Acinetobacter baumannii Is Dependent on the Type II Secretion System and Its Substrate LipA for Lipid Utilization and In Vivo Fitness

ABSTRACT

Gram-negative bacteria express a number of sophisticated secretion systems to transport virulence factors across the cell envelope, including the type II secretion (T2S) system. Genes for the T2S components GspC through GspN and PilD are conserved among isolates of Acinetobacter baumannii, an increasingly common nosocomial pathogen that is developing multidrug resistance at an alarming rate. In contrast to most species, however, the T2S genes are dispersed throughout the genome rather than linked into one or two operons. Despite this unique genetic organization, we show here that the A. baumannii T2S system is functional. Deletion of gspD or gspE in A. baumannii ATCC 17978 results in loss of secretion of LipA, a lipase that breaks down long-chain fatty acids. Due to a lack of extracellular lipase, the gspD mutant, the gspE mutant, and a lipA deletion strain are incapable of growth on long-chain fatty acids as a sole source of carbon, while their growth characteristics are indistinguishable from those of the wild-type strain in nutrient-rich broth. Genetic inactivation of the T2S system and its substrate, LipA, also has a negative impact on in vivo fitness in a neutropenic murine model for bacteremia. Both the gspD and lipA mutants are outcompeted by the wild-type strain as judged by their reduced numbers in spleen and liver following intravenous coinoculation. Collectively, our findings suggest that the T2S system plays a hitherto-unrecognized role in in vivo survival of A. baumannii by transporting a lipase that may contribute to fatty acid metabolism.

IMPORTANCE Infections by multidrug-resistant Acinetobacter baumannii are a growing health concern worldwide, underscoring the need for a better understanding of the molecular mechanisms by which this pathogen causes disease. In this study, we demonstrated that A. baumannii expresses a functional type II secretion (T2S) system that is responsible for secretion of LipA, an extracellular lipase required for utilization of exogenously added lipids. The T2S system and the secreted lipase support in vivo colonization and thus contribute to the pathogenic potential of A. baumannii.

INTRODUCTION

Acinetobacter baumannii, an increasingly common nosocomial Gram-negative pathogen, is responsible for a wide range of infections, including pneumonia, urinary tract infections, bacteremia, meningitis, and skin and wound infections (1–4). Immunocompromised and severely ill patients in the intensive care unit, individuals with extensive wounds or invasive devices, and those who are undergoing or have recently undergone antibiotic regimens are particularly susceptible to A. baumannii infections (5–7). Ventilator-associated pneumonia and bloodstream infections are the most severe, resulting in 25% to 35% mortality rates (8, 9).

The pathogenic success of A. baumannii is likely multifactorial, but its ability to persist on dry surfaces, form biofilms, resist complement-mediated killing, and survive antibiotic treatment are of importance (1, 10–13). The escalating frequency of multidrug-resistant (MDR) strains of A. baumannii is of particular concern. In the past 10 years, there has been an alarming 60% increase in the number of MDR clinical isolates reported (http://www.cddep.org). An important and clinically relevant aspect of bacterial infections is the ability of bacteria to grow as a biofilm (14). These matrix-encased, multilayer bacterial communities are exceptionally resistant to antibiotic treatment and are prone to spreading antibiotic resistance through horizontal gene transfer (15, 16). Clinical isolates that form biofilms survive for long periods of time on dry surfaces and are able to colonize common hospital equipment, such as ventilator tubes (17). Several factors have been shown to be necessary for abiotic biofilm formation, including a pilus assembly system, the OmpA outer membrane protein, and capsular polysaccharide, which is also protective against complement-mediated killing (13, 18–20).

