The Flagellum of Pseudomonas aeruginosa Is Required for Resistance to Clearance by Surfactant Protein A

Surfactant protein A (SP-A) is an important lung innate immune protein that kills microbial pathogens by opsonization and membrane permeabilization. We investigated the basis of SP-A-mediated pulmonary clearance of Pseudomonas aeruginosa using genetically-engineered SP-A mice and a library of signature-tagged P. aeruginosa mutants. A mutant with an insertion into flgE, the gene that encodes flagellar hook protein, was preferentially cleared by the SP-A+/+ mice, but survived in the SP-A−/− mice. Opsonization by SP-A did not play a role in flgE clearance. However, exposure to SP-A directly permeabilized and killed the flgE mutant, but not the wild-type parental strain. P. aeruginosa strains with mutation in other flagellar genes, as well as mucoid, nonmotile isolates from cystic fibrosis patients, were also permeabilized by SP-A. Provision of the wild-type fliC gene restored the resistance to SP-A-mediated membrane permeabilization in the fliC-deficient bacteria. In addition, non-mucoid, motile revertants of CF isolates reacquired resistance to SP-A-mediated membrane permeability. Resistance to SP-A was dependent on the presence of an intact flagellar structure, and independent of flagellar-dependent motility. We provide evidence that flagellar-deficient mutants harbor inadequate amounts of LPS required to resist membrane permeabilization by SP-A and cellular lysis by detergent targeting bacterial outer membranes. Thus, the flagellum of P. aeruginosa plays an indirect but important role resisting SP-A-mediated clearance and membrane permeabilization.


INTRODUCTION
Pseudomonas aeruginosa is a ubiquitous Gram-negative pathogen and a major cause of morbidity in patients with nosocomial pneumonia and cystic fibrosis (CF). Environmental strains and most clinical isolates of P. aeruginosa involved in acute infections are flagellated and motile [1][2][3]. In contrast, approximately 39% of P. aeruginosa isolated from chronically infected CF patients are nonmotile [4,5].
Other studies have suggested that exposure to airway surface liquid from CF patients quickly represses the expression of P. aeruginosa flagellin, the major structural protein of bacterial flagella [6]. This repression is mediated by the AlgT-dependent transcriptional regulator AmrZ [7]. The P. aeruginosa flagellum is an important virulence factor [3] and a potent activator of host immunity [8]. A number of studies have shown that P. aeruginosa strains lacking flagella are much less readily phagocytosed by alveolar macrophages and polymorphonuclear leukocytes [4,5]. Thus, the loss of flagellar expression within the CF airway may represent an adaptive response that allows P. aeruginosa to evade detection and clearance by host defense mechanisms during the chronic phase of CF lung infection [9].
SP-A is a large oligomeric protein composed of 18 subunits, each containing a collagen-like sequence and a calcium-dependent carbohydrate recognition domain, that is intimately associated with surfactant phospholipids lining the air-liquid interface of the lung [10,11]. Upon encountering an inhaled bacterial pathogen, SP-A dissociates from the phospholipid membrane and binds to the microbe by calcium dependent ligation of surface bacterial glycoconjugates, including lipopolysaccharide. The association of SP-A with the bacteria enhances uptake by phagocytes, and modulates inflammatory responses including cytokine production and the oxidant burst [11][12][13][14]. Recent data from our group indicates that SP-A also directly inhibits the growth of gramnegative bacteria such as Escherichia coli, Bordetella pertussis, and P. aeruginosa by permeabilizing the bacterial membrane [15][16][17]. In general, rough bacterial mutants decorated with truncated LPS species were more readily permeabilized by SP-A than smooth strains [15][16][17], and incorporation of rough but not smooth LPS multilamellar liposomes composed of bacterial phospholipids conferred susceptibility to SP-A-mediated permeabilization [18].
We recently adapted signature tagged mutagenesis (STM) to comparatively identify microbial factors that confer resistance or susceptibility to SP-A [17]. Subsets of a P. aeruginosa library composed of 6912 mutants containing individual gene insertions were instilled into SP-A +/+ and SP-A 2/2 mice. Mutants which are cleared in an SP-A-dependent manner were tested for susceptibility to SP-A-mediated opsonization, phagocytic killing and permeabilization. In this paper, we characterize a SP-A sensitive P. aeruginosa mutant identified by comparative STM, the flagellar mutant flgE.

