ABSTRACT
The siderophores staphyloferrin A (SA) and staphyloferrin B (SB) of Staphylococcus aureus are essential for iron acquisition in the iron-restricted environment of the host, such as in subcutaneous abscesses. SA and SB are secreted by SfaA and SbnD transporters, respectively. To assess the further function of SfaA and SbnD in S. aureus fitness, we tested its effect on murine abscess models and intracellular replication in epithelial cells. Bacterial fitness in abscesses and in epithelial cells was studied, by comparing the parental strains RN6390 and MW2 and their ΔsfaA and ΔsbnD mutants using competition assays in a murine abscess model and invasion and replication assays with human lung adenocarcinoma cell line A549. In the murine abscess model using equal inocula of a ΔsfaA or ΔsbnD mutant and the wild-type RN6390 strain, the ΔsfaA mutant exhibited growth defects of 2.2-fold. Additionally, replication of the ΔsfaA mutant within A549 cells was decreased 3.0-fold. In complementation experiments, the ΔsfaA mutant carrying plasmid-borne sfaA restored the growth fitness in abscesses and epithelial cells. The ΔsbnD mutant, in contrast, showed no growth defect in either abscesses or epithelial cells. Our findings demonstrate that the efflux transporter of the siderophore SA contributes to the ability of S. aureus to replicate in abscesses and epithelial cells. Furthermore, fitness of S. aureus in these sites of replication is not compromised by the absence of transporter SbnD.
KEYWORDS: bacterial fitness, SbnD, SfaA, siderophore transporter, Staphylococcus aureus
INTRODUCTION
Staphylococcus aureus is a common colonizer of the anterior nares and skin in humans and often causes skin infections. A number of efflux transporters conferring antimicrobial resistance are present in S. aureus. Of these transporters, we previously demonstrated that the expression levels of norB (resistance to quinolones), norD, and tet38 (resistance to tetracycline) genes were increased in murine subcutaneous abscesses and that these genes contribute to S. aureus survival in abscesses and on skin (1, 2). Furthermore, we found that antimicrobial fatty acids are natural substrates of Tet38 (3). NorB, NorD, and Tet38 are members of the major facilitator superfamily (MFS)-type efflux pump encoded on the S. aureus chromosome and are energized by proton motive force (2, 4).
Iron is required by living organisms, including S. aureus and other bacterial pathogens (5). The vertebrate host tightly regulates free-iron levels and sequesters this valuable nutrient as a mechanism to prevent bacterial proliferation. The host is immediately able to remove iron from those sites if a pathogen is detected (6). Extracellular free iron is rapidly removed by transferrin and lactoferrin, proteins with a high affinity for iron (7). To combat this sequestration, S. aureus elaborates siderophores, small molecules with exceptionally high affinity for iron. Siderophores target and remove iron that is bound to transferrin and lactoferrin (8). S. aureus synthesizes two siderophores, staphyloferrin A (SA) and staphyloferrin B (SB), and both are secreted into the extracellular milieu. The biosynthesis of SA and SB has been well characterized within the last decade, and the enzymes for the generation and export of the two staphyloferrins are encoded by the sfaABCD and sbnABCDEFGHI operons, respectively (9, 10). Efflux of SA is mediated by SfaA, a membrane transport protein of the MFS; disruption of sfaA abrogates export of the siderophore, resulting in its intracellular accumulation (8, 11). SB is secreted into the extracellular milieu in part by SbnD, another MFS efflux protein, and by an alternate as-yet-unidentified exporter. Sheldon and Heinrichs demonstrated that S. aureus strains unable to synthesize SA are significantly impaired for macroscopic subcutaneous abscess formation (8). However, the contribution of SfaA and SbnD transporters to the ability of S. aureus cells to survive and replicate within abscesses and epithelial cells has not been studied.
To assess further the function of SfaA and SbnD transporters in S. aureus fitness in vivo, we compared isogenic sfaA and sbnD mutants and their parental strain in murine abscess models and for cellular invasion and intracellular replication in the A549 epithelial cell line.
RESULTS
Antimicrobial susceptibility of overexpressors and knockout mutants of sfaA and sbnD.
Because some efflux pumps have broad substrate profiles and include antibiotics in addition to their natural substrates (3), we determined the effects of the S. aureus efflux transporters encoded by sfaA and sbnD on the antimicrobial activity of 20 antimicrobials by comparing the wild type, overexpressors, and knockout mutants (Table 1). Disruption of sfaA or sbnD resulted in a ≤2-fold increase in susceptibility to all of the tested antimicrobials, including fatty acids and antimicrobial polyamines, which are substrates of some resistance pumps. We used S. aureus strain MW2 as the host of overexpressors, because we were unable to construct overexpressors of sfaA and sbnD in RN6390. A growth defect was found not only in MW2 containing plasmids overexpressing sfaA or sbnD but also in MW2 carrying the empty plasmid vector, causing decreased MICs of ampicillin, gentamicin, polymyxin B, colistin, rifampin, undecanoic acid, and palmitoleic acid. No increase in the resistance to any of the tested antimicrobials was found in the overexpressors of sfaA and sbnD.
TABLE 1.
