|Home | About | Journals | Submit | Contact Us | Français|
Although Acinetobacter baumannii has emerged as a significant cause of nosocomial infections worldwide, there have been few investigations describing the factors important for A. baumannii persistence and pathogenesis. This paper describes the first reported identification of a glycosyltransferase, LpsB, involved in lipopolysaccharide (LPS) biosynthesis in A. baumannii. Mutational, structural, and complementation analyses indicated that LpsB is a core oligosaccharide glycosyl transferase. Using a genetic approach, lpsB was compared with the lpsB homologues of several A. baumannii strains. These analyses indicated that LpsB is highly conserved among A. baumannii isolates. Furthermore, we developed a monoclonal antibody, monoclonal antibody 13C11, which reacts to an LPS core epitope expressed by approximately one-third of the A. baumannii clinical isolates evaluated to date. Previous studies describing the heterogeneity of A. baumannii LPS were limited primarily to structural analyses; therefore, studies evaluating the correlation between these surface glycolipids and pathogenesis were warranted. Our data from an evaluation of LpsB mutant 307::TN17, which expresses a deeply truncated LPS glycoform consisting of only two 3-deoxy-d-manno-octulosonic acid residues and lipid A, suggest that A. baumannii LPS is important for resistance to normal human serum and confers a competitive advantage for survival in vivo. These results have important implications for the role of LPS in A. baumannii infections.
Acinetobacter baumannii, an opportunistic bacterial pathogen whose medical importance is increasing, is a significant cause of nosocomial infections worldwide and has recently emerged as a community-acquired pathogen as well (29). A. baumannii has been cultured from moist skin of healthy humans, but increased colonization of skin and the respiratory and gastrointestinal tracts occurs in individuals in long-term care and hospital facilities. Nosocomial colonization or infection typically occurs at surgical sites or via long-lasting invasive medical devices, such as endotracheal breathing tubes (resulting in ventilator-associated pneumonia), indwelling urinary catheters (resulting in urinary tract infections), or ventricular shunts or pressure-monitoring devices (resulting in meningitis or bacteremia). In addition to causing sporadic infections, A. baumannii causes epidemic nosocomial outbreaks, as this organism is adept at surviving and persisting for extended periods of time under a wide range of environmental conditions (9, 34). A. baumannii has been associated with uncommonly severe, rapidly progressive community-acquired pneumonia (with mortality rates of 40% to 65%), usually in patients with compromised host defenses (4, 17). Furthermore, the incidence of multidrug-resistant A. baumannii infections, which are associated with significant morbidity and mortality, is increasing worldwide (for reviews, see references 23 and 29).
The importance of A. baumannii infections in war-related injuries is well established. A. baumannii was the most common Gram-negative bacillus recovered from traumatic injuries to the lower extremities during the Vietnam War (40). Most recently, A. baumannii-associated soft tissue infections, osteomyelitis, pneumonia, and bacteremia have been reported in United States service personnel injured in the Iraq, Kuwait, and Afghanistan regions during Operation Iraqi Freedom and Operation Enduring Freedom (5, 31, 38). A. baumannii also emerged as an important pathogen in survivors of the Asian tsunami in 2004 (11, 22). It is the ability of this organism to colonize hospital equipment and to resist typical disinfection methods that perpetuates the outbreak cycle. Moreover, infections caused by this organism are particularly challenging due to its impressive repertoire of intrinsic and acquired antibiotic resistance determinants (30). The changing epidemiology and increasing incidence of infections due to A. baumannii demonstrate that the medical importance of this pathogen is increasing.
Compared to the virulence mechanisms of other human pathogens, the virulence mechanisms of A. baumannii have not been well characterized (for a review, see reference 29). Thus, there is a significant gap in our knowledge and understanding of the factors involved in A. baumannii pathogenesis. However, one of the virulence factors that have been shown to be involved in multiple steps of the disease process for other Gram-negative human pathogens is lipopolysaccharide (LPS).
