Thymidine-Dependent Staphylococcus aureus Small-Colony Variants Are Associated with Extensive Alterations in Regulator and Virulence Gene Ex
http://www.100md.com
感染与免疫杂志 2005年第7期
Institute of Medical of Microbiology, University of Muenster, Muenster 48149, Germany
Department of Microbiology, Dartmouth Medical School, Hanover, New Hampshire
Institute of Medical Microbiology and Hygiene, University of Saarland Hospital and Clinics, Homburg, Germany
ABSTRACT
Chronic airway infection is a hallmark of cystic fibrosis (CF) and many CF patients are infected persistently by Staphylococcus aureus. Thymidine-dependent trimethoprim-sulfamethoxazole (SXT)-resistant S. aureus small-colony variants (SCVs), often in combination with isogenic normal S. aureus phenotypes, are highly prevalent and persistent in airway secretions of CF patients due to long-term SXT therapy (B. Kahl, M. Herrmann, A. S. Everding, H. G. Koch, K. Becker, E. Harms, R. A. Proctor, and G. Peters, J. Infect. Dis. 177:1023-1029, 1998). In this report, SCVs were compared to normal S. aureus by transcription analysis of important regulator (sigB, sarA, and agr) and virulence (-hemolysin, hla, and protein A, spa) genes. Growth curve analyses revealed longer doubling times and lower final densities for SCVs than for normal strains. sigB activity was measured by transcription analysis of the sigB target gene asp23. For nearly all SCVs, expression of all regulators was decreased as assessed by asp23 reverse transcription-PCR for sigB and Northern analysis for sarA and agr. These results are in agreement with diminished hla signals in all SCVs and increased spa signals in 5 of 10 SCVs compared to the isogenic normal S. aureus. Both supplementation of SCVs with thymidine and activation of the agr quorum-sensing system by the supernatant of the isogenic normal strain reversed transcription to almost normal levels. In conclusion, multiple changes in growth characteristics and in regulator and virulence gene expression render SCVs less virulent and allow them to survive in the hostile environment present in the airways of CF patients, thereby illustrating adaptation of the bacteria during long-term persistence.
INTRODUCTION
Staphylococcus aureus is a major human pathogen responsible for a variety of community-acquired and nosocomial infections ranging from mild skin diseases, abscesses, and soft-tissue diseases to life-threatening infections such as osteomyelitis, endocarditis, pneumonia, and sepsis (19). Cystic fibrosis (CF) is characterized by chronic suppurative airway disease with progressive pulmonary insufficiency (16). Depending on the study, 50 to 80% of children and adolescents with CF in particular are chronically colonized or infected by S. aureus (4, 5, 9, 13, 15, 23). However, the pathogenesis of persistent S. aureus infection in CF patients is complex and not fully elucidated.
The persistence of S. aureus in CF and other persistent infections such as osteomyelitis and device-related infections has been associated with the isolation of S. aureus small-colony variants (SCVs) (11, 15, 27). In contrast to the normal S. aureus phenotype, SCVs grow as tiny, nonpigmented, and nonhemolytic colonies. SCVs (i) produce greatly reduced amounts of -hemolysin; (ii) persist within host cells in in vitro assays (26); (iii) are auxotrophic for substrates such as menadione, hemin, thiamine, or thymidine; and (iv) can revert to their normal phenotype (1, 28). In a 6-year prospective study, persistence of clonal SCVs and/or normal S. aureus has been shown in the airways of CF patients over extended periods, as disclosed by pulsed-field gel electrophoresis (15). From this study, 212 SCVs were available for auxotrophism testing. A total of 122 SCVs were thymidine dependent, and 26 SCVs were both thymidine and hemin dependent. As a result of long-term trimethoprim-sulfamethoxazole (SXT) treatment, SXT-resistant and thymidine-dependent SCVs emerged, while the corresponding normal strains remained SXT susceptible (13). While SXT interferes with the tetrahydrofolic acid pathway, tetrahydrofolic acid acts as a coenzyme for thymidylate synthetase, which catalyzes the synthesis of dTMP from dUMP (25). Since dTMP is essential for DNA synthesis, susceptible S. aureus strains are affected by SXT therapy. However, thymidine-dependent SCVs are resistant to SXT and survive if extracellular thymidine is provided. Phenotypically, thymidine-dependent SCVs display two different colony types, (i) fried-egg SCVs with translucent edges surrounding a smaller, elevated pigmented center and (ii) pinpoint colonies, which are nearly 10 times smaller than the normal S. aureus colony (14).
S. aureus has the capability to synthesize a variety of extracellular and cell wall-bound proteins, many of which are involved in pathogenesis (22). The expression of most of these virulence factors is under the control of global regulators such as agr (accessory gene regulator) (17) and sarA (staphylococcal accessory regulator) (7). The agr system of S. aureus represents a quorum-sensing system. A secreted autoinducing peptide (AIP) induces activation of the agr operon at a given threshold, thereby up-regulating extracellular protein production while down-regulating synthesis of cell wall-associated proteins, which serve as adhesins for the pathogen to host tissues (17). The sarA locus encodes the DNA binding protein SarA, which activates the synthesis of both extracellular and cell wall-associated proteins via agr-dependent and agr-independent pathways (8). The sarA locus is composed of three overlapping transcripts with sarA P1 (0.58 kb) and sarA P2 (1.15 kb) maximally transcribed during the exponential phase (tapering toward the stationary phase), while sarA P3 (0.84 kb) is expressed primarily during the late-exponential and stationary phases (2). Based on transcriptional analysis, the sarA P3 promoter has been shown to be sigB dependent (10, 20). In contrast to the primary sigma factor SigA, which is responsible for the transcription of housekeeping genes, the alternative sigma factor SigB (sigB) plays an important role in regulating gene expression upon environmental stresses (e.g., stationary phase, heat, and osmotic stresses).
While the phenotypic characteristics of clinical and laboratory SCVs have been described by several groups (21, 28, 29), a detailed analysis of alteration in gene regulation putatively contributing to or resulting in the SCV phenotype has not been attempted. Therefore, in this report, we characterized the transcription patterns of important regulator (sigB, sarA, and agr) and virulence genes (hla and spa) of long-term persisting thymidine-dependent S. aureus SCVs isolated from the airways of CF patients in comparison to the normal isogenic strain, which was isolated at the same time.
MATERIALS AND METHODS
Bacterial strains and growth conditions. Ten normal and S. aureus SCV strain pairs previously described were used for further analysis (15). These strain pairs were cultured from individual CF patients persistently infected by S. aureus. All SCVs were thymidine dependent and clonal to their normal counterpart, as determined by pulsed-field gel electrophoresis as described previously (13). To ensure stability of the SCV phenotype, the first four strain pairs were subcultured at least 10 times in the laboratory. To confirm the transcription patterns of these pairs, which have been repetitively subcultured in the laboratory, six other isogenic thymidine-dependent SCV-normal strain pairs from other patients were chosen for further analysis. These strain pairs were subcultured only once after primary storage at –70°C. Five SCVs displayed the fried-egg phenotype with translucent edges surrounding a smaller, elevated pigmented center; the other five SCVs exhibited pinpoint colonies, which are nearly 10 times smaller than the normal S. aureus colony size.
Normal S. aureus isolates and SCVs were grown in brain heart infusion (BHI; Merck, Darmstadt, Germany) broth, which supports the growth of SCVs, while Escherichia coli was grown in Luria-Bertani broth. E. coli InvF' (Invitrogen, Karlsruhe, Germany) was used as a host strain for the TA cloning vectors pCRII and pCR 2.1. For the growth curve analysis, cultures were grown at 35°C on a rotary shaker. For the determination of the optical density at 578 nm (OD578) beyond a value of 2, the value of the ODs was calculated from readings of diluted cultures. Ampicillin was used at a concentration of 50 μg/ml. For supplementation studies, thymidine (Fluka Chemie, Buchs, Switzerland) was added to BHI broth at a concentration of 100 μg/ml.
DNA probes. The internal fragments of the following genes were amplified by PCR (Perkin Elmer 2400, überlingen, Germany): sarA, nucleotides (nt) 864 to 1218 (GenBank accession no. U46541) RNAIII, nt 1260 to 1571 (GenBank accession no. X52543); hla, nt 583 to 1211 (GenBank accession no. X01645); and spa, nt 776 to 1160 (GenBank accession no. X61307). PCR products were cloned into pCRII or pCR 2.1 (Invitrogen), sequenced for control with the ABI PRISM 310 Genetic Analyser (Perkin Elmer, Weiterstadt, Germany), and used as probes, which were generated by the Dig Probe Synthesis kit (Roche, Mannheim, Germany), for Northern blot analysis. Plasmid DNA was extracted with the QIAprep Spin Mini Kit (QIAGEN, Hilden, Germany).