Antimicrobial susceptibility of overexpressors and knockout mutants of sfaA and sbnD
| Strain | MIC (μg/ml)a |
|||||||||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| Amp | Cef | Gen | Nor | Cip | Tet | Ery | Cm | Pol | Col | Rif | Van | Lin | Spe | Und | Pal | Chl | Tpp | Rho | Et | |
| RN6390 | 0.125 | 0.125 | 0.5 | 0.5 | 0.25 | 0.5 | 0.5 | 8 | 128 | 2,048 | 0.004 | 1 | 2 | 1,024 | 1,024 | 256 | 0.5 | 32 | 1 | 8 |
| RN6390 ΔsfaA | 0.125 | 0.125 | 1 | 1 | 0.25 | 0.5 | 0.5 | 8 | 128 | 2,048 | 0.008 | 1 | 2 | 1,024 | 1,024 | 512 | 0.5 | 32 | 2 | 4 |
| RN6390 ΔsbnD | 0.125 | 0.25 | 0.25 | 0.5 | 0.25 | 0.5 | 0.5 | 8 | 64 | 1,024 | 0.008 | 1 | 2 | 1,024 | 1,024 | 512 | 0.5 | 16 | 2 | 8 |
| MW2 | 512 | 256 | 4 | 1 | 0.5 | 0.5 | 0.5 | 16 | 128 | 1,024 | 0.008 | 1 | 4 | 2,048 | 1,024 | 128 | 1 | 16 | 1 | 8 |
| MW2(vector) | 128 | 256 | 4 | 1 | 0.25 | 0.25 | 0.5 | 16 | 64 | 512 | 0.004 | 1 | 4 | 1,024 | 256 | 32 | 1 | 16 | 0.5 | 4 |
| MW2 sfaA++ | 256 | 256 | 1 | 1 | 0.25 | 0.25 | 0.5 | 8 | 64 | 512 | 0.002 | 1 | 2 | 2,048 | 256 | 32 | 0.5 | 16 | 0.5 | 4 |
| MW2 sbnD++ | 256 | 128 | 1 | 1 | 0.5 | 0.25 | 0.25 | 8 | 32 | 256 | 0.002 | 1 | 2 | 1,024 | 512 | 32 | 0.5 | 16 | 0.5 | 4 |
Amp, ampicillin; Cef, cephalexin; Gen, gentamicin; Nor, norfloxacin; Cip, ciprofloxacin; Tet, tetracycline; Ery, erythromycin; Cm, chloramphenicol; Pol, polymyxin B; Col, colistin; Rif, rifampin; Van, vancomycin; Lin, linezolid; Spe, spermidine; Und, undecanoic acid; Pal, palmitoleic acid; Chl, chlorhexidine; Tpp, tetraphenyl phosphonium; Rho, rhodamine; Et, ethidium bromide.
Expression of sfaA and sbnD in subcutaneous abscesses.
We previously demonstrated that the expression of some efflux pump genes in S. aureus was upregulated in murine subcutaneous abscesses and that these genes contribute to S. aureus survival in abscesses and on skin (1, 2). Thus, we infected mice with strain RN6390 and assessed the relative expression of sfaA and sbnD in total bacterial RNA recovered from an abscess versus that of in vitro cultures using real-time quantitative reverse transcription-PCR (qRT-PCR). Compared to their expression in vitro in Trypticase soy broth (TSB), sfaA and sbnD were significantly upregulated 58.1- and 16.7-fold in abscesses, respectively (Fig. 1). The upregulation level of sfaA was 3.5-fold higher than that of sbnD.
FIG 1.
Expression of sfaA and sbnD in S. aureus RN6390 recovered from murine subcutaneous abscesses relative to in vitro gene expression. All values are the mean of results ± standard error of the mean (SEM) from three independent experiments.
Competition and single mutant assays in subcutaneous abscesses.
To compare the fitness in murine abscesses between wild-type RN6390 and ΔsfaA or ΔsbnD mutants, we carried out competition assays. In order to score the survival of mutants versus wild-type strains, we used as a competitor a rifampin-resistant strain of RN6390 (RN6390RIF), which exhibited a slight fitness defect in competition assays with RN6390 when the two strains were injected in roughly equal numbers in mouse flanks and harvested at 24 h (competitive index [CI] = 1.71) but which was almost comparable in vitro (CI = 1.13) (Fig. 2).
FIG 2.
Competitive growth of sfaA and sbnD mutants with wild-type S. aureus RN6390 in murine subcutaneous abscesses. (A) Fitness competition between S. aureus RN6390 (wild type), the sfaA-deficient strain (ΔsfaA), the ΔsfaA strain carrying empty vector [ΔsfaA (vector)], and the sfaA-complemented strain [ΔsfaA (pSfaA)]. (B) Fitness competition between S. aureus RN6390 (wild type), the sbnD-deficient strain (ΔsbnD), the ΔsbnD strain carrying empty vector [ΔsbnD (vector)], and the sbnD-complemented strain [ΔsbnD (pSbnD)]. The competitive index (CI) was calculated as the output ratio of two strains divided by the input ratio. Each circle represents the CI for one abscess, and the horizontal bars represent medians for the group. The data shown in panels A and B are the results compiled from three independent experiments. An asterisk indicates that the median CI is significantly different from that of wild-type strain (P < 0.01).
When the competition assay was performed between RN6390RIF and the RN6390 ΔsfaA mutant, the CI (0.79) of the ΔsfaA mutant was 2.2-fold lower than that of RN6390RIF (P < 0.01) (Fig. 2A). In vitro competition assays using TSB showed no difference between RN6390RIF and the ΔsfaA mutant (CI = 1.13). The growth defect was also observed in the ΔsfaA mutant carrying an empty plasmid vector (CI = 0.76). The data strongly suggest that the CI will be further decreased if the competition assay is performed between wild-type RN6390 and the ΔsfaA mutant, because the growth of RN6390RIF in vivo was less than that of the wild type. To complement the fitness defect of the ΔsfaA mutant, we constructed pSfaA and pSbnD carrying the sfaA and sbnD genes, respectively. Both genes were stably expressed from pSfaA and pSbnD, as confirmed by qRT-PCR, and both plasmids were stable in RN6390 and its derivatives when grown in TSB for at least 48 h in the absence of chloramphenicol and during infection for up to 24 h (data not shown). In the competition assay, plasmid-encoded SfaA complemented (CI = 1.64) the fitness defect of the ΔsfaA mutant. In contrast, the ΔsbnD mutant (CI = 2.10) showed no growth defect in vitro or in vivo (Fig. 2B).
We also performed a set of single-strain assays to investigate the fitness difference between RN6390, the ΔsfaA mutant, and the ΔsbnD mutant, in which each of two strains with the same inocula was singly injected on the opposite flanks of mice (Fig. 3). At 24 h after injection, the median ratios of output CFU to input CFU were 47.7 for RN6390 and 26.2 for the ΔsfaA mutant, a 1.8-fold difference (Fig. 3A; P < 0.01). The same result was observed for the ΔsfaA mutant (fold increase = 29.5) carrying the empty plasmid vector. Plasmid-encoded SfaA complemented (fold increase = 48.1) the fitness defect of the ΔsfaA mutant as in the competition assays. In contrast, the ΔsbnD mutant (fold increase = 41.6) showed no growth defect (Fig. 3B).
FIG 3.
