our microarray analysis. Earlier studies suggested that pneumococcal pili were mainly involved in the host cell adhesion [21].  case, also supported by our negative findings, the rlrA pilus locus genes are not likely to be involved in the early stage of interaction with host epithelial cells.

Conclusion
The data presented here provide the first assessment of S. pneumoniae early response genes to human lung epithelial cells. It revealed gene expression changes that might be associated with bacterial adaptation, survival, growth and colonization. Up-regulation of several cell envelope genes, such as bgaA and nanA, and the genes with unknown functions, is likely required for a successful colonization. The specific roles of the identified genes and the functions of coordinated regulation of multiple genes have yet to be further investigated.