Transcriptional Profiling of Mycoplasma hyopneumoniae during Heat Shock Using Microarrays
http://www.100md.com
感染与免疫杂志 2006年第1期
Departments of Veterinary Microbiology and Preventive Medicine
Statistics, Iowa State University, Ames, Iowa
Fellowship for Interpretation of Genomes and San Diego State University, San Diego, California
ABSTRACT
Bacterial pathogens undergo stress during host colonization and disease processes. These stresses result in changes in gene expression to compensate for potentially lethal environments developed in the host during disease. Mycoplasma hyopneumoniae colonizes the swine epithelium and causes a pneumonia that predisposes the host to enhanced disease from other pathogens. How M. hyopneumoniae responds to changing environments in the respiratory tract during disease progression is not known. In fact, little is known concerning the capabilities of mycoplasmas to respond to changing growth environments. With limited genes, mycoplasmas are thought to possess only a few mechanisms for gene regulation. A microarray consisting of 632 of the 698 open reading frames of M. hyopneumoniae was constructed and used to study gene expression differences during a temperature shift from 37°C to 42°C, a temperature swing that might be encountered during disease. To enhance sensitivity, a unique hexamer primer set was employed for generating cDNA from only mRNA species. Our analysis identified 91 genes that had significant transcriptional differences in response to heat shock conditions (P < 0.01) with an estimated false-discovery rate of 4 percent. Thirty-three genes had a change threshold of 1.5-fold or greater. Many of the heat shock proteins previously characterized in other bacteria were identified as significant in this study as well. A proportion of the identified genes (54 of 91) currently have no assigned function.
INTRODUCTION
Mycoplasma hyopneumoniae is the causative agent of porcine enzootic pneumonia and is a major contributor to the porcine respiratory disease complex (26, 31). Like other mycoplasmas, M. hyopneumoniae has a small genome with limited biosynthetic potential (22). Mycoplasmas in general are rarely exposed to severe environmental changes; they are transmitted from animal to animal and do not survive long outside their host. They do, however, encounter changing environmental conditions within the host as the immune response develops and responds with the recruitment of large numbers of neutrophils and macrophages to the infection site and the host becomes febrile (32). Some invasive mycoplasma species, e.g., Mycoplasma bovis, encounter different environments within the colonized tissues after systemic spread. Thus, it is of interest to study potential growth conditions that might impact the ability of M. hyopneumoniae to adapt to and survive in changing environmental conditions.
Heat shock proteins (Hsps) are a group of proteins induced during stressful conditions that result in denaturation of polypeptides. Temperature extremes, high pressure, and exposure to toxic compounds can induce expression of these proteins, conditions that result in protein unfolding and denaturation (12). The Hsps are generally thought of as chaperones, proteins involved in the correct folding of polypeptides as they emerge from the ribosome. They operate by binding to exposed hydrophobic residues in improperly folded protein structures. They can also be involved in the assembly or disassembly of multimeric protein structures and translocation of proteins across membranes. An additional class of Hsps includes proteases that degrade misfolded or abnormal proteins (11, 12). The Hsps are highly conserved and appear to be universal, having functions required for normal cell growth.
The regulation of bacterial heat shock proteins in the Firmicutes falls into four general classes. Genes regulated by the repressor HrcA are designated class I, those regulated by the alternate sigma factor B are designated class II, and class III genes are regulated by CtsR (15). Genes that do not fall within these groupings are placed into class IV. In class I genes, the repressor HrcA binds to a specific inverted repeat sequence, the CIRCE element, within the promoter (14). Mycoplasmas do not have the typical sigma factors involved in responses to high-temperature shifts; only a single general sigma factor has been identified in the mycoplasma genomes so far sequenced (2, 8, 10, 16, 22, 24, 27, 34, 36). Also, those genes that undergo regulation are thought to be controlled by the limited number of transcriptional regulators found in the minimal genome (23). It is likely that one of these regulators is HrcA, whose gene in M. hyopneumoniae has been designated mhp010 (22).
Heat shock has long been considered a stress condition that many pathogens encounter in their pathogen-host relationship. While much has been learned about the Hsp genes and their regulation in many bacterial species (12, 39), few studies have examined mycoplasmas for heat shock responses. In 1990, two reports of heat shock studies involved identification of heat shock proteins (Hsps) by gel electrophoresis in Mycoplasma capricolum subsp. capricolum and Acholeplasma laidlawii (1, 3, 4). More recently, transcriptional profiling was performed on Mycoplasma pneumoniae (35) to identify a set of genes that responded to changing growth temperatures. The results of that study indicated that 47 genes were up-regulated and 30 genes were down-regulated during a temperature shift from 32°C to 42°C at a P value <0.001. This report is also notable in that it was the first genome-wide transcriptional profiling study performed for a mycoplasma species.
Despite the importance of M. hyopneumoniae to swine production, there have been few studies of potential molecular mechanisms of its pathogenesis in the face of and in response to environmental changes. This has probably been due to the difficulty in growing the organism and its recalcitrance to genetic manipulation. The few reports that exist have focused on mechanisms of adherence to swine cilia (17, 18, 21, 40), with no reports of potential stress-related proteins. Analysis of the genome sequence has identified putative Hsps DnaK (mhp072), DnaJ (mhp073), ClpB (mhp278), GrpE (mhp011), and Lon (mhp541) and putative heat shock regulator HrcA (mhp010) (22). To better understand the mechanisms involved in stress responses, we have determined the response of M. hyopneumoniae to heat shock using a microarray approach. Our studies show significant changes in transcriptional activities in numerous genes, suggesting that M. hyopneumoniae, like other bacteria, tightly regulates its genes in response to the environment.
MATERIALS AND METHODS
Mycoplasma strains and culture conditions. Pathogenic M. hyopneumoniae strain 232, a derivative of strain 11, was used in this study (20). Cultures used were passaged fewer than 15 times in vitro in Friis media as previously described (9). One hundred twenty-five-milliliter cultures in 250-ml flasks were grown at 37°C to early log phase as determined by medium color change and optical density. Flasks were shifted to 30°C for 2 h, and then paired flasks were shifted to 37 or 42°C for 30 min. There were six replicates. Cells were pelleted by centrifugation at 24,000 x g, 500 μl of RNAlater (Ambion, Inc., Austin, Tex.) was added to the pellets, and pellets were stored at –70°C until the total RNA was isolated.
Microarray construction. The M. hyopneumoniae microarray consists of PCR products (probes) spotted to glass substrates. To identify primers for PCRs, a FASTA file containing the DNA sequences for all putative open reading frames (ORFs) (22) was analyzed with Primer3 software (29) using a set of scripts to allow for iteration and unique primer design across the genome. The software began the design process by reading and storing the sequences for each of the ORFs. The program then iterated through each ORF, determining those for which valid primers had been identified. At this point, a mispriming library file was created, which consisted of the sequences of all ORFs except for the one for which primers were to be designed. This ensured that any primers found would not match any other ORFs besides the ORF requiring primers. Next, the Primer3 parameters were written to a file (see Table S1 in the supplemental material), and this file and the mispriming library were passed to Primer3, with the results appended to another file. The program then moved on to the next ORF. Once each ORF had been examined, any ORFs still needing primers were identified. The parameters passed to Primer3 (i.e., melting temperature, primer product size, mispriming score, etc.) were relaxed, and the process was then repeated a total of three times, by which time any ORFs for which unique primers could not be identified were not included in the microarray design. Two sets of primers were identified for each ORF and used to generate PCR products for array spotting (see Table S2 in the supplemental material).
PCRs were performed in 96-well plates in an MJ Research Dyad thermal cycler (Bio-Rad Laboratories, Inc., Waltham, Mass.) on all sets of PCR primers in 100-μl reaction mixtures using the following conditions: 1x PCR buffer; 2.5 mM MgCl2; 2.0 mM each dATP, dCTP, dTTP, and dGTP (Roche Diagnostics Corp., Indianapolis, Ind.); 2 pmol of each primer, forward and reverse; 2.5 units Taq polymerase (New England Biolabs, Beverly, Mass.); and 100 ng M. hyopneumoniae strain 232 chromosomal DNA, which had been isolated by phenol-chloroform extraction. The thermal cycling conditions were denaturation at 95°C for 5 min, and then 30 cycles of denaturation at 94°C for 1 min, annealing at 50°C for 1 min, and elongation at 72°C for 30 s, with a final elongation at 72°C for 10 min. Products were confirmed by 2% agarose gel electrophoresis and purified using FilterEX 96-well glass filter plates (model 3510; Corning, Inc., Big Flats, N.Y.) according to the manufacturer's protocol. Purified PCR products were quantified by UV absorbance, dried by vacuum centrifugation, and resuspended to an approximate concentration of 200 ng per μl in spotting buffer (Corning). A 1-μl volume of probe was analyzed by electrophoresis in 2% agarose gels to confirm the concentration and purity of each product.
The microarrays were printed using a BioRobotics MicroGrid (Genomic Solutions, Ann Arbor, Mich.) onto Corning UltraGAPS substrates. Each probe was printed in triplicate, in nonadjacent spots to reduce nonuniform substrate errors in replicated probes. The array consisted of 16 subarrays, with a 12-column by 12-row configuration. Each slide was divided into two regions (upper and lower), and each region contained the full array of spots. Slides were UV cross-linked at 450 mJ in a Stratalinker (Stratagene, La Jolla, Calif.) to immobilize DNA. To determine the optimal cross-linking energy needed for each batch of arrays, test slides were cross-linked at 350, 450, and 550 mJ and stained with SpotQC (Integrated DNA Technologies) per the manufacturer's instructions. The energy giving the highest signal strength was chosen for the batch treatment. Prior to hybridization, arrays were prehybridized with sodium borohydride to reduce background according to methods outlined by Raghavachari et al. (25).