Comparison of the growth ratios for S. aureus strains in murine subcutaneous abscesses. (A) Fold change of the input (0 h) and output (24 h) numbers of CFU of RN6390 (wild-type), the sfaA-deficient strain (ΔsfaA), the ΔsfaA strain carrying empty vector [ΔsfaA (vector)], and the sfaA-complemented strain [ΔsfaA (pSfaA)]. (B) Fold change of the input (0 h) and output (24 h) numbers of CFU of RN6390 (wild type) and the sbnD-deficient strain (ΔsbnD). All values are the mean of results ± SEM from three independent experiments. An asterisk indicates that the fold change is significantly different from that of the wild-type strain (P < 0.05).
Effects of sfaA and sbnD on S. aureus fitness in renal abscesses.
In a murine kidney abscess model, an S. aureus sbnE mutant, which is defective for the production of SB, was attenuated compared to its wild-type strain, indicating that siderophore production is important to the survival of S. aureus in murine kidney abscesses (12). To determine whether sfaA or sbnD contributes to S. aureus survival in a murine kidney as well as a subcutaneous abscess, we determined the survival of MW2 and its isogenic mutants, the ΔsfaA and ΔsbnD mutants, in murine kidney abscesses. Renal abscesses formed reliably by 4 days after injection, and kidneys of individual mice injected with S. aureus contained approximately 108 cells. The average numbers for the wild type and the ΔsfaA and ΔsbnD mutants were 8.1 × 108, 2.3 × 108, and 8.4 × 108 CFU/kidney, respectively (Fig. 4). With respect to the average fold change relative to the wild type, that of the ΔsfaA mutant was significantly lower than that of ΔsbnD mutant (P < 0.05). In contrast, no significant difference was found between the ΔsbnD mutant and the parental strain. As in the subcutaneous abscess model, the ΔsbnD mutant did not show a growth defect in the renal abscess model, a finding distinct from earlier findings with a ΔsbnE mutant (12). These results suggest that there is redundancy in SB transport mechanisms and that SbnD is not the only exporter of SB.
FIG 4.
Comparison of the relative fold changes in number of CFU for sfaA and sbnD mutants and wild-type S. aureus MW2 in murine kidney abscesses. All values are the mean of results ± SEM from three independent experiments. An asterisk indicates that the values are significantly different (P < 0.05).
Effects of sfaA and sbnD on S. aureus replication within epithelial cells.
To determine whether SfaA and SbnD affect the ability of S. aureus to replicate within epithelial cells, we developed an intracellular replication assay using epithelial cell line A549. At 45 min of contact with a multiplicity of infection (MOI) of 10:1, wild-type RN6390 initiated growth after 60 min and reached 464 ± 43 CFU/well after 5 h. In contrast, the ΔsfaA mutant had limited replication, ranging from 106 ± 12 to 153 ± 0 CFU/well, and the count was 3.0-fold lower than that of RN6390 (Fig. 5A). The sfaA-complemented strain showed the same replication within epithelial cells as the wild-type strain. On the other hand, the intracellular replication of the ΔsbnD mutant was similar to that of the wild-type strain (Fig. 5B).
FIG 5.
Intracellular replication of S. aureus strains in A549 cells. (A) Viable bacterial counts for RN6390 (wild type) and the sfaA-complemented strain [ΔsfaA (pSfaA)] increased after 60 min. Viable bacterial counts for the sfaA-deficient strain (ΔsfaA) increased less than did those for the wild-type and ΔsfaA (pSfaA) strains. (B) Viable bacterial counts for both wild-type and sbnD-deficient (ΔsbnD) strains increased after 60 min. All values are the mean of results ± standard deviation (SD) from three independent experiments. An asterisk indicates that the numbers of CFU are significantly different from those of the wild-type strain (P < 0.05).
Siderophore bioassay.
To assess more directly the effects of SfaA and SbnD on the export of SA and SB, respectively, we performed siderophore bioassays on the culture supernatants of S. aureus and its ΔsfaA and ΔsbnD mutants. The growth of RN6390 Δsir and Δhts strains depends on the presence of SA and SB, respectively, when added to iron-depleted culture medium (13). The results showed that the supernatant of RN6390 and MW2 wild-type strains could promote the growth of both Δhts and Δsir strains in iron-restricted medium (Tris minimal succinate with Chelex-100 [C-TMS]) (Fig. 6). The supernatant of ΔsfaA mutants promoted the growth of the Δhts strain but not the Δsir strain, indicating that there was very limited SA but an abundance of SB in the supernatant of ΔsfaA mutants. In contrast, the supernatant of ΔsbnD mutants improved the growth of the Δsir mutant as much as its parental strains but also partially improved the growth of the Δhts mutant to 0.3- to 0.5-fold, suggesting that SB secretion is only partially decreased in the supernatant of the ΔsbnD mutant compared to the wild-type strain, a finding consistent with the presence of mechanisms for the export of SB in addition to SbnD.
FIG 6.
Siderophore bioassays for sfaA and sbnD mutants and wild-type S. aureus RN6390 (A) and MW2 (B). In both panels A and B, the left side [labeled SA (Δsir)] is a measure of the presence of staphyloferrin A (SA) in culture supernatants, since spent culture supernatants were supplied to a sir mutant impregnated in broth medium bioassays. The sir mutant is defective for staphyloferrin B (SB) uptake, and therefore growth promotion in this iron-restricted medium relies on supplying SA to the bacteria. In both panels A and B, the right side [labeled SB (Δhts)] is a measure of the presence of SB in culture supernatants, since spent culture supernatants were supplied to an hts mutant impregnated in broth medium bioassays. The hts mutant is defective for SA uptake, and therefore growth promotion in this iron-restricted medium relies on supplying SB to the bacteria. All values are the mean of results ± SEM from three independent experiments. An asterisk indicates that the values are significantly different (P < 0.05).
DISCUSSION
We have demonstrated that the SfaA transporter of SA, but not the SbnD transporter of SB, in S. aureus contributes to bacterial fitness in abscesses and epithelial cells.