While research has focused on the mechanisms of antibiotic resistance and biofilm formation and the epidemiology of A. baumannii, our understanding of A. baumannii pathogenesis is lagging and little is known about the contribution of secreted proteins to A. baumannii survival and propagation during infection. Sequencing of several A. baumannii genomes has revealed that A. baumannii contains genes for a variety of transport systems, including the assembly and translocation system for type IV pilus and type VI and type IV secretion systems. The type IV pilus supports the twitching motility of A. baumannii (21) but may also contribute to adhesion, colonization, biofilm formation, and transformation like type IV pili in other Gram-negative pathogens, while the type VI secretion system is utilized for bacterial competition and the type IV secretion system may be required for virulence (22, 23). In addition, A. baumannii possesses genes for a type II secretion (T2S) system (24, 25). Bacteria that use the T2S system are typically environmental bacteria; however, they also include pathogens such as Vibrio cholerae, enterotoxigenic Escherichia coli, Pseudomonas aeruginosa, and Legionella pneumophila (26–29). The T2S system mediates the secretion of toxins and hydrolytic enzymes, including proteases, lipases, lipoproteins, and enzymes that break down complex carbohydrates, and has been found to be required for in vivo survival and virulence (30–39). Following inner membrane translocation via the Sec or TAT pathways, T2S substrates engage with the T2S system for transport across the outer membrane. This multiprotein secretion system is encoded by 12 to 16 general secretion pathway (gsp) genes (40–42). With rare exceptions, mutations in any of the core gsp genes, gspC through gspM and pilD, prevent extracellular secretion (41).

The gsp genes are scattered throughout the A. baumannii genome instead of being organized into one or two operons (27). Due to their unusual arrangement, it was unclear whether the gsp genes of A. baumannii encode a functional secretion system. Here, we demonstrate that the T2S system in A. baumannii is functional and identify a lipase as one of its secreted substrates. We show that the extracellular lipase, LipA, and the T2S system that transports this enzyme across the outer membrane are required for utilization of exogenously added lipids and support colonization of A. baumannii in a murine model of bacteremia.

MATERIALS AND METHODS

Bacterial strains and growth conditions.

All strains listed in Table 1 were cultured in Luria-Bertani (LB) broth or on LB agar at 37°C. Carbenicillin (100 μg/ml) was used for plasmid maintenance.

TABLE 1.

Plasmids and bacterial strains

Strain or plasmid	Relevant characteristic(s)a	Reference or source
Plasmids
pK18mobsacB	Suicide vector containing sacB (Kmr)	77
pCVD442	Suicide vector containing sacB (Apr)	78
pMMB67EH	Low-copy-number, IPTG-inducible vector (Apr)	79
pgspD	pMMB67EH-gspD	This study
pgspE1	pMMB67EH-gspE1	This study
pgspE2	pMMB67EH-gspE2	This study
pgspN	pMMB67EH-gspN	This study
plipA	pMMB67EH-lipA	This study
plipBA	pMMB67EH-lipBA	This study
Strains
E. coli MC1061	F− lac mutant; K-12 laboratory strain	80
E. coli MM294/pRK2013	Helper strain for conjugation	81
E. coli SY327λpir	λpir lysogen; permits replication of pCVD442	82
A. baumannii AYE	Clinical strain	83
A. baumannii AB0057	Clinical strain	84
A. baumannii AB5075	Clinical strain	85
A. baumannii 17978	Wild type for T2S	ATCC
A. baumannii ΔgspD	Replacement of gspD with aph-3 (Kmr)	This study
A. baumannii ΔgspE1	Replacement of gspE1 with aph-3 (Kmr)	This study
A. baumannii ΔgspE2	Replacement of gspE2 with aph-3 (Kmr)	This study
A. baumannii ΔgspN	Replacement of gspN with aph-3 (Kmr)	This study
A. baumannii ΔlipA	Replacement of lipA with aph-3 (Kmr)	This study

^aAp^r, ampicillin resistance; Km^r, kanamycin resistance.

Construction of ΔlipA, ΔgspD, ΔgspE1, ΔgspE2, and ΔgspN strains.

Chromosomal DNA isolated from the wild-type (WT) A. baumannii ATCC 17978 strain was used as the template for PCR. PCRs were carried out with Phusion DNA polymerase. Primers were synthesized by IDT Technologies.

To generate the ΔgspD strain, we used the primers indicated in Table 2 to amplify and clone 500 bp of DNA upstream and downstream of the gspD gene as well as the aph-3 kanamycin cassette into pCVD442. Transconjugates in which pCVD442 had recombined into the A. baumannii genome were selected on LB agar containing carbenicillin and chloramphenicol. To select for the second recombination event, individual colonies were cultured overnight in LB broth, diluted, cultured to the late log phase, and spread on LB agar containing 3% sucrose. Sucrose and kanamycin-resistant, carbenicillin-sensitive isolates were screened for loss of growth on lipid agar. Strains that were kanamycin resistant and carbenicillin sensitive were designated ΔgspD mutants. The deletion was verified by PCR. All other gene deletion strains (the ΔlipA, ΔgspE1, ΔgspE2, and ΔgspN mutants) were constructed in a similar manner using the appropriate primers shown in Table 2.