Isolation of RNA and Northern analysis. Prior to RNA extraction, growth curves were performed for normal isolates, SCVs, and SCVs supplemented with thymidine. RNA was extracted according to early log, late log, and stationary phases for the particular strains. Overnight cultures of normal colonies and S. aureus SCVs were set to 0.1 at an optical density at 578 nm (OD578) in BHI and were then allowed to grow to the respective growth phases. For supplementation of SCVs, 100 μg/ml thymidine was added to the BHI broth.
The cells were pelleted, and total cellular RNA was isolated with the FastPrep system (Fast Prep FP120 instrument; Qbiogene, Heidelberg, Germany) (2). Since SCVs have a tendency to revert to normal growth, the growth of SCVs after the shaking liquid culture was controlled by streaking the respective cultures onto Columbia blood agar plates, which were incubated for 24 h at 35°C. The extracted RNA was processed further only if the SCVs did not revert to the normal phenotype or if <30% of the cultures reverted. Ten micrograms of RNA was electrophoresed through a 1.5% 0.66 M formaldehyde gel in morpholinepropanesulfonic acid (MOPS) running buffer (20 mM MOPS, 10 mM sodium acetate, 2 mM EDTA [pH 7.0]). Every gel was controlled for equal loading of RNA. RNA was transferred onto a Hybond N+ membrane (Amersham Pharmacia Biotech Europe GmbH, Braunschweig, Germany) under mild alkaline conditions with a turboblotter system (Schleicher and Schuell, Dassel, Germany). RNA was fixed to the membrane by baking (80°C for 0.5 h), hybridized under aqueous conditions at 65°C overnight with the digoxigenin-labeled DNA probes, washed, and detected by chemiluminescence (Roche, Mannheim, Germany). One fried-egg SCV and one pinpoint SCV with their respectie isogenic normal strains were chosen for optimal representation and reproducibility of the transcription patterns of the different tested strain pairs.
asp23 transcription by reverse transcription-PCR. For RNA isolation from culture, S. aureus was grown in BHI medium to the desired growth phase and controlled for potential reversion of the phenotype as described above. Bacteria were mechanically disrupted (Fast Prep FP120 instrument; Qbiogene, Heidelberg, Germany), and RNA was isolated (RNeasy mini kit; QIAGEN, Hilden, Germany). After treatment with RNase-free DNase I (QIAGEN), total RNA samples were amplified in an ABI PRISM 7000 SDS with SYBR Green PCR Master Mix (Applied Biosystems, Weiterstadt, Germany) and gyrB primers (gyrB f1, 5'-GACTGATGCCGATGTGGA-3'; gyrB r1, 5'-AACGGTGGCTGTGCAATA-3') to check for the absence of genomic DNA. Previously transcribed cDNA served as a positive control. RNA was then reverse transcribed (High Capacity cDNA Archive kit; Applied Biosystems). cDNA was used for real-time amplification with asp23-specific primers (asp23 fl, 5'-CAAGAACAAAATCAAGAGCCTCAAT-3'; asp23 r1, 5'-CTTCACGTGCAGCGATACCA-3') and 100 ng of cDNA per reaction mixture. The level of mRNA expression of asp23 was normalized against the internal control gyrB expression, which is constitutively expressed (12). The amount of asp23 transcript was expressed as the n-fold expression difference of asp23 relative to the expression of the control gene gyrB (2–CT, where CT represents the difference in threshold cycle between the target and control genes). All determinations were performed in duplicate.
agr activation of S. aureus SCVs by the isogenic normal S. aureus. For the activation of SCVs by the AIP of the isogenic normal strain, the culture supernatant (SN) from the normal strain was prepared after overnight culture and sterile filtration. Bacteria were inoculated from an overnight culture to an initial OD578 of 0.1 in BHI and grown for 1 h at 35°C. One-tenth of the sterile SN or 1/10 of BHI (as a control) was added to 9/10 of the culture, after which the bacteria were allowed to grow for another 2 h or 4 h. RNA was prepared as described above. The activity of agr was assessed by Northern blot analysis with an RNAIII probe as described above.
RESULTS
SCVs have an extended lag phase and do not reach the densities of normal strains. For each normal colony/SCV strain pair, growth curves were analyzed. Five of the tested SCVs belonged to the fried-egg phenotype, while the other five SCVs exhibited pinpoint colonies (Table 1). All SCVs differed in their growth phase characteristics from normal strains (Fig. 1A and 2A) by extended lag phases (mean difference from the normal S. aureus colony, 2.85 h; range, 1 to 6 h) and lower final densities (mean OD at 578 nm [OD578], 4.5; range, 2 to 8 compared to a mean OD578 of 12.3; range, 9 to 14 for the normal S. aureus colony) (Table 1). The maximum densities for pinpoint SCVs were even lower than for fried-egg SCVs (Table 1; Fig. 1A and 2A). If SCVs were supplemented with thymidine, the bacteria yielded higher densities than nonsupplemented SCVs. There was a difference in growth after supplementation between the fried-egg and pinpoint phenotypes. While the maximum density of most fried-egg SCVs was reached within <12 h (four of five SCVs) upon supplementation with thymidine, the maximum optical density of all pinpoint SCVs (five of five) was not reached within this period and was determined for two pinpoint SCVs at 19 h, as shown in Fig. 2A for one particular pinpoint SCV.
Expression of sigB, sarA, and agr is decreased in SCVs. As SigB represents an alternate sigma factor, which reacts to environmental stress especially in response to growth phase, heat shock, or starvation, we evaluated transcription of this important regulator. Since the activity of SigB is determined by the amount of free SigB dissociated from its anti-sigma factor RsbW, SigB activity is best determined by expression of asp23, a SigB target gene (3). asp23 transcription was expressed as fold increase compared to the expression of the constitutively transcribed gyrB gene. The expression of asp23 was decreased in eight of nine SCVs tested, compared to isogenic normal strains (Table 1). Only SCV 2 expressed nearly the same levels of asp23 as the normal strain. Interestingly, there was a high rate of variation of asp23 activity among the different normal strains, ranging from 1.5- to 178-fold increase over gyrB gene expression during the stationary phase of normal strains. Since sigB activity is important for the pigmentation of S. aureus, we determined the color of the strains grown on Columbia blood agar plates (Table 1). Seven of 10 SCVs displayed fewer pigmented colonies than the normal strains. These differences in pigmentation between normal and SCVs were consistent with low levels of asp23 in SCVs. However, normal strain 9, which had the highest activity of asp23, displayed white colonies on the agar plate, indicating that additional factors also have an impact on pigmentation (Table 1). The asp23 activity of nine SCVs was additionally tested by supplementing the medium with 100 μg/ml of thymidine. Five of these SCVs were fully restored and three were partially restored in their asp23 activity, while only one SCV was not supplemented (data not shown).
To assess the activity of the sarA regulon in SCVs versus normal strains, we analyzed sarA transcription by Northern blot analysis. The normal clinical strains displayed quite typical sarA transcription patterns: sarA P1 and sarA P2 were maximally expressed during early log phase, declining towards stationary phase, while sarA P3 was most active during stationary phase (Fig. 1B and 2B). sarA P3 expression of the SCVs was not detectable or maximal during stationary phase and always decreased compared to the level of the normal S. aureus (Fig. 1B and 2B) in 9 of 10 SCVs tested (Table 1). Only SCV 4 displayed the same sarA P3 transcription level as the normal S. aureus (Table 1). The differential effect of SCVs and normal S. aureus phenotypes was less apparent on sarA P1 and sarA P2 transcription. Upon supplementation with thymidine, transcription for sarA P3 was partially restored in two of seven SCVs and fully restored in five of seven SCVs (Fig. 1B and 2B).