Experimental design. Six independent RNA samples from heat-shocked cells were paired with six independent RNA samples from control cells for hybridization on six two-color microarrays. For three of the six arrays, the control sample was labeled with Alexa 555 dye and compared to the heat-shocked sample labeled with Alexa 647 dye. The dye assignment to control and treated samples was reversed for the other three arrays to account for variation due to labeling efficiencies. The three slides were hybridized under identical conditions as described below.
RNA isolation. RNA was isolated from frozen cell pellets using the Versagene RNA purification system (Gentra Systems, Minneapolis, Minn.) according to the manufacturer's protocol. The optional step of DNase treatment was routinely performed on the column. With a cutoff of 150 bp, 5S rRNA and tRNAs were removed from the samples, limiting interference in downstream manipulations. Samples were quantified and checked for purity using an ND-1000 spectrophotometer (Nanodrop, Wilmington, Del.). If necessary, samples were concentrated using Microcon YM-30 microconcentrators (Millipore, Billerica, Mass.) for optimal cDNA generation.
Random primer set for M. hyopneumoniae. A set of 129 hexamer oligonucleotide primers was designed to generate cDNA sequences from M. hyopneumoniae mRNA (see Table S3 in the supplemental material). The 4,096 possible hexamer oligonucleotide sequences were screened in silico to remove any sequences that would potentially amplify rRNA or tRNA, which makes up 95% or more of the total RNA preparation. The remaining 327 primers were compared to the 698 ORFs in the annotated sequence. Primers were then identified that would theoretically hybridize to the coding strand of the mRNA for each ORF. One hundred picomoles of each primer (Integrated DNA Technologies, Coralville, Iowa) was then combined to form a primer mixture for cDNA generation.
Target generation and hybridization. Targets were generated from total RNA extracted from cell pellets as described above. Fluorescently labeled cDNA targets were generated and purified using the SuperScript indirect cDNA labeling system (Invitrogen Corp., Carlsbad, Calif.). Targets were labeled with Alexa Fluor 555 reactive dye or Alexa Fluor 647 reactive dye (Molecular Probes, Inc., Eugene, Oreg.). Following purification of the labeled cDNA, samples were dried in a vacuum centrifuge and then resuspended in 10 μl Pronto! cDNA/long oligonucleotide hybridization solution (Corning). Targets were denatured at 95°C for 5 min and centrifuged at 13,000 x g for 2 min at room temperature. Targets from control and heat-shocked cultures were then combined, pipetted to an array, and covered with a 22- by 22-mm HybriSlip (Schleicher & Schuell, Keene, N.H.). Slides were placed in a Corning hybridization chamber and incubated in a 42°C water bath for 12 to 16 h. Slides were washed according to Corning's UltraGAPS protocol and dried by centrifugation.
Image acquisition and data analysis. Each dye channel of each array was scanned three times under varying laser power and photomultiplier tube gain settings using a ScanArray Express laser scanner (Applied BioSystems, Inc., Foster City, Calif.) to increase the dynamic range of expression measurement (6). The images were quantified using the softWorRx Tracker analysis software package (Applied Precision, Inc., Issaquah, Wash.). Spot-specific mean signals were corrected for local background by subtracting spot-specific median background intensities. The natural logarithm of the background-corrected signals from a single scan was adjusted by an additive constant so that all scans of the same array-by-dye combination would have a common median. The median of these adjusted-log-background-corrected signals across multiple scans was then computed for each spot to obtain one value for each combination of spot, array, and dye channel. These data for the two dye channels on any given array were normalized using LOWESS normalization to adjust for intensity-dependent dye bias (7, 38). Following LOWESS adjustment, the data from each channel were adjusted by an additive constant so that the median for any combination of array and dye would be the same for all array-by-dye combinations. The normalized values for triplicate spots were averaged within each array to produce one normalized measure of expression for each of the 632 probe sequences and each of the 12 RNA samples.
Statistical analysis. A separate mixed-linear-model analysis was conducted for each probe sequence using the normalized data (37). Each mixed model included fixed effects for treatment (heat shock versus control), slide region (upper versus lower), and dye (Alexa 555 versus 647) as well as random effects for slide and slide-by-region interaction. A t test for differential expression across treatments was conducted for each probe as part of our mixed-linear-model analyses. The 632 P values from these t tests were converted to Q values using the method of Storey and Tibshirani (30). These Q values can be used to obtain approximate control of the false-discovery rate (FDR) at a specified value. For example, declaring probes with Q values less than or equal to 0.05 to be differentially expressed produces a list of significant results for which the FDR is estimated to be approximately 5%. Along with Q values, estimates of change (n-fold) were computed for each probe by taking the inverse natural log of the mean treatment difference estimated as part of our mixed-linear-model analyses.
Validation of microarray data. To confirm significant transcriptional differences between genes, semiquantitative RT-PCR analysis was performed on four genes randomly chosen from the pool of genes shown to have significant transcriptional differences during heat shock using the 16S RNA as a control. PCR was performed in 96-well plates in 20-μl reaction mixtures with the six study replicates for each ORF from control and heat-shocked samples, using the following conditions: 1x PCR buffer; 2 mM MgCl2; 1.0 mM each dATP, dCTP, dTTP and dGTP (Fisher Bioreagents, Fairlawn, NJ); 5 pmol of each primer, forward and reverse; 2.5 units Taq polymerase (New England Biolabs); and 2 μl cDNA from heat-shocked samples diluted 1:50 or 2 μl cDNA from control samples diluted 1:50. The thermal cycling conditions were denaturation at 95°C for 5 min, followed by 25 cycles of denaturation at 94°C for 1 min, annealing at 50°C for 1 min, and elongation at 72°C for 30 s, with a final elongation at 72°C for 10 min. The reaction mixtures were mixed with 4 μl 6x loading dye, and 10 μl was analyzed by 1.5% agarose gel electrophoresis and then stained with 0.5 μg/ml ethidium bromide to confirm that a single product of the correct size was present. The gel was visualized and digitized, and band density was measured with FlourChem 8000 software version 2.0 (Alpha Innotech Corp., San Leandro, Calif.).
Analysis of variance was performed to determine differences between band density values of heat-shocked versus control samples. Band densities were background corrected by subtracting the background density within the reaction lane from the sample band density. Estimated differences were considered significant if the P value of a t test was <0.01.
RESULTS
Primer design and array construction. The primer design software was successful in designing primers that could amplify unique products from 632 of 698 putative ORFs in the M. hyopneumoniae genome sequence. Following purification, PCR products were analyzed by agarose gel for consistency in concentration and product size (data not shown).
Heat shock studies. The total RNA concentrations postpurification were approximately 7 to 18 μg. In each cDNA reaction, 7 to 12 μg of total RNA was added, yielding cDNA concentrations of 2 to 7 μg. Control and heat shock samples were paired and had the same concentration of total RNA added to the cDNA reaction mixtures. Three samples from each treatment were labeled with Alexa Fluor 555 and three with Alexa Fluor 647. The label incorporation of fluorescent dyes ranged from 20 to 81 bp/dye molecule for each dye, which is well within the recommended range of label incorporation (http://probes.invitrogen.com/resources/calc/basedyeratio.html). To account for any variation due to dye incorporation, the experimental design included a dye swap. Data from each of the six replicates were used in the statistical analysis.
Statistical analysis indicated that 91 genes had significant transcriptional differences, with a P value < 0.01 and an estimated false-discovery rate of 4 percent. Results are presented in Tables 1 and 2. Thirty-three of the statistically significant changes in expression exhibited greater-than-1.5-fold up- or down-regulation. The results are also displayed as a volcano plot and genes with significant transcriptional differences indicated (Fig. 1). Examples of the normalized data for specific genes are shown in Fig. 2. The quality of the data is reflected by the similarity in the slopes of the lines for each gene. Each line connects points representing the normalized measures of expression for the two treatments from a single array. The two different plotting symbols, circle and square, represent Alexa Fluor 555 and 647 dyes, respectively.
Validation of the microarray data. The primers used in generating the RT-PCR products are given in Table 3. For all except the 16S RNA sequence, these primers were used to generate the PCR products spotted to the microarray (see Table S2 in the supplemental material). The results are shown in Fig. 3. Analysis indicated that the four genes verified were significant using semiquantitative PCR in support of the microarray results. As expected, the 16S RNA concentration did not vary (P = 0.474).
DISCUSSION
Little is known about the mechanisms M. hyopneumoniae uses to colonize the respiratory epithelium and circumvent the host immune response. During infection, M. hyopneumoniae must encounter different environments as the numbers of organisms increase and the host responds accordingly with an influx of macrophages and neutrophils. If mycoplasmas are like other bacterial pathogens, they have mechanisms to respond to the changing in vivo environment, ensuring the organism's survival during active host immune responses.
The host employs both innate and adaptive immune response strategies to protect it against pathogens. These environmental changes are generally thought of as stress conditions because they are often detrimental to the pathogen. For instance, cytokine release by host immune effector cells can lead to temperature fluxes in the host that may impact the organism's ability to maintain itself in the respiratory tract. Most bacteria have several genes that are induced to respond to this "heat shock." These genes usually encode proteins involved in protein folding and the prevention of protein aggregation, and in protein degradation of improperly folded proteins. The increased transcription of these genes during temperature shifts is indicative of an important role in survival. Denatured or nascent proteins do not function properly and would eventually cause bacterial death.