Upon entering host tissues, S. aureus responds to the dearth of available free iron by dramatically altering its gene expression profile. This change in gene expression is often mediated by the iron-dependent repressor Fur (ferric uptake regulator) (14). The release of Fur-mediated repression during times of iron starvation results in a coordinated upregulation of a number of genes involved in iron acquisition, glycolysis, and virulence. S. aureus fur mutants exhibit a severe virulence defect in animal models of infection (15). Abscesses represent iron-starved microenvironments inside the host (16). In this study, we demonstrated that the expression levels of sfaA and sbnD were significantly increased in subcutaneous abscesses. Furthermore, we also confirmed that the expression of both sfaA and sbnD was upregulated in iron-restricted medium (TSB supplemented with 600 μM dipyridyl) (data not shown), indicating that the expression of sfaA and sbnD is regulated by iron. Expression of the norA efflux pump gene has also been reported to be regulated by iron. norA, like sfaA and sbnD, encodes an MFS-type efflux pump and is regulated by Fur (9, 12, 17). Fur also indirectly regulates the expression of norD encoding MFS-type efflux pump NorD (2). norD also contributes substantially to S. aureus fitness in the abscess infection model, but its native function is unknown. As in the norD-deficient strain (2), no change in antimicrobial susceptibilities was found in the sfaA- and sbnD-deficient strains. We could not determine accurate antimicrobial susceptibilities for strains overexpressing sfaA or sbnD from plasmid pLZ113, because they grew poorly. The reasons for the failure of the overexpressor constructs in RN6390 and the growth defect of the overexpressors in MW2 are unclear, but it appears that overexpression of SfaA and SbnD is toxic for the bacterial cell. Further study is necessary to determine if SfaA and SbnD export compounds other than SA and SB, respectively.
Glucose in concentrations found in human serum represses SA synthesis in vitro (13, 18). Thus, SA may not be the primary siderophore expressed during staphylococcal bacteremia (8). Recent observations support a role for SA that is more important in skin colonization and infections than in bloodstream infections. Sheldon and Heinrichs showed that S. aureus strains unable to synthesize SA produce smaller abscesses than both the wild-type and SB-deficient strains (8), but they targeted whole sfaABCD and sbnABCDEFGHI operons and the relative numbers of CFU were not reported. Our findings now show that the absence of SfaA transporter alone produces a growth defect in murine abscesses as well as in epithelial cells and that when wild-type and mutant bacteria are both present in an abscess, the fitness defect of the sfaA mutant is reduced but still detectable, likely resulting from the secretion of SA by the parental strain. It has been suggested that intracellular accumulation of SA in an sfa mutant might account for its growth defect (8). The difference in growth defects in competition versus single mutant experiments, however, argues that it is the absence of secreted SA that accounts for the growth defect of the sfaA mutant in the abscess environment.
In contrast to the SfaA-deficient strain, the SbnD-deficient mutant exhibited little or no growth defect in fitness in either abscesses or epithelial cells, suggesting that there is redundancy in the transporter systems, i.e., that other transporters may be able to export SB (11). Since the Sir-deficient mutants could not grow in iron-restricted medium containing the supernatant of an SfaA-deficient strain but the Hts-deficient mutants could grow in that medium with the supernatant of an SbnD-deficient strain, our data indicated that SfaA plays a key role in secretion of SA but suggested that secretion of SB is mediated not only by SbnD but also by the other transporters. The lack of a requirement for SB was also found in mutants deficient in the synthesis of SB (8). Therefore, the SfaA transporter appears to have a particularly important role in iron acquisition in in vivo iron-restricted environments, such as a subcutaneous abscess or kidney abscess or within epithelial cells.
Although many efflux transporters have been shown to have broad substrate profiles and to function to export diverse toxic substances in the environment (19–21), we could find no evidence that SfaA or SbnD can confer resistance by export of natural product or other antimicrobials and disinfectants that have been shown to be substrates of other transporters. The ability of the parental strain to reduce the growth defect of the sfaA mutant in the abscess model in direct competition experiments, in comparison to that in separate comparisons of abscesses of mutant and parental strains, also supports the concept that SfaA exports a beneficial compound rather than removing a toxic one.
MATERIALS AND METHODS
Bacterial strains and growth conditions.
The strains and plasmids used in this study are shown in Table 2. S. aureus strains were cultivated in brain heart infusion (BHI; Difco, Sparks, MD) or Trypticase soy broth (TSB; Difco), and Escherichia coli strains were grown in Luria-Bertani (LB; Difco) broth. Bacteria were grown at 37°C unless stated otherwise. The following antibiotics were obtained from Sigma (St. Louis, MO) and used at the concentrations indicated: ampicillin, 100 μg/ml; chloramphenicol, 10 μg/ml; kanamycin, 50 μg/ml; anhydrotetracycline, 1 μg/ml; and rifampin, 10 μg/ml.
TABLE 2.
Bacterial strains and plasmids used in this study
| Strain or plasmid | Genotype or relevant characteristic(s)a | Reference or source |
|---|---|---|
| Staphylococcus aureus strains | ||
| RN4220 | Restriction-deficient laboratory strain | 29 |
| RN6390 | 8325-4 wild type | 30 |
| RN6390RIF | Rifampin-resistant derivative | This study |
| MW2 | Community-acquired methicillin-resistant strain | 1 |
| MW2(pLZ113) | MW2 carrying empty vector, MW2 (vector) | This study |
| MW2(pLZ113::sfaA) | sfaA overexpressor (sfaA++) | This study |
| MW2(pLZ113::sbnD) | sbnD overexpressor (sbnD++) | This study |
| MW2 ΔsfaA | sfaA knockout mutant | This study |
| MW2 ΔsbnD | sbnD knockout mutant | This study |
| RN6390 ΔsfaA | sfaA knockout mutant | This study |
| RN6390 ΔsbnD | sbnD knockout mutant | This study |
| RN6390 Δsir | Staphyloferrin B transport-deficient mutant | 13 |
| RN6390 Δhts | Staphyloferrin A transport-deficient mutant | 13 |
| RN6390 ΔsfaA+pTZN10 | sfaA knockout mutant carrying empty vector, ΔsfaA (vector) | This study |
| RN6390 ΔsfaA+pTZN10::sfaA | sfaA-complemented strain, ΔsfaA (pSfaA) | This study |
| RN6390 ΔsbnD+pTZN10 | sbnD knockout mutant carrying empty vector, ΔsbnD (vector) | This study |
| RN6390 ΔsbnD+pTZN10::sbnD | sbnD-complemented strain, ΔsbnD (pSbnD) | This study |
| Escherichia coli strains | ||
| TOP10 | hsdR mcrA lacZΔM15 endA1 recA1 | Life Technologies |
| DC10B | DH10B Δdcm; does not methylate DNA on cytosine | 22 |
| Plasmids | ||
| pLZ113 | 6.6-kb overexpression shuttle vector for S. aureus; Ampr Kmr | 23 |
| pIMAY | 5.7-kb allelic exchange vector for S. aureus; Cmr | 22 |
| pTZN10 | 6.4-kb low-copy-number shuttle vector for S. aureus; Ampr Cmr | 4, 24 |
Ampr, ampicillin resistance; Kmr, kanamycin resistance; Cmr, chloramphenicol resistance.
