|Home | About | Journals | Submit | Contact Us | Français|
Mycoplasma genitalium is the causative agent of non-gonococcal, chlamydia-negative urethritis in men and has been linked to reproductive tract disease syndromes in women. As with other mycoplasmas, M. genitalium lacks many regulatory genes because of its streamlined genome and total dependence on a parasitic existence. Therefore, it is important to understand how gene regulation occurs in M. genitalium, particularly in response to environmental signals likely to be encountered in vivo. In this study, we developed an oligonucleotide-based microarray to investigate transcriptional changes in M. genitalium following osmotic shock. Using a physiologically relevant osmolarity condition (0.3 M sodium chloride), we identified 39 upregulated and 72 downregulated genes. Of the upregulated genes, 21 were of unknown function and 15 encoded membrane-associated proteins. The majority of downregulated genes encoded enzymes involved in energy metabolism and components of the protein translation process. These data provide insights into the in vivo response of M. genitalium to hyperosmolarity conditions and identify candidate genes that may contribute to mycoplasma survival in the urogenital tract.
Mycoplasma genitalium is an important sexually transmitted pathogen that causes non-gonococcal, chlamydia-negative urethritis in men and reproductive tract disease syndromes in women (Baseman & Tully, 1997; Jensen, 2004). It is considered to be a parasite of the urogenital tract that colonizes the surface of epithelial cells through multiple adhesin-mediated cytadherence mechanisms (Alvarez et al., 2003; Baseman, 1993). Although M. genitalium is often regarded as an extracellular pathogen, it has been reported to invade host cells and establish long-term persistence (Baseman et al., 1995; Blaylock et al., 2004; Ueno et al., 2008). This observation may account for the chronic infections associated with M. genitalium, despite extensive antibiotic therapies and an intact and functional host immune system.
M. genitalium represents the smallest known self-replicating micro-organism that can be passaged in artificial medium. Also, M. genitalium has a streamlined genome (0.58 Mbp) containing only 482 protein-encoding genes, which approximates the minimal gene set essential to sustain independent life (Fraser et al., 1995; Glass et al., 2006). Despite the limitation of genome content, M. genitalium must be able to adapt to changing in vivo environmental conditions given its intimate interaction with host cells. Unfortunately, little information is available concerning the mechanisms by which M. genitalium regulates gene expression. Transcriptional control in mycoplasmas appears unique, as these prokaryotes possess relatively few regulatory factors compared with other bacteria, possibly due to reduced genome size (Himmelreich et al., 1997) and their successful adaptation to host environments (Weiner et al., 2000). The availability of DNA microarrays facilitates the study of transcriptional regulation in mycoplasmas by examining global transcriptional changes. The first evidence of differential expression of genes in pathogenic mycoplasmas was revealed using microarrays in Mycoplasma pneumoniae during heat shock (Weiner et al., 2003). Later, additional Mycoplasma species were tested under conditions of heat shock (Madsen et al., 2006a), oxidative stress (Schafer et al., 2007), depletion of iron (Madsen et al., 2006b), association with eukaryotic cells (Cecchini et al., 2007) or during infection (Madsen et al., 2008). To date, only a subset of conserved heat-shock genes of M. genitalium has been shown by macroarray to be differentially expressed in response to elevated temperature (Musatovova et al., 2006) and no global transcriptional analysis has been performed on this smallest self-replicating prokaryote.
In this study, we developed an oligonucleotide-based microarray specific for M. genitalium, which represents all known ORFs annotated in the genome. Utilizing this microarray, we examined the transcriptional changes of M. genitalium in response to osmotic stress, an environmental change that M. genitalium encounters in the human urinary tract where osmotic pressure fluctuations are created by passing and residual urine. With this approach, we identified a set of genes that displayed differential expression under osmotic stress.
M. genitalium type strain G37 was routinely grown in Spiroplasma (SP)-4 medium at 37 °C. To assess the growth of M. genitalium in the presence of various concentrations of NaCl, we measured the incorporation of 14C-labelled lysine into protein over time. Equal numbers of strain G37 cells (1×106 colour change units) were inoculated into three 12-well plates (1 ml SP-4 medium per well). After incubation at 37 °C for 60 h, NaCl was added to final concentrations of 0.1, 0.2 and 0.3 M to triplicate wells of each plate. The remaining three wells were untreated and served as controls. Then, 0.1 μCi (0.37 kBq) l-[U-14C]-lysine (Amersham) was added to each well, and cultures were reincubated at 37 °C. At each time point (72, 96 and 120 h), one plate was processed by aspirating the spent medium, washing wells once with PBS, lysing mycoplasma cells with 0.5 ml 10% SDS, and determining radioactivity values by scintillation counting.