To determine the activity of the agr regulon, which is partly dependent on sarA activity, we consecutively used an RNAIII probe for Northern analysis. While agr transcription was highest in late exponential and stationary phases in normal S. aureus strains, agr transcription was absent or decreased in eight of nine SCVs (Fig. 1B and 2B). SCV 6 displayed a stronger RNAIII transcript than the normal strain 6, whereas no RNAIII transcript was detectable for either phenotype of strain pair 1 (Table 1). In four of seven SCVs supplemented with thymidine, RNAIII transcription was fully restored, as shown in the representative strain pairs (Fig. 1B and 2B) and partially restored in SCV 2 (data not shown). Upon thymidine supplementation of SCV 6, which displayed a higher level of RNAIII transcript than the normal S. aureus, the RNAIII transcript was highly expressed during all growth phases (data not shown).
-Hemolysin and protein A transcription are altered in SCVs compared to normal S. aureus. The presence and transcription of sarA- and agr-dependent virulence genes such as hla and spa were analyzed. Despite the presence of the -hemolysin gene (hla) in all studied SCVs and normal strains as determined by Southern blot analysis (data not shown), transcription of hla was found to be significantly lower in all SCVs than in the normal cohort strains (Table 1; Fig. 1B and 2B). As anticipated, transcription of hla was maximal in late logarithmic and stationary phases in normal strains, while transcription of hla was decreased or almost absent in all SCVs (Fig. 1B and 2B; Table 1). Upon supplementation with thymidine, hla expression was fully restored in all SCVs (Fig. 1B and 2B).
Southern blot analysis disclosed the presence of spa in all strains (data not shown). In exponential growth phase, transcription of spa in SCVs was comparable to the normal strains (Fig. 1B and 2B). spa transcription of the normal strains declined towards late log phase, and no spa signal was detectable during stationary phase. In contrast, spa signals of many SCVs increased toward late log phase and were still present or even increased during stationary phase in five of nine SCVs (Fig. 1B and 2B). In SCVs 6 and 8, spa signals were weaker than in the normal strains, while in SCVs 1 and 2 no difference in spa transcription between the SCV and normal phenotype was apparent (Table 1). Strain pair 3 did not express spa (Table 1). The transcription results concerning spa in SCV and normal strain pairs were consistent with the expression of Spa, as determined by Western blot analysis after extraction of cell wall-associated proteins, which were probed with chicken anti-protein A antibodies (data not shown). Upon supplementation with thymidine, transcription of spa was fully restored for SCVs 5 and 10 (Fig. 1B and 2B) and partially restored for SCV 7 (data not shown). No difference in transcription was seen in SCVs 1, 2, and 8. In strain pair 6 with the enhanced spa signal in the normal strain, the thymidine-supplemented SCV expressed spa only during the early log phase, while the SCV expressed spa at all time points.
agr activation of an SCV by the AIP of the normal S. aureus strain. At a certain threshold, S. aureus secretes an AIP growth phase dependently into the medium to communicate and synchronize the expression of virulence genes. During growth phases, SCVs reached only low cell densities. Therefore, we hypothesized that the threshold for the activation of the agr regulon may not be reached in SCVs. Accordingly, we analyzed whether the AIP of the normal strain is able to activate the agr response of SCVs. To test this, the SN of an isogenic normal S. aureus was prepared after overnight culture and added to an early growth phase culture of normal and S. aureus SCVs, which were further grown in the presence or absence of the AIP. The activity of agr was determined by Northern blot analysis. While no RNAIII signal was detected after 2 h or 4 h in the SCV grown in BHI only, the addition of the AIP of the normal strain activated the agr response of the SCV strain at 2 h and 4 h (Fig. 3), indicating a potentially functional agr locus.
DISCUSSION
Thymidine-dependent SCVs are highly prevalent and persistent in airway secretions of CF patients if the patients are treated or have been treated by long-term SXT (11, 13, 15). In this report, we demonstrated profound alterations of growth characteristics and regulator and effector gene expression in clinical thymidine-dependent SCVs compared to their isogenic normal S. aureus strains isolated at the same time. Both fried-egg and pinpoint SCV phenotypes displayed slower growth curves than the isogenic normal S. aureus strains. Interestingly, the two phenotypes of SCVs could be distinguished not only by their different morphotypes but also by their growth characteristics, because pinpoint SCVs had even longer doubling times and reached much lower densities than the fried-egg SCVs. Additionally, differences between these phenotypes were also obvious upon supplementation with thymidine, since the growth curves of most fried-egg SCVs reverted earlier to the growth characteristics of a normal phenotype upon supplementation than the pinpoint SCVs, probably due to faster growth and earlier uptake of external thymidine. However, transcription profiles of both phenotypes were comparable, depending on the respective background of the normal strain.
The alternative sigma factor SigB represents a powerful regulatory effector in response to environmental stress (18). Using asp23 transcription as a marker for sigB activity, we assessed sigB activity of SCVs versus isogenic normal S. aureus strains. Since environmental changes such as long-term antibiotic therapy have resulted in the SCV phenotype, we expected an increased stress response in SCVs. In contrast to our expectations, the expression of sigB was decreased compared to the normal S. aureus strains, probably due to the action of other upstream-located regulators or as an indication that limited sources of thymidine, which are crucial for the survival of thymidine-dependent SCVs, are not sensed by sigB. Low levels of sigB expression were consistent with the nonpigmentation of SCVs in comparison to the normal S. aureus. The sarA P3 promoter has been shown to be under the control of sigB (2, 10, 20). The regulatory network of the sigB-sarA-agr regulators is very complex. While Chien et al. have shown that functional SarA controls agr expression (8), Bischoff et al. demonstrated that sigB increases sarA expression while RNAIII levels were reduced simultaneously (3). The data of these two groups indicate that sigB has sarA-dependent as well as sarA-independent effects upon agr transcription. Our data are consistent in view of the sigB-sarA hierarchy resulting in a decreased level of sarA P3 transcription, presumably due to low sigB activity in SCVs compared to the normal S. aureus strains, as shown for 9 of 10 SCVs. Furthermore, low levels of agr transcription in SCVs are consistent with low levels of sarA P3 expression. Bischoff et al. demonstrated that low sigB transcription resulted in up-regulation of agr. However, agr transcription in SCVs was almost absent, indicating sigB-independent regulation of agr in SCVs.
The fact that thymidine supplementation of SCV 6 resulted in enhanced RNAIII transcription compared to the very weak RNAIII transcript present in the respective isogenic normal strain, which was isolated at the same time, suggests that this SCV emerged earlier from an isogenic normal strain, which presumably displayed strong agr activity. Interestingly, the isogenic normal strain, which was isolated at the same time as the SCV, apparently underwent changes of the agr regulon during long-term persistence, resulting in an almost negative agr status. Changes in the agr regulon have been shown to occur during serial passages in vitro, due to mutations in the nucleotide sequence of agr (24). It is conceivable that such changes also occur due to long-term persistence of S. aureus in the host. The assumptions that the SCV emerged from an earlier normal strain with an active agr regulon are also consistent concerning the spa transcription, because after supplementation of the SCV, the spa transcript was strongest during the early log phase and not present in the later growth phases (in contrast to the spa transcript of the isogenic normal strain isolated at the same time, which also expressed a strong signal during stationary phase).
As with SCVs associated with hemin and/or menadione auxotrophy (1, 28), thymidine-dependent SCVs do not readily produce -hemolysin on Columbia agar (14). Although hla was present in SCVs as confirmed by Southern blotting, the expression of hla was strongly down-regulated in SCVs at the transcriptional level. These results are not surprising, considering the decrease in sarA activity, as well as in agr activity, which concertedly controls hla expression. Although it has been shown that sigB affects hla transcription, resulting in increased hemolysis in a sigB mutant strain, low levels of sigB activity in SCVs were not associated with a hemolytic phenotype (6), indicating sigB-independent regulation of hla in these SCVs.
Taken together, thymidine-dependent SCVs display an agr mutant phenotype in light of an inactive agr locus (RNAIII) associated with increased spa transcription and decreased hla transcription. Interestingly, upon thymidine supplementation of SCVs, transcription of the agr locus and the agr-dependent virulence genes spa and hla were fully or partially restored in most SCVs, confirming that the agr locus was potentially functional in SCVs. Moreover, it was possible to activate the agr response of SCVs by the AIP of the isogenic normal S. aureus strain, suggesting that the agr system was inactive, probably due to the decreased growth rate and low density of AIP present in SCVs. The fact that sigB and sarA expression, which is independent of agr, was also restored after thymidine supplementation points to other regulatory mechanisms responsible for the various alterations that occur in thymidine-dependent SCVs.