The P. aeruginosa Flagellar Gene flgE Confers Resistance to SP-A
We utilized a P. aeruginosa PA01 STM mutant library [17,19] and a SP-A deficient mouse model to identify putative microbial targets for SP-A. We identified an STM mutant that was able to grow in the lungs of SP-A 2/2 but not SP-A +/+ mice, as indicated by the presence of a PCR amplified oligo tag in bacterial output from SP-A 2/2 mouse lungs, and the absence of a PCR amplified oligo tag in bacterial output from SP-A +/+ mouse lungs ( Figure 1A).
The disrupted genetic locus was cloned by plasmid rescue and sequenced. The DNA sequences flanking the transposon, pUTminiTn5Km2 (Table 1) Figure 1C) [20]. The flgE mutant was determined to harbor a single transposon insertion after hybridization with the transposon vector pUTminiTn5Km2 (Table 1) (data not shown). A DNA fragment that flanked both ends of the transposon integration in the flgE gene was cloned from the wild-type PA01. A radiolabeled-probe that was synthesized from the flanking DNA fragment hybridized to a higher molecular weight Pst I restricted DNA fragment in the chromosome of the flgE mutant than the parent strain PA01 ( Figure 1D). These results confirmed that the transposon inserted singly into the flgE gene.
The SP-A-sensitive STM Mutant, flgE, Competes Poorly With the Parental PA01 Strain in the Lungs of SP-A +/+ Mice Apart from confronting pulmonary host defenses, the STM screening strategy used requires individual mutants to compete with 71 other STM mutants within individual library pools. Competitive mixed infection assays have been used to assess the fitness of individual mutants versus their parental strains during in vivo infection [17,21]. When inoculated with equal numbers of withtype PA01 into the lungs of mice, the flgE mutant was only 33% as competitive as PA01 in SP-A +/+ animals ( Figure 2A). In contrast, the flgE mutant competed comparably, if not better than, the wild-type PA01 in SP-A 2/2 mice (Figure 2A). We also tested the flgE mutant in single infection studies. Eighteen hours after intranasal inoculation, the viable bacterial counts of flgE mutant bacteria were 1.3 log higher in lung homogenates from SP-A 2/2 mice compared to SP-A +/+ mice ( Figure 2B). Our results indicate that flagellar function plays an important role against SP-A-mediated clearance of P. aeruginosa.

Opsonization by SP-A Does Not Play A Role in the Preferential Clearance of the flgE Mutant In Vivo
Previous studies have demonstrated that flagella act as a major ligand for nonopsonic phagocytosis of P. aeruginosa, and are required to trigger internalization [4][5]. Consequently, P.  aeruginosa lacking flagella are resistant to nonopsonic phagocytosis by macrophages and neutrophils. In multiple studies, SP-A has been reported to opsonize pathogens and enhance their phagocytosis by alveolar macrophages and neutrophils [22][23][24]. To rule out SP-A-mediated opsonization contributed to the differential clearance of flgE mutant in SP-A mouse models, we compared the phagocytosis of flgE to PA01 bacteria in SP-A +/+ and SP-A 2/2 mice after intranasal inoculation in vivo, and in the presence and absence of SP-A, in vitro. Macrophages with internalized GFPexpressing flgE and PA01 were enumerated by phase contrast fluorescence microscopy. Uptake of PA01 bacteria was noted in 42.167.6% of alveolar macrophages isolated 2 hr after intranasal inoculation from SP-A +/+ mice, compared to 28.563.0% of alveolar macrophages with PA01 from SP-A 2/2 mice ( Figure 3A). These results suggest that SP-A facilitates the in vivo macrophage-mediated phagocytosis of PA01 bacteria, which have intact flagella. In contrast, the loss of flagella in flgE mutant bacteria results in lower total uptake by alveolar macrophages when compared to wild-type PA01 bacteria. Specifically, flgE bacteria were internalized in only 19.865.7% of alveolar macrophages from SP-A +/+ mice compared with 12.765.1% from SP-A 2/2 mice ( Figure 3A). Thus, in vitro, SP-A enhanced phagocytosis of PA01 by approximately 50% but did not have any effect on the phagocytosis of flgE bacteria ( Figure 3B). Similarly, as shown in Figure 3C, the uptake of PA01 and the flgE mutant by neutrophils were not increased in the presence of SP-A ( Fig. 3C). In fact, SP-A-opsonized PA01 was phagocytosed less efficiently than untreated bacteria by neutrophils. However, the uptake of PA01 was significantly greater than flgE, irrespective of the presence or absence of SP-A. Collectively, our results suggest that while flagella enhance internalization of P. aeruginosa into macrophages and neutrophils,  Macrophages were recovered by bronchoalveolar lavage 2 hr later. The percentage of macrophages with internalized bacteria was determined using a phase contrast fluorescence microscope. Two hundred macrophages were counted for each mouse. Data are the mean6SE, n = 3. *p,0.05. (B) Analysis of in vitro phagocytosis was performed using live, GFP-expressing PA01 and flgE bacteria with alveolar macrophages isolated from SP-A 2/2 mice. The percentage of macrophages with internalized bacteria was determined using a phase contrast fluorescence microscope. Two hundred macrophages were counted for each mouse. Data are the mean6SE, n = 3. *p,0.01. (C) In vitro phagocytosis of PA01 and flgE by human neutrophils isolated from the blood of healthy volunteers. Phagocytosis experiments were performed as described for (A). Two hundred neutrophils were counted. Data are the mean6SE, n = 3. *p,0.05. PMN = Neutrophils. M = Macrophages. doi:10.1371/journal.pone.0000564.g003 SP-A-mediated macrophage and neutrophil phagocytosis are not responsible for the preferential clearance of flgE mutant bacteria from SP-A +/+ mice. These results also confirm data from other studies that flagella serve as critical ligands for the internalization of bacteria [4,5].