LPS typically consists of a hydrophobic anchor domain termed lipid A (or endotoxin) that makes up the outer leaflet of the Gram-negative bacterial outer membrane, a nonrepeated core oligosaccharide structure, which can be divided into “inner core” (lipid A proximal) and “outer core” regions, and a distal polysaccharide comprised of repeat-unit structures having variable lengths termed the O antigen (OAg). In general, the lipid A region is considered the most toxic or inflammatory region of LPS, although the polysaccharide portion of the molecule also has potent immunomodulating and immunostimulating properties (24, 33). LPS has been shown to contribute both to bacterial evasion of host immune responses, affecting both innate and acquired host responses to infection, and to initiation of an overwhelming host inflammatory response that significantly correlates with the morbidity and mortality of infected patients (15, 24, 32). Moreover, the cell surface location of LPS contributes to the interaction between the bacterium and its environment. To begin to understand how A. baumannii LPS impacts virulence and immune responses to this pathogen, insight into the chemical structure and biologic activities of LPS is essential.
The LPS structures expressed by a variety of A. baumannii isolates have been characterized in previous studies. These studies demonstrated that the A. baumannii OAg region is highly heterogeneous as numerous LPS glycoforms have been defined using both structural and antibody-based studies (21, 26-28, 41-44). Although the structural analyses of A. baumannii LPS determined that this organism expresses an S-form LPS, phenotypic analyses of the molecules has been challenging because the OAg cannot be visualized following gel electrophoresis. This problem appears to be a result of the absence of unsubstituted vicinal OH groups, which makes the polysaccharides nonoxidizable for detection (41). Recent biologic studies indicated that A. baumannii LPS may be an important immunostimulatory molecule involved in Toll-like receptor 4 (TLR4) signaling in a mouse pneumonia model (14). This immunostimulatory capacity of LPS was substantiated using human cells in vitro, although the latter study also identified TLR2 as an important signaling factor (8). Despite these elegant structural and biologic studies, there is currently no definitive link between A. baumannii LPS and virulence. In addition, there have not been any reports defining any of the genes involved in the biosynthesis of the A. baumannii LPS molecule. Thus, detailed studies evaluating the correlation between the principal surface glycolipids and pathogenesis are warranted.
In this paper, we describe isolation and characterization of an A. baumannii LPS mutant, 307::TN17, that was identified using transposon mutagenesis and an antibody specific to a core LPS glycoform expressed by A. baumannii strain 307-0294. Our data indicate that LpsB, an LPS glycosyl transferase involved in biosynthesis of the LPS core, is highly conserved among clinical isolates. Furthermore, our studies comparing an lpsB mutant to the wild-type strain demonstrated that A. baumannii LPS is important for serum resistance in vitro and for survival in vivo.
A. baumannii strain 307-0294, a blood isolate from a patient hospitalized at Erie County Medical Center, Buffalo, NY, is classified as a sequence type 15, clonal group 1 strain (7, 18, 35, 36). A. baumannii strains ATCC 19606, ATCC 15308, and ATCC 17978 were purchased from the American Type Culture Collection (Manassas, VA). In addition, we used 74 A. baumannii clinical strains isolated from infected military personnel serving in Iraq and Afghanistan (kindly provided by David Craft and Paul Scott of the Walter Reed Army Medical Center). All strains were routinely cultured using Mueller-Hinton (MH) medium or agar supplemented with antibiotics when appropriate (50 μg/ml kanamycin, 200 μg/ml carbenicillin, 5 μg/ml tetracycline). Escherichia coli XLI-Blue was used for plasmid cloning and was cultured at 37°C on Luria-Bertani agar plates or in Luria-Bertani broth in the presence of antibiotics when they were required.