In conclusion, multiple changes in growth characteristics and in regulator and virulence gene expression render SCVs less virulent and allow them to survive in the hostile environment present in the airways of CF patients, thereby illustrating adaptation of the bacteria during long-term persistence.
ACKNOWLEDGMENTS
This work was supported by the Innovative Medical Research Grant of the Medical Faculty, University of Muenster, grant Ka-1-6-I/98-45 (to B.C.K.); by a grant from the German Minister of Education and Research for the Interdisciplinary Center for Clinical Research, grant C8 (to G.P. and M.H.); by grants from the University of Saarland (2003) and from the Medical Faculty of the University of Saarland (2004) (to M.H.); by grants SPP1047 and SPP1130 from the Deutsche Forschungsgemeinschaft, (to M.H.); and by a grant from the NIH, grant RO1-AI37142 (to A.L.C.).
We thank Barbara Ritzerfeld and Barbara Grünastel for expert and enthusiastic technical assistance.
REFERENCES
1. Balwit, J. M., P. van Langefelde, J. M. Vann, and R. A. Proctor. 1994. Gentamicin-resistant menadione and hemin auxotrophic Staphylococcus aureus persist within cultured endothelial cells. J. Infect. Dis. 170:1033-1037.
2. Bayer, M. G., J. H. Heinrichs, and A. L. Cheung. 1996. The molecular architecture of the sar locus in Staphylococcus aureus. J. Bacteriol. 178:4563-4570.
3. Bischoff, M., J. M. Entenza, and P. Giachino. 2001. Influence of a functional sigB operon on the global regulators sar and agr in Staphylococcus aureus. J. Bacteriol. 183:5171-5179.
4. Branger, C., C. Gardye, and N. Lambert-Zechovsky. 1996. Persistence of Staphylococcus aureus strains among cystic fibrosis patients over extended periods of time. J. Med. Microbiol. 45:294-301.
5. Burns, J. L., J. Emerson, J. R. Stapp, D. L. Yim, J. Krzewinski, L. Louden, B. W. Ramsey, and C. R. Clausen. 1998. Microbiology of sputum from patients at cystic fibrosis centers in the United States. Clin. Infect. Dis. 27:158-163.
6. Cheung, A. L., Y. T. Chien, and A. S. Bayer. 1999. Hyperproduction of alpha-hemolysin in a sigB mutant is associated with elevated SarA expression in Staphylococcus aureus. Infect. Immun. 67:1331-1337.
7. Cheung, A. L., J. M. Koomey, C. A. Butler, and S. J. Projan. 1992. Regulation of exoprotein expression in Staphylcoccus aureus by a locus (sar) distinct from agr. Proc. Natl. Acad. Sci. USA 89:6462-6466.
8. Chien, Y. T., and A. L. Cheung. 1998. Molecular interactions between two global regulators, sar and agr, in Staphlococcus aureus. J. Biol. Chem. 273:2645-2652.
9. Cystic Fibrosis Foundation. Patient registry 2003 annual report. Cystic Fibrosis Foundation, Bethesda, Md.
10. Deora, R., T. Tseng, and T. K. Misra. 1997. Alternative transcription vector SB of Staphylococcus aureus: characterization and role in transcription of the global regulatory locus sar. J. Bacteriol. 179:6355-6359.
11. Gilligan, P. H., P. A. Gage, D. F. Welch, M. J. Muszynski, and K. R. Wait. 1987. Prevalence of thymidin-dependent Staphylococcus aureus in patients with cystic fibrosis. J. Clin. Microbiol. 25:1258-1261.
12. Goerke, C., S. Campana, M. G. Bayer, G. Dring, K. Botzenhart, and C. Wolz. 2000. Direct quantitative transcript analysis of the agr regulon of Staphylococcus aureus during human infection in comparison to the expression profile in vitro. Infect. Immun. 68:1304-1311.
13. Kahl, B., M. Herrmann, A. S. Everding, H. G. Koch, K. Becker, E. Harms, R. A. Proctor, and G. Peters. 1998. Persistent infection with small colony variant strains of Staphylococcus aureus in patients with cystic fibrosis. J. Infect. Dis. 177:1023-1029.
14. Kahl, B. C., G. Belling, R. Reichelt, M. Herrmann, R. A. Proctor, and G. Peters. 2003. Thymidine-dependent small-colony variants of Staphylococcus aureus exhibit gross morphological and ultrastructural changes consistent with impaired cell separation. J. Clin. Microbiol. 41:410-413.
15. Kahl, B. C., A. Duebbers, G. Lubritz, J. Haeberle, H. G. Koch, B. Ritzerfeld, M. Reilly, E. Harms, R. A. Proctor, M. Herrmann, and G. Peters. 2003. Population dynamics of persistent Staphylococcus aureus isolated from the airways of cystic fibrosis patients during a 6-year prospective study. J. Clin. Microbiol. 41:4424-4427.
16. Knowles, M. R., P. H. Gilligan, and R. C. Boucher. 2000. Cystic fibrosis, p. 767-772. In G. L. Mandell, J. E. Bennet, and R. Dolin (ed.), Principles and practice of infectious diseases. Churchill Livingstone, New York, N.Y.
17. Kornblum, J., B. N. Kreiswirth, S. J. Projan, H. Ross, and R. P. Novick. 1990. agr: a polycistronic locus regulating exoprotein synthesis in Staphylococcus aureus, p. 373-402. In R. P. Novick (ed.), Molecular biology of the staphylococci. VCH Publishers, Inc., New York. N.Y.
18. Kullik, I., and P. Giachino. 1997. The alternative sigma factor B in Staphylococcus aureus: regulation of the sigB operon in response to growth phase and heat shock. Arch. Microbiol. 167:151-159.
19. Lowy, F. D. 1998. Staphylococcus aureus infections. N. Engl. J. Med. 339:520-532.
20. Manna, A. C., M. G. Bayer, and A. L. Cheung. 1998. Transcriptional analysis of different promoters in the sar locus in Staphylococcus aureus. J. Bacteriol. 180:3828-3836.
21. Proctor, R. A., J. M. Balwit, and O. Vesga. 1994. Variant subpopulations of Staphylococcus aureus as cause of persistent and recurrent infections. Infect. Agents Dis. 3:302-312.
22. Projan, S. J., and R. P. Novick. 1997. The molecular basis of pathogenicity, p. 55-81. In K. B. Crossley and G. L. Archer (ed.), The staphylococci in human disease. Churchill Livingstone, New York, N.Y.
23. Renders, N. H. M., A. van Belkum, S. E. Overbeek, J. W. Mouton, and H. A. Verbrugh. 1997. Molecular epidemiology of Staphylococcus aureus strains colonizing the lungs of related and unrelated cystic fibrosis patients. Clin. Microbiol. Infect. 3:216-221.
24. Somerville, G. A., S. B. Beres, J. R. Fitzgerald, F. R. DeLeo, R. L. Cloe, J. S. Hoff, and J. M. Musser. 2002. In vitro serial passage of Staphylococcus aureus: changes in physiology, virulence factor production, and agr nucleotide sequence. J. Bacteriol. 184:1430-1437.
25. Stryer, L. 1995. Biosynthesis of nucleotides, p. 739-762. In L. Stryer (ed.), Biochemistry. W. H. Freeman and Company, New York, N.Y.
26. von Eiff, C., K. Becker, D. Metze, G. Lubritz, J. Hockmann, T. Schwarz, and G. Peters. 2001. Intracellular persistence of Staphylococcus aureus small-colony variants within keratinocytes: a cause for antibiotic treatment failure in a patient with Darier's disease. Clin. Infect. Dis. 32:1643-1647.
27. von Eiff, C., D. Bettin, R. A. Proctor, B. Rolaufs, N. Lindner, W. Winkelmann, and G. Peters. 1997. Recovery of small colony variants of Staphylococcus aureus following gentamicin bead placement for osteomyelitis. Clin. Infect. Dis. 25:1250-1251.
28. von Eiff, C., C. Heilmann, R. A. Proctor, C. Woltz, G. Peters, and F. Goetz. 1997. A site-directed Staphylococcus aureus hemB mutant is a small-colony variant which persists intracellularly. J. Bacteriol. 179:4706-4712.