An Intact Flagellum is Required to Confer Resistance of P. aeruginosa to SP-A-mediated Membrane Permeabilization
We have recently shown that SP-A is capable of directly killing Gram-negative bacteria and fungi by membrane permeabilization and inhibition of macro-molecular synthesis, independent of macrophage-mediated phagocytosis [15,17,25]. We examined whether direct membrane permeabilization contributed to the preferential clearance of the flagellar-deficient flgE mutant. The flgE mutant bacteria were permeabilized by SP-A to an extent that was 3.4 fold greater than the wild-type strain PA01 after exposure to SP-A for 120 minutes ( Figure 4A).
We postulate that the susceptibility of the flgE mutant to SP-Amediated membrane permeabilization is due to its inability to assemble a flagellum. We tested additional flagellar mutants, fliC (PAO-fliC) and fliD (PAO-fliD) [26,27], for their susceptibility to membrane permeabilization by SP-A. The fliC gene encodes flagellin, which comprises the filament of a flagellum whereas the fliD gene encodes the cap of a flagellum ( Figure 1C). Both PAO-fliC and PAO-fliD mutant bacteria exhibited increased susceptibility to SP-A-mediated membrane permeabilization to an extent that was 3.3 and 4.4 fold greater than the wild-type strain PA01, respectively, at 120 min post-exposure to SP-A (Figures 4, B and C). The levels of PAO-fliC and PAO-fliD bacteria susceptibility were comparable to the flgE mutant ( Figure 4A), and to the E. coli K12 [15]. These results indicate that an intact flagellum or flagellar function is essential for resistance to SP-A-mediated membrane permeabilization, and is consistent with previous finding that flagellum is required for acute lung infection by P. aeruginosa [3].
To confirm that the loss of flagella confers susceptibility of P. aeruginosa to SP-A, we determined if provision of the wild-type fliC gene in trans to the fliC-deficient mutant bacteria could restore the resistance to SP-A-mediated membrane permeabilization. Importantly, while PAO-fliC mutant bacteria were susceptible to SP-Amediated membrane permeabilization, the complemented PAOC-fliC bacteria, were as resistant to SP-A-mediated membrane permeabilization as the wild-type PA01 ( Figure 4D). These results indicate that an intact flagellar structure or function is important in protecting P. aeruginosa from membrane destabilization by SP-A.

Membrane Permeabilization by SP-A Directly Kills Flagellar Mutants
We next determined if SP-A-mediated membrane permeabilization kills flagellar mutants. Live/dead staining, based on exclusion of propidium iodide from the live cells, was immediately performed on bacterial cells following 120 min of exposure to SP-A. Viable cells are stained green while dead cells are stained red. As shown in Figure 5A, SP-A did not kill wild-type PA01 (green stained). In contrast, mixtures of green-and red-stained cells of flgE were visible following treatment with SP-A. Approximately 11% of the flgE bacteria were killed within 60 minutes of exposure to SP-A. Bacterial aggregation can affect viability. However, in contrast to the robust SP-A induced aggregation of E. coli K12, SP-A did not aggregate PA01 or flgE bacteria ( Figure 5B). These results suggest that the loss of flagella render the P. aeruginosa vulnerable to membrane permeabilization and killing by SP-A. Furthermore, aggregation of P. aeruginosa is unlikely to play a role in the preferential clearance of the flagellar mutants.
Mucoid, Nonmotile P. aeruginosa Isolated from CF Patients are Susceptible to SP-A-mediated Membrane Permeabilization As discussed earlier, significant numbers (39%) of P. aeruginosa strains isolated from chronically-infected CF patients are both mucoid and nonmotile [4]. The loss of motility is due, in part, to repression of flagellar biosynthesis by the AlgT-dependent transcriptional regulator AmrZ [7] induced by the CF airway fluids [6]. However, mucoid, nonmotile CF P. aeruginosa isolates frequently revert to nonmucoid, motile phenotypes when they are grown in static culture under laboratory conditions [28]. We examined the effect of SP-A on two mucoid, nonflagellated strains of P. aeruginosa isolated from CF patients. Strains CF51 and CF90 were nonmotile ( Figure 6, A and B), and membrane permeabilized by SP-A to an extent that was 3.2 fold and 3.6 fold greater than the wild-type strain PA01, respectively, after 120 min exposure ( Figure 6, C and D). The levels of membrane permeabilization in these two CF strains were similar to flagellum-deficient flgE, fliC and fliD mutants. The overall correlation of flagellar motility and resistance to membrane permeability are summarized in Table 2.