Sixty-one percent of the up-regulated genes with a change threshold of 1.5-fold or greater (Table 1) have been assigned a function based on their sequence homology. These include functions involved in protein folding, metabolism, and translation. dnaK (mhp072), a member of the Escherichia coli hsp70 gene family, was identified as the most significantly up-regulated gene. DnaK's role as a chaperone protecting proteins from improper folding is well characterized, and it also serves a regulatory function by sequestering 32 from the cytoplasm and presenting it to proteases. During heat shock, DnaK binds to denatured proteins, thus releasing 32 to interact with RNA polymerase and up-regulate heat shock genes (39). It should be noted, however, that M. hyopneumoniae does not make 32, and thus, some of the functions ascribed to it in other species are not applicable. Other heat shock-related genes, parC (mhp034), dnaJ (mhp073), and clpB (mhp278), were also significantly up-regulated in M. hyopneumoniae. Proteins encoded by these genes function to assist in protein folding, in degrading improperly folded proteins, and in cell partitioning activities associated with DNA structure. Both dnaK and dnaJ have upstream CIRCE inverted repeats that bind HrcA to up-regulate genes during heat shock (14). Some of the genes appear to be members of operons, e.g., mhp144 through mhp153, whose products are thought to be involved in carbohydrate metabolism, and mhp211 through mhp214, which code for ribosomal proteins and RpoA (mhp213), the subunit of RNA polymerase. The structure of this operon is different than that found in E. coli; the M. hyopneumoniae operon lacks S4, the transcriptional regulator that binds to the leader sequence upstream of S13 (19). Thus, the mode of regulation for this operon in M. hyopneumoniae is unknown, but it appears different than in other bacteria. The other member of this operon, mhp213, had a P value of 0.013266 and just missed our chosen cutoff (P < 0.01) for inclusion in Table 1. Thirty-nine percent of the genes with a change threshold of 1.5-fold or greater had unassigned functions and may represent important stress-related functions required for host colonization and persistence.
One of the more interesting genes up-regulated during heat stress was ffh, encoding the 54-kilodalton subunit homologue of the eukaryotic signal recognition particle. This is the first report of its regulation in response to heat stress, and this regulation may represent a need to increase protein translocation in M. hyopneumoniae either to stabilize the membrane or to maintain the integrity of membrane proteins. ffh is normally thought not to undergo altered expression, but recent evidence suggests that in gram-positive bacteria it may be regulated in response to acid stress (13). In addition, ffh was not regulated in M. pneumoniae during heat stress (35), suggesting that different mycoplasma species may control different sets of genes during heat shock as a consequence of host adaptation.
Of 41 genes that showed significantly lower transcript levels in response to elevated temperatures, 9 genes had a change threshold of 1.5-fold or greater. Interestingly, the factor-of-change values are lower for these genes than for those up-regulated. Many of these genes are involved in translation and DNA replication, suggesting that temperature stress slows translation and replication, reducing energy needs, and slows cellular physiology. Down-regulated genes fall into the classes of transporter genes, mgtE (mhp485), gatB (mhp030), and proS (mhp397). This is similar to other bacterial pathogens and suggests that a reduction in energy-requiring processes in the bacterial cell during heat stress is a fundamental property of pathogenesis and bacterial physiology.
Little is known about the response of mycoplasmas to heat shock, particularly at the transcriptional level. Several studies have identified heat shock proteins or specific gene products in mycoplasmas (4, 28, 35). Weiner et al. analyzed the response of M. pneumoniae to heat shock conditions on a global scale using microarrays and identified 47 up-regulated genes (35). No other studies have taken a global approach to identifying heat shock-related genes in mycoplasmas, however. A report that Hsp60 was present in M. hyopneumoniae (28) could not be substantiated by the genome sequence of strain 232, the strain used in these studies (22). In fact, both GroEL and GroES are missing in the three M. hyopneumoniae genomes sequenced (22, 34). These data indicate that 91 M. hyopneumoniae genes undergo significant transcriptional changes in response to a 37 to 42°C temperature shift at a significance (P) level of <0.01 (Tables 1 and 2). This suggests that M. hyopneumoniae responds to temperature changes by altering transcriptional activities of specific genes through unknown mechanisms. It also suggests that mycoplasmas respond differently depending on the species and possibly host.
A large percentage of heat shock-responsive genes (60%) have no functional assignment. Future studies will be required to identify the roles of the associated gene products in physiology and pathogenesis. How these genes are controlled in the absence of typical heat shock regulators (e.g., rpoH) is not known, but perhaps mechanisms operative in related gram-positive bacteria might be involved. However, now that these genes have been identified, studies of upstream sequences may provide insight into regulatory mechanisms despite the lack of genetic tools in M. hyopneumoniae. These studies focused on transcriptional changes, but other mechanisms involving ribosome conformation (33) or posttranslational processing (5) might also be involved.
ACKNOWLEDGMENTS
We thank Hui-Hsien Chou for assistance with primer design. We also thank Nancy Upchurch and Barb Erickson for assistance with mycoplasma cultures. Monica Perez contributed to the construction of the microarray by performing PCRs. Mike Carruthers assisted with spot finding. Josh Pitzer assisted in the RT-PCRs validating the microarray results.
Funding for this project was provided in part by the Iowa Healthy Livestock Advisory Council.
Supplemental material for this article may be found at http://iai.asm.org/.
REFERENCES
1. Borchsenius, S. N., E. V. Budantseva, and M. S. Vonsky. 1990. The heat-shock proteins of Acholeplasma laidlawii. Zentralbl. Bakteriol. Suppl. 20:657-658.
2. Chambaud, I., R. Heilig, S. Ferris, V. Barbe, D. Samson, F. Galisson, I. Moszer, K. Dybvig, H. Wroblewski, A. Viari, E. P. Rocha, and A. Blanchard. 2001. The complete genome sequence of the murine respiratory pathogen Mycoplasma pulmonis. Nucleic Acids Res. 29:2145-2153.
3. Dascher, C. C., and J. Maniloff. 1992. Heat shock response, p. 349-354. In J. Maniloff, R. N. McElhaney, L. R. Finch, and J. B. Baseman (ed.), Mycoplasmas: molecular biology and pathogenesis. American Society for Microbiology, Washington, D.C.
4. Dascher, C. C., S. K. Poddar, and J. Maniloff. 1990. Heat shock response in mycoplasmas, genome-limited organisms. J. Bacteriol. 172:1823-1827.
5. Djordjevic, S. P., S. J. Cordwell, M. A. Djordjevic, J. Wilton, and F. C. Minion. 2004. Proteolytic processing of the Mycoplasma hyopneumoniae cilium adhesin. Infect. Immun. 72:2791-2802.
6. Dudley, A. M., J. Aach, M. A. Steffen, and G. M. Church. 2002. Measuring absolute expression with microarrays with a calibrated reference sample and an extended signal intensity range. Proc. Natl. Acad. Sci. USA 99:7554-7559.
7. Dudoit, S., Y. H. Yang, M. J. Callow, and T. P. Speed. 2000. Statistical methods for identifying differentially expressed genes in replicated cDNA microarray experiments. [Online.] http://www.stat.Berkeley.edu/users/terry/zarray/Html/papersindex.html.
8. Fraser, C. M., J. D. Gocayne, O. White, M. D. Adams, R. A. Clayton, R. D. Fleischmann, C. J. Bult, A. R. Kerlavage, G. Sutton, J. M. Kelley, J. L. Fritchman, J. F. Weidman, K. V. Small, M. Sandusky, J. Fuhrmann, D. Nguyen, T. R. Utterback, D. M. Saudek, C. A. Phillips, J. M. Merrick, J.-F. Tomb, B. A. Dougherty, K. A. Bott, P.-C. Hu, T. S. Lucier, S. N. Peterson, H. O. Smith, C. A. Hutchison, and J. C. Venter. 1995. The minimal gene complement of Mycoplasma genitalium. Science 270:397-403.
9. Friis, N. F. 1975. Some recommendations concerning primary isolation of Mycoplasma hyopneumoniae and Mycoplasma flocculare, a survey. Nord. Vetmed. 27:337-339.
10. Glass, J. I., E. J. Lefkowitz, J. S. Glass, C. R. Heiner, E. Y. Chen, and G. H. Cassell. 2000. The complete sequence of the mucosal pathogen Ureaplasma urealyticum. Nature 407:757-762.
11. Gottesman, S., and M. R. Maurizi. 1992. Regulation by proteolysis: energy-dependent proteases and their targets. Microbiol. Rev. 56:592-621.
12. Gross, C. A. 1996. Function and regulation of the heat shock proteins, p. 1382-1399. In F. C. Neidhardt, R. I. Curtiss, J. L. Ingraham, E. C. C. Lin, K. B. Low, B. Magasanik, W. S. Reznikoff, M. Riley, M. Schaechter, and H. E. Umbarger (ed.), Escherichia coli and Salmonella typhimurium: cellular and molecular biology, vol. 1. ASM Press, Washington, D.C.
13. Gutierrez, J. A., P. J. Crowley, D. G. Cvitkovitch, L. J. Brady, I. R. Hamilton, J. D. Hillman, and A. S. Bleiweis. 1999. Streptococcus mutans ffh, a gene encoding a homologue of the 54 kDa subunit of the signal recognition particle, is involved in resistance to acid stress. Microbiology 145:357-366.
14. Hecker, M., W. Schumann, and U. Volker. 1996. Heat-shock and general stress response in Bacillus subtilis. Mol. Microbiol. 19:417-428.
15. Helmann, J. D., M. F. Wu, P. A. Kobel, F. J. Gamo, M. Wilson, M. M. Morshedi, M. Navre, and C. Paddon. 2001. Global transcriptional response of Bacillus subtilis to heat shock. J. Bacteriol. 183:7318-7328.
16. Himmelreich, R., H. Hilbert, H. Plagens, E. Pirkl, B.-C. Li, and R. Herrmann. 1996. Complete sequence analysis of the genome of the bacterium Mycoplasma pneumoniae. Nucleic Acids Res. 24:4420-4449.
17. Hsu, T., S. Artiushin, and F. C. Minion. 1997. Cloning and functional analysis of the P97 swine cilium adhesin gene of Mycoplasma hyopneumoniae. J. Bacteriol. 179:1317-1323.