Standard molecular biology reagents were obtained from New England Biolabs, Inc. (Beverly, MA). Plasmid isolation and amplicon purification were performed using kits manufactured by Qiagen (Santa Clarita, CA). Chromosomal DNA was isolated using standard methods. PCR amplification was performed for 25 cycles with the GeneAMP 9700 PCR system (P.E. Applied Biosystems, Foster City, CA) using Platinum Taq High Fidelity DNA polymerase (Invitrogen, Carlsbad, CA) and primer set-dependent annealing temperatures and extension times. DNA nucleotide sequences of all constructs were obtained by automated DNA sequencing (RPCI Biopolymer Facility, Roswell Park Cancer Institute, Buffalo, NY) and were analyzed with MacVector 10.5. Total RNA was isolated using an RNAProtect and RNeasy minikit (Qiagen) and was subjected to RQ1 RNase-free DNase treatment (Promega). Reverse transcription-PCR (RT-PCR) analyses were performed using a OneStep RT-PCR kit (Qiagen) to ensure that transcription of the genes flanking the insertionally inactivated coding region in the mutant strain remained unaffected.
A. baumannii LPS core-specific monoclonal antibody (MAb) 13C11 was developed by injecting BALB/c mice intraperitoneally with viable A. baumannii 307-0294, using previously described methods (2, 18). Hybridoma supernatants were screened by immunodotting, Western blot assays, and flow cytometry using standard procedures to determine the presence of surface carbohydrate-reactive antibodies (18).
A. baumannii 307-0294 transposon (TN) mutant derivatives were generated using the Epicentre EZ::TN<KAN-2> Transposome system as described previously (18, 36). A mutant derivative of A. baumannii 307-0294 that lost reactivity to MAb 13C11 was identified by a colony lift assay and designated 307::TN17. Chromosomal DNA from 307::TN17 that was purified was restriction enzyme digested with Sau3AI and rescue cloned into BamHI-digested pUC18. Recombinant plasmids containing the TN and flanking A. baumannii DNA were isolated by selecting for resistance to both ampicillin and kanamycin after electroporation into E. coli XLI-Blue and were subjected to automated DNA sequencing for determination of the TN insertion site. A derivative that was complemented in trans, 307::TN17/pSS11, was generated using the recently described A. baumannii shuttle vector pNLAC1 developed by our group (36). pNLAC1 was generated by subcloning the ori-repM region of pMAC (a recently described 9.5-kb mobilizable extrachromosomal element isolated from A. baumannii ATCC 19606 ) into pBR322, which allowed replication in both E. coli (for subcloning) and A. baumannii (for complementation). pNLAC1 constructs (with and without the lpsB coding region plus 200 bp of flanking DNA) were transformed into the wild-type and mutant strains and assessed to determine restoration of the phenotype, as described previously.
LPS samples used for SDS-PAGE and Western blot analyses were prepared using proteinase K (PK)-treated whole-cell lysates or enriched sheared surface preparations, prepared as described previously with additional incubation for 2 h at 60°C with PK (0.4 mg/ml) when necessary (20). LPS was resolved by SDS-16% polyacrylamide gel electrophoresis and visualized by silver staining (19, 37). Mild periodate oxidation at an acidic pH to destroy carbohydrate determinants was performed in the presence (final concentration, 20 mM) and absence (control reactions) of sodium meta-periodate as described previously (3). In vitro susceptibility assays were performed as described previously (19).
Compositional and mass spectrometric (MS) analyses were performed at the National Research Council of Canada (Ottawa, Canada). Bacterial cells were killed with 2% (wt/vol) (final concentration) phenol and pelleted by centrifugation. LPS was isolated by standard methods (45). In brief, killed cells were freeze-dried and washed once with ethanol and then twice each with acetone and light petroleum ether to remove lipids and other lipophilic components. Washed cells were extracted by the hot phenol-water method, and the retentate was dialyzed and freeze-dried. A 2% solution of this crude LPS preparation was treated with DNase and RNase at 37°C for 4 h and then with proteinase K at 37°C for 4 h. Small peptides were removed by dialysis. After freeze-drying, the retentate was again rehydrated to obtain a 2% solution and centrifuged at 8,000 × g for 15 min, and then the supernatant was centrifuged at 100,000 × g for 5 h. The pellet from the high-speed centrifugation, which contained purified LPS, was redissolved in water and freeze-dried. The LPS was treated with anhydrous hydrazine to prepare O-deacylated LPS as described previously (12, 37). KOH-treated LPS was isolated by treating LPS with 4 N KOH (10 mg/ml, 125°C, 30 h) and neutralizing the preparation with HCl after cooling; salts were removed by centrifugation (5,000 × g) and by elution from a Sephadex G-25 gel permeation chromatography column. Purified carbohydrate-containing material was assessed to determine its sugar content as derived alditol acetates and by performing compositional analyses using capillary electrophoresis-electrospray mass spectrometry (CE-ES-MS) as previously described (39).