TABLE 2.

Primers used for plasmid construction

Primer	Sequence (5′–3′)	Plasmid construct(s) generated
KanUp	CCGGAATTGCCAGCTGGG	ΔgspD, ΔgspE1, ΔgspE2, ΔgspN, and ΔlipA strains
KanDown	TTCAGAAGAACTCGTCAAG	ΔgspD, ΔgspE1, ΔgspE2, ΔgspN, and ΔlipA strains
gspD1	GGTTCAACTTCTTACCAATT	ΔgspD vector
gspD2	CCCAGCTGGCAATTCCGGTAAAGCCATAACTCGCGA	ΔgspD vector
gspD3	CTTGACGAGTTCTTCTGAAGCGCCGTAGTAGCATGTTA	ΔgspD vector
gspD4	CGGTGCGGGTTTTGGCACAG	ΔgspD vector
gspD5	CGCGGATCCGTCACCATAAGAGTTAGGAA	pgspD
gspD6	CGCGCATGCCATAACATGCTACTACGGCGCTG	pgspD
gspE1-1	GTTAAACAGACTTCACGCTG	ΔgspE1 vector
gspE1-2	CCCAGCTGGCAATTCCGGGTTTCAGG	ΔgspE1 vector
gspE1-3	CTTGACGAGTTCTTCTGAAAGTGAAG	ΔgspE1 vector
gspE1-4	GGTGCTGTAACTAACCCAG	ΔgspE1 vector
gspE1-5	GACGAGCTCCTCATCATTATAAATTGT	pgspE1
gspE1-6	GACGTCGACGCATTTTTTATATCTTAC	pgspE1
gspE2-1	GAGCCTAGTCTCTTTTTTAA	ΔgspE2 vector
gspE2-2	CTCTTGCGACATGACTTGCTTTTTCTTCATTCAGCC	ΔgspE2 vector
gspE2-3	CTTGACGAGTTCTTCTGACCATTAAATTAAATTTTT	ΔgspE2 vector
gspE2-4	CTACACGTTTTAAAGGCTTATAATC	ΔgspE2 vector
gspE2-5	GACGTCGACATAGATGAGGTGAATCTT	pgspE2
gspE2-6	GACGAATTCATATATTGGGGAAAACAC	pgspE2
gspN1	GTTGAACAGCTTCTAGAATTTGGCG	ΔgspN vector
gspN2	CCCAGCTGGCAATTCCGGCTTTTTCTTCATTCAGCC	ΔgspN vector
gspN3	CTTGACGAGTTCTTCTGAGGTGGTAACTAATGAAAG	ΔgspN vector
gspN4	GCTCTGTAGGTTGAGACGGTGTAGC	ΔgspN vector
gspN5	CACGAATTCCATGTTGGTAAGGCTGAATG	pgspN
gspN6	GCGAAGCTTCCATACTTTCATTAGTT	pgspN
lipA1	AAGCTTGTCGACTTACACACACGTAC	ΔlipA vector
lipA2	GTTGCATGCCGGTTAAAACCCGCCAT	ΔlipA vector
lipA3	TTAGAGCTCCAAGGATTATAAGCTTT	ΔlipA vector
lipA4	TTACCCGGGTTGATATGCGCTTTA	ΔlipA vector
lipA5	GAGGAATTCAGTAAAAAATGAAAAGG	plipA
lipA6	GAGGTCGACTAAAGCGTAAGCTTATA	plipA
lipBA1	CAACGAGCTCAAACTTAAGGAAGATA	plipBA
lipBA2	GAGGTCGACTAAAGCGTAAGCTTATA	plipBA

Construction of plipBA, plipA, pgspE1, pgspE2, pgspN, and pgspD plasmids.