Nonmucoid, Motile Revertants of CF P. aeruginosa Isolates Reacquire Resistance to SP-A-mediated Membrane Permeabilization
If P. aeruginosa flagella play critical roles in resistance to SP-Amediated membrane permeability, we predicted that nonmucoid, motile revertants of CF isolates CF51 and CF90, named CF51N and CF90N, would have increased resistance to SP-A-mediated membrane permeabilization. Both revertants were nonmucoid, and determined to be motile by swimming motility assay ( Figure 6, A and B). The levels of resistance to SP-A-mediated membrane permeabilization were increased by 2.0 and 1.7 fold in CF51N and CF90N, respectively, when compared to parental strains CF51 and CF90 ( Figure 6, C and D; Table 2). Collectively, these results confirm that P. aeruginosa flagellum plays a crucial role in conferring resistance to SP-A-mediated membrane permeabilization.  aeruginosa strains CF51 and CF90 were nonmotile, whereas their nonmucoid motile revertants CF51N and CF90N had regained their flagellar-based motility on swimming motility agar plates. The wild-type strain PA01was used as control for motility assay. (C-D) Both CF51N and CF90N were more resistant to SP-A-mediated membrane permeabilization than their parental isolates, CF51 and CF90, respectively. Data are the mean6SE, n = 8 for each bacterial strain. The differences in membrane permeabilization between CF51 versus PA01; CF51 versus CF51N; and CF51N versus PA01 were significant from 43 rd -min, 53 rd -min and 53 rd -min onward, respectively. *p,0.05. The difference in membrane permeabilization between CF90 versus PA01; CF90 versus CF90N; and CF90N versus PA01 was significant from 38 th -min, 28 th -min, and 43 rd -min onward, respectively. *p,0.05. doi:10.1371/journal.pone.0000564.g006 Susceptibility to SP-A-mediated Membrane Permeabilization is Dependent On An Intact Flagellum Structure but Independent of Flagellar-mediated Swimming Motility The susceptibility of flagellum-deficient mutants to SP-A could be due to changes in membrane fluidity and rigidity, or due to the loss of motility function. We examined whether the nonmotile P. aeruginosa mutants that harbor intact flagella were resistant to membrane permeabilization by SP-A. We utilized P. aeruginosa mutants with deletions in motAB, motCD or motAB motCD genes, all of which encode the stator for flagellum [29]. The stator conducts protons across the inner membrane and couples proton transport to rotation of the flagellar motor [30]. The phenotypic analysis of the deletion mutants demonstrated that the MotAB and MotCD proteins are functionally redundant for swimming (Fig. 7A), as well as for swarming in relatively low agar concentrations [29]. However, under higher agar concentrations, the motCD mutant is unable to swarm [29]. The double mutant dmotAB dmotCD can neither swim ( Figure 7A) nor swarm under any conditions tested, but still makes a flagellum [29]. Importantly, dmotAB, dmotCD and dmotAB, dmotCD were all as resistant to SP-A-mediated membrane permeabilization as their parental wild-type PA14 ( Figure 7B). As a control, we demonstrated that flgK mutant, which was derived from PA14, and unable to assemble a flagellum due to the lack of the flagellar hook-associated protein, was nonmotile ( Figure 7C) and membrane permeabilized by SP-A to an extent that was 7.9-fold greater than its wild-type parental strain, PA14. These results indicate that resistance to SP-A-mediated membrane permeabilization is dependent on the presence of an intact flagellar structure, rather than the motility function mediated by flagellum.
The Loss of Flagellum Reduces the Ability of P. aeruginosa to Express Adequate LPS and Destabilize Its Outer Membranes We have previously shown that LPS is essential for resistance to SP-A-mediated membrane permeabilization in P. aeruginosa [17]. We determined whether the loss of flagellum negatively affected LPS levels on the outer membrane of P. aeruginosa, which could lead to susceptibility to SP-A. As shown in Figure 8A, the amounts of the LPS inner core sugars on P. aeruginosa was dramatically reduced after 15 min exposure to SP-A. However, the reduction is much more severe in the flgE mutant (230%) when compared to wild-type PA01 (213.5%). At 30 min post-SP-A exposure, PA01 increased the synthesis of inner core sugars by +23.6% in comparison to +14.1% for flgE. Furthermore, using an antibody against the B-band O-antigen, the flgE mutant was found to harbor 48% and 50% less LPS than the wild-type PA01 during both logarithmic and stationary phases of growth, respectively, in the absence of SP-A ( Figure 8B). Collectively, these results suggest that the flagellar-deficient flgE mutant bacteria express lower levels of LPS on its outer membrane.
Gram-negative bacteria with defects in their outer membranes are well known to be more sensitive to lysis by detergents, including sodium dodecyl sulfate (SDS). We examined the flagellar-deficient mutants flgE, fliC and fliD for their increased susceptibility to 0.25% SDS. As shown in Figure 8C, all three mutants were more sensitive to lysis by SDS. In contrast, the wild-type PA01, and PAOC-fliC (complemented fliC mutant) were equally more resistant to SDS.  These results support the notion that that the loss of flagellum reduces the expression of LPS and the stability of outer membranes on P. aeruginosa, rendering them susceptible to SP-A-mediated membrane permeabilization and lysis SDS.