An oligonucleotide-based microarray specific for M. genitalium was designed and synthesized by Qiagen Operon. Oligonucleotides (70-mers) representing 482 putative ORFs of type strain G37 (ATCC 33530) were printed on slides by Microarrays Inc. In addition, 24 empty spots were included as negative controls. All spots were printed in triplicate on one slide to improve the reproducibility of array data.
To examine the effects of hyperosmolarity on M. genitalium transcription, four 50 ml cultures of strain G37 in 75 cm2 tissue culture flasks (Corning) were grown to exponential phase, as determined by medium colour change and colony density. Then, NaCl was added to three flasks to achieve final concentrations of 0.1, 0.2 and 0.3 M. Parallel cultures of M. genitalium in the absence of NaCl served as controls. All cultures were incubated for 1 h at 37 °C prior to RNA extraction. Experiments were repeated six times, which produced six independent RNA sample pairs from NaCl-treated cultures and control cultures for each NaCl condition. Dye swap was performed on three of six RNA pairs to minimize effects caused by biased labelling efficiencies.
Tri Reagent (Sigma) was used to extract total M. genitalium RNA according to manufacturer's guidelines. Briefly, surface-attached M. genitalium cells were washed twice with ice-cold sterile PBS, and Tri Reagent (8 ml per 75 cm2 flask surface) was added to lyse cells. For microarray analysis, total RNA preparations were further processed with an RNeasy mini kit (Qiagen) to remove RNAs less than 200 nt. To stabilize RNA samples, 200 units RNase inhibitor RNaseOUT (Invitrogen) was added to each RNA preparation. For real-time PCR analysis, total RNA preparations were further treated with amplification-grade DNase I (Invitrogen) to remove residual genomic DNA.
For probe preparation, hybridization and slide washing, we followed microarray protocols (SOP no. M007 and M008) from the Pathogen Functional Genomics Resource Center of the J. Craig Venter Institute with slight modification (http://pfgrc.jcvi.org/index.php/microarray/protocols.html). Hybridization probes were generated with a set of 482 gene-specific primers (designed and synthesized by Operon Biotechnologies) by reverse transcription of 4 μg total M. genitalium RNA in the presence of aminoallyl-labelled dUTP (2:1 ratio of aa-dUTP:dTTP). After clean-up, the aminoallyl-labelled cDNA preparations were coupled to either cyanine (Cy)3 or Cy5 fluorescent molecules (Amersham Biosciences). Hybridization and slide washing followed protocol SOP no. M008.
Microarray slides were scanned with a GenePix 4000B microarray scanner (Axon Instruments) at 10 μm resolution, and laser power was kept constant for Cy3/Cy5 scans of all slides. Photomultiplier tube (PMT) voltages were adjusted using GenePix Pro 6.0 (Axon Instruments) to obtain maximal signal intensities of each feature with a saturation of <0.005% and count ratio close to 1.0. The resulting images were gridded and processed using GenePix Pro 6.0 to obtain the signal intensity for each spot. The data were further imported into and analysed by Acuity 4.0 (Axon Instruments). The fluorescence levels for each feature of individual slides were first normalized using the ratio-based method and a dataset was created. The relative expression level of each ORF was calculated, and the P-values were determined across the dataset using Student's t test. Then, the 482 P-values from the t test were imported into QVALUE (Storey & Tibshirani, 2003) and false discovery rate (Q) values were computed at an estimated false discovery rate of 5%. Along with Q values, differentially expressed genes were identified by filtering features that were up- or downregulated more than twofold (P<0.01).