The lipA and lipB genes were amplified from chromosomal DNA using the appropriate primers shown in Table 2. The product was ligated into a low-copy-number, broad-host-range vector, pMMB67EH, to make plipBA. This broad-host-range expression vector has been used in many Gram-negative species, including P. aeruginosa and V. cholerae (43), and is stably maintained in A. baumannii. The construct was verified by sequencing and conjugated from the E. coli MC1061 strain into WT and mutant A. baumannii strains. The plasmids overexpressing lipA, gspD, gspE1, gspE2, and gspN were constructed in the same manner.

Lipid agar.

Selective agar was utilized to detect extracellular lipase activity. The medium was prepared as described previously (44) with modified minimal medium (47.8 mM Na₂HPO₄, 22 mM KH₂PO₄, 8.5 mM NaCl, 18.7 mM NH₄Cl₂, 0.1% Tween 20, 0.2 mM CaCl₂), 40 μg/ml of neutral red, and 0.5% filtered-sterilized olive oil.

Lipase assay.

The strains were grown in LB broth supplemented with IPTG (isopropyl-β-d-thiogalactopyranoside) at a 50 μM final concentration to induce the expression of plasmid-encoded LipA. Following 16 h of growth, supernatants and cells were separated by centrifugation at 3,500 rpm for 10 min. A spectrophotometric assay was used to measure lipase activity by incubating culture supernatant with 0.9 mM 4-nitrophenyl myristate–80 mM Tris-HCl (pH 8.0)–0.15% Triton X-100 buffer at 37°C and measuring the release of 4-nitrophenol at 415 nm over time. All assays were performed in triplicate, and means and standard deviations of the results are presented.

SDS-PAGE and immunoblotting.

Culture supernatants were concentrated by precipitation utilizing pyrogallol red-molybdate-methanol as described previously (30). The samples were normalized to equivalent optical densities at 600 nm (OD₆₀₀) and subjected to SDS-PAGE and immunoblot analysis using antibodies raised against A. calcoaceticus LipA (1:100) (45) and goat anti-rabbit IgG–horseradish peroxidase (HRP). Immunoblots were imaged using a Typhoon Trio (Amersham Biosciences).

Serine hydrolase probe.

Overnight cultures of WT/plipBA and ΔgspD/plipBA strains were grown in LB broth–50 μM IPTG at 37°C. Supernatants and cells were separated by centrifugation at 3,500 rpm for 10 min. ActiveX FP serine hydrolase probe (0.5 μl) (Thermo Scientific) was added to 25 μl of culture supernatants and incubated at room temperature for 60 min. Samples were matched by equivalent OD₆₀₀ values, boiled in SDS sample buffer, subjected to SDS-PAGE on 4% to 12% Bis-Tris polyacrylamide gels (NuPAGE; Invitrogen), and visualized using a Typhoon Trio variable mode imager system and ImageQuant software.

In vitro competition assay.

WT and mutant A. baumannii strains were cultured separately overnight in LB broth at 37°C. Strains were diluted 1:100, and equivalent numbers of WT and mutant strains were pooled and cultured together at 37°C. At 0 (input), 8, and 24 h, aliquots of the mixed culture were diluted and plated on LB agar with or without kanamycin. CFU counts were determined after 24 h of incubation at 37°C. The competitive index (CI) was determined after 24 h as follows: CI = (mutant CFU/WT CFU)/(mutant input CFU/WT input CFU).

In vivo competition assay.

Eight-week-old female CBA/J mice (Jackson Laboratory) were injected intravenously with 150 μl and 100 μl of 20 mg/ml cyclophosphamide for 4 and 3 days, respectively, before the start of the experiment. Overnight cultures of WT, ΔlipA, and ΔgspD A. baumannii bacteria were diluted in phosphate-buffered saline (PBS). Inocula of 10⁷ cells at 1:1 ratios of WT:ΔlipA or WT:ΔgspD bacteria were administered via tail-vein injection. After 24 h, mice were euthanized. Spleens and livers were removed and homogenized in PBS, and CFU counts were determined after plating on LB agar with and without kanamycin and 24 h of incubation at 37°C. CI values were calculated as described above.

Statistical tests.

A Wilcoxon signed-rank test was calculated for the CIs obtained for the in vitro and in vivo competition assays. A Student t test was performed for the lipase activity assay. Values were considered significant at P values of ≤0.05.