DISCUSSION
In this study, we screened a library of P. aeruginosa mutant bacteria in SP-A null and sufficient mice for mutations with increased susceptibility to pulmonary clearance by SP-A. One SP-A sensitive P. aeruginosa mutant with a disruption in flgE, a gene required for flagellar function, was identified. Enhanced uptake by phagocytes did not play a role in the preferential clearance of flgE bacteria, consistent with previous findings that flagellum-deficient mutants are resistant to phagocytosis by macrophages and neutrophils [4,5]. Rather, flgE bacteria were more readily membrane permeabilized and killed by SP-A than its wild-type parental strain, PA01. The importance of flagella in resisting the membrane destabilizing activities of SP-A was verified by multiple independent methods: Firstly, we showed that additional flagellar mutants fliC, fliD and flgK were also susceptible to SP-A-mediated membrane permeabilization. Secondly, genetic complementation of the fliC mutation with a copy of wild-type fliC gene restored resistance to SP-A-mediated membrane permeabilization. Thirdly, mucoid, nonmotile CF isolates of P. aeruginosa were also membrane permeabilized by SP-A. Finally, motile, nonmucoid revertants derived from the SP-Asusceptible CF isolates were able to reacquire resistance to SP-A. These results reveal that flagellum plays critical roles in conferring resistance to membrane permeabilization by SP-A. Our genetic analysis suggests that the presence of the flagellar structure rather than the flagellar-dependent motility is largely responsible for the resistance to SP-A-mediated membrane permeabilization. This conclusion is derived from our experimental evidence showing that the dmotAB, dmotCD mutant, which has a flagellum but is nonmotile [29], is as resistant to SP-A-mediated membrane permeabilization as its isogenic wild-type parental strain, PA14. Thus, it is likely that susceptibility of flagellardeficient mutants to SP-A-mediated membrane permeabilization is caused by alterations to the outer cell membrane, such as LPS density, charge distribution or phosphohlipid packing density. Several lines of evidence support our prediction: Firstly, we found that exposure to SP-A induced higher levels of LPS in the wildtype P. aeruginosa PA01 than its isogenic flgE mutant bacteria. Secondly, flgE bacteria contain reduced levels of LPS during both logarithmic and stationary phases of growth. Thirdly, flagellardeficient mutants flgE, fliC and fliD were all more susceptible to 0.25% SDS, indicative of weakened outer membranes.
Because the nonmotile dmotAB, dmotCD mutant is resistant to SP-A, the susceptibility to SP-A appears to be independent of flagellar-mediated motility functions. However, we are unable to rule out other flagellum-dependent mechanisms that may protect P. aeruginosa from SP-A-mediated membrane permeabilization, including swimming chemotaxis and swarming motility [31], and initiation of biofilm formation through attachment to substrates [32]. Swarming motility is a flagellum-dependent, coordinated, multicellular behavior promoting rapid bacterial colony migration and expansion on more viscous surface, such as the epithelial lining layer that P. aeruginosa encounters after inhalation into the alveolar space. In numerous species of Gram-negative bacteria including P. aeruginosa, swarmer cells alter their cellular morphology as well as induce the expression of PmrA-B regulated LPS modification genes pmrHFIJKLMN, and other LPS biosynthesis genes that are required for successful swarming [33], probably by increasing wettability. Overexpression and modification of LPS contributes to higher levels of resistance to antimicrobial peptides and antibiotics in swarming Gram-nagative bacteria [31,33]. As such, swarming P. aeruginosa cells may be more resistant to SP-Amediated membrane permeabilization. It is also conceivable that the loss of swimming chemotaxis in the flagellum-deficient mutants render them unable to escape alveolar microenvironments in which SP-A is enriched, such as in surfactant-containing compartments. Another possible explanation, though less likely, is the inability of flagellar mutant to form biofilms. Flagellummediate attachment to substrate is the first step in the initiation of biofilm formation [32], a mode of growth which is known to contribute to the resistance of bacteria to antibiotics and other antimicrobial agents [34,35]. However, biofilms mode of growth is thought to be more important in chronic model of infection, as in the airways of CF patients [9,36], rather than in the acute pneumonia model of infection.