Real-time PCR analysis was performed with the ABI Prism 7900 sequence detection system and SYBR green chemistry (Applied Biosystems). Primers were designed using Primer Express 2.0 (Applied Biosystems; Table 1). All primers were tested to be specific for individual genes, as indicated by a single peak after dissociation of each amplification product and a single band after agarose gel electrophoresis. Efficiencies of the target amplification and reference (endogenous control) amplification were confirmed to be equal (R2 >0.99) using five serial 10-fold dilutions of M. genitalium genomic DNA (108–104 copies per reaction) as templates according to User Bulletin No.2 (Applied Biosystems). To prepare the cDNA template, 1 μg DNase I-treated RNA was reverse transcribed with SuperScript Reverse Transcriptase (RT) II (Invitrogen). Then, cDNA templates were diluted and mixed with SYBR green PCR master mix and corresponding primers. Also, RT-negative RNAs were included as templates to estimate the residual contamination of genomic DNA. Amplifications were carried out under default conditions provided by the manufacturer. We applied the comparative threshold cycle method to compare amounts of transcripts under different experimental conditions. The relative levels of transcripts were expressed as fold changes (n-fold) compared with control values, and calculations were completed using RQ Manager 1.2 (Applied Biosystems).
Microarray data were submitted to the Gene Expression Omnibus database under the accession number GSE22661.
To examine the transcriptomes of M. genitalium in response to osmotic shock, we added NaCl to exponential-phase cultures of M. genitalium to final concentrations of 0.1, 0.2 and 0.3 M and continued incubation for 1 h. The addition of NaCl raised the osmotic pressure of SP-4 medium from 394 mOsm kg−1 to 584, 771 and 930 mOsm kg−1, respectively, which are within the physiological range of human urine osmolarity (50–1400 mOsm kg−1). The growth of M. genitalium was not affected in the presence of 0.1 M NaCl, as indicated by the rate of protein synthesis (Fig. 1). At high NaCl concentrations (0.2 and 0.3 M NaCl), although M. genitalium protein synthesis was reduced, the viability of M. genitalium was not significantly affected based upon similar numbers of c.f.u. before and after each NaCl treatment (data not shown).
Microarray analysis was performed to compare the global transcriptome profiles of mycoplasma cultures grown in the presence of NaCl with control cultures. Genes exhibiting similar expression patterns under osmotic stress were grouped by hierarchical clustering analysis (Fig. 2). We observed that the expression of genes was not significantly changed when M. genitalium was treated with 0.1 M NaCl compared with controls. In contrast, we observed significant differential expression of genes when mycoplasma cells were exposed to 0.2 M and 0.3 M NaCl. Although some genes were preferentially expressed under 0.2 M NaCl treatment (see Supplementary Table S1, available with the online version of this paper), more genes tended to be differentially expressed under 0.3 M NaCl treatment. Subsequently, differentially expressed genes under 0.3 M NaCl treatment were identified by the criteria of increases or decreases in signal greater than twofold, P-value less than 0.01 and an estimated false discovery rate of 5%. A total of 39 upregulated genes (Table 2) and 72 downregulated genes (Table 3) met the criteria. The location of these genes and their status of expression are presented in Supplementary Fig. S1 (available with the online version of this paper). To verify that these genes were differentially expressed due to osmotic upshift rather than from salt-specific effects, sucrose was used to increase osmolarity, and similar results were obtained (data not shown).
Real-time PCR was employed and eight genes, which exhibited highly differential expression by microarray and represented important physiological roles, were chosen to confirm microarray results. MG_151, which encodes a ribosomal protein exhibiting relatively stable expression under osmotic shock, was used as the endogenous control. The primers selected for these genes are listed in Table 1. RNA samples from two biological replicates were analysed. As shown in Fig. 3, the expression patterns of all eight genes were consistent with the microarray results, although the fold changes of these osmotically upregulated genes were lower than those observed by microarray analysis. This discrepancy could be due to variations in the expression of MG_151 expression upon osmotic shock.