Complement-mediated bactericidal assays were performed by measuring the change in bacterial titer over time in the presence of 90% active or inactive (heated at 56°C for 30 min) human serum. An input bacterial titer of approximately 1 × 105 CFU was used, and titers were measured at 0, 1, 2, and 3 h as described previously (19, 35, 36). Numbers of surviving CFU were determined by plating duplicate serial 10-fold dilutions in duplicate or triplicate. A minimum of three independent assays were performed for each strain.
The rat soft tissue infection model has recently been established by our group as a clinically relevant animal model for assessing A. baumannii infection (35, 36). In brief, 1% croton oil (1 ml) was injected into a space created by subcutaneous injection of 30 to 50 ml of air on the back of anesthetized Long-Evans rats. This space matured into an encapsulated, fluid-filled “pouch” within 7 days, mimicking a subcutaneous abscess replete with exudative fluid. On day 8, approximately 106 CFU of A. baumannii was injected into the pouches of anesthetized animals. Competition assays were performed by introducing equal numbers of CFU of the mutant and wild type. Plating the organisms on the appropriate selective media was used to distinguish between the two strains. Serial dilutions of pouch fluid aliquots (0.5 ml) obtained from anesthetized animals at 0, 6, 24, and 48 h after bacterial challenge were plated to determine bacterial titers. Assays were performed using cohorts of three rats per strain on at least two separate occasions.
Data are expressed below as means ± standard errors of the means. To normalize in vitro and in vivo data, log10-transformed values were used. The area under each curve was calculated from log10-transformed values, and the areas were compared using the two-tailed unpaired t test for statistical significance, using GraphPad Prism 5 software.
Mice were immunized with A. baumannii 307-0294, and fusion was performed using standard methods. The resulting hybridomas were screened by an immunodot assay versus whole-organism and proteinase K (PK)-treated whole-cell lysates. Clone 13C11 reacted to a PK-resistant, periodate-sensitive, low-molecular-mass bacterial cell surface component (Fig. (Fig.1)1) and was selected for further analysis. Based on the data, we hypothesized that MAb 13C11 was specific for a carbohydrate epitope in the A. baumannii 307-0294 LPS core. To further characterize the putative LPS epitope recognized by MAb 13C11, a panel of A. baumannii clinical isolates from infected military personnel were evaluated by colony lift analysis to determine their reactivity to MAb 13C11. MAb 13C11 reacted to a conserved surface-exposed epitope expressed by 34.2% (26/78) of the A. baumannii isolates in our collection. The surface reactivity of MAb 13C11 was confirmed by flow cytometry, and additional Western blot analyses indicated that MAb 13C11 reacted to a single, conserved band at approximately 7-kDa for all positive strains (data not shown). Furthermore, we performed immunodot assays with a variety of Gram-negative organisms, and MAb 13C11 did not react with any of the isolates, suggesting that this antibody may be specific to A. baumannii LPS (data not shown). As anti-core LPS antibodies have been shown to protect against lethal Gram-negative sepsis in animal models, we hypothesized that further characterization of this LPS epitope may result in a better understanding of the mechanism of pathogenesis used by A. baumannii to establish infections.
A. baumannii 307-0294 was subjected to random TN mutagenesis, and approximately 500 kanamycin-resistant transformants were immunoscreened for loss of reactivity to MAb 13C11. One of the mutants that did not react to MAb 13C11, designated 307::TN17, was selected for further study, and the TN insertion point was determined by DNA sequencing. Analysis of the subcloned fragment comprised of DNA flanking the site of TN insertion indicated that the TN had inserted into the 1,101-bp lpsB open reading frame (at bp 686) that was predicted to encode a 366-amino-acid protein (GenBank accession number ACJ59103, locus tag ABBFA_003104 ).