Ethics statement.

All mouse experiments were performed according to the protocol (PRO00005052) approved by the University Committee on Use and Care of Animals at the University of Michigan. This protocol is in complete compliance with the guidelines for humane use and care of laboratory animals mandated by the National Institutes of Health.

RESULTS

A. baumannii encodes a functional T2S system.

We analyzed the sequenced genome of several A. baumannii strains and performed homology searches with the T2S genes of V. cholerae and P. aeruginosa. We identified genes for each of the T2S components GspC through GspN and PilD (Fig. 1). In contrast to most species, these genes are scattered around the genome in six noncontiguous segments rather than being linked into one or two operons, perhaps as a result of genome plasticity and the remarkable ability of A. baumannii to acquire foreign DNA. To determine whether A. baumannii has a functional T2S system, we utilized A. baumannii ATCC 17978, a strain that was originally isolated from a 4-month-old child with fatal meningitis and that is amenable to genetic inactivation and plasmid-based complementation studies. Using allelic exchange, we inserted a kanamycin resistance gene cassette in place of gspD, gspE1, gspE2, and gspN. The gspD gene was chosen because in all studied T2S systems, GspD, the outer membrane pore that serves as the conduit through which proteins are transported, is absolutely essential for the T2S. GspE contributes energy for the secretion process by hydrolyzing ATP and is also indispensable for secretion. However, as A. baumannii carries two potential gspE genes, gspE1 and gspE2, we deleted both genes to resolve whether the gspE1 gene or the gspE2 gene or both are required for secretion in A. baumannii. In contrast to the roles of GspD and GspE, the role of GspN has not yet been determined and its gene is not present in every species with a functional T2S system. As gspN is localized in the same operon as gspD in A. baumannii, we wanted to determine whether it is also required for T2S in A. baumannii. To complement the deletion strains, we constructed expression vectors encoding the wild-type (WT) copy of each mutant gene and expressed them in trans.

FIG 1.

T2S genes in A. baumannii. Putative T2S components are encoded by gspN (N; A1S_0269), gspC (C; A1S_0270), gspD (D; A1S_0271), gspE1 (E1; A1S_0616), gspE2 (E2; A1S_2290), gspF (F; A1S_0369), gspG (G; A1S_0370), gspH (H; A1S_1562), gspI (I; A1S_1563), gspJ (J; A1S_1564), gspK (K; A1S_1565), gspL (L; A1S_2255), gspM (M; A1S_2254), and gspO (also named pilD) (O; A1S_0327). The genes shown in gray were deleted using allelic exchange technology.

While the ΔgspD and ΔgspE1 mutants grew on LB agar as well as the WT, ΔgspE2, and ΔgspN strains, they were unable to grow on minimal agar with olive oil as the sole carbon source (Fig. 2A), a phenotype previously observed for T2S mutants of P. aeruginosa and V. cholerae due to their inability to secrete lipase (31, 44). Growth was restored when the ΔgspD and ΔgspE1 mutants were complemented with the appropriate expression plasmids (Fig. 2A).

FIG 2.

Growth on lipid agar requires LipA and a functional T2S system. Growth of strains on lipid agar, lipid agar with Orlistat, LB agar, and LB agar with Orlistat is shown from top to bottom. (A) Agar included carbenicillin and IPTG for plasmid maintenance and induction of expression of cloned genes. (B) Growth of reference strain ATCC 17978 (WT) and three clinical isolates.

Sequence analysis of A. baumannii ATCC 17978 identified genes with high homology to the genes of the P. aeruginosa and V. cholerae T2S substrate, LipA (Fig. 3), and its chaperone, LipB (46), suggesting that A. baumannii LipA may also be a T2S substrate that is capable of hydrolyzing lipids and generating nutrients for growth. To verify that LipA is the secreted substrate responsible for growth of A. baumannii on lipid agar, we constructed a lipA deletion strain by substituting the lipA gene for a gene encoding kanamycin resistance through homologous recombination. Similarly to the ΔgspD and ΔgspE1 mutants, the ΔlipA strain was unable to grow on the minimal lipid agar (Fig. 2A). Growth was restored when the lipA mutant was complemented with a plasmid encoding lipA and lipB. We also expressed the plasmid-encoded lipBA genes in the ΔgspD and ΔgspE1 deletion strain; however, despite overexpression of lipBA, no growth was observed on lipid agar (Fig. 2A). This suggests that the T2S system is active in A. baumannii and is responsible for the extracellular secretion of LipA.