18. Hsu, T., and F. C. Minion. 1998. Identification of the cilium binding epitope of the Mycoplasma hyopneumoniae P97 adhesin. Infect. Immun. 66:4762-4766.
19. Keener, J., and M. Nomura. 1996. Regulation of ribosome synthesis, p. 1417-1431. In F. C. Neidhardt, R. I. Curtiss, J. L. Ingraham, E. C. C. Lin, K. B. Low, B. Magasanik, W. S. Reznikoff, M. Riley, M. Schaechter, and H. E. Umbarger (ed.), Escherichia coli and Salmonella typhimurium: cellular and molecular biology. ASM Press, Washington, D.C.
20. Mare, C. J., and W. P. Switzer. 1965. New species: Mycoplasma hyopneumoniae, a causative agent of virus pig pneumonia. Vet. Med. Small Anim. Clin. 60:841-846.
21. Minion, F. C., C. Adams, and T. Hsu. 2000. R1 region of P97 mediates adherence of Mycoplasma hyopneumoniae to swine cilia. Infect. Immun. 68:3056-3060.
22. Minion, F. C., E. L. Lefkowitz, M. L. Madsen, B. J. Cleary, S. M. Swartzell, and G. G. Mahairas. 2004. The genome sequence of Mycoplasma hyopneumoniae strain 232, the agent of swine mycoplasmosis. J. Bacteriol. 186:7123-7133.
23. Muto, A., and C. Ushida. 2002. Transcription and translation, p. 323-345. In S. Razin and R. Herrmann (ed.), Molecular biology and pathogenicity of mycoplasmas. Plenum Press, New York, N.Y.
24. Papazisi, L., T. S. Gorton, G. Kutish, P. F. Markham, G. F. Browning, D. K. Nguyen, S. Swartzell, A. Madan, G. Mahairas, and S. J. Geary. 2003. The complete genome sequence of the avian pathogen Mycoplasma gallisepticum strain R(low). Microbiology 149:2307-2316.
25. Raghavachari, N., Y. P. Bao, G. Li, X. Xie, and U. R. Meuller. 2003. Reduction of autofluorescence on DNA microarrays and slide surfaces by treatment with sodium borohydride. Anal. Biochem. 312:101-105.
26. Ross, R. F. 1993. Mycoplasma-animal pathogens, p. 69-109. In I. Kahane and A. Adoni (ed.), Rapid diagnosis of mycoplasmas. Plenum Press, New York, N.Y.
27. Sasaki, Y., J. Ishikawa, A. Yamashita, K. Oshima, T. Kenri, K. Furuya, C. Yoshino, A. Horino, T. Shiba, T. Sasaki, and M. Hattori. 2002. The complete genomic sequence of Mycoplasma penetrans, an intracellular bacterial pathogen in humans. Nucleic Acids Res. 30:5293-5300.
28. Scherm, B., G.-F. Gerlach, and M. Runge. 2002. Analysis of heat shock protein 60 encoding genes of mycoplasmas and investigations concerning their role in immunity and infection. Vet. Microbiol. 89:141-150.
29. Skaletsky, R. S. 2000. Primer3 on the WWW for general users and for biologist programmers, p. 365-386. In S. Krawetz and S. Misener (ed.), Bioinformatics methods and protocols: methods in molecular biology. Humana Press, Totowa, N.J.
30. Storey, J. D., and R. Tibshirani. 2003. Statistical significance for genomewide studies. Proc. Natl. Acad. Sci. USA 100:9440-9445.
31. Thacker, E. L., P. G. Halbur, R. F. Ross, R. Thanawongnuwech, and B. J. Thacker. 1999. Mycoplasma hyopneumoniae potentiation of porcine reproductive and respiratory syndrome virus-induced pneumonia. J. Clin. Microbiol. 37:620-627.
32. Thanawongnuwech, R., B. Thacker, P. Halbur, and E. L. Thacker. 2004. Increased production of proinflammatory cytokines following infection with porcine reproductive and respiratory syndrome virus and Mycoplasma hyopneumoniae. Clin. Diagn. Lab. Immunol. 11:901-908.
33. Vanbogelen, R. A., and F. C. Neidhardt. 1990. Ribosomes as sensors of heat and cold shock in Escherichia coli. Proc. Natl. Acad. Sci. USA 87:5589-5593.
34. Vasconcelos, A. T., H. B. Ferreira, C. V. Bizarro, S. L. Bonatto, M. O. Carvalho, P. M. Pinto, D. F. Almeida, L. G. Almeida, R. Almeida, L. Alves-Filho, E. N. Assunecao, V. A. Azevedo, M. R. Bogo, M. M. Brigido, M. Brocchi, H. A. Burity, A. A. Camargo, S. S. Camargo, M. S. Carepo, D. M. Carraro, J. C. de Mattos Cascardo, L. A. Castro, G. Cavalcanti, G. Chemale, R. G. Collevatti, C. W. Cunha, B. Dallagiovanna, B. P. Dambros, O. A. Dellagostin, C. Falcao, F. Fantinatti-Garboggini, M. S. Felipe, L. Fiorentin, G. R. Franco, N. S. Freitas, D. Frias, T. B. Grangeiro, E. C. Grisard, C. T. Guimaraes, M. Hungria, S. N. Jardim, M. A. Krieger, J. P. Laurino, L. F. Lima, M. I. Lopes, E. L. Loreto, H. M. Madeira, G. P. Manfio, A. Q. Maranhao, C. T. Martinkovics, S. R. Medeiros, M. A. Moreira, M. Neiva, C. E. Ramalho-Neto, M. F. Nicolas, S. C. Oliveira, R. F. Paixao, F. O. Pedrosa, S. D. Pena, M. Pereira, L. Pereira-Ferrari, I. Piffer, L. S. Pinto, D. P. Potrich, A. C. Salim, F. R. Santos, R. Schmitt, M. P. Schneider, A. Schrank, I. S. Schrank, A. F. Schuck, H. N. Seuanez, D. W. Silva, R. Silva, S. C. Silva, C. M. Soares, K. R. Souza, R. C. Souza, C. C. Staats, M. B. Steffens, S. M. Teixeira, T. P. Urmenyi, M. H. Vainstein, L. W. Zuccherato, A. J. Simpson, and A. Zaha. 2005. Swine and poultry pathogens: the complete genome sequences of two strains of Mycoplasma hyopneumoniae and a strain of Mycoplasma synoviae. J. Bacteriol. 187:5568-5577.
35. Weiner, J., III, C.-U. Zimmerman, H. W. H. Gohlmann, and R. Herrmann. 2003. Transcription profiles of the bacterium Mycoplasma pneumoniae grown at different temperatures. Nucleic Acids Res. 31:6306-6320.
36. Westberg, J., A. Persson, A. Holmberg, A. Goesmann, J. Lundegerb, K.-E. Johansson, B. Pettersson, and M. Uhlaen. 2004. The genome sequence of Mycoplasma mycoides subsp. mycoides SC type strain PG1T, the causative agent of contagious bovine pleuropneumonia (CBPP). Genome Res. 14:221-227.
37. Wolfinger, R. D., G. Gibson, E. D. Wolfinger, L. Bennett, H. Hamadeh, P. Bushel, C. Afshari, and R. S. Paules. 2001. Assessing gene significance from cDNA microarray expression data via mixed models. J. Comp. Biol. 8:625-637.
38. Yang, Y. H., S. Dudoit, P. Luu, D. M. Lin, V. Peng, J. Ngai, and T. P. Speed. 2002. Normalization for cDNA microarray data: a robust composite method for addressing single and multiple slide systematic variation. Nucleic Acids Res. 30:15. [Online.]
39. Yura, T., M. Kanemori, and M. Morita. 2000. The heat shock response: regulation and function, p. 3-18. In R. Storz and R. Hengge-Aronis (ed.), Bacterial stress response. ASM Press, Washington, D.C.