We found that motile, nonmucoid revertants derived from CF P. aeruginosa strains, CF51N and CF90N, were more resistant to SP-A-mediated membrane permeabilization than their mucoid, nonmotile parental isolates CF51 and CF90. This transition correlated with the reacquisition of swimming motility. However, the resistance of CF51N and CF90N to SP-A-mediated membrane permeabilization was not as prominent when compared to the laboratory wild-type strain PA01. There are several possible explanations for this discrepancy. Firstly, it is likely that PA01 and CF51 and CF90 have different levels of intrinsic resistance to SP-A as these are not isogenic strains. Secondly, it is well known that many CF isolates of P. aeruginosa harbor mutated forms of LPS. In particular, CF isolates often undergo mutations in genes that participate in the biosynthesis of the highly antigenic B-band Oantigen of LPS [37], which is important for P. aeruginosa's resistance to SP-A [17]. Thus, LPS mutations in the CF51N and CF90N revertants may explain the incomplete restoration of resistance to SP-A-mediated membrane permeabilization even though the flagellar expression is reacquired.
Reduction in the SP-A levels of the alveolar lining fluid, as has been reported in bacterial pneumonia, adult respiratory distress syndrome and CF [38][39][40], may play a role in the susceptibility to pulmonary microbial colonization and infection. Proteases released by neutrophil [41] and P. aeruginosa [42,43] are the most likely causes of SP-A depletion. Thus, the much-reduced SP-A levels within diseased CF airways may create an environment for P. aeruginosa to safely repress the synthesis and assembly of the energy consuming and highly proinflammatory flagella [8]. In addition, the loss of flagellum will also decrease the chance of P. aeruginosa being phagocytized by SP-A-facilitated macrophage and neutrophils.
Our results suggest that the killing of flagellum-deficient mutants is more robust in vivo than in vitro. One likely explanation for the discrepancy is that our in vitro membrane permeability conditions do not accurately model the environment in the lung. It is also possible that other antimicrobial factors within alveolar space, including SP-A, SP-D, lactoferrin, b-defensins and other antimicrobial peptides and proteins [44,45], collectively contribute to the killing of P. aeruginosa in vivo. We are currently examining whether these peptides act synergistically in clearance of P. aeruginosa.
In summary, our data suggest that P. aeruginosa flagella play an important role in resistance to pulmonary clearance by SP-A and in particular to SP-A-mediated membrane permeabilization. Therapeutic strategies aimed at blocking flagellar biosynthesis or formation, or aerosolizing SP-A or SP-A fragments into septic airways where SP-A levels are severely depleted, may be reasonable therapeutic approaches to explore in the future. Bacterial strains, plasmids, media and growth conditions All bacterial strains, plasmid vectors, and their derivatives are described in Table 1. Two P. aeruginosa PA01 strains [46,47] and PA14 [48] were the wild-type strains used in the present study. The P. aeruginosa STM library [19] was created by random insertion of the transposon pUTmini-Tn5 into the genome of PA01 strain from Holloway et al [46]. The PAO-fliC [27] and PAO-fliD [26] mutants, and the complemented fliC strain PAOC-fliC [49], and their isogenic parental PA01 (originally from Mike Vasil, see Table 1) [47] were the same as previously described. The flgK, dmotAB, dmotCD and dmotAB, dmotCD mutants, which were derived from PA14, was as previously described [29,32]. Two mucoid clinical isolates of P. aeruginosa, CF51 and CF90, were obtained from the University of Cincinnati Adult Cystic Fibrosis Center following approval by the Institutional Review Board at the University of Cincinnati. Escherichia coli K12 used as a control in some experiments was transfected with the plasmid PVL 1393 (Invitrogen, Carlsbad, CA) to confer ampicillin resistance. Plasmid pUCP19GFP (GFP + Ap r ), which constitutively expresses gfp, was used to label the strains for microscopic studies. Bacterial strains were grown in Luria broth (LB) for 16 hr at 37uC, and then suspended in LB with 20% glycerol and frozen in aliquots at 280uC. Before each assay, aliquots of the bacteria were cultured from frozen stocks in liquid LB or on LB agar plates (1.5% agar) with or without antibiotics for 3-16 hr at 37uC with continuous shaking, and then suspended in Tris-buffered saline, pH 7.4. Bacteria were cultured to stationary growth phase (OD600<3.0) as indicated. The optical density at 600 nm was determined by spectrophotometry and correlated with numbers of viable bacteria by colony-forming units (cfu) after plating serial dilutions on agar plates. The antibiotic concentrations used were as follows: for E. coli, ampicillin (100 mg/ml); for P. aeruginosa, carbenicillin (300 mg/ ml), gentamycin (100 mg/ml), kanamycin (100 mg/ml), rifampicin (50 mg/ml) and tetracycline (100 mg/ml).