Construction of sfaA- and sbnD-deficient mutants.
We constructed an in-frame deletion of sfaA and sbnD in RN6390 and MW2 based on the technique described by Monk et al. (22). The sequences upstream of the genes to be deleted were amplified with primers A and B (A/B) (up to the start codon) and the downstream sequences with the primers C/D (down from the stop codon) separately (Table 3). The upstream and downstream PCR products were diluted and used as the template in a second spliced overlap extension (SOE) PCR with the A/D primers (Table 3). Deletion constructs were cleaved at restriction sites introduced into primers A and D during PCR and ligated into pIMAY cut with the same enzymes and then transformed into E. coli DC10B. The inserts of the plasmid were confirmed by DNA sequencing. The plasmids were then electroporated into RN6390 and plated onto BHI agar containing chloramphenicol at 28°C. To integrate pIMAY into the chromosome, a single colony from the transformation plates was homogenized in 200 μl of TSB. The suspensions were a series of 10-fold dilutions, and 100 μl of each dilution was spread onto BHI agar containing chloramphenicol and incubated overnight. Large colonies were selected, and the absence of extrachromosomal plasmid was verified by PCR (22). Additionally, we verified whether plasmid integration had occurred upstream or downstream of the gene (OUT F/D Rev primers or OUT R/A Fwd primers) (Table 3). Overnight cultures of both the upstream and downstream crossovers that were free of replicating plasmid were grown at 28°C without chloramphenicol and then plated onto BHI agar containing anhydrotetracycline. The plates were incubated at 28°C for 48 h. Large colonies were patched on BHI agar containing anhydrotetracycline and BHI agar containing chloramphenicol and grown overnight. Chloramphenicol-sensitive colonies were screened by colony PCR with primers (OUT F/OUT R) (Table 3) to identify clones containing the desired mutation. Putative mutants were validated by PCR amplification of genomic DNA flanking the deletion and DNA sequencing.
TABLE 3.
Primers used in this study
| Purpose and primer or amplified gene | Direction | Sequencea | Fragment size (bp)b |
|---|---|---|---|
| sfaA and sbnD knockout | |||
| sfaA-AFwd (KpnI) | Forward | TCTGGTACCAAACTCTCAGAATAAGTTAAATCATCTGTCTTATC | 477 (1,105) |
| sfaA-B | Reverse | TGAATGGGCAAAGGCAGATAGA | |
| sfaA-C | Forward | TCTATCTGCCTTTGCCCATTCACATAAAACTTACACCCGCATTCTGTT | 644 |
| sfaA-DRev (SacI) | Reverse | ATAGAGCTCGCTTGATAGCTAATTGCAAATGTCATATTAG | |
| sfaA-OUT F | Forward | TCAATAATAAATTGATCATTACCATCC | 1,529 |
| sfaA-OUT R | Reverse | AGTTGTATCATAATCTTTAGACAGTGACA | |
| sbnD-AFwd (KpnI) | Forward | AACGGTACCAAGATCAATTAGGTCATCTAACTGTTC | 723 (1,436) |
| sbnD-B | Reverse | CATATGTCACCATTGGATTTGG | |
| sbnD-C | Forward | CCAAATCCAATGGTGACATATGTAAAGAATTAATACAACATGCAGCG | 713 |
| sbnD-DRev (EcoRI) | Reverse | ATAGAATTCCACTATATTGCATATACTGTGGATGC | |
| sbnD-OUT F | Forward | GGAGAACAAGCAGAACAATTATTAC | 1,732 |
| sbnD-OUT R | Reverse | TCAATTGTCCGTTCAATCTGTT | |
| sfaA and sbnD overexpression | |||
| sfaAOE-F (KpnI) | Forward | AACGGTACCATGACAAAATATTTTTTTAGCAGTTCT | 1,192 |
| sfaAOE-R (SacI) | Reverse | ATAGAGCTCTCAAGATATACCATGATTACTTTCCTC | |
| sbnDOE-F (KpnI) | Forward | AACGGTACCATGATTAATCAGTCTATATGGCGC | 1,257 |
| sbnDOE-R (EcoRI) | Reverse | ATAGAATTCTTATTTTGCACTTTTTTGTTTCAAC | |
| sfaA and sbnD complementation | |||
| sfaAcomp-F (PstI) | Forward | AACCTGCAGCCAATCGCTATTATGTTGTTGTAA | 1,391 |
| sfaAcomp-R (SacI) | Reverse | ATAGAGCTCTCAAGATATACCATGATTACTTTCCTC | |
| sbnDcomp-F (PstI) | Forward | AACCTGCAGGTTACAACTGATCAGCTAGCTGTCA | 1,411 |
| sbnDcomp-R (SacI) | Reverse | ATAGAGCTCTTATTTTGCACTTTTTTGTTTCAAC | |
| Real-time qRT-PCR | |||
| sfaA | Forward | CCTGAAGATGCAACAAGATACATGC | 144 |
| Reverse | TGTCAAAACTGGTAAAAGTGTCGTATATG | ||
| sbnD | Forward | GTTACAGCTTCGATATTAGGTTTTAGTGC | 163 |
| Reverse | GACATTGAAATGAACGGCGAATAC |
Restriction sites are indicated by underlining. Regions of homology for spliced overlap extension (SOE) PCR with the B primer are shown in italics.
The fragment sizes of SOE PCR with the A/D primers are shown in parentheses.
Cloning of sfaA and sbnD and construction of overexpressors and complemented strains of sfaA and sbnD mutants.