40. Zhang, Q., T. F. Young, and R. F. Ross. 1995. Identification and characterization of a Mycoplasma hyopneumoniae adhesin. Infect. Immun. 63:1013-1019.(Melissa L. Madsen, Dan Ne)
濠电儑绲藉ú鐘诲礈濠靛洤顕遍柛娑卞枤椤╃兘鏌涘☉鍗炲閺夆晜妫冮弻娑樷枎韫囨挴鍋撴禒瀣劦妞ゆ巻鍋撻柛鐘崇〒濡叉劕鈹戦崶鈹炬灃閻庡箍鍎卞Λ娑㈠焵椤掑鐏︽鐐差儔楠炲洭顢旈崨顓炵哎濠电偠鎻徊鎯洪幋鐘典笉闁挎繂鎷嬮崵鍫澪旈敂绛嬪劌闁哥偞鎸抽弻鏇㈠幢閺囩姴濡介柣銏╁灠缂嶅﹪骞婇敓鐘茬疀妞ゆ挾鍋熸禒鎰版⒑閸︻厐鐟懊洪妶鍥潟闁冲搫鎳庤繚闂佺ǹ鏈粙鎺楁倵椤斿墽纾奸柡鍐ㄥ€稿暩婵犫拃鍕垫疁鐎殿喖鐖煎畷姗€濡歌閸撴垶绻涚€涙ḿ鐭婂Δ鐘叉憸閺侇噣顢曢敂钘夘€涘┑锛勫仜婢х晫绮欐繝鍥ㄧ厸濠㈣泛锕ら弳鏇熸叏閻熼偊妯€闁轰礁绉撮悾婵嬪礃椤垳鎴烽梻浣筋嚃閸犳捇宕濊箛娑辨晣缂備焦岣块埢鏃堟煟閹寸儑渚涢柛鏂垮暣閺岋繝宕掑顓犵厬缂備焦顨呴ˇ閬嶅焵椤掑喚娼愮紒顔肩箻閿濈偤鏁冮崒姘卞摋闁荤娀缂氬▍锝囩矓閸喓鈧帒顫濋鐘闂侀潧娲ゅú銊╁焵椤掑偆鏀版繛澶嬬洴瀹曘垽濡堕崶銊ヮ伕閻熸粎澧楃敮妤咃綖婢舵劖鍋ｉ柛銉娑撹尙绱掓潏銊х畼闁归濞€閹粓鎸婃径澶岀梾濠电偛顕慨楣冨春閺嶎厼鍨傞柕濞炬櫆閸嬨劌霉閿濆懎鏆熸俊顖氱墦濮婃椽顢曢敐鍡欐闂佺粯鎼换婵嬬嵁鐎ｎ喖绠ｆ繝濠傚閹枫劑姊洪幐搴ｂ槈闁哄牜鍓熷畷鐟扳堪閸曨収娴勫銈嗗笂閻掞箓寮抽鍫熺厱闁瑰搫绉村畵鍡涙煃瑜滈崜姘潩閵娾晜鍋傞柨鐔哄Т鐟欙箓骞栭幖顓炵仯缂佲偓婢跺⊕褰掑礂閸忚偐娈ら梺缁樼箖閻╊垰鐣烽敓鐘茬闁肩⒈鍓氶鎴︽⒑鐠団€虫灁闁告柨楠搁埢鎾诲箣閻愭潙顎撳┑鐘诧工閸燁垶骞嗛崒姣綊鎮╅幓鎺濆妷濠电姭鍋撻柟娈垮枤绾鹃箖鏌熺€电ǹ啸鐟滅増鐓￠弻娑㈠箳閺傚簱鏋呭┑鐐叉噹闁帮絾淇婇幘顔芥櫢闁跨噦鎷�Statistics, Iowa State University, Ames, Iowa
Fellowship for Interpretation of Genomes and San Diego State University, San Diego, California
ABSTRACT
Bacterial pathogens undergo stress during host colonization and disease processes. These stresses result in changes in gene expression to compensate for potentially lethal environments developed in the host during disease. Mycoplasma hyopneumoniae colonizes the swine epithelium and causes a pneumonia that predisposes the host to enhanced disease from other pathogens. How M. hyopneumoniae responds to changing environments in the respiratory tract during disease progression is not known. In fact, little is known concerning the capabilities of mycoplasmas to respond to changing growth environments. With limited genes, mycoplasmas are thought to possess only a few mechanisms for gene regulation. A microarray consisting of 632 of the 698 open reading frames of M. hyopneumoniae was constructed and used to study gene expression differences during a temperature shift from 37°C to 42°C, a temperature swing that might be encountered during disease. To enhance sensitivity, a unique hexamer primer set was employed for generating cDNA from only mRNA species. Our analysis identified 91 genes that had significant transcriptional differences in response to heat shock conditions (P < 0.01) with an estimated false-discovery rate of 4 percent. Thirty-three genes had a change threshold of 1.5-fold or greater. Many of the heat shock proteins previously characterized in other bacteria were identified as significant in this study as well. A proportion of the identified genes (54 of 91) currently have no assigned function.
INTRODUCTION
Mycoplasma hyopneumoniae is the causative agent of porcine enzootic pneumonia and is a major contributor to the porcine respiratory disease complex (26, 31). Like other mycoplasmas, M. hyopneumoniae has a small genome with limited biosynthetic potential (22). Mycoplasmas in general are rarely exposed to severe environmental changes; they are transmitted from animal to animal and do not survive long outside their host. They do, however, encounter changing environmental conditions within the host as the immune response develops and responds with the recruitment of large numbers of neutrophils and macrophages to the infection site and the host becomes febrile (32). Some invasive mycoplasma species, e.g., Mycoplasma bovis, encounter different environments within the colonized tissues after systemic spread. Thus, it is of interest to study potential growth conditions that might impact the ability of M. hyopneumoniae to adapt to and survive in changing environmental conditions.
Heat shock proteins (Hsps) are a group of proteins induced during stressful conditions that result in denaturation of polypeptides. Temperature extremes, high pressure, and exposure to toxic compounds can induce expression of these proteins, conditions that result in protein unfolding and denaturation (12). The Hsps are generally thought of as chaperones, proteins involved in the correct folding of polypeptides as they emerge from the ribosome. They operate by binding to exposed hydrophobic residues in improperly folded protein structures. They can also be involved in the assembly or disassembly of multimeric protein structures and translocation of proteins across membranes. An additional class of Hsps includes proteases that degrade misfolded or abnormal proteins (11, 12). The Hsps are highly conserved and appear to be universal, having functions required for normal cell growth.
The regulation of bacterial heat shock proteins in the Firmicutes falls into four general classes. Genes regulated by the repressor HrcA are designated class I, those regulated by the alternate sigma factor B are designated class II, and class III genes are regulated by CtsR (15). Genes that do not fall within these groupings are placed into class IV. In class I genes, the repressor HrcA binds to a specific inverted repeat sequence, the CIRCE element, within the promoter (14). Mycoplasmas do not have the typical sigma factors involved in responses to high-temperature shifts; only a single general sigma factor has been identified in the mycoplasma genomes so far sequenced (2, 8, 10, 16, 22, 24, 27, 34, 36). Also, those genes that undergo regulation are thought to be controlled by the limited number of transcriptional regulators found in the minimal genome (23). It is likely that one of these regulators is HrcA, whose gene in M. hyopneumoniae has been designated mhp010 (22).
Heat shock has long been considered a stress condition that many pathogens encounter in their pathogen-host relationship. While much has been learned about the Hsp genes and their regulation in many bacterial species (12, 39), few studies have examined mycoplasmas for heat shock responses. In 1990, two reports of heat shock studies involved identification of heat shock proteins (Hsps) by gel electrophoresis in Mycoplasma capricolum subsp. capricolum and Acholeplasma laidlawii (1, 3, 4). More recently, transcriptional profiling was performed on Mycoplasma pneumoniae (35) to identify a set of genes that responded to changing growth temperatures. The results of that study indicated that 47 genes were up-regulated and 30 genes were down-regulated during a temperature shift from 32°C to 42°C at a P value <0.001. This report is also notable in that it was the first genome-wide transcriptional profiling study performed for a mycoplasma species.
Despite the importance of M. hyopneumoniae to swine production, there have been few studies of potential molecular mechanisms of its pathogenesis in the face of and in response to environmental changes. This has probably been due to the difficulty in growing the organism and its recalcitrance to genetic manipulation. The few reports that exist have focused on mechanisms of adherence to swine cilia (17, 18, 21, 40), with no reports of potential stress-related proteins. Analysis of the genome sequence has identified putative Hsps DnaK (mhp072), DnaJ (mhp073), ClpB (mhp278), GrpE (mhp011), and Lon (mhp541) and putative heat shock regulator HrcA (mhp010) (22). To better understand the mechanisms involved in stress responses, we have determined the response of M. hyopneumoniae to heat shock using a microarray approach. Our studies show significant changes in transcriptional activities in numerous genes, suggesting that M. hyopneumoniae, like other bacteria, tightly regulates its genes in response to the environment.
MATERIALS AND METHODS
Mycoplasma strains and culture conditions. Pathogenic M. hyopneumoniae strain 232, a derivative of strain 11, was used in this study (20). Cultures used were passaged fewer than 15 times in vitro in Friis media as previously described (9). One hundred twenty-five-milliliter cultures in 250-ml flasks were grown at 37°C to early log phase as determined by medium color change and optical density. Flasks were shifted to 30°C for 2 h, and then paired flasks were shifted to 37 or 42°C for 30 min. There were six replicates. Cells were pelleted by centrifugation at 24,000 x g, 500 μl of RNAlater (Ambion, Inc., Austin, Tex.) was added to the pellets, and pellets were stored at –70°C until the total RNA was isolated.
Microarray construction. The M. hyopneumoniae microarray consists of PCR products (probes) spotted to glass substrates. To identify primers for PCRs, a FASTA file containing the DNA sequences for all putative open reading frames (ORFs) (22) was analyzed with Primer3 software (29) using a set of scripts to allow for iteration and unique primer design across the genome. The software began the design process by reading and storing the sequences for each of the ORFs. The program then iterated through each ORF, determining those for which valid primers had been identified. At this point, a mispriming library file was created, which consisted of the sequences of all ORFs except for the one for which primers were to be designed. This ensured that any primers found would not match any other ORFs besides the ORF requiring primers. Next, the Primer3 parameters were written to a file (see Table S1 in the supplemental material), and this file and the mispriming library were passed to Primer3, with the results appended to another file. The program then moved on to the next ORF. Once each ORF had been examined, any ORFs still needing primers were identified. The parameters passed to Primer3 (i.e., melting temperature, primer product size, mispriming score, etc.) were relaxed, and the process was then repeated a total of three times, by which time any ORFs for which unique primers could not be identified were not included in the microarray design. Two sets of primers were identified for each ORF and used to generate PCR products for array spotting (see Table S2 in the supplemental material).