29. von Eiff, C., P. E. Vaudaux, B. C. Kahl, D. P. Lew, S. Emler, A. Schmidt, G. Peters, and R. A. Proctor. 1999. Bloodstream infections caused by small-colony variants of coagulase-negative staphylococci following pacemaker implantation. Clin. Infect. Dis. 29:932-934.(Thymidine-Dependent Staph)
Department of Microbiology, Dartmouth Medical School, Hanover, New Hampshire
Institute of Medical Microbiology and Hygiene, University of Saarland Hospital and Clinics, Homburg, Germany
ABSTRACT
Chronic airway infection is a hallmark of cystic fibrosis (CF) and many CF patients are infected persistently by Staphylococcus aureus. Thymidine-dependent trimethoprim-sulfamethoxazole (SXT)-resistant S. aureus small-colony variants (SCVs), often in combination with isogenic normal S. aureus phenotypes, are highly prevalent and persistent in airway secretions of CF patients due to long-term SXT therapy (B. Kahl, M. Herrmann, A. S. Everding, H. G. Koch, K. Becker, E. Harms, R. A. Proctor, and G. Peters, J. Infect. Dis. 177:1023-1029, 1998). In this report, SCVs were compared to normal S. aureus by transcription analysis of important regulator (sigB, sarA, and agr) and virulence (-hemolysin, hla, and protein A, spa) genes. Growth curve analyses revealed longer doubling times and lower final densities for SCVs than for normal strains. sigB activity was measured by transcription analysis of the sigB target gene asp23. For nearly all SCVs, expression of all regulators was decreased as assessed by asp23 reverse transcription-PCR for sigB and Northern analysis for sarA and agr. These results are in agreement with diminished hla signals in all SCVs and increased spa signals in 5 of 10 SCVs compared to the isogenic normal S. aureus. Both supplementation of SCVs with thymidine and activation of the agr quorum-sensing system by the supernatant of the isogenic normal strain reversed transcription to almost normal levels. In conclusion, multiple changes in growth characteristics and in regulator and virulence gene expression render SCVs less virulent and allow them to survive in the hostile environment present in the airways of CF patients, thereby illustrating adaptation of the bacteria during long-term persistence.
INTRODUCTION
Staphylococcus aureus is a major human pathogen responsible for a variety of community-acquired and nosocomial infections ranging from mild skin diseases, abscesses, and soft-tissue diseases to life-threatening infections such as osteomyelitis, endocarditis, pneumonia, and sepsis (19). Cystic fibrosis (CF) is characterized by chronic suppurative airway disease with progressive pulmonary insufficiency (16). Depending on the study, 50 to 80% of children and adolescents with CF in particular are chronically colonized or infected by S. aureus (4, 5, 9, 13, 15, 23). However, the pathogenesis of persistent S. aureus infection in CF patients is complex and not fully elucidated.
The persistence of S. aureus in CF and other persistent infections such as osteomyelitis and device-related infections has been associated with the isolation of S. aureus small-colony variants (SCVs) (11, 15, 27). In contrast to the normal S. aureus phenotype, SCVs grow as tiny, nonpigmented, and nonhemolytic colonies. SCVs (i) produce greatly reduced amounts of -hemolysin; (ii) persist within host cells in in vitro assays (26); (iii) are auxotrophic for substrates such as menadione, hemin, thiamine, or thymidine; and (iv) can revert to their normal phenotype (1, 28). In a 6-year prospective study, persistence of clonal SCVs and/or normal S. aureus has been shown in the airways of CF patients over extended periods, as disclosed by pulsed-field gel electrophoresis (15). From this study, 212 SCVs were available for auxotrophism testing. A total of 122 SCVs were thymidine dependent, and 26 SCVs were both thymidine and hemin dependent. As a result of long-term trimethoprim-sulfamethoxazole (SXT) treatment, SXT-resistant and thymidine-dependent SCVs emerged, while the corresponding normal strains remained SXT susceptible (13). While SXT interferes with the tetrahydrofolic acid pathway, tetrahydrofolic acid acts as a coenzyme for thymidylate synthetase, which catalyzes the synthesis of dTMP from dUMP (25). Since dTMP is essential for DNA synthesis, susceptible S. aureus strains are affected by SXT therapy. However, thymidine-dependent SCVs are resistant to SXT and survive if extracellular thymidine is provided. Phenotypically, thymidine-dependent SCVs display two different colony types, (i) fried-egg SCVs with translucent edges surrounding a smaller, elevated pigmented center and (ii) pinpoint colonies, which are nearly 10 times smaller than the normal S. aureus colony (14).
S. aureus has the capability to synthesize a variety of extracellular and cell wall-bound proteins, many of which are involved in pathogenesis (22). The expression of most of these virulence factors is under the control of global regulators such as agr (accessory gene regulator) (17) and sarA (staphylococcal accessory regulator) (7). The agr system of S. aureus represents a quorum-sensing system. A secreted autoinducing peptide (AIP) induces activation of the agr operon at a given threshold, thereby up-regulating extracellular protein production while down-regulating synthesis of cell wall-associated proteins, which serve as adhesins for the pathogen to host tissues (17). The sarA locus encodes the DNA binding protein SarA, which activates the synthesis of both extracellular and cell wall-associated proteins via agr-dependent and agr-independent pathways (8). The sarA locus is composed of three overlapping transcripts with sarA P1 (0.58 kb) and sarA P2 (1.15 kb) maximally transcribed during the exponential phase (tapering toward the stationary phase), while sarA P3 (0.84 kb) is expressed primarily during the late-exponential and stationary phases (2). Based on transcriptional analysis, the sarA P3 promoter has been shown to be sigB dependent (10, 20). In contrast to the primary sigma factor SigA, which is responsible for the transcription of housekeeping genes, the alternative sigma factor SigB (sigB) plays an important role in regulating gene expression upon environmental stresses (e.g., stationary phase, heat, and osmotic stresses).
While the phenotypic characteristics of clinical and laboratory SCVs have been described by several groups (21, 28, 29), a detailed analysis of alteration in gene regulation putatively contributing to or resulting in the SCV phenotype has not been attempted. Therefore, in this report, we characterized the transcription patterns of important regulator (sigB, sarA, and agr) and virulence genes (hla and spa) of long-term persisting thymidine-dependent S. aureus SCVs isolated from the airways of CF patients in comparison to the normal isogenic strain, which was isolated at the same time.
MATERIALS AND METHODS
Bacterial strains and growth conditions. Ten normal and S. aureus SCV strain pairs previously described were used for further analysis (15). These strain pairs were cultured from individual CF patients persistently infected by S. aureus. All SCVs were thymidine dependent and clonal to their normal counterpart, as determined by pulsed-field gel electrophoresis as described previously (13). To ensure stability of the SCV phenotype, the first four strain pairs were subcultured at least 10 times in the laboratory. To confirm the transcription patterns of these pairs, which have been repetitively subcultured in the laboratory, six other isogenic thymidine-dependent SCV-normal strain pairs from other patients were chosen for further analysis. These strain pairs were subcultured only once after primary storage at –70°C. Five SCVs displayed the fried-egg phenotype with translucent edges surrounding a smaller, elevated pigmented center; the other five SCVs exhibited pinpoint colonies, which are nearly 10 times smaller than the normal S. aureus colony size.
Normal S. aureus isolates and SCVs were grown in brain heart infusion (BHI; Merck, Darmstadt, Germany) broth, which supports the growth of SCVs, while Escherichia coli was grown in Luria-Bertani broth. E. coli InvF' (Invitrogen, Karlsruhe, Germany) was used as a host strain for the TA cloning vectors pCRII and pCR 2.1. For the growth curve analysis, cultures were grown at 35°C on a rotary shaker. For the determination of the optical density at 578 nm (OD578) beyond a value of 2, the value of the ODs was calculated from readings of diluted cultures. Ampicillin was used at a concentration of 50 μg/ml. For supplementation studies, thymidine (Fluka Chemie, Buchs, Switzerland) was added to BHI broth at a concentration of 100 μg/ml.
DNA probes. The internal fragments of the following genes were amplified by PCR (Perkin Elmer 2400, überlingen, Germany): sarA, nucleotides (nt) 864 to 1218 (GenBank accession no. U46541) RNAIII, nt 1260 to 1571 (GenBank accession no. X52543); hla, nt 583 to 1211 (GenBank accession no. X01645); and spa, nt 776 to 1160 (GenBank accession no. X61307). PCR products were cloned into pCRII or pCR 2.1 (Invitrogen), sequenced for control with the ABI PRISM 310 Genetic Analyser (Perkin Elmer, Weiterstadt, Germany), and used as probes, which were generated by the Dig Probe Synthesis kit (Roche, Mannheim, Germany), for Northern blot analysis. Plasmid DNA was extracted with the QIAprep Spin Mini Kit (QIAGEN, Hilden, Germany).