BlastP queries using the NCBI database detected conserved domains that placed the predicted coding region in the GTB-type superfamily of nucleotide-sugar-dependent glycosyltransferases (GTs). The deduced protein sequence is most closely related to sequences of the WavL-like group 1 GT (GT1) family. WavL has been shown to be involved in the biosynthesis of the LPS core oligosaccharide in Vibrio cholerae (25). Members of the GT1 family transfer activated UDP-, ADP-, GDP-, or CMP-linked sugars to the LPS core. LpsB exhibited the highest levels of homology (40% to 48% identity [ID]) to other putative GT1 homologues identified by genome sequencing of predominately environmental bacteria, including Thiobacillus denitrificans ATCC 25259, Nitrosococcus oceani ATCC 19707, and Psychrobacter arcticus 273-4. Structural searches revealed that the highest levels of homology were the levels of homology to the crystal structure of the Bacillus anthracis strain Ames family GT4 enzyme designated ORF Ba1558 (24% ID) and the UDP-glucose:(heptosyl)-LPS-α-1,3-glucosyltransferase WaaG (formerly RfaG) of E. coil (22% ID). WaaG, a well-described GT involved in LPS biosynthesis in many bacteria, is responsible for addition of the first glucose moiety to the LPS core via an α-1,3-glycosidic linkage to heptose II, the distal heptose residue of the inner core Hep-Hep-3-deoxy-d-manno-octulosonic acid (Kdo) trisaccharide. The homology to WaaG is particularly interesting in light of the recent determination of the structure of the LPS carbohydrate backbone of A. baumannii strain ATCC 19606 (a MAb 13C11-positive strain), which showed that the A. baumannii LPS core is heptose deficient (42). Additional sequence analysis of the entire 4.4-kb subcloned insert revealed the presence of a lipid A biosynthetic lauroyl acyltransferase (LpxL) coding region upstream of lpsB (GenBank accession number ACJ56415.1) and a coding region for a glutamate-aspartate symport protein GltP homologue downstream in the opposite orientation (GenBank accession number ACJ57010.1) (Fig. (Fig.2).2). BlastN analyses of the entire coding region for all three genes revealed 97% to 100% ID in the six publically available annotated A. baumannii genomes currently accessible in the GenBank database (strains A. baumannii 307-0294 [accession number CP001172], AB0057 [accession number CP001182], AYE [accession number CU459141], SDF [accession number CU468230], ACICU [accession number CP000863], and ATCC17978 [accession number CP000521]).
SDS-PAGE analysis demonstrated that the more slowly migrating approximately 7- kDa LPS glycoform of A. baumannii wild-type strain 307-0294 is not present in mutant 307::TN17 (Fig. (Fig.3A).3A). A corresponding Western blot was probed with MAb 13C11, which reacted only to the wild-type LPS and specifically to the 7-kDa band (Fig. (Fig.3B).3B). These data suggest that the loss of functional LpsB activity specifically affected the A. baumannii 307-0294 LPS glycoform and that the mutant did not express the LPS epitope recognized by MAb 13C11.
LPS was isolated from flask-grown cells by using well-documented, standard protocols that included an initial organic solvent wash to delipidate the cells in order to enhance the efficiency of the subsequent phenol extraction step. Phenol extraction is the universally accepted method for isolating LPS from Gram-negative bacterial cells (45). The resulting crude LPS extracts are then treated with DNase, RNase, and proteinase K in order to remove nucleic acids and any residual proteins, which results in a pure LPS preparation.