PCRs were performed in 96-well plates in an MJ Research Dyad thermal cycler (Bio-Rad Laboratories, Inc., Waltham, Mass.) on all sets of PCR primers in 100-μl reaction mixtures using the following conditions: 1x PCR buffer; 2.5 mM MgCl2; 2.0 mM each dATP, dCTP, dTTP, and dGTP (Roche Diagnostics Corp., Indianapolis, Ind.); 2 pmol of each primer, forward and reverse; 2.5 units Taq polymerase (New England Biolabs, Beverly, Mass.); and 100 ng M. hyopneumoniae strain 232 chromosomal DNA, which had been isolated by phenol-chloroform extraction. The thermal cycling conditions were denaturation at 95°C for 5 min, and then 30 cycles of denaturation at 94°C for 1 min, annealing at 50°C for 1 min, and elongation at 72°C for 30 s, with a final elongation at 72°C for 10 min. Products were confirmed by 2% agarose gel electrophoresis and purified using FilterEX 96-well glass filter plates (model 3510; Corning, Inc., Big Flats, N.Y.) according to the manufacturer's protocol. Purified PCR products were quantified by UV absorbance, dried by vacuum centrifugation, and resuspended to an approximate concentration of 200 ng per μl in spotting buffer (Corning). A 1-μl volume of probe was analyzed by electrophoresis in 2% agarose gels to confirm the concentration and purity of each product.
The microarrays were printed using a BioRobotics MicroGrid (Genomic Solutions, Ann Arbor, Mich.) onto Corning UltraGAPS substrates. Each probe was printed in triplicate, in nonadjacent spots to reduce nonuniform substrate errors in replicated probes. The array consisted of 16 subarrays, with a 12-column by 12-row configuration. Each slide was divided into two regions (upper and lower), and each region contained the full array of spots. Slides were UV cross-linked at 450 mJ in a Stratalinker (Stratagene, La Jolla, Calif.) to immobilize DNA. To determine the optimal cross-linking energy needed for each batch of arrays, test slides were cross-linked at 350, 450, and 550 mJ and stained with SpotQC (Integrated DNA Technologies) per the manufacturer's instructions. The energy giving the highest signal strength was chosen for the batch treatment. Prior to hybridization, arrays were prehybridized with sodium borohydride to reduce background according to methods outlined by Raghavachari et al. (25).
Experimental design. Six independent RNA samples from heat-shocked cells were paired with six independent RNA samples from control cells for hybridization on six two-color microarrays. For three of the six arrays, the control sample was labeled with Alexa 555 dye and compared to the heat-shocked sample labeled with Alexa 647 dye. The dye assignment to control and treated samples was reversed for the other three arrays to account for variation due to labeling efficiencies. The three slides were hybridized under identical conditions as described below.
RNA isolation. RNA was isolated from frozen cell pellets using the Versagene RNA purification system (Gentra Systems, Minneapolis, Minn.) according to the manufacturer's protocol. The optional step of DNase treatment was routinely performed on the column. With a cutoff of 150 bp, 5S rRNA and tRNAs were removed from the samples, limiting interference in downstream manipulations. Samples were quantified and checked for purity using an ND-1000 spectrophotometer (Nanodrop, Wilmington, Del.). If necessary, samples were concentrated using Microcon YM-30 microconcentrators (Millipore, Billerica, Mass.) for optimal cDNA generation.
Random primer set for M. hyopneumoniae. A set of 129 hexamer oligonucleotide primers was designed to generate cDNA sequences from M. hyopneumoniae mRNA (see Table S3 in the supplemental material). The 4,096 possible hexamer oligonucleotide sequences were screened in silico to remove any sequences that would potentially amplify rRNA or tRNA, which makes up 95% or more of the total RNA preparation. The remaining 327 primers were compared to the 698 ORFs in the annotated sequence. Primers were then identified that would theoretically hybridize to the coding strand of the mRNA for each ORF. One hundred picomoles of each primer (Integrated DNA Technologies, Coralville, Iowa) was then combined to form a primer mixture for cDNA generation.
Target generation and hybridization. Targets were generated from total RNA extracted from cell pellets as described above. Fluorescently labeled cDNA targets were generated and purified using the SuperScript indirect cDNA labeling system (Invitrogen Corp., Carlsbad, Calif.). Targets were labeled with Alexa Fluor 555 reactive dye or Alexa Fluor 647 reactive dye (Molecular Probes, Inc., Eugene, Oreg.). Following purification of the labeled cDNA, samples were dried in a vacuum centrifuge and then resuspended in 10 μl Pronto! cDNA/long oligonucleotide hybridization solution (Corning). Targets were denatured at 95°C for 5 min and centrifuged at 13,000 x g for 2 min at room temperature. Targets from control and heat-shocked cultures were then combined, pipetted to an array, and covered with a 22- by 22-mm HybriSlip (Schleicher & Schuell, Keene, N.H.). Slides were placed in a Corning hybridization chamber and incubated in a 42°C water bath for 12 to 16 h. Slides were washed according to Corning's UltraGAPS protocol and dried by centrifugation.
Image acquisition and data analysis. Each dye channel of each array was scanned three times under varying laser power and photomultiplier tube gain settings using a ScanArray Express laser scanner (Applied BioSystems, Inc., Foster City, Calif.) to increase the dynamic range of expression measurement (6). The images were quantified using the softWorRx Tracker analysis software package (Applied Precision, Inc., Issaquah, Wash.). Spot-specific mean signals were corrected for local background by subtracting spot-specific median background intensities. The natural logarithm of the background-corrected signals from a single scan was adjusted by an additive constant so that all scans of the same array-by-dye combination would have a common median. The median of these adjusted-log-background-corrected signals across multiple scans was then computed for each spot to obtain one value for each combination of spot, array, and dye channel. These data for the two dye channels on any given array were normalized using LOWESS normalization to adjust for intensity-dependent dye bias (7, 38). Following LOWESS adjustment, the data from each channel were adjusted by an additive constant so that the median for any combination of array and dye would be the same for all array-by-dye combinations. The normalized values for triplicate spots were averaged within each array to produce one normalized measure of expression for each of the 632 probe sequences and each of the 12 RNA samples.
Statistical analysis. A separate mixed-linear-model analysis was conducted for each probe sequence using the normalized data (37). Each mixed model included fixed effects for treatment (heat shock versus control), slide region (upper versus lower), and dye (Alexa 555 versus 647) as well as random effects for slide and slide-by-region interaction. A t test for differential expression across treatments was conducted for each probe as part of our mixed-linear-model analyses. The 632 P values from these t tests were converted to Q values using the method of Storey and Tibshirani (30). These Q values can be used to obtain approximate control of the false-discovery rate (FDR) at a specified value. For example, declaring probes with Q values less than or equal to 0.05 to be differentially expressed produces a list of significant results for which the FDR is estimated to be approximately 5%. Along with Q values, estimates of change (n-fold) were computed for each probe by taking the inverse natural log of the mean treatment difference estimated as part of our mixed-linear-model analyses.
Validation of microarray data. To confirm significant transcriptional differences between genes, semiquantitative RT-PCR analysis was performed on four genes randomly chosen from the pool of genes shown to have significant transcriptional differences during heat shock using the 16S RNA as a control. PCR was performed in 96-well plates in 20-μl reaction mixtures with the six study replicates for each ORF from control and heat-shocked samples, using the following conditions: 1x PCR buffer; 2 mM MgCl2; 1.0 mM each dATP, dCTP, dTTP and dGTP (Fisher Bioreagents, Fairlawn, NJ); 5 pmol of each primer, forward and reverse; 2.5 units Taq polymerase (New England Biolabs); and 2 μl cDNA from heat-shocked samples diluted 1:50 or 2 μl cDNA from control samples diluted 1:50. The thermal cycling conditions were denaturation at 95°C for 5 min, followed by 25 cycles of denaturation at 94°C for 1 min, annealing at 50°C for 1 min, and elongation at 72°C for 30 s, with a final elongation at 72°C for 10 min. The reaction mixtures were mixed with 4 μl 6x loading dye, and 10 μl was analyzed by 1.5% agarose gel electrophoresis and then stained with 0.5 μg/ml ethidium bromide to confirm that a single product of the correct size was present. The gel was visualized and digitized, and band density was measured with FlourChem 8000 software version 2.0 (Alpha Innotech Corp., San Leandro, Calif.).
Analysis of variance was performed to determine differences between band density values of heat-shocked versus control samples. Band densities were background corrected by subtracting the background density within the reaction lane from the sample band density. Estimated differences were considered significant if the P value of a t test was <0.01.
RESULTS
Primer design and array construction. The primer design software was successful in designing primers that could amplify unique products from 632 of 698 putative ORFs in the M. hyopneumoniae genome sequence. Following purification, PCR products were analyzed by agarose gel for consistency in concentration and product size (data not shown).
Heat shock studies. The total RNA concentrations postpurification were approximately 7 to 18 μg. In each cDNA reaction, 7 to 12 μg of total RNA was added, yielding cDNA concentrations of 2 to 7 μg. Control and heat shock samples were paired and had the same concentration of total RNA added to the cDNA reaction mixtures. Three samples from each treatment were labeled with Alexa Fluor 555 and three with Alexa Fluor 647. The label incorporation of fluorescent dyes ranged from 20 to 81 bp/dye molecule for each dye, which is well within the recommended range of label incorporation (http://probes.invitrogen.com/resources/calc/basedyeratio.html). To account for any variation due to dye incorporation, the experimental design included a dye swap. Data from each of the six replicates were used in the statistical analysis.
Statistical analysis indicated that 91 genes had significant transcriptional differences, with a P value < 0.01 and an estimated false-discovery rate of 4 percent. Results are presented in Tables 1 and 2. Thirty-three of the statistically significant changes in expression exhibited greater-than-1.5-fold up- or down-regulation. The results are also displayed as a volcano plot and genes with significant transcriptional differences indicated (Fig. 1). Examples of the normalized data for specific genes are shown in Fig. 2. The quality of the data is reflected by the similarity in the slopes of the lines for each gene. Each line connects points representing the normalized measures of expression for the two treatments from a single array. The two different plotting symbols, circle and square, represent Alexa Fluor 555 and 647 dyes, respectively.
Validation of the microarray data. The primers used in generating the RT-PCR products are given in Table 3. For all except the 16S RNA sequence, these primers were used to generate the PCR products spotted to the microarray (see Table S2 in the supplemental material). The results are shown in Fig. 3. Analysis indicated that the four genes verified were significant using semiquantitative PCR in support of the microarray results. As expected, the 16S RNA concentration did not vary (P = 0.474).