The entire 1,194 bp and 1,257 bp of the sfaA and sbnD coding sequences, respectively, were amplified from RN6390 chromosomal DNA using primers shown in Table 3. The amplicons were cloned into appropriate restriction sites of shuttle vector pLZ113 (23). sfaA and sbnD located on pLZ113 are overexpressed in response to anhydrotetracycline. These plasmids were electroporated into E. coli TOP10. The constructs pLZ113::sfaA (sfaA++; with ++ indicating overexpression) and pLZ113::sbnD (sbnD++) were extracted and reintroduced into S. aureus RN4220 and then into MW2 and maintained by the addition of kanamycin to the culture medium. The plasmid inserts were verified by DNA sequencing.
From approximately 200 bp upstream, regions of the sfaA and sbnD genes extending to the stop codons of their structural genes were amplified from RN6390 chromosomal DNA using primers shown in Table 3. The amplicons were cloned into appropriate restriction sites of shuttle vector pTZN10, which is a stable low-copy-number vector for efflux pump expression (4, 24). These plasmids were electroporated into E. coli TOP10. The constructs pTZN10::sfaA (pSfaA) and pTZN10::sbnD (pSbnD) were extracted and reintroduced into S. aureus RN4220 and then into RN6390 ΔsfaA and RN6390 ΔsbnD and maintained by the addition of chloramphenicol to the culture medium. The inserts of the plasmid were verified by DNA sequencing.
MIC determination.
Ampicillin, cephalexin, gentamicin, norfloxacin, ciprofloxacin, tetracycline, erythromycin, chloramphenicol, polymyxin B, colistin, rifampin, vancomycin, linezolid, spermidine, undecanoic acid, palmitoleic acid, chlorhexidine, tetraphenylphosphonium chloride, rhodamine, and ethidium bromide were purchased from Sigma and used in the MIC determinations. The MICs of these antibiotics were determined using the broth microdilution method according to Clinical and Laboratory Standards Institute (CLSI) guidelines (25).
Mouse subcutaneous abscess model.
Male Swiss Webster mice aged 4 to 6 weeks (Charles River Laboratories, Woburn, MA) were used for the subcutaneous abscess model, as previously described (1, 2). Briefly, exponential-phase S. aureus cultures were prepared by diluting overnight cultures 1:100 into BHI broth and incubating them with shaking until the culture medium reached an optical density at 600 nm (OD600) of 0.8. Cells of a single strain or a 1:1 mixture of wild-type and mutant strains were washed and diluted 1:20 in phosphate-buffered saline (PBS). The cell suspension was then mixed with an equal volume of autoclaved Cytodex 1 microcarrier beads (60 to 87 μm; Sigma) in PBS. The suspension of cells and beads (1 × 106 to 9 × 106 CFU/0.2 ml) was injected subcutaneously in each shaved flank of a mouse anesthetized with ketamine and xylazine. After 24 h, the abscesses were excised and homogenized, and the bacterial burden was determined by enumerating the CFU on TSB agar plates. For in vitro experiments, 10 μl of the bacterial suspension was inoculated into 10 ml of TSB and grown for 24 h with shaking. For fitness competition experiments, to distinguish the two strains, we used a rifampin-resistant derivative of RN6390 designated RN6390RIF. RN6390RIF is a spontaneous mutant selected for its resistance to 50 μg/ml rifampin, and its growth rate is the same as that of wild-type RN6390 (data not shown). The total CFU numbers recovered from abscesses (or culture) of the two competing bacterial strains were evaluated using the competitive index (CI), which is defined as the output/input ratio. The ratio was calculated as follows: the number of CFU of the competing strain was divided by the number of CFU of RN6390RIF. For in vivo comparisons, a CI near 1 indicates similar numbers of CFU of the two competing strains recovered from abscesses, suggesting comparable fitness of the two strains. A CI of <1 indicates the fitness advantage of the wild-type RN6390RIF over the mutant competitor. All samples were analyzed in two abscesses of two to three mice, and the assays were repeated with three independent biological samples. The difference between paired strains was analyzed using Wilcoxon signed-rank test by comparing their CIs to 1, a theoretical median. For comparison of two strains injected separately, the median output/input ratios were analyzed using Mann-Whitney U tests.
Preparation of bacterial RNA and real-time qRT-PCR.
Total S. aureus RNA was isolated using an RNeasy minikit (Qiagen, Valencia, CA). For the in vitro control, 4 × 108 to 5 × 108 cells were collected and lysed with lysostaphin (Sigma). Bacterial RNA expressed in vivo was isolated from abscesses as previously described (2). All RNA samples were digested with DNase I (Qiagen). The real-time quantitative reverse-transcription PCRs (qRT-PCRs) were carried out as previously described (2). Primers designed for the qRT-PCR assays are listed in Table 3. All samples were analyzed in triplicate, and expression levels were normalized against gmk gene expression (2). The assays were repeated with six independent biological samples. The difference in relative expression measured by RT-PCR was analyzed using Mann-Whitney U tests.
Mouse renal abscess model.
The mouse renal abscess model was generated as previously described (12). Female Swiss Webster mice aged 6 to 8 weeks were used. Briefly, cultures of exponential-phase S. aureus MW2 and its ΔsfaA and ΔsbnD isogenic mutants were prepared by diluting overnight cultures 1:100 into TSB and incubating them with shaking until the culture medium reached an OD600 of 0.6. Bacteria were washed and suspended in PBS. The inoculum was verified by enumerating the CFU on TSB agar plates. The suspension of cells (approximately 107 CFU/0.1 ml) was injected intravenously via the tail vein of a mouse anesthetized with ketamine and xylazine. After 4 days, the kidneys were removed and homogenized in 0.01% Triton X-100 using a homogenizer. The homogenate dilutions were plated on TSB agar to enumerate the recovered bacteria. All samples were analyzed in one to three mice, and the assays were repeated with four independent biological samples. The difference in relative fold change of number of CFU/kidney was analyzed using Mann-Whitney U tests.
Intracellular replication assays.