Sugar analyses of both LPS and KOH-treated LPS from A. baumannii 307-0294 identified glucose, N-acetylglucosamine, and Kdo, whereas glucose was not present in LPS and KOH-treated LPS from the 307::TN17 lpsB mutant in the same analyses. CE-ES-MS analyses were performed to compare the compositions of the LPS elaborated by 307::TN17 and A. baumannii 307-0294. CE-ES-MS is a well-established technique for ascertaining the compositions of isolated LPS-derived samples, such as the O-deacylated and completely deacylated samples described here (39). Analyses of the O-deacylated LPS of A. baumannii 307-0294 suggested a composition of lipid A-OH, 3 Kdo residues, 2 HexN residues, HexNAcA, and 4 Hex residues, which is consistent with the LPS composition observed previously for A. baumannii strain ATCC 19606 (43). In contrast, analysis of both LPS-OH and fully deacylated (KOH-treated) LPS preparations obtained from the 307::TN17 mutant revealed a highly truncated molecule containing just 2 Kdo residues and the lipid A region (Table (Table11).
lpsB, along with approximately 200 bases of flanking DNA, was subcloned into the A. baumannii complementation plasmid pNLAC1, and the sequence was confirmed to be correct by a sequence analysis (36). The resulting construct, designated pSS11, was introduced into 307::TN17 to generate the in trans complemented strain 307::TN17/pSS11. Analyses of the chromosomal lpsB coding region and extraction of the plasmid from 307::TN17/pSS11 indicated that the complementing plasmid was maintained as an independent replicon in 307::TN17. SDS-PAGE and Western blot assays, as well as colony lift analyses of the wild type, the LPS-deficient TN mutant 307::TN17, and the complemented derivative 307::TN17/pSS11, demonstrated that the MAb 13C11-reactive LPS moiety was restored (Fig. 3A and B, lane 3, and 3C).
A. baumannii 307-0294 is resistant to the bactericidal activity of normal human serum (NHS), exhibiting complete resistance even in the presence of 90% NHS after 3 h (35, 36). To evaluate whether the loss of functional LpsB enzymatic activity and corresponding production of a severely truncated LPS molecule affected serum sensitivity, comparative bactericidal assays were performed. These assays indicated that 307::TN17 exhibited significantly (P = 0.0001) increased serum sensitivity compared to the parental strain (Fig. (Fig.4).4). In contrast to the survival of A. baumannii 307-0294 in 90% serum after 3 h, only 12.8% of the 307::TN17 inoculum survived under these conditions. Importantly, in the complemented derivative 307::TN17/pSS11 serum resistance was completely restored, and the levels of resistance were comparable to wild-type levels. All strains were also assessed in the presence of 90% heat-inactivated (56°C for 30 min) NHS and in MH broth; all growth and survival rates were comparable (data not shown).
To investigate if LPS expression contributes to A. baumannii infection in vivo, we examined the growth and survival of the isogenic LPS mutant 307::TN17, both in competition with A. baumannii wild-type strain 307-0294 and by itself, in a rat soft tissue infection model. Compared to A. baumannii 307-0294, 307::TN17 alone did not exhibit a significant decrease (P = 0.1185) in survival in this model (Fig. (Fig.5A).5A). In contrast, when the challenge inoculum contained A. baumannii 307-0294 and 307::TN17 at a ratio of 1:1 (Fig. (Fig.5B),5B), the expression of a truncated LPS molecule by the mutant affected the growth and survival of 307::TN17 compared to the growth and survival of A. baumannii 307-0294 (P = 0.0282). Since the growth of A. baumannii 307-0294 and the growth of 307::TN17 were similar in laboratory media in vitro (data not shown), these data support the conclusion that the expression of a deep rough LPS phenotype is disadvantageous in vivo. These results suggest that the full-length LPS molecule is important for the survival of A. baumannii 307-0294 and provides a competitive advantage for establishment and persistence of A. baumannii infections in vivo.