Bacteria respond to stressful environmental conditions by altering the expression of genes that enhance their survival. In the case of M. genitalium, a cell-wall-free bacterium that colonizes the genito-urinary tract, no information was available about its transcriptional response to osmotic stress, a condition likely to be encountered in vivo. By utilizing a whole-genomic microarray, we observed differential expression of genes when M. genitalium was exposed to 0.2 and 0.3 M NaCl for 1 h. In total, we identified 39 upregulated and 72 downregulated genes after 0.3 M NaCl treatment, which represents about 23% (111/482) of predicted genes encoded by the genome. The microarray data were further confirmed by selecting eight genes for real-time PCR analysis, with all genes showing consistent patterns of expression. Of 39 genes upregulated by osmotic shock, 15 genes encode membrane or membrane-associated proteins, as suggested by the presence of a transmembrane domain(s) in their primary amino acid sequences (Table 2). It is noteworthy that five (MG_067, MG_068, MG_149, MG_439 and MG_440) of the 21 putative lipoprotein genes encoded by the genome were significantly induced, with MG_149 being the most highly upregulated. This is consistent with the idea that pathogenic bacteria adapt to various host environments by varying the synthesis of surface components (Mekalanos, 1992). Together, these results indicate that membrane remodelling may occur in M. genitalium in order to cope with high-osmolarity environments. Since the majority of these induced genes are annotated as being hypothetical, their upregulation under osmotic stress implies their functionally important role in maintaining mycoplasma cellular integrity during osmotically stressful conditions in vivo. An unexpected finding is that MG_454, which encodes an osmotically inducible OsmC-like protein that is upregulated in Escherichia coli and other bacteria (Atichartpongkul et al., 2001), was downregulated (Table 3). The downregulation of MG_454 was validated by both real-time PCR (Fig. 3) and Northern blot analysis (data not shown). Although MG_454 shares sequence homology with OsmC, it is likely that transcription of MG_454 in M. genitalium is different from transcription in other bacteria given the paucity of factors involved in transcription regulation.
Cells typically accumulate osmotically active compounds referred to as compatible solutes, such as proline and glycine betaine, which are highly congruous with cellular functions that offset the deleterious effects of high osmolarity. Therefore, genes involved in the biosynthesis or uptake of compatible solutes are often induced by osmotic shock (Kempf & Bremer, 1998). As M. genitalium relies entirely on exogenous supplies of nutrients for growth, the pathways for synthesizing compatible solutes are not easily identifiable from the genome. Interestingly, we noticed that MG_064 and MG_289, two genes encoding putative ABC transporters, showed significant upregulation under osmotic shock (Table 2). MG_064 encodes a putative permease, which has been revealed to be indispensable under in vitro growth conditions (Glass et al., 2006). MG_289 was annotated as a gene encoding a putative substrate binding protein of phosphonate ABC transporter, which shares sequence homology with p37 (recently designated Cypl) of Mycoplasma hyorhinis, an important factor involved in mycoplasma virulence, tumorigenic transformation and putative extracytoplasmic thiamine-binding (Sippel et al., 2009). Since SP-4 medium is enriched with amino acids that could serve as compatible solutes, it is interesting to speculate that MG_064 and MG_289 might play a role in their uptake. It remains to be investigated which substrate(s) might be transported by MG_064 and MG_289 when M. genitalium encounters osmotic stress.
Of a total of 72 genes significantly downregulated by osmotic shock, many were metabolism-related, including those involved in glycolysis (fba, pgi, pgk, gap, eno, gpml, ldh, pta, ackA, pdhA and pdhB). Moreover, the expression of genes involved in protein translation was also repressed, including those encoding ribosomal proteins (rpsP, rplL, rplS, rplK, rplA, rpsD and rpmF), translation factors (tuf and tsf) and tRNA synthesis (trmD). The downregulation of these genes is consistent with the observed reduction in mycoplasma protein synthesis under osmotic shock (Fig. 1), similar to other bacteria (Weber & Jung, 2002). We noticed that eight genes (atpA, atpB, atpC, aptD, atpE, atpF, atpG and atpH), which encode proteins involved in the formation of F1/F0-ATP synthase complex, were repressed as well. In most prokaryotes, the function of the F1/F0-ATP synthase complex is to maintain a proton gradient through catabolism and hydrolysis of ATP. Since the cytochrome-dependent electron transport chain is absent in mycoplasmas, the function of the synthase complex is believed to be to chiefly maintain the electrochemical gradient in mycoplasmas. Decreased expression of genes encoding F1/F0-ATP synthase was observed in the early response of E. coli during osmotic shock, suggesting severe inhibition of respiration (Meury, 1994; Weber & Jung, 2002). Downregulation of four of the eight ATP synthase subunits has been observed in Mycoplasma gallisepticum when co-incubated with eukaryotic cells (Cecchini et al., 2007). The functional implication of this downregulation of F1/F0-ATP synthase genes in M. genitalium under osmotic stress is unknown.
