RESEARCH ARTICLE

Kdo Hydrolase Is Required for Francisella tularensis Virulence and Evasion of TLR2-Mediated Innate Immunity
Nihal A. Okan, Sabina Chalabaev, Tae-Hyun Kim, Avner Fink, Robin A. Ross, Dennis L. Kasper
Department of Microbiology and Immunobiology, Harvard Medical School, Boston, Massachusetts, USA
N.A.O. and S.C. contributed equally in this work.

ABSTRACT The highly virulent Francisella tularensis subsp. tularensis has been classiﬁed as a category A bioterrorism agent. A

live vaccine strain (LVS) has been developed but remains unlicensed in the United States because of an incomplete understanding of its attenuation. Lipopolysaccharide (LPS) modiﬁcation is a common strategy employed by bacterial pathogens to avoid innate immunity. A novel modiﬁcation enzyme has recently been identiﬁed in F. tularensis and Helicobacter pylori. This enzyme, a two-component Kdo (3-deoxy-D-manno-octulosonic acid) hydrolase, catalyzes the removal of a side chain Kdo sugar from LPS precursors. The biological signiﬁcance of this modiﬁcation has not yet been studied. To address the role of the twocomponent Kdo hydrolase KdhAB in F. tularensis pathogenesis, a kdhAB deletion mutant was constructed from the LVS strain. In intranasal infection of mice, the kdhAB mutant strain had a 50% lethal dose (LD50) 2 log10 units higher than that of the parental LVS strain. The levels of the proinﬂammatory cytokines tumor necrosis factor alpha (TNF- ) and interleukin-1 (IL-1 ) in bronchoalveolar lavage ﬂuid were signiﬁcantly higher (2-fold) in mice infected with the kdhAB mutant than in mice infected with LVS. In vitro stimulation of bone marrow-derived macrophages with the kdhAB mutant induced higher levels of TNF- and IL-1 in a TLR2-dependent manner. In addition, TLR2 / mice were more susceptible than wild-type mice to kdhAB bacterial infection. Finally, immunization of mice with kdhAB bacteria elicited a high level of protection against the highly virulent F. tularensis subsp. tularensis strain Schu S4. These ﬁndings suggest an important role for the Francisella Kdo hydrolase system in virulence and offer a novel mutant as a candidate vaccine.
IMPORTANCE The ﬁrst line of defense against a bacterial pathogen is innate immunity, which slows the progress of infection and

allows time for adaptive immunity to develop. Some bacterial pathogens, such as Francisella tularensis, suppress the early innate immune response, killing the host before adaptive immunity can mature. To avoid an innate immune response, F. tularensis enzymatically modiﬁes its lipopolysaccharide (LPS). A novel LPS modiﬁcation—Kdo (3-deoxy-D-manno-octulosonic acid) saccharide removal— has recently been reported in F. tularensis. We found that the kdhAB mutant was signiﬁcantly attenuated in mice. Additionally, the mutant strain induced an early innate immune response in mice both in vitro and in vivo. Immunization of mice with this mutant provided protection against the highly virulent F. tularensis strain Schu S4. Thus, our study has identiﬁed a novel LPS modiﬁcation important for microbial virulence. A mutant lacking this modiﬁcation may be used as a live attenuated vaccine against tularemia.
Received 30 December 2012 Accepted 7 January 2013 Published 12 February 2013 Citation Okan NA, Chalabaev S, Kim T, Fink A, Ross RA, Kasper DL. 2013. Kdo hydrolase is required for Francisella tularensis virulence and evasion of TLR2-mediated innate immunity. mBio 4(1):e00638-12. doi:10.1128/mBio.00638-12. Editor Rino Rappuoli, Novartis Vaccines and Diagnostics Copyright © 2013 Okan et al. This is an open-access article distributed under the terms of the Creative Commons Attribution-Noncommercial-ShareAlike 3.0 Unported license, which permits unrestricted noncommercial use, distribution, and reproduction in any medium, provided the original author and source are credited. Address correspondence to Dennis L. Kasper, dennis_kasper@hms.harvard.edu.