Isolation of RNA and Northern analysis. Prior to RNA extraction, growth curves were performed for normal isolates, SCVs, and SCVs supplemented with thymidine. RNA was extracted according to early log, late log, and stationary phases for the particular strains. Overnight cultures of normal colonies and S. aureus SCVs were set to 0.1 at an optical density at 578 nm (OD578) in BHI and were then allowed to grow to the respective growth phases. For supplementation of SCVs, 100 μg/ml thymidine was added to the BHI broth.
The cells were pelleted, and total cellular RNA was isolated with the FastPrep system (Fast Prep FP120 instrument; Qbiogene, Heidelberg, Germany) (2). Since SCVs have a tendency to revert to normal growth, the growth of SCVs after the shaking liquid culture was controlled by streaking the respective cultures onto Columbia blood agar plates, which were incubated for 24 h at 35°C. The extracted RNA was processed further only if the SCVs did not revert to the normal phenotype or if <30% of the cultures reverted. Ten micrograms of RNA was electrophoresed through a 1.5% 0.66 M formaldehyde gel in morpholinepropanesulfonic acid (MOPS) running buffer (20 mM MOPS, 10 mM sodium acetate, 2 mM EDTA [pH 7.0]). Every gel was controlled for equal loading of RNA. RNA was transferred onto a Hybond N+ membrane (Amersham Pharmacia Biotech Europe GmbH, Braunschweig, Germany) under mild alkaline conditions with a turboblotter system (Schleicher and Schuell, Dassel, Germany). RNA was fixed to the membrane by baking (80°C for 0.5 h), hybridized under aqueous conditions at 65°C overnight with the digoxigenin-labeled DNA probes, washed, and detected by chemiluminescence (Roche, Mannheim, Germany). One fried-egg SCV and one pinpoint SCV with their respectie isogenic normal strains were chosen for optimal representation and reproducibility of the transcription patterns of the different tested strain pairs.
asp23 transcription by reverse transcription-PCR. For RNA isolation from culture, S. aureus was grown in BHI medium to the desired growth phase and controlled for potential reversion of the phenotype as described above. Bacteria were mechanically disrupted (Fast Prep FP120 instrument; Qbiogene, Heidelberg, Germany), and RNA was isolated (RNeasy mini kit; QIAGEN, Hilden, Germany). After treatment with RNase-free DNase I (QIAGEN), total RNA samples were amplified in an ABI PRISM 7000 SDS with SYBR Green PCR Master Mix (Applied Biosystems, Weiterstadt, Germany) and gyrB primers (gyrB f1, 5'-GACTGATGCCGATGTGGA-3'; gyrB r1, 5'-AACGGTGGCTGTGCAATA-3') to check for the absence of genomic DNA. Previously transcribed cDNA served as a positive control. RNA was then reverse transcribed (High Capacity cDNA Archive kit; Applied Biosystems). cDNA was used for real-time amplification with asp23-specific primers (asp23 fl, 5'-CAAGAACAAAATCAAGAGCCTCAAT-3'; asp23 r1, 5'-CTTCACGTGCAGCGATACCA-3') and 100 ng of cDNA per reaction mixture. The level of mRNA expression of asp23 was normalized against the internal control gyrB expression, which is constitutively expressed (12). The amount of asp23 transcript was expressed as the n-fold expression difference of asp23 relative to the expression of the control gene gyrB (2–CT, where CT represents the difference in threshold cycle between the target and control genes). All determinations were performed in duplicate.
agr activation of S. aureus SCVs by the isogenic normal S. aureus. For the activation of SCVs by the AIP of the isogenic normal strain, the culture supernatant (SN) from the normal strain was prepared after overnight culture and sterile filtration. Bacteria were inoculated from an overnight culture to an initial OD578 of 0.1 in BHI and grown for 1 h at 35°C. One-tenth of the sterile SN or 1/10 of BHI (as a control) was added to 9/10 of the culture, after which the bacteria were allowed to grow for another 2 h or 4 h. RNA was prepared as described above. The activity of agr was assessed by Northern blot analysis with an RNAIII probe as described above.
RESULTS
SCVs have an extended lag phase and do not reach the densities of normal strains. For each normal colony/SCV strain pair, growth curves were analyzed. Five of the tested SCVs belonged to the fried-egg phenotype, while the other five SCVs exhibited pinpoint colonies (Table 1). All SCVs differed in their growth phase characteristics from normal strains (Fig. 1A and 2A) by extended lag phases (mean difference from the normal S. aureus colony, 2.85 h; range, 1 to 6 h) and lower final densities (mean OD at 578 nm [OD578], 4.5; range, 2 to 8 compared to a mean OD578 of 12.3; range, 9 to 14 for the normal S. aureus colony) (Table 1). The maximum densities for pinpoint SCVs were even lower than for fried-egg SCVs (Table 1; Fig. 1A and 2A). If SCVs were supplemented with thymidine, the bacteria yielded higher densities than nonsupplemented SCVs. There was a difference in growth after supplementation between the fried-egg and pinpoint phenotypes. While the maximum density of most fried-egg SCVs was reached within <12 h (four of five SCVs) upon supplementation with thymidine, the maximum optical density of all pinpoint SCVs (five of five) was not reached within this period and was determined for two pinpoint SCVs at 19 h, as shown in Fig. 2A for one particular pinpoint SCV.
Expression of sigB, sarA, and agr is decreased in SCVs. As SigB represents an alternate sigma factor, which reacts to environmental stress especially in response to growth phase, heat shock, or starvation, we evaluated transcription of this important regulator. Since the activity of SigB is determined by the amount of free SigB dissociated from its anti-sigma factor RsbW, SigB activity is best determined by expression of asp23, a SigB target gene (3). asp23 transcription was expressed as fold increase compared to the expression of the constitutively transcribed gyrB gene. The expression of asp23 was decreased in eight of nine SCVs tested, compared to isogenic normal strains (Table 1). Only SCV 2 expressed nearly the same levels of asp23 as the normal strain. Interestingly, there was a high rate of variation of asp23 activity among the different normal strains, ranging from 1.5- to 178-fold increase over gyrB gene expression during the stationary phase of normal strains. Since sigB activity is important for the pigmentation of S. aureus, we determined the color of the strains grown on Columbia blood agar plates (Table 1). Seven of 10 SCVs displayed fewer pigmented colonies than the normal strains. These differences in pigmentation between normal and SCVs were consistent with low levels of asp23 in SCVs. However, normal strain 9, which had the highest activity of asp23, displayed white colonies on the agar plate, indicating that additional factors also have an impact on pigmentation (Table 1). The asp23 activity of nine SCVs was additionally tested by supplementing the medium with 100 μg/ml of thymidine. Five of these SCVs were fully restored and three were partially restored in their asp23 activity, while only one SCV was not supplemented (data not shown).
To assess the activity of the sarA regulon in SCVs versus normal strains, we analyzed sarA transcription by Northern blot analysis. The normal clinical strains displayed quite typical sarA transcription patterns: sarA P1 and sarA P2 were maximally expressed during early log phase, declining towards stationary phase, while sarA P3 was most active during stationary phase (Fig. 1B and 2B). sarA P3 expression of the SCVs was not detectable or maximal during stationary phase and always decreased compared to the level of the normal S. aureus (Fig. 1B and 2B) in 9 of 10 SCVs tested (Table 1). Only SCV 4 displayed the same sarA P3 transcription level as the normal S. aureus (Table 1). The differential effect of SCVs and normal S. aureus phenotypes was less apparent on sarA P1 and sarA P2 transcription. Upon supplementation with thymidine, transcription for sarA P3 was partially restored in two of seven SCVs and fully restored in five of seven SCVs (Fig. 1B and 2B).
To determine the activity of the agr regulon, which is partly dependent on sarA activity, we consecutively used an RNAIII probe for Northern analysis. While agr transcription was highest in late exponential and stationary phases in normal S. aureus strains, agr transcription was absent or decreased in eight of nine SCVs (Fig. 1B and 2B). SCV 6 displayed a stronger RNAIII transcript than the normal strain 6, whereas no RNAIII transcript was detectable for either phenotype of strain pair 1 (Table 1). In four of seven SCVs supplemented with thymidine, RNAIII transcription was fully restored, as shown in the representative strain pairs (Fig. 1B and 2B) and partially restored in SCV 2 (data not shown). Upon thymidine supplementation of SCV 6, which displayed a higher level of RNAIII transcript than the normal S. aureus, the RNAIII transcript was highly expressed during all growth phases (data not shown).