DISCUSSION
Little is known about the mechanisms M. hyopneumoniae uses to colonize the respiratory epithelium and circumvent the host immune response. During infection, M. hyopneumoniae must encounter different environments as the numbers of organisms increase and the host responds accordingly with an influx of macrophages and neutrophils. If mycoplasmas are like other bacterial pathogens, they have mechanisms to respond to the changing in vivo environment, ensuring the organism's survival during active host immune responses.
The host employs both innate and adaptive immune response strategies to protect it against pathogens. These environmental changes are generally thought of as stress conditions because they are often detrimental to the pathogen. For instance, cytokine release by host immune effector cells can lead to temperature fluxes in the host that may impact the organism's ability to maintain itself in the respiratory tract. Most bacteria have several genes that are induced to respond to this "heat shock." These genes usually encode proteins involved in protein folding and the prevention of protein aggregation, and in protein degradation of improperly folded proteins. The increased transcription of these genes during temperature shifts is indicative of an important role in survival. Denatured or nascent proteins do not function properly and would eventually cause bacterial death.
Sixty-one percent of the up-regulated genes with a change threshold of 1.5-fold or greater (Table 1) have been assigned a function based on their sequence homology. These include functions involved in protein folding, metabolism, and translation. dnaK (mhp072), a member of the Escherichia coli hsp70 gene family, was identified as the most significantly up-regulated gene. DnaK's role as a chaperone protecting proteins from improper folding is well characterized, and it also serves a regulatory function by sequestering 32 from the cytoplasm and presenting it to proteases. During heat shock, DnaK binds to denatured proteins, thus releasing 32 to interact with RNA polymerase and up-regulate heat shock genes (39). It should be noted, however, that M. hyopneumoniae does not make 32, and thus, some of the functions ascribed to it in other species are not applicable. Other heat shock-related genes, parC (mhp034), dnaJ (mhp073), and clpB (mhp278), were also significantly up-regulated in M. hyopneumoniae. Proteins encoded by these genes function to assist in protein folding, in degrading improperly folded proteins, and in cell partitioning activities associated with DNA structure. Both dnaK and dnaJ have upstream CIRCE inverted repeats that bind HrcA to up-regulate genes during heat shock (14). Some of the genes appear to be members of operons, e.g., mhp144 through mhp153, whose products are thought to be involved in carbohydrate metabolism, and mhp211 through mhp214, which code for ribosomal proteins and RpoA (mhp213), the subunit of RNA polymerase. The structure of this operon is different than that found in E. coli; the M. hyopneumoniae operon lacks S4, the transcriptional regulator that binds to the leader sequence upstream of S13 (19). Thus, the mode of regulation for this operon in M. hyopneumoniae is unknown, but it appears different than in other bacteria. The other member of this operon, mhp213, had a P value of 0.013266 and just missed our chosen cutoff (P < 0.01) for inclusion in Table 1. Thirty-nine percent of the genes with a change threshold of 1.5-fold or greater had unassigned functions and may represent important stress-related functions required for host colonization and persistence.
One of the more interesting genes up-regulated during heat stress was ffh, encoding the 54-kilodalton subunit homologue of the eukaryotic signal recognition particle. This is the first report of its regulation in response to heat stress, and this regulation may represent a need to increase protein translocation in M. hyopneumoniae either to stabilize the membrane or to maintain the integrity of membrane proteins. ffh is normally thought not to undergo altered expression, but recent evidence suggests that in gram-positive bacteria it may be regulated in response to acid stress (13). In addition, ffh was not regulated in M. pneumoniae during heat stress (35), suggesting that different mycoplasma species may control different sets of genes during heat shock as a consequence of host adaptation.
Of 41 genes that showed significantly lower transcript levels in response to elevated temperatures, 9 genes had a change threshold of 1.5-fold or greater. Interestingly, the factor-of-change values are lower for these genes than for those up-regulated. Many of these genes are involved in translation and DNA replication, suggesting that temperature stress slows translation and replication, reducing energy needs, and slows cellular physiology. Down-regulated genes fall into the classes of transporter genes, mgtE (mhp485), gatB (mhp030), and proS (mhp397). This is similar to other bacterial pathogens and suggests that a reduction in energy-requiring processes in the bacterial cell during heat stress is a fundamental property of pathogenesis and bacterial physiology.
Little is known about the response of mycoplasmas to heat shock, particularly at the transcriptional level. Several studies have identified heat shock proteins or specific gene products in mycoplasmas (4, 28, 35). Weiner et al. analyzed the response of M. pneumoniae to heat shock conditions on a global scale using microarrays and identified 47 up-regulated genes (35). No other studies have taken a global approach to identifying heat shock-related genes in mycoplasmas, however. A report that Hsp60 was present in M. hyopneumoniae (28) could not be substantiated by the genome sequence of strain 232, the strain used in these studies (22). In fact, both GroEL and GroES are missing in the three M. hyopneumoniae genomes sequenced (22, 34). These data indicate that 91 M. hyopneumoniae genes undergo significant transcriptional changes in response to a 37 to 42°C temperature shift at a significance (P) level of <0.01 (Tables 1 and 2). This suggests that M. hyopneumoniae responds to temperature changes by altering transcriptional activities of specific genes through unknown mechanisms. It also suggests that mycoplasmas respond differently depending on the species and possibly host.
A large percentage of heat shock-responsive genes (60%) have no functional assignment. Future studies will be required to identify the roles of the associated gene products in physiology and pathogenesis. How these genes are controlled in the absence of typical heat shock regulators (e.g., rpoH) is not known, but perhaps mechanisms operative in related gram-positive bacteria might be involved. However, now that these genes have been identified, studies of upstream sequences may provide insight into regulatory mechanisms despite the lack of genetic tools in M. hyopneumoniae. These studies focused on transcriptional changes, but other mechanisms involving ribosome conformation (33) or posttranslational processing (5) might also be involved.
ACKNOWLEDGMENTS
We thank Hui-Hsien Chou for assistance with primer design. We also thank Nancy Upchurch and Barb Erickson for assistance with mycoplasma cultures. Monica Perez contributed to the construction of the microarray by performing PCRs. Mike Carruthers assisted with spot finding. Josh Pitzer assisted in the RT-PCRs validating the microarray results.
Funding for this project was provided in part by the Iowa Healthy Livestock Advisory Council.
Supplemental material for this article may be found at http://iai.asm.org/.
REFERENCES
1. Borchsenius, S. N., E. V. Budantseva, and M. S. Vonsky. 1990. The heat-shock proteins of Acholeplasma laidlawii. Zentralbl. Bakteriol. Suppl. 20:657-658.
2. Chambaud, I., R. Heilig, S. Ferris, V. Barbe, D. Samson, F. Galisson, I. Moszer, K. Dybvig, H. Wroblewski, A. Viari, E. P. Rocha, and A. Blanchard. 2001. The complete genome sequence of the murine respiratory pathogen Mycoplasma pulmonis. Nucleic Acids Res. 29:2145-2153.
3. Dascher, C. C., and J. Maniloff. 1992. Heat shock response, p. 349-354. In J. Maniloff, R. N. McElhaney, L. R. Finch, and J. B. Baseman (ed.), Mycoplasmas: molecular biology and pathogenesis. American Society for Microbiology, Washington, D.C.
4. Dascher, C. C., S. K. Poddar, and J. Maniloff. 1990. Heat shock response in mycoplasmas, genome-limited organisms. J. Bacteriol. 172:1823-1827.
5. Djordjevic, S. P., S. J. Cordwell, M. A. Djordjevic, J. Wilton, and F. C. Minion. 2004. Proteolytic processing of the Mycoplasma hyopneumoniae cilium adhesin. Infect. Immun. 72:2791-2802.
6. Dudley, A. M., J. Aach, M. A. Steffen, and G. M. Church. 2002. Measuring absolute expression with microarrays with a calibrated reference sample and an extended signal intensity range. Proc. Natl. Acad. Sci. USA 99:7554-7559.
7. Dudoit, S., Y. H. Yang, M. J. Callow, and T. P. Speed. 2000. Statistical methods for identifying differentially expressed genes in replicated cDNA microarray experiments. [Online.] http://www.stat.Berkeley.edu/users/terry/zarray/Html/papersindex.html.
8. Fraser, C. M., J. D. Gocayne, O. White, M. D. Adams, R. A. Clayton, R. D. Fleischmann, C. J. Bult, A. R. Kerlavage, G. Sutton, J. M. Kelley, J. L. Fritchman, J. F. Weidman, K. V. Small, M. Sandusky, J. Fuhrmann, D. Nguyen, T. R. Utterback, D. M. Saudek, C. A. Phillips, J. M. Merrick, J.-F. Tomb, B. A. Dougherty, K. A. Bott, P.-C. Hu, T. S. Lucier, S. N. Peterson, H. O. Smith, C. A. Hutchison, and J. C. Venter. 1995. The minimal gene complement of Mycoplasma genitalium. Science 270:397-403.
9. Friis, N. F. 1975. Some recommendations concerning primary isolation of Mycoplasma hyopneumoniae and Mycoplasma flocculare, a survey. Nord. Vetmed. 27:337-339.
10. Glass, J. I., E. J. Lefkowitz, J. S. Glass, C. R. Heiner, E. Y. Chen, and G. H. Cassell. 2000. The complete sequence of the mucosal pathogen Ureaplasma urealyticum. Nature 407:757-762.
11. Gottesman, S., and M. R. Maurizi. 1992. Regulation by proteolysis: energy-dependent proteases and their targets. Microbiol. Rev. 56:592-621.