FIG 3.

Sequence alignment of LipA from different species. The alignment of LipA from A. baumannii, V. cholerae, and P. aeruginosa is shown. The predicted N-terminal signal sequence is underlined. Yellow highlights indicate the catalytic residues.

When A. baumannii is cultured in the absence of lipids in LB broth under standard laboratory conditions, the lipBA genes are likely not expressed. This is consistent with our inability to detect significant differences in lipase activity between culture supernatants of the WT, ΔgspD, ΔgspE1, ΔgspE2, ΔgspN, and ΔlipA strains (Fig. 4). Therefore, to quantitatively measure the level of LipA secretion, we induced expression of plasmid-encoded lipBA genes in the WT, ΔgspD, ΔgspE1, ΔgspE2, and ΔgspN strains. This resulted in detectable lipase activity in the WT/plipBA, ΔgspE2/plipBA, and ΔgspN/plipBA culture supernatants using 4-nitrophenyl myristate, while no activity was observed in the ΔgspD/plipBA and ΔgspE1/plipBA culture supernatants (Fig. 4A). In addition, immunoblotting with antibodies directed against A. calcoaceticus LipA resulted in the detection of extracellular LipA when overexpressed in WT A. baumannii but not in the ΔgspD mutant, providing further support for the idea of a requirement of a functional T2S system in extracellular secretion of LipA (Fig. 4B).

FIG 4.

Extracellular secretion of A. baumannii lipase is dependent on an intact T2S system. (A) Enzymatic activity of overnight culture supernatants against the lipase substrate 4-nitrophenyl myristate was measured as a change in absorbance at 415 nm per min and normalized to the absorbance of the culture at 600 nm. Bars show standard deviations from the means (*, P ≤ 0.05; **, P ≤ 0.001). (B) Concentrated stationary-phase WT/plipBA (lane 1) and ΔgspD/plipBA (lane 2) culture supernatants were subjected to SDS-PAGE and immunoblot analysis using LipA antibodies. Molecular mass markers are shown on the left. The position of LipA is indicated.

Characterization of the secreted substrate, LipA.

The A. baumannii T2S substrate, LipA, is 54% identical to the well-characterized P. aeruginosa LipA, which is a member of proteobacterial lipase homology group I (47). Lipases within this group require helper proteins (chaperones) for their proper folding (48, 49). This relationship appears to be true in A. baumannii as well. The ΔlipA/plipA strain, which overexpresses only the lipase, exhibited lower extracellular lipase activity than the ΔlipA/plipBA strain, where both the lipase and the chaperone are overexpressed (Fig. 4). This difference is likely due to a reduction in secretion of LipA, as it has been shown that proper folding of LipA is required for the outer membrane translocation of LipA via the T2S system in Pseudomonas glumae (49). Members of lipase homology group I have been classified as serine hydrolases with active sites containing the residues Ser-His-Asp/Glu (50, 51). Consistent with the overall high sequence homology, sequence alignments indicate that the A. baumannii LipA has the catalytic residues Ser-His-Asp (Fig. 3). We used the ActiveX serine hydrolase probe, which binds covalently to the serine nucleophile in the active site of serine hydrolases, to further examine the catalytic property of LipA. As LipA activity is not detected in the supernatant of WT A. baumannii, we incubated supernatants of WT/plipBA and ΔgspD/plipBA cultures induced with increasing amounts of IPTG with this probe. In the WT/plipBA samples, we observed a band increasing in intensity with increasing IPTG induction that corresponded to a protein of ∼32 kDa, the expected size of LipA (Fig. 5). In contrast, this band was not present in any of the ΔgspD/plipBA samples. Taken together with the sequence homology between LipA of A. baumannii and P. aeruginosa, these findings indicate that A. baumannii LipA is a secreted serine hydrolase belonging to lipase homology group I.

FIG 5.