Purification of human SP-A
Human SP-A was purified from the lung washings of patients with alveolar proteinosis by a modification of the method of Suwabe et al, [50] and stored in 5 mM Tris, 150 mM NaC, pH 7.4, at 220uC. The preparations were deemed free of EDTA by a modified spectrophotometric assay of Kratochvil and White, using ß-phenanthrolene-disulfonic acid as the indicator [51].

Protein assays
Routine protein concentrations were determined with a bicinchoninic acid protein assay kit (BCA; Pierce Chemical Co., Rockford, Illinois, USA) using BSA as a standard. Protein samples were separated on 8-16% SDS-PAGE gel and stained with Coomassie blue or silver.

Animal husbandry
Swiss Black SP-A 2/2 mice (a gift of J. Whitsett/T. Korfhagen) were developed from embryonic stem cells after disruption of the mouse SP-A gene by homologous recombination and maintained by breeding with Swiss Black mice, as previously reported [52]. The SP-A null allele was bred into the C3H/HeN background through nine generations, as described [15]. C3H/HeN control (SP-A +/+ ) mice were purchased from Charles River Laboratory (Boston). All comparisons made with the SP-A 2/2 mice were with age and strain-matched C3H/HEN controls. All animals were housed in positively ventilated microisolator cages with automatic recirculating water located in a room with laminar, high efficiency particulate-filtered air. The animals received autoclaved food, water, and bedding. Mice were handled in accordance with approved protocols through the Institutional Animal Care and Use Committee at the University of Cincinnati School of Medicine.
Comparative screening of the P. aeruginosa STM library in SP-A +/+ and SP-A 2/2 mice The P. aeruginosa PA01 STM library and screening methods were previously described [19]. Briefly, SP-A +/+ and SP-A 2/2 mice were anaesthetized using isofluorane and inoculated intranasally with 1610 7 cells of individual STM pools of 72 mutants in 50 ml of phosphate buffered saline buffer. After 16 hr of in vivo selection, lungs were removed from sacrificed mice, homogenized, and plated on LB agar. At least 10,000 bacterial colonies were harvested for multiplex PCR as described previously [17,19]. Twenty sets of 72 mutants (total 1440) had been screened at the time of this writing. An additional 240 mutants previously shown to be attenuated in a chronic model of rat lung infection were also screened [19]. The mutants that were differentially cleared by SP-A during the in vivo passaging were subsequently retested by single PCR of bacterial cells recovered from three separate mice as previously described [17].

Competitive and single infection assays in vivo
For competition assays, mice were inoculated intranasally with 1610 7 bacterial cells comprised of a 1:1 ratio of wild-type PA01 and flgE. Lungs were harvested 16 hr after infection for bacterial load determination. The competitive index (CI) is defined as the output ratio of mutant to wild-type bacteria divided by the input ratio of mutant to wild-type bacteria [21]. Thus, for a mutant strain that was less competitive than the parental strain from which it was derived, a CI value of ,1 would be achieved. Single infections with 1610 7 flgE cells were performed by the intranasal route in SP-A +/+ and SP-A 2/2 mice (group of five). Attenuation was defined as the log 10 difference in CFU of flgE recovered from the lung tissues of SP-A +/+ and SP-A 2/2 mice lung tissue 16 hr after inoculation.

Alveolar macrophage and neutrophil phagocytosis assays in vivo and in vitro
For in vivo phagocytosis studies, 2.5610 8 GFP-expressing PA01 or flgE bacteria were intranasally inoculated into SP-A +/+ and SP-A 2/2 mice. Two hours later, infected lungs were lavaged with PBS to obtain alveolar macrophages. For in vitro studies, freshly recovered mouse macrophages (,5610 5 cells from 3 uninfected SPA 2/2 mice) were adhered to chamber plastic culture slides, coated with 5% poly-D-lysine in 200 ml RPMI (Dulbecco's, containing 2.5 mg/liter gentamycin and 0.1% BSA), for 2 hr at 37uC in 5% CO 2 and used immediately. Live GFP-expressing PA01 or flgE in RPMI were preincubated with or without 25 mg/ ml human SP-A for 1 hr at 37uC with rotation. The media was removed from each well, replaced with 250 ml of opsonized PA01 or flgE, and incubated for 1 hr at 37uC in 5% CO 2 at a ratio of 100 bacteria to 1 macrophage. Chamber slides were washed 3 times with PBS containing 1 mM CaCl 2 . Fluorescence from extracellular bacteria was quenched with crystal violet (0.8 mg/ ml). Following two additional washes, cells were fixed with 1% paraformaldehyde in PBS plus 1 mM CaCl 2 for 10 min, and stained with Evans blue for 2 min. The percentage of macrophages with engulfed bacteria was quantified by phase contrast fluorescence microscopy. At least 200 macrophages were counted. Human neutrophils were purified and cultured as previously described [53]. Neutrophil phagocytosis was performed as described for macrophages.