Intracellular replication assays were performed using the ATCC CCL-185 human lung adenocarcinoma cell line A549 (26). The S. aureus wild-type RN6390 was compared with its ΔsfaA and ΔsbnD mutants. The sfaA-complemented strain was used as an additional control for the effect of sfaA on bacterial internalization. The assays were based on a technique described previously (27, 28). A 20-ml fresh culture of S. aureus was prepared from an overnight culture and grown to an OD600 of 0.4. The bacteria were washed twice with Dulbecco's modified Eagle's medium (DMEM) (Life Technologies) supplemented with 10% fetal bovine serum (FBS) (Fisher Scientific, Waltham, MA), referred to here as assay medium. The bacterial pellet was resuspended in a 20-ml volume of fresh assay medium. A549 cells were cultured in assay medium in a 75-ml tissue culture flask until 90% confluence was reached, as previously described by the manufacturer (ATCC), and were then seeded into 24-well plates (Costar) in assay medium and grown again to 90% confluence. After 45 min of bacterium-epithelial cell contact, gentamicin (200 μg/ml) was added to each well, followed by incubation at 37°C to eliminate the extracellular bacteria. For samples obtained at each time point from 10 min to 5 h, gentamicin was removed, and the cells were washed twice with assay medium and then lysed with 250 μl of Triton X-100. The bacteria were diluted in PBS and plated on TSB agar plates, and colony counts were performed to determine the numbers of viable intracellular bacteria. All samples were analyzed in triplicate, and the assays were repeated with three independent biological samples. The difference in number of CFU in A549 cells was analyzed using Welch's t test.
Siderophore broth bioassays.
To measure siderophore production in S. aureus RN6390, MW2, and their ΔsfaA and ΔsbnD isogenic mutants, siderophore bioassays were performed as previously described, with some modifications (13). To prepare concentrated supernatants, all strains were grown with aeration in iron-chelating medium, Tris minimal succinate with Chelex-100 (C-TMS), for 24 h at 37°C. Culture growth was measured at OD600, and densities were normalized to approximately 1.0. Bacterial cells were removed by centrifugation, and supernatants were filtered and lyophilized. Dried materials were resuspended in sterile double distilled water (ddH2O; 1/10 of the original culture volume). Concentrated culture supernatants (250 μl) were added to 2.5-ml volumes of the reporter RN6390 Δsir and Δhts strains (OD600 of 0.0001) in C-TMS medium (RN6390 Δsir and Δhts mutants were kindly provided by David Heinrichs, University of Western Ontario, Canada). Growth promotion was assessed by measurement at OD600 after incubation for 24 h at 37°C.
ACKNOWLEDGMENTS
We thank Timothy Foster (Trinity College, Dublin) for providing us with the pIMAY vector and its host strain DC10B. We also thank David Heinrichs (University of Western Ontario, Canada) for providing us the RN6390 Δsir and Δhts mutants.
We have no conflicts of interest that are directly relevant to the content of this article.
This work was supported by U.S. Public Health Service grants no. R37-AI23988 and P01-AI083214 (M. Gilmore, principal investigator) from the National Institutes of Health to D.C.H. Additionally, this work was supported by a Research Grant from The Nagai Foundation Tokyo, Japan.
REFERENCES
- 1.Ding Y, Onodera Y, Lee JC, Hooper DC. 2008. NorB, an efflux pump in Staphylococcus aureus strain MW2, contributes to bacterial fitness in abscesses. J Bacteriol 190:7123–7129. doi: 10.1128/JB.00655-08. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Ding Y, Fu Y, Lee JC, Hooper DC. 2012. Staphylococcus aureus NorD, a putative efflux pump coregulated with the Opp1 oligopeptide permease, contributes selectively to fitness in vivo. J Bacteriol 194:6586–6593. doi: 10.1128/JB.01414-12. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 3.Truong-Bolduc QC, Villet RA, Estabrooks ZA, Hooper DC. 2014. Native efflux pumps contribute resistance to antimicrobials of skin and the ability of Staphylococcus aureus to colonize skin. J Infect Dis 209:1485–1493. doi: 10.1093/infdis/jit660. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Nakaminami H, Noguchi N, Sasatsu M. 2010. Fluoroquinolone efflux by the plasmid-mediated multidrug efflux pump QacB variant QacBIII in Staphylococcus aureus. Antimicrob Agents Chemother 54:4107–4111. doi: 10.1128/AAC.01065-09. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Haley KP, Skaar EP. 2012. A battle for iron: host sequestration and Staphylococcus aureus acquisition. Microbes Infect 14:217–227. doi: 10.1016/j.micinf.2011.11.001. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Carpenter BM, Whitmire JM, Merrell DS. 2009. This is not your mother's repressor: the complex role of Fur in pathogenesis. Infect Immun 77:2590–2601. doi: 10.1128/IAI.00116-09. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Bullen JJ, Rogers HJ, Spalding PB, Ward CG. 2005. Iron and infection: the heart of the matter. FEMS Immunol Med Microbiol 43:325–330. doi: 10.1016/j.femsim.2004.11.010. [DOI] [PubMed] [Google Scholar]
- 8.Sheldon JR, Heinrichs DE. 2015. Recent developments in understanding the iron acquisition strategies of gram positive pathogens. FEMS Microbiol Rev 39:592–630. doi: 10.1093/femsre/fuv009. [DOI] [PubMed] [Google Scholar]
- 9.Beasley FC, Vines ED, Grigg JC, Zheng Q, Liu S, Lajoie GA, Murphy ME, Heinrichs DE. 2009. Characterization of staphyloferrin A biosynthetic and transport mutants in Staphylococcus aureus. Mol Microbiol 72:947–963. doi: 10.1111/j.1365-2958.2009.06698.x. [DOI] [PubMed] [Google Scholar]