The ActiveX serine protease probe binds to LipA. Culture supernatants isolated from overnight WT/plipBA and ΔgspD/plipBA cultures induced with increasing concentrations (μM) of IPTG were incubated with the ActiveX serine protease probe for 1 h. Samples were subjected to SDS-PAGE and Typhoon image analysis as described in Materials and Methods. A molecular mass marker is shown on the right.

Orlistat is a semisynthetic, therapeutic drug prescribed for the treatment of obesity. It binds to the active site of human pancreatic lipase, thereby preventing the enzyme from breaking down dietary lipids in the intestine (52, 53). By incubating WT/plipBA culture supernatants with Orlistat, we learned that it also inhibits A. baumannii LipA (Fig. 4). Through its ability to inhibit LipA, Orlistat was therefore capable of inhibiting growth of A. baumannii when added to lipid agar but not when added to LB agar (Fig. 2). Similarly, the recently isolated clinical MDR strains of A. baumannii AYE, 0057, and 5075 were analyzed and shown to able to grow on lipid agar in the absence but not in the presence of Orlistat (Fig. 2B), suggesting that these MDR isolates also secrete lipases that can break down lipids and support fatty acid metabolism.

The T2S system and its secreted substrate, LipA, are required for A. baumannii colonization in a murine bacteremia model.

We utilized a modified murine bacteremia model originally developed for uropathogenic E. coli to determine if the T2S system and its substrate, LipA, are required for A. baumannii in vivo fitness (54). Due to the low virulence of A. baumannii ATCC 17978 in mice (55), we used an experimentally induced leukopenic mouse model (56), treating the mice with cyclophosphamide prior to infection. To diminish the potential of confounding factors between mice, we coinfected each mouse with equal numbers (1 × 10⁷ CFU/ml) of either WT and ΔgspD bacteria or WT and ΔlipA bacteria. After 24 h, mice were euthanized and the spleens and livers were harvested and homogenized. Homogenates were plated on LB agar with and without kanamycin, CFU counts were determined, and the competitive index (CI) was calculated for each mutant strain as described in Materials and Methods (Fig. 6). A CI value below 1 indicates a colonization defect of the mutant relative to the WT strain. Both mutants were outcompeted by the WT strain and exhibited colonization defects in both the spleen and the liver. The reduction in colonization was not simply due to the presence of the kanamycin resistance cassette, as a strain with the Tn5 transposon containing the kanamycin resistance gene inserted into the ddc gene of A. baumannii showed no colonization defect in the spleen compared to the WT strain in a recent study using the same bacteremia model (56). In addition, the reduction in colonization of the ΔlipA and ΔgspD mutants is unique to the in vivo environment, as the ΔgspD and ΔlipA mutants grow as well as WT A. baumannii in LB broth (see Fig. S1 in the supplemental material) and in vitro competition experiments in LB broth did not show a difference between WT and ΔlipA or WT and ΔgspD strains (CI = 0.98 ± 0.76 and CI = 0.84 ± 0.58, respectively).

FIG 6.

The gspD and lipA genes are required for A. baumannii fitness in a mouse model of bacteremia. Mice were coinoculated with equal numbers of either WT and ΔgspD or WT and ΔlipA strains. Mice were euthanized 24 h postinfection, the organs were harvested, and the CFU levels were determined for each strain. The competitive indexes were calculated as follows: (mutant CFU/WT CFU)/(mutant input CFU/WT input CFU). For all four competitions, P values = <0.001.

DISCUSSION

The T2S system controls the secretion of toxins and hydrolytic enzymes required for virulence in a variety of pathogens. Here, we present the results of the first study that identifies a functional T2S system in A. baumannii. Our study shows not only that A. baumannii possesses all the genes for a T2S system but also that there is a measurable reduction in extracellular LipA activity and loss of growth on lipid agar when gspD and gspE1, two of the core T2S genes, are deleted. In contrast, deletion of gspE2 has no effect on lipase secretion in A. baumannii. While GspE1 supports T2S and the homologous ATPase A1S_0329 likely provides energy for type IV pilus assembly, it is not clear what process GspE2 participates in, if any. Perhaps it affects T2S or type IV pilus biogenesis under very specific conditions. The finding that deletion of gspN had no effect on lipase secretion is consistent with its absence in many bacteria with functional T2S systems. In kind, the gspN homolog pulN is not required for pullulanase secretion via the T2S system in Klebsiella oxytoca (57). Similarly to GspE2, GspN may support T2S under only very specific growth conditions. Alternatively, it may substitute for GspC to support secretion of T2S substrates other than LipA (see below).