In this paper, we describe the use of random TN mutagenesis for identification of an LPS core GT in A. baumannii. This LPS core GT is highly conserved, as the sequence of A. baumannii 307-0294 lpsB exhibits significant homology to the lpsB coding regions in six other A. baumannii strains (96 to 100% ID); in addition, this gene was identified by PCR amplification in all A. baumannii isolates evaluated regardless of the MAb 13C11 reactivity phenotype (data not shown). The LpsB-deficient mutant 307::TN17 did not produce a 7-kDa carbohydrate moiety detectable by SDS-PAGE analysis and also did not exhibit reactivity to the LPS core-specific MAb 13C11. Furthermore, compositional analyses of this mutant indicated that it expresses a deeply truncated LPS molecule consisting of only 2 Kdo residues and lipid A. In contrast, wild-type A. baumannii 307-0294 expresses an LPS core glycoform with the same composition as the previously published LPS core of A. baumannii ATCC 19606, which is also an MAb 13C11-reactive strain, comprised of lipid A-OH, 3 Kdo residues, 2 HexN residues, HexNAcA, and 4 Hex residues (43).
The results of this study indicate that an A. baumannii LpsB mutant defective in biosynthesis of full-length LPS is sensitive to the bactericidal activity of NHS. When it was compared to A. baumannii wild-type strain 307-0294, 307::TN17 exhibited a survival rate that was 29.6% of the wild-type rate at 1 h and was 4.6% of the wild-type rate at 3 h. Not only did a previous study report that clinical isolates of A. baumannii obtained from bacteremic patients exhibited intrinsic resistance to 90% NHS, but the authors also postulated that the expression of LPS might be the essential factor for this resistance and suggested that this surface molecule might be important for allowing the strains to survive in blood and contribute to the pathogenesis of this species in human infections (10). Our results obtained using a defined mutant expressing a truncated LPS molecule and a derivative complemented in trans, which restored the LPS glycoform to the wild-type phenotype, confirm this hypothesis and conclusively demonstrate that LPS is essential for resistance to NHS. Moreover, detergent sensitivity assays indicated that the deep core LPS mutant 307::TN17 exhibits increased sensitivity (P < 0.05) to SDS, Triton X-100, and deoxycholate compared to the wild type, further confirming that the complete core LPS glycoform is essential for the functional integrity of the A. baumannii outer membrane (data not shown). Importantly, the growth of 307::TN17 was comparable to the growth of A. baumannii wild-type strain 307-0294 in vitro, indicating that the deep core LPS truncation exhibited by the mutant did not affect the cell viability or growth rate.
The ability of the mutant to survive in vivo was assessed by performing direct challenge assays, as well as competitive survival assays, using the rat soft tissue infection model. We recently established this animal model, which is clinically relevant to A. baumannii since this organism is increasingly recognized as a cause of soft tissue infections, as a method for evaluating the importance of A. baumannii genes in vivo (35, 36). These analyses demonstrated that the 307::TN17 lpsB mutant could survive and grow in the rat soft tissue infection model for at least 48 h; however, the expression of a truncated LPS molecule was disadvantageous compared to the expression of the native molecule in this model system. Interestingly, lpsB mutants of Sinorhizobium meliloti also exhibit deep core LPS biosynthetic defects and are competitively compromised for alfalfa nodulation compared to wild-type strains (13, 16). The data describing defective infections by LpsB mutants support the conclusion that production of full-length LPS plays an important functional role in survival in vivo.
Although structural analyses of the wild-type LPS molecules expressed by A. baumannii strains have been described previously, this is the first report describing the identification, characterization, and mutagenesis of a gene involved in the biosynthesis of A. baumannii LPS. Serum susceptibility assays demonstrated that the expression of full-length LPS on the surface of A. baumannii is critical for protection against the bactericidal effects of NHS. Further, in vivo studies using a rat soft tissue infection model indicated that LPS contributes to the survival and fitness of A. baumannii. Taken together, the data suggest that full-length LPS plays a critical role in the pathogenesis of A. baumannii infections and supports the hypothesis that LPS-deficient mutants are likely attenuated for virulence in vivo. These findings strongly support the conclusion that additional studies investigating the contribution of the biosynthesis and cell surface expression of LPS to infections caused by A. baumannii should be performed.
We thank Amy Howlett for skilled technical assistance.
Editor: S. R. Blanke
Published ahead of print on 1 March 2010.