Static growth experiments
Previous studies have demonstrated that mucoid strains of P. aeruginosa are capable of spontaneously reverting to a nonmucoid phenotype and acquiring flagellum based motility in a static growth conditions [28]. In this study, overnight cultures of two mucoid CF isolates of P. aeruginosa strains, CF51 and CF90, were inoculated into 5 ml LB in 186150 mM culture tubes. These cultures were incubated at 37uC vertically in test tube racks without shaking for 72 hr, and then plated on regular LB agar. Individual nonmucoid colonies were replicated on swimming motility agar plates (LB solidified with 0.3% agar) to confirm reversion to motility phenotype.

Flagellum-dependent swimming motility
Flagellum-dependent swimming motility was measured in fresh swimming motility agar plates with or without antibiotics. Plates were inoculated with 5 ml of an overnight culture of each strain of same cell density (as determined by spectrophotometer) and then incubated at 30uC for 24-48 hr. Motility was quantitatively assessed by examining the radius of opacity from the inoculation point, indicating the extent of bacterial swimming, compared to that of the wild-type PA01.

P. aeruginosa membrane permeability assays
The effect of the SP-A or melittin on P. aeruginosa cell wall integrity was assessed by determining permeability to a phosphatase substrate, Enzyme-Labeled Fluorescence 97 (ELF-97, Molecular Probes) as described [17]. SP-A (50 mg/ml) or melittin (5 mg/ml) was incubated with 1610 8 logarithmic or stationary phase bacterial cells/ml in 100 ml of 5 mM Tris and 150 mM NaCl for 15 min at 37uC, and 100 mM ELF97 phosphatase substrate was added. Fluorescence was measured at excitation and emission wavelengths of 355 and 535 nm, respectively, for a period of 120 min.

LIVE/DEAD bacterial staining
The viability of SP-A exposed bacteria was determined by LIVE/ DEAD baclight TM Bacterial Viability Kit (Molecular Probes, Cat. No. L-7012). Briefly, stationary phase E. coli or P. aeruginosa bacteria were washed three times with 5 mM Tris, 150 mM NaCl, and adjusted to an OD 600 of 1.0. A 100 ml aliquot of the cells was added to culture tubes containing SP-A (100 mg/ml) in 100 ml of 5mM Tris, 150mM NaCl, or to buffer alone as control. The mixtures were incubated at 37uC while shaking at 300 rpm for 1 hr, and stained according to the instructions provided by the supplier. The viability of the cells was checked under a fluorescence microscope. At least 1, 000 cells were counted.
LPS purification and western blot analysis P. aeruginosa PA01 and flgE cells were grown to late logarithmic or stationary phase in LB at 37uC and concentrated to A 560 1.5 in PBS, and verified to contain the same viable cell counts (data not shown). The resultant cell pellet from 500 ml of sample was resuspended in 100 ml of PBS and incubated with or without 100 mg/ml of SP-A at 37uC for 30 min. An equal volume of 26sample buffer (6% SDS; 6% 2-mercaptoethanol; 10 mM dithiothreitol; 46% glycerol; 60 mM Tris, pH 8.0; 0.1% bromophenol blue) was added. The samples were boiled for 10 min and protein was digested by the addition of proteinase K to a final concentration of 50 mg/ml at 37uC overnight. The samples were boiled again for 10 min and cooled down to room temperature. A second volume of proteinase K, equal to the first, was added and incubated at 55uC for 3 h. Gross LPS profiles were determined by silver stain after SDS-PAGE electrophoresis.
Western blot analyses were performed as described [54]. Briefly, LPS samples (15 ml) were resolved by SDS-PAGE and electroblotted to Immobilon P polyvinylidene difluoride membranes (Millipore, Bedford, MA). The membranes were then incubated for 60 min at room temperature in blocking solution (PBS containing 3% bovine serum albumin), followed by a 4-hr incubation with monoclonal antibodies against the inner core (MAb 7-4) and B-band O-antigen (MAb MF15-4) (generously provided by Dr. Joseph Lam, Guelph University), respectively. The membranes were hybridized with horseradish peroxidaseconjugated goat anti-mouse IgG secondary antibody. The immune complexes were visualized using the ECL Western Blotting Detection System (Amershan Biosciences, Piscataway, New Jersey) and Kodak BIOMAX (Kodak, Rochester, New York) X-ray films.

Statistical analyses
Statistical analysis was performed using the Student's t-test and one-way analyses of variance (ANOVA). A significant difference was considered to be p,0.05.