As mentioned earlier, the T2S genes are distributed throughout the genome in multiple operons in A. baumannii rather than in the one or two operons typical of other organisms encoding T2S systems. These smaller operons contain 1 to 4 gsp genes and, in some cases, may represent functional units that encode T2S components that directly interact. For example, GspC and GspD interact via the periplasmic domain of GspC and mutations that interfere with their interaction have a negative impact on secretion (58, 59). Interestingly, Xanthomonas campestris does not have a gspC gene; instead, it expresses GspN, which is required for secretion and has been shown to interact with GspD (60). Two proteins of the inner membrane platform, GspM and GspL, are encoded on another operon, and we have demonstrated in previous studies that these proteins stabilize each other within the cytoplasmic membrane and can be coimmunoprecipitated from V. cholerae (61). The minor pseudopilins GspI, GspJ and GspK interact and may form a subcomplex at the tip of the T2S pseudopilus (62), and they are encoded by the same operon that also contains the minor pseudopilin gspH gene. The genes for the inner membrane protein GspF and the major pseudopilin GspG are located in their own operon. Currently, there is no known interaction between GspF and GspG; however, the colocalization of their respective genes may suggest a potential functional interaction between these proteins in the T2S complex. Finally, while the cytoplasmic ATPase GspE forms a stable complex with GspL at the cytoplasmic membrane (63, 64), gspE1 and gspE2 are encoded separately from gspL on their own individual operons. The reason for this is not known, but, as discussed above, the data may suggest that GspE1 and GspE2 may be used under different conditions or for the secretion of different substrates. The unusual arrangement of the T2S genes presents an interesting avenue for further study.

While the exact role of LipA is currently unknown, it may be required for nutrient acquisition, breaking down lipids or long-chain fatty acids into shorter forms that are imported by the bacterium and used as sources of carbon and energy through β-oxidation (65). Besides being consumed as nutrients, fatty acids derived through lipid hydrolysis may be used as signaling molecules. Specifically, fatty acid signals are capable of restoring persister cells to a metabolically active state (66). The signaling molecule cis-2-dodecenoic acid contributes to virulence in cystic fibrosis Burkholderia cenocepacia infections (67) (68). While lipases may not have been considered virulence factors in the past, the P. aeruginosa LipA and LipC lipases contribute to motility, biofilm formation, pyoverdine production, and rhamnolipid production (69–71). Other lipases cleave host lipids to generate fatty acids that are integrated into the pathogens' own membranes (72). In addition, the extracellular esterase of group A Streptococcus reduces phagocyte recruitment by hydrolyzing platelet-activating factor (PAF) (73), and lipases produced and secreted by the opportunistic fungal pathogen Candida albicans support colonization and penetration of host cells (74, 75). Whether A. baumannii LipA is involved in nutrient acquisition or signaling or plays another important role in virulence will be explored in future studies.

The remarkable ability of A. baumannii to develop resistance to multiple antibiotics underscores the necessity for novel treatment approaches. Therapeutics that disarm A. baumannii or reduce its in vivo fitness are promising alternatives. One possibility is that of targeting LipA with a lipase inhibitor such as Orlistat. Alternatively, therapeutic targeting of secretion systems that transport multiple virulence factors to the exterior of the bacterium may be of particular interest, as their inactivation should have a greater impact than the targeting of individual virulence factors. It is quite possible that A. baumannii secretes additional proteins via the T2S besides LipA based on the following observations. First, the T2S system is active and supports secretion whether exogenous lipids are present or not, while lipA expression appears to require lipids/fatty acids (Fig. 2A and 4). Second, the T2S system is commonly required for the secretion of several proteins in other species. For example, proteomic analyses of culture supernatants of L. pneumophila, V. cholerae, and Burkholderia pseudomallei indicate that more than 20 different proteins may be dependent on the T2S system for extracellular release (30, 39, 76). Future work will focus on identifying additional T2S substrates and determining their role in A. baumannii pathogenesis.

Funding Statement

The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.