-Hemolysin and protein A transcription are altered in SCVs compared to normal S. aureus. The presence and transcription of sarA- and agr-dependent virulence genes such as hla and spa were analyzed. Despite the presence of the -hemolysin gene (hla) in all studied SCVs and normal strains as determined by Southern blot analysis (data not shown), transcription of hla was found to be significantly lower in all SCVs than in the normal cohort strains (Table 1; Fig. 1B and 2B). As anticipated, transcription of hla was maximal in late logarithmic and stationary phases in normal strains, while transcription of hla was decreased or almost absent in all SCVs (Fig. 1B and 2B; Table 1). Upon supplementation with thymidine, hla expression was fully restored in all SCVs (Fig. 1B and 2B).
Southern blot analysis disclosed the presence of spa in all strains (data not shown). In exponential growth phase, transcription of spa in SCVs was comparable to the normal strains (Fig. 1B and 2B). spa transcription of the normal strains declined towards late log phase, and no spa signal was detectable during stationary phase. In contrast, spa signals of many SCVs increased toward late log phase and were still present or even increased during stationary phase in five of nine SCVs (Fig. 1B and 2B). In SCVs 6 and 8, spa signals were weaker than in the normal strains, while in SCVs 1 and 2 no difference in spa transcription between the SCV and normal phenotype was apparent (Table 1). Strain pair 3 did not express spa (Table 1). The transcription results concerning spa in SCV and normal strain pairs were consistent with the expression of Spa, as determined by Western blot analysis after extraction of cell wall-associated proteins, which were probed with chicken anti-protein A antibodies (data not shown). Upon supplementation with thymidine, transcription of spa was fully restored for SCVs 5 and 10 (Fig. 1B and 2B) and partially restored for SCV 7 (data not shown). No difference in transcription was seen in SCVs 1, 2, and 8. In strain pair 6 with the enhanced spa signal in the normal strain, the thymidine-supplemented SCV expressed spa only during the early log phase, while the SCV expressed spa at all time points.
agr activation of an SCV by the AIP of the normal S. aureus strain. At a certain threshold, S. aureus secretes an AIP growth phase dependently into the medium to communicate and synchronize the expression of virulence genes. During growth phases, SCVs reached only low cell densities. Therefore, we hypothesized that the threshold for the activation of the agr regulon may not be reached in SCVs. Accordingly, we analyzed whether the AIP of the normal strain is able to activate the agr response of SCVs. To test this, the SN of an isogenic normal S. aureus was prepared after overnight culture and added to an early growth phase culture of normal and S. aureus SCVs, which were further grown in the presence or absence of the AIP. The activity of agr was determined by Northern blot analysis. While no RNAIII signal was detected after 2 h or 4 h in the SCV grown in BHI only, the addition of the AIP of the normal strain activated the agr response of the SCV strain at 2 h and 4 h (Fig. 3), indicating a potentially functional agr locus.
DISCUSSION
Thymidine-dependent SCVs are highly prevalent and persistent in airway secretions of CF patients if the patients are treated or have been treated by long-term SXT (11, 13, 15). In this report, we demonstrated profound alterations of growth characteristics and regulator and effector gene expression in clinical thymidine-dependent SCVs compared to their isogenic normal S. aureus strains isolated at the same time. Both fried-egg and pinpoint SCV phenotypes displayed slower growth curves than the isogenic normal S. aureus strains. Interestingly, the two phenotypes of SCVs could be distinguished not only by their different morphotypes but also by their growth characteristics, because pinpoint SCVs had even longer doubling times and reached much lower densities than the fried-egg SCVs. Additionally, differences between these phenotypes were also obvious upon supplementation with thymidine, since the growth curves of most fried-egg SCVs reverted earlier to the growth characteristics of a normal phenotype upon supplementation than the pinpoint SCVs, probably due to faster growth and earlier uptake of external thymidine. However, transcription profiles of both phenotypes were comparable, depending on the respective background of the normal strain.
The alternative sigma factor SigB represents a powerful regulatory effector in response to environmental stress (18). Using asp23 transcription as a marker for sigB activity, we assessed sigB activity of SCVs versus isogenic normal S. aureus strains. Since environmental changes such as long-term antibiotic therapy have resulted in the SCV phenotype, we expected an increased stress response in SCVs. In contrast to our expectations, the expression of sigB was decreased compared to the normal S. aureus strains, probably due to the action of other upstream-located regulators or as an indication that limited sources of thymidine, which are crucial for the survival of thymidine-dependent SCVs, are not sensed by sigB. Low levels of sigB expression were consistent with the nonpigmentation of SCVs in comparison to the normal S. aureus. The sarA P3 promoter has been shown to be under the control of sigB (2, 10, 20). The regulatory network of the sigB-sarA-agr regulators is very complex. While Chien et al. have shown that functional SarA controls agr expression (8), Bischoff et al. demonstrated that sigB increases sarA expression while RNAIII levels were reduced simultaneously (3). The data of these two groups indicate that sigB has sarA-dependent as well as sarA-independent effects upon agr transcription. Our data are consistent in view of the sigB-sarA hierarchy resulting in a decreased level of sarA P3 transcription, presumably due to low sigB activity in SCVs compared to the normal S. aureus strains, as shown for 9 of 10 SCVs. Furthermore, low levels of agr transcription in SCVs are consistent with low levels of sarA P3 expression. Bischoff et al. demonstrated that low sigB transcription resulted in up-regulation of agr. However, agr transcription in SCVs was almost absent, indicating sigB-independent regulation of agr in SCVs.
The fact that thymidine supplementation of SCV 6 resulted in enhanced RNAIII transcription compared to the very weak RNAIII transcript present in the respective isogenic normal strain, which was isolated at the same time, suggests that this SCV emerged earlier from an isogenic normal strain, which presumably displayed strong agr activity. Interestingly, the isogenic normal strain, which was isolated at the same time as the SCV, apparently underwent changes of the agr regulon during long-term persistence, resulting in an almost negative agr status. Changes in the agr regulon have been shown to occur during serial passages in vitro, due to mutations in the nucleotide sequence of agr (24). It is conceivable that such changes also occur due to long-term persistence of S. aureus in the host. The assumptions that the SCV emerged from an earlier normal strain with an active agr regulon are also consistent concerning the spa transcription, because after supplementation of the SCV, the spa transcript was strongest during the early log phase and not present in the later growth phases (in contrast to the spa transcript of the isogenic normal strain isolated at the same time, which also expressed a strong signal during stationary phase).
As with SCVs associated with hemin and/or menadione auxotrophy (1, 28), thymidine-dependent SCVs do not readily produce -hemolysin on Columbia agar (14). Although hla was present in SCVs as confirmed by Southern blotting, the expression of hla was strongly down-regulated in SCVs at the transcriptional level. These results are not surprising, considering the decrease in sarA activity, as well as in agr activity, which concertedly controls hla expression. Although it has been shown that sigB affects hla transcription, resulting in increased hemolysis in a sigB mutant strain, low levels of sigB activity in SCVs were not associated with a hemolytic phenotype (6), indicating sigB-independent regulation of hla in these SCVs.
Taken together, thymidine-dependent SCVs display an agr mutant phenotype in light of an inactive agr locus (RNAIII) associated with increased spa transcription and decreased hla transcription. Interestingly, upon thymidine supplementation of SCVs, transcription of the agr locus and the agr-dependent virulence genes spa and hla were fully or partially restored in most SCVs, confirming that the agr locus was potentially functional in SCVs. Moreover, it was possible to activate the agr response of SCVs by the AIP of the isogenic normal S. aureus strain, suggesting that the agr system was inactive, probably due to the decreased growth rate and low density of AIP present in SCVs. The fact that sigB and sarA expression, which is independent of agr, was also restored after thymidine supplementation points to other regulatory mechanisms responsible for the various alterations that occur in thymidine-dependent SCVs.
In conclusion, multiple changes in growth characteristics and in regulator and virulence gene expression render SCVs less virulent and allow them to survive in the hostile environment present in the airways of CF patients, thereby illustrating adaptation of the bacteria during long-term persistence.