- 10.Cheung J, Beasley FC, Liu S, Lajoie GA, Heinrichs DE. 2009. Molecular characterization of staphyloferrin B biosynthesis in Staphylococcus aureus. Mol Microbiol 74:594–608. doi: 10.1111/j.1365-2958.2009.06880.x. [DOI] [PubMed] [Google Scholar]
- 11.Hannauer M, Sheldon JR, Heinrichs DE. 2015. Involvement of major facilitator superfamily proteins SfaA and SbnD in staphyloferrin secretion in Staphylococcus aureus. FEBS Lett 589:730–737. doi: 10.1016/j.febslet.2015.02.002. [DOI] [PubMed] [Google Scholar]
- 12.Dale SE, Doherty-Kirby A, Lajoie G, Heinrichs DE. 2004. Role of siderophore biosynthesis in virulence of Staphylococcus aureus: identification and characterization of genes involved in production of a siderophore. Infect Immun 72:29–37. doi: 10.1128/IAI.72.1.29-37.2004. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Sheldon JR, Marolda CL, Heinrichs DE. 2014. TCA cycle activity in Staphylococcus aureus is essential for iron-regulated synthesis of staphyloferrin A, but not staphyloferrin B: the benefit of a second citrate synthase. Mol Microbiol 92:824–839. doi: 10.1111/mmi.12593. [DOI] [PubMed] [Google Scholar]
- 14.Friedman DB, Stauff DL, Pishchany G, Whitwell CW, Torres VJ, Skaar EP. 2006. Staphylococcus aureus redirects central metabolism to increase iron availability. PLoS Pathog 2:e87. doi: 10.1371/journal.ppat.0020087. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Torres VJ, Attia AS, Mason WJ, Hood MI, Corbin BD, Beasley FC, Anderson KL, Stauff DL, McDonald WH, Zimmerman LJ, Friedman DB, Heinrichs DE, Dunman PM, Skaar EP. 2010. Staphylococcus aureus fur regulates the expression of virulence factors that contribute to the pathogenesis of pneumonia. Infect Immun 78:1618–1628. doi: 10.1128/IAI.01423-09. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16.Reniere ML, Skaar EP. 2008. Staphylococcus aureus haem oxygenases are differentially regulated by iron and haem. Mol Microbiol 69:1304–1315. doi: 10.1111/j.1365-2958.2008.06363.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Deng X, Sun F, Ji Q, Liang H, Missiakas D, Lan L, He C. 2012. Expression of multidrug resistance efflux pump gene norA is iron responsive in Staphylococcus aureus. J Bacteriol 194:1753–1762. doi: 10.1128/JB.06582-11. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Psychogios N, Hau DD, Peng J, Guo AC, Mandal R, Bouatra S, Sinelnikov I, Krishnamurthy R, Eisner R, Gautam B, Young N, Xia J, Knox C, Dong E, Huang P, Hollander Z, Pedersen TL, Smith SR, Bamforth F, Greiner R, McManus B, Newman JW, Goodfriend T, Wishart DS. 2011. The human serum metabolome. PLoS One 6:e16957. doi: 10.1371/journal.pone.0016957. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Konings WN, Lolkema JS, Bolhuis H, van Veen HW, Poolman B, Driessen AJ. 1997. The role of transport processes in survival of lactic acid bacteria. Energy transduction and multidrug resistance. Antonie Van Leeuwenhoek 71:117–128. doi: 10.1023/A:1000143525601. [DOI] [PubMed] [Google Scholar]
- 20.Konings WN. 2002. The cell membrane and the struggle for life of lactic acid bacteria. Antonie Van Leeuwenhoek 82:3–27. doi: 10.1023/A:1020604203977. [DOI] [PubMed] [Google Scholar]
- 21.Konings WN, Albers SV, Koning S, Driessen AJ. 2002. The cell membrane plays a crucial role in survival of bacteria and archaea in extreme environments. Antonie Van Leeuwenhoek 81:61–72. doi: 10.1023/A:1020573408652. [DOI] [PubMed] [Google Scholar]
- 22.Monk IR, Shah IM, Xu M, Tan MW, Foster TJ. 2012. Transforming the untransformable: application of direct transformation to manipulate genetically Staphylococcus aureus and Staphylococcus epidermidis. mBio 3:e00277–. doi: 10.1128/mBio.00277-11. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23.Zhang L, Fan F, Palmer LM, Lonetto MA, Petit C, Voelker LL, St John A, Bankosky B, Rosenberg M, McDevitt D. 2000. Regulated gene expression in Staphylococcus aureus for identifying conditional lethal phenotypes and antibiotic mode of action. Gene 255:297–305. doi: 10.1016/S0378-1119(00)00325-5. [DOI] [PubMed] [Google Scholar]
- 24.Nakaminami H, Noguchi N, Nishijima S, Kurokawa I, Sasatsu M. 2008. Characterization of the pTZ2162 encoding multidrug efflux gene qacB from Staphylococcus aureus. Plasmid 60:108–117. doi: 10.1016/j.plasmid.2008.04.003. [DOI] [PubMed] [Google Scholar]
- 25.CLSI. 2009. Methods for dilution antimicrobial susceptibility tests for bacteria that grow aerobically; approved standard M7-A8. Clinical and Laboratory Standards Institute, Wayne, PA. [Google Scholar]
- 26.Garzoni C, Francois P, Huyghe A, Couzinet S, Tapparel C, Charbonnier Y, Renzoni A, Lucchini S, Lew DP, Vaudaux P, Kelley WL, Schrenzel J. 2007. A global view of Staphylococcus aureus whole genome expression upon internalization in human epithelial cells. BMC Genomics 8:171. doi: 10.1186/1471-2164-8-171. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27.Cheung AL, Bayles KW. 2007. Tissue culture assays used to analyze invasion by Staphylococcus aureus. Curr Protoc Microbiol Chapter 9:Unit 9C.4. doi: 10.1002/9780471729259.mc09c04s4. [DOI] [PubMed] [Google Scholar]
- 28.Truong-Bolduc QC, Bolduc GR, Medeiros H, Vyas JM, Wang Y, Hooper DC. 2015. Role of the Tet38 efflux pump in Staphylococcus aureus internalization and survival in epithelial cells. Infect Immun 83:4362–4372. doi: 10.1128/IAI.00723-15. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 29.Traber K, Novick R. 2006. A slipped-mispairing mutation in AgrA of laboratory strains and clinical isolates results in delayed activation of agr and failure to translate delta- and alpha-haemolysins. Mol Microbiol 59:1519–1530. doi: 10.1111/j.1365-2958.2006.04986.x. [DOI] [PubMed] [Google Scholar]
- 30.Manna A, Cheung AL. 2001. Characterization of sarR, a modulator of sar expression in Staphylococcus aureus. Infect Immun 69:885–896. doi: 10.1128/IAI.69.2.885-896.2001. [DOI] [PMC free article] [PubMed] [Google Scholar]