12. Gross, C. A. 1996. Function and regulation of the heat shock proteins, p. 1382-1399. In F. C. Neidhardt, R. I. Curtiss, J. L. Ingraham, E. C. C. Lin, K. B. Low, B. Magasanik, W. S. Reznikoff, M. Riley, M. Schaechter, and H. E. Umbarger (ed.), Escherichia coli and Salmonella typhimurium: cellular and molecular biology, vol. 1. ASM Press, Washington, D.C.
13. Gutierrez, J. A., P. J. Crowley, D. G. Cvitkovitch, L. J. Brady, I. R. Hamilton, J. D. Hillman, and A. S. Bleiweis. 1999. Streptococcus mutans ffh, a gene encoding a homologue of the 54 kDa subunit of the signal recognition particle, is involved in resistance to acid stress. Microbiology 145:357-366.
14. Hecker, M., W. Schumann, and U. Volker. 1996. Heat-shock and general stress response in Bacillus subtilis. Mol. Microbiol. 19:417-428.
15. Helmann, J. D., M. F. Wu, P. A. Kobel, F. J. Gamo, M. Wilson, M. M. Morshedi, M. Navre, and C. Paddon. 2001. Global transcriptional response of Bacillus subtilis to heat shock. J. Bacteriol. 183:7318-7328.
16. Himmelreich, R., H. Hilbert, H. Plagens, E. Pirkl, B.-C. Li, and R. Herrmann. 1996. Complete sequence analysis of the genome of the bacterium Mycoplasma pneumoniae. Nucleic Acids Res. 24:4420-4449.
17. Hsu, T., S. Artiushin, and F. C. Minion. 1997. Cloning and functional analysis of the P97 swine cilium adhesin gene of Mycoplasma hyopneumoniae. J. Bacteriol. 179:1317-1323.
18. Hsu, T., and F. C. Minion. 1998. Identification of the cilium binding epitope of the Mycoplasma hyopneumoniae P97 adhesin. Infect. Immun. 66:4762-4766.
19. Keener, J., and M. Nomura. 1996. Regulation of ribosome synthesis, p. 1417-1431. In F. C. Neidhardt, R. I. Curtiss, J. L. Ingraham, E. C. C. Lin, K. B. Low, B. Magasanik, W. S. Reznikoff, M. Riley, M. Schaechter, and H. E. Umbarger (ed.), Escherichia coli and Salmonella typhimurium: cellular and molecular biology. ASM Press, Washington, D.C.
20. Mare, C. J., and W. P. Switzer. 1965. New species: Mycoplasma hyopneumoniae, a causative agent of virus pig pneumonia. Vet. Med. Small Anim. Clin. 60:841-846.
21. Minion, F. C., C. Adams, and T. Hsu. 2000. R1 region of P97 mediates adherence of Mycoplasma hyopneumoniae to swine cilia. Infect. Immun. 68:3056-3060.
22. Minion, F. C., E. L. Lefkowitz, M. L. Madsen, B. J. Cleary, S. M. Swartzell, and G. G. Mahairas. 2004. The genome sequence of Mycoplasma hyopneumoniae strain 232, the agent of swine mycoplasmosis. J. Bacteriol. 186:7123-7133.
23. Muto, A., and C. Ushida. 2002. Transcription and translation, p. 323-345. In S. Razin and R. Herrmann (ed.), Molecular biology and pathogenicity of mycoplasmas. Plenum Press, New York, N.Y.
24. Papazisi, L., T. S. Gorton, G. Kutish, P. F. Markham, G. F. Browning, D. K. Nguyen, S. Swartzell, A. Madan, G. Mahairas, and S. J. Geary. 2003. The complete genome sequence of the avian pathogen Mycoplasma gallisepticum strain R(low). Microbiology 149:2307-2316.
25. Raghavachari, N., Y. P. Bao, G. Li, X. Xie, and U. R. Meuller. 2003. Reduction of autofluorescence on DNA microarrays and slide surfaces by treatment with sodium borohydride. Anal. Biochem. 312:101-105.
26. Ross, R. F. 1993. Mycoplasma-animal pathogens, p. 69-109. In I. Kahane and A. Adoni (ed.), Rapid diagnosis of mycoplasmas. Plenum Press, New York, N.Y.
27. Sasaki, Y., J. Ishikawa, A. Yamashita, K. Oshima, T. Kenri, K. Furuya, C. Yoshino, A. Horino, T. Shiba, T. Sasaki, and M. Hattori. 2002. The complete genomic sequence of Mycoplasma penetrans, an intracellular bacterial pathogen in humans. Nucleic Acids Res. 30:5293-5300.
28. Scherm, B., G.-F. Gerlach, and M. Runge. 2002. Analysis of heat shock protein 60 encoding genes of mycoplasmas and investigations concerning their role in immunity and infection. Vet. Microbiol. 89:141-150.
29. Skaletsky, R. S. 2000. Primer3 on the WWW for general users and for biologist programmers, p. 365-386. In S. Krawetz and S. Misener (ed.), Bioinformatics methods and protocols: methods in molecular biology. Humana Press, Totowa, N.J.
30. Storey, J. D., and R. Tibshirani. 2003. Statistical significance for genomewide studies. Proc. Natl. Acad. Sci. USA 100:9440-9445.
31. Thacker, E. L., P. G. Halbur, R. F. Ross, R. Thanawongnuwech, and B. J. Thacker. 1999. Mycoplasma hyopneumoniae potentiation of porcine reproductive and respiratory syndrome virus-induced pneumonia. J. Clin. Microbiol. 37:620-627.
32. Thanawongnuwech, R., B. Thacker, P. Halbur, and E. L. Thacker. 2004. Increased production of proinflammatory cytokines following infection with porcine reproductive and respiratory syndrome virus and Mycoplasma hyopneumoniae. Clin. Diagn. Lab. Immunol. 11:901-908.
33. Vanbogelen, R. A., and F. C. Neidhardt. 1990. Ribosomes as sensors of heat and cold shock in Escherichia coli. Proc. Natl. Acad. Sci. USA 87:5589-5593.
34. Vasconcelos, A. T., H. B. Ferreira, C. V. Bizarro, S. L. Bonatto, M. O. Carvalho, P. M. Pinto, D. F. Almeida, L. G. Almeida, R. Almeida, L. Alves-Filho, E. N. Assunecao, V. A. Azevedo, M. R. Bogo, M. M. Brigido, M. Brocchi, H. A. Burity, A. A. Camargo, S. S. Camargo, M. S. Carepo, D. M. Carraro, J. C. de Mattos Cascardo, L. A. Castro, G. Cavalcanti, G. Chemale, R. G. Collevatti, C. W. Cunha, B. Dallagiovanna, B. P. Dambros, O. A. Dellagostin, C. Falcao, F. Fantinatti-Garboggini, M. S. Felipe, L. Fiorentin, G. R. Franco, N. S. Freitas, D. Frias, T. B. Grangeiro, E. C. Grisard, C. T. Guimaraes, M. Hungria, S. N. Jardim, M. A. Krieger, J. P. Laurino, L. F. Lima, M. I. Lopes, E. L. Loreto, H. M. Madeira, G. P. Manfio, A. Q. Maranhao, C. T. Martinkovics, S. R. Medeiros, M. A. Moreira, M. Neiva, C. E. Ramalho-Neto, M. F. Nicolas, S. C. Oliveira, R. F. Paixao, F. O. Pedrosa, S. D. Pena, M. Pereira, L. Pereira-Ferrari, I. Piffer, L. S. Pinto, D. P. Potrich, A. C. Salim, F. R. Santos, R. Schmitt, M. P. Schneider, A. Schrank, I. S. Schrank, A. F. Schuck, H. N. Seuanez, D. W. Silva, R. Silva, S. C. Silva, C. M. Soares, K. R. Souza, R. C. Souza, C. C. Staats, M. B. Steffens, S. M. Teixeira, T. P. Urmenyi, M. H. Vainstein, L. W. Zuccherato, A. J. Simpson, and A. Zaha. 2005. Swine and poultry pathogens: the complete genome sequences of two strains of Mycoplasma hyopneumoniae and a strain of Mycoplasma synoviae. J. Bacteriol. 187:5568-5577.
35. Weiner, J., III, C.-U. Zimmerman, H. W. H. Gohlmann, and R. Herrmann. 2003. Transcription profiles of the bacterium Mycoplasma pneumoniae grown at different temperatures. Nucleic Acids Res. 31:6306-6320.
36. Westberg, J., A. Persson, A. Holmberg, A. Goesmann, J. Lundegerb, K.-E. Johansson, B. Pettersson, and M. Uhlaen. 2004. The genome sequence of Mycoplasma mycoides subsp. mycoides SC type strain PG1T, the causative agent of contagious bovine pleuropneumonia (CBPP). Genome Res. 14:221-227.
37. Wolfinger, R. D., G. Gibson, E. D. Wolfinger, L. Bennett, H. Hamadeh, P. Bushel, C. Afshari, and R. S. Paules. 2001. Assessing gene significance from cDNA microarray expression data via mixed models. J. Comp. Biol. 8:625-637.
38. Yang, Y. H., S. Dudoit, P. Luu, D. M. Lin, V. Peng, J. Ngai, and T. P. Speed. 2002. Normalization for cDNA microarray data: a robust composite method for addressing single and multiple slide systematic variation. Nucleic Acids Res. 30:15. [Online.]
39. Yura, T., M. Kanemori, and M. Morita. 2000. The heat shock response: regulation and function, p. 3-18. In R. Storz and R. Hengge-Aronis (ed.), Bacterial stress response. ASM Press, Washington, D.C.
40. Zhang, Q., T. F. Young, and R. F. Ross. 1995. Identification and characterization of a Mycoplasma hyopneumoniae adhesin. Infect. Immun. 63:1013-1019.(Melissa L. Madsen, Dan Ne)