F

rancisella tularensis, one of the deadliest respiratory pathogens in existence, is considered a potential biological weapon because it is readily aerosolized and exhibits a high degree of infectivity and lethality in humans. An attenuated live vaccine strain (LVS) has been developed but has not been licensed as a vaccine because of an incomplete understanding of the basis for its attenuated virulence and associated side effects. There has been a widespread research effort aimed at elucidating Francisella pathogenesis and identifying components for rational vaccine design. F. tularensis is a successful pathogen that can evade and suppress innate immune responses. After intranasal infection of mice, F. tularensis initially colonizes the lung, infecting several cell types but failing to stimulate proinﬂammatory cytokine production for

the ﬁrst 2 to 3 days (1–5). This initial delay in inﬂammation contributes to the ability of F. tularensis to replicate in the lung, spread into systemic organs, and cause sepsis leading to the host’s death (4). To suppress the early proinﬂammatory cytokine response, F. tularensis employs a variety of strategies, such as avoiding disclosure of itself to the innate immune system (6, 7). Lipopolysaccharide (LPS) of F. tularensis plays a signiﬁcant role in evasion of host immune responses (8–11). LPS, the major component of Gram-negative bacterial outer membranes (12, 13), itself has three components: (i) lipid A, a glucosamine-based glycolipid; (ii) a core oligosaccharide ketosidically linked to lipid A by an eight-carbon sugar, 3-deoxy-D-manno-octulosonic acid (Kdo); and (iii) the O-antigen polysaccharide comprising multi-

January/February 2013 Volume 4 Issue 1 e00638-12

®

mbio.asm.org 1

Okan et al.

FIG 1 LPSs from F. tularensis LVS and kdhAB mutant strain differ from
each other by a side chain Kdo. The structure of core oligosaccharide synthesized by LVS (55) and the proposed structure of core oligosaccharide synthesized by the kdhAB mutant, as determined by HPAEC (14), are shown.

ple repeating units of a complex tetrasaccharide. The lipid A of many bacterial species such as Escherichia coli is a powerful stimulant of Toll-like receptor 4 (TLR4)-mediated innate immune responses. In contrast, F. tularensis lipid A triggers no TLRmediated innate response because of a number of key structural modiﬁcations (9); speciﬁcally, despite initial similarity to E. coli lipid A, the lipid A precursor of F. tularensis undergoes dephosphorylation, deacylation, and carbohydrate incorporation that render the molecule undetectable by the TLR4 receptor. Differences in the length of the fatty acid chain of lipid A (16 to 18 carbons for F. tularensis as opposed to 12 to 14 carbons for E. coli) further enhance the ability of F. tularensis to evade TLR4mediated innate immunity. These modiﬁcations contribute to bacterial virulence (8, 10). A novel LPS-modifying two-component KdhAB Kdo hydrolase enzyme complex has recently been identiﬁed in F. tularensis (14, 15). A core LPS precursor containing two Kdo residues is ﬁrst synthesized. One Kdo in the backbone of the core polysaccharide is ketosidically linked to lipid A and glycosidically linked to the remainder of the core polysaccharide, and the second Kdo residue is glycosidically attached as a side chain. KdhAB removes the side chain Kdo, leaving an LPS molecule containing only one Kdo residue (Fig. 1) (14–16). The KdhAB Kdo hydrolase is a twocomponent enzyme composed of a sialidase-like protein (the catalytic subunit) and a small inner membrane protein (perhaps a membrane-anchoring subunit) (15). This novel enzyme complex has been reported in two other pathogens, Helicobacter pylori and Legionella pneumophila (14, 17, 18). Its biological signiﬁcance—in particular, its role in pathogenicity— has not yet been investigated. Here, we show that the F. tularensis kdhAB LVS mutant is attenuated in a mouse model of infection and stimulates—in a TLR2-dependent manner—a stronger proinﬂammatory cytokine response than the parental LVS. Furthermore, immunization of mice with the kdhAB mutant provides signiﬁcant protection against fully virulent F. tularensis type A strain Schu S4.
RESULTS

The F. tularensis LVS kdhAB mutant is attenuated in mice. To fully evaluate the importance of the KdhAB Kdo hydrolase in the virulence of F. tularensis LVS, kdhA and kdhB genes were deleted by homologous recombination in an LVS background, giving rise to the kdhAB mutant. Previously described analysis by highperformance anion-exchange chromatography (HPAEC) (14)

conﬁrmed that, unlike the LPS of parental LVS, which has a single Kdo residue, the LPS of the kdhAB mutant has two Kdo residues (Fig. 1; see Table S1 in the supplemental material). To evaluate the role of Kdo hydrolase activity in F. tularensis virulence, we infected BALB/cByJ mice with graded doses of LVS or the kdhAB mutant via the intranasal route. All mice challenged with 103 CFU of LVS succumbed to the infection within 9 days (Fig. 2A). In contrast, all mice challenged with a similar dose of the kdhAB mutant survived (Fig. 2A). At a dose of 105 CFU of the kdhAB mutant, 17% of the infected mice survived (Fig. 2A). The observed phenotypes in the kdhAB mutant