ACKNOWLEDGMENTS
This work was supported by the Innovative Medical Research Grant of the Medical Faculty, University of Muenster, grant Ka-1-6-I/98-45 (to B.C.K.); by a grant from the German Minister of Education and Research for the Interdisciplinary Center for Clinical Research, grant C8 (to G.P. and M.H.); by grants from the University of Saarland (2003) and from the Medical Faculty of the University of Saarland (2004) (to M.H.); by grants SPP1047 and SPP1130 from the Deutsche Forschungsgemeinschaft, (to M.H.); and by a grant from the NIH, grant RO1-AI37142 (to A.L.C.).
We thank Barbara Ritzerfeld and Barbara Grünastel for expert and enthusiastic technical assistance.
REFERENCES
1. Balwit, J. M., P. van Langefelde, J. M. Vann, and R. A. Proctor. 1994. Gentamicin-resistant menadione and hemin auxotrophic Staphylococcus aureus persist within cultured endothelial cells. J. Infect. Dis. 170:1033-1037.
2. Bayer, M. G., J. H. Heinrichs, and A. L. Cheung. 1996. The molecular architecture of the sar locus in Staphylococcus aureus. J. Bacteriol. 178:4563-4570.
3. Bischoff, M., J. M. Entenza, and P. Giachino. 2001. Influence of a functional sigB operon on the global regulators sar and agr in Staphylococcus aureus. J. Bacteriol. 183:5171-5179.
4. Branger, C., C. Gardye, and N. Lambert-Zechovsky. 1996. Persistence of Staphylococcus aureus strains among cystic fibrosis patients over extended periods of time. J. Med. Microbiol. 45:294-301.
5. Burns, J. L., J. Emerson, J. R. Stapp, D. L. Yim, J. Krzewinski, L. Louden, B. W. Ramsey, and C. R. Clausen. 1998. Microbiology of sputum from patients at cystic fibrosis centers in the United States. Clin. Infect. Dis. 27:158-163.
6. Cheung, A. L., Y. T. Chien, and A. S. Bayer. 1999. Hyperproduction of alpha-hemolysin in a sigB mutant is associated with elevated SarA expression in Staphylococcus aureus. Infect. Immun. 67:1331-1337.
7. Cheung, A. L., J. M. Koomey, C. A. Butler, and S. J. Projan. 1992. Regulation of exoprotein expression in Staphylcoccus aureus by a locus (sar) distinct from agr. Proc. Natl. Acad. Sci. USA 89:6462-6466.
8. Chien, Y. T., and A. L. Cheung. 1998. Molecular interactions between two global regulators, sar and agr, in Staphlococcus aureus. J. Biol. Chem. 273:2645-2652.
9. Cystic Fibrosis Foundation. Patient registry 2003 annual report. Cystic Fibrosis Foundation, Bethesda, Md.
10. Deora, R., T. Tseng, and T. K. Misra. 1997. Alternative transcription vector SB of Staphylococcus aureus: characterization and role in transcription of the global regulatory locus sar. J. Bacteriol. 179:6355-6359.
11. Gilligan, P. H., P. A. Gage, D. F. Welch, M. J. Muszynski, and K. R. Wait. 1987. Prevalence of thymidin-dependent Staphylococcus aureus in patients with cystic fibrosis. J. Clin. Microbiol. 25:1258-1261.
12. Goerke, C., S. Campana, M. G. Bayer, G. Dring, K. Botzenhart, and C. Wolz. 2000. Direct quantitative transcript analysis of the agr regulon of Staphylococcus aureus during human infection in comparison to the expression profile in vitro. Infect. Immun. 68:1304-1311.
13. Kahl, B., M. Herrmann, A. S. Everding, H. G. Koch, K. Becker, E. Harms, R. A. Proctor, and G. Peters. 1998. Persistent infection with small colony variant strains of Staphylococcus aureus in patients with cystic fibrosis. J. Infect. Dis. 177:1023-1029.
14. Kahl, B. C., G. Belling, R. Reichelt, M. Herrmann, R. A. Proctor, and G. Peters. 2003. Thymidine-dependent small-colony variants of Staphylococcus aureus exhibit gross morphological and ultrastructural changes consistent with impaired cell separation. J. Clin. Microbiol. 41:410-413.
15. Kahl, B. C., A. Duebbers, G. Lubritz, J. Haeberle, H. G. Koch, B. Ritzerfeld, M. Reilly, E. Harms, R. A. Proctor, M. Herrmann, and G. Peters. 2003. Population dynamics of persistent Staphylococcus aureus isolated from the airways of cystic fibrosis patients during a 6-year prospective study. J. Clin. Microbiol. 41:4424-4427.
16. Knowles, M. R., P. H. Gilligan, and R. C. Boucher. 2000. Cystic fibrosis, p. 767-772. In G. L. Mandell, J. E. Bennet, and R. Dolin (ed.), Principles and practice of infectious diseases. Churchill Livingstone, New York, N.Y.
17. Kornblum, J., B. N. Kreiswirth, S. J. Projan, H. Ross, and R. P. Novick. 1990. agr: a polycistronic locus regulating exoprotein synthesis in Staphylococcus aureus, p. 373-402. In R. P. Novick (ed.), Molecular biology of the staphylococci. VCH Publishers, Inc., New York. N.Y.
18. Kullik, I., and P. Giachino. 1997. The alternative sigma factor B in Staphylococcus aureus: regulation of the sigB operon in response to growth phase and heat shock. Arch. Microbiol. 167:151-159.
19. Lowy, F. D. 1998. Staphylococcus aureus infections. N. Engl. J. Med. 339:520-532.
20. Manna, A. C., M. G. Bayer, and A. L. Cheung. 1998. Transcriptional analysis of different promoters in the sar locus in Staphylococcus aureus. J. Bacteriol. 180:3828-3836.
21. Proctor, R. A., J. M. Balwit, and O. Vesga. 1994. Variant subpopulations of Staphylococcus aureus as cause of persistent and recurrent infections. Infect. Agents Dis. 3:302-312.
22. Projan, S. J., and R. P. Novick. 1997. The molecular basis of pathogenicity, p. 55-81. In K. B. Crossley and G. L. Archer (ed.), The staphylococci in human disease. Churchill Livingstone, New York, N.Y.
23. Renders, N. H. M., A. van Belkum, S. E. Overbeek, J. W. Mouton, and H. A. Verbrugh. 1997. Molecular epidemiology of Staphylococcus aureus strains colonizing the lungs of related and unrelated cystic fibrosis patients. Clin. Microbiol. Infect. 3:216-221.
24. Somerville, G. A., S. B. Beres, J. R. Fitzgerald, F. R. DeLeo, R. L. Cloe, J. S. Hoff, and J. M. Musser. 2002. In vitro serial passage of Staphylococcus aureus: changes in physiology, virulence factor production, and agr nucleotide sequence. J. Bacteriol. 184:1430-1437.
25. Stryer, L. 1995. Biosynthesis of nucleotides, p. 739-762. In L. Stryer (ed.), Biochemistry. W. H. Freeman and Company, New York, N.Y.
26. von Eiff, C., K. Becker, D. Metze, G. Lubritz, J. Hockmann, T. Schwarz, and G. Peters. 2001. Intracellular persistence of Staphylococcus aureus small-colony variants within keratinocytes: a cause for antibiotic treatment failure in a patient with Darier's disease. Clin. Infect. Dis. 32:1643-1647.
27. von Eiff, C., D. Bettin, R. A. Proctor, B. Rolaufs, N. Lindner, W. Winkelmann, and G. Peters. 1997. Recovery of small colony variants of Staphylococcus aureus following gentamicin bead placement for osteomyelitis. Clin. Infect. Dis. 25:1250-1251.
28. von Eiff, C., C. Heilmann, R. A. Proctor, C. Woltz, G. Peters, and F. Goetz. 1997. A site-directed Staphylococcus aureus hemB mutant is a small-colony variant which persists intracellularly. J. Bacteriol. 179:4706-4712.
29. von Eiff, C., P. E. Vaudaux, B. C. Kahl, D. P. Lew, S. Emler, A. Schmidt, G. Peters, and R. A. Proctor. 1999. Bloodstream infections caused by small-colony variants of coagulase-negative staphylococci following pacemaker implantation. Clin. Infect. Dis. 29:932-934.(Thymidine-Dependent Staph)