Presently, we cannot explain what mechanisms regulate the differential expression of genes under osmotic shock, given the presence of only a single σ factor in M. genitalium. However, the increased expression of genes encoding DNA gyrase (gyrA and gyrB) and GTP pyrophosphokinase (relA) during osmotic shock is of special interest (Table 2). DNA gyrase is involved in introducing negative supercoiling into DNA and maintaining the homeostasis of DNA topology of bacterial chromosomes together with DNA topoisomerase IV (Wang, 1996). The expression of gyrase is regulated by DNA supercoiling, which is often altered in response to environmental changes such as osmolarity and temperature, and has been identified as an important factor in globally influencing gene expression in many bacteria (Dorman, 1991). Previously, a microarray study identified a regulatory role of DNA supercoiling in controlling the osmotic stress response in E. coli (Cheung et al., 2003). Therefore, the significant induction of gyrB and gyrA in M. genitalium under 0.3 M NaCl treatment (Table 2) suggests that DNA supercoiling of the chromosome is perturbed and may be essential in regulating genes under osmotic shock. Also, MG_278 (relA; Table 2) encodes a conserved enzyme responsible for the synthesis of ppGpp, which is an important signal molecule involved in stringent response. Interestingly, the overexpression of MPN_397, a homologue of MG_278 in M. pneumoniae, leads to transcriptional changes mainly related to stringent responses, indicating that the biological conservation of this signalling mechanism occurred despite genome reduction (Yus et al., 2009). Induction of relA under osmotic shock has also been demonstrated in Lactococcus lactis. Moreover, the disruption of rel, a homologue of relA in Listeria monocytogenes, reduced osmotolerance, suggesting that appropriate intracellular concentrations of ppGpp are essential for full osmotolerance (Okada et al., 2002). Therefore, it appears that DNA gyrase and GTP pyrophosphokinase contribute to mechanisms of transcriptional regulation in M. genitalium during osmotic stress.
Another noteworthy feature of M. genitalium genes differentially expressed under osmotic shock is that many appear to be co-transcribed and comprise operons (Supplementary Fig. S1). For example, MG_003 and MG_004 encode two subunits of DNA gyrase (gyrA and gyrB) and both are induced under osmotic shock. A similar example among downregulated genes is the gene cluster encoding F1/F0-ATP synthase complex (MG_405 to MG_398). Operon structure remains unclear in M. genitalium, as genes are often clustered in the same orientation with small or no intergenic spaces. Since M. genitalium lacks rho factor, a rho-independent transcriptional terminator is predicted to operate; so far, few have been successfully identified (Kingsford et al., 2007; Washio et al., 1998). Consequently, genes of unrelated functions could be co-transcribed due to run-on expression (Benders et al., 2005). A recent comprehensive analysis of M. pneumoniae transcriptomes describes the preferential use of hairpin formation to tightly regulate gene expression (Güell et al., 2009). Furthermore, operon structure seems to be dynamic and modulated by environmental changes, with the expression of genes within an operon being linked to specific environmental conditions. Therefore, further analysis of gene clusters differentially regulated under osmotic shock in M. genitalium may help define operon structure and provide hints for assigning functional roles to hypothetical genes.
In conclusion, using a DNA microarray we provide the first assessment, to our knowledge, of global transcriptional responses of M. genitalium to osmotic stress. Our data offer insights as to how M. genitalium responds to physiologically relevant environmental cues. Moreover, genes differentially expressed under osmotic shock provide candidates for further investigation of mechanisms involved in transcription regulation in M. genitalium.
This work was supported by the National Institutes of Health grant no. AI045429 and the Frost Bank Trusts. The authors are grateful to Dr Oxana Musatovova for guidance with real-time PCR and helpful discussion in preparing this manuscript.
Microarray data have been submitted to the Gene Expression Omnibus database with the accession number GSE22661.
A supplementary figure, showing an expression map of M. genitalium in the presence of 0.3 M NaCl, and a supplementary table, listing M. genitalium genes preferentially upregulated in the presence of 0.2 M NaCl, are available with the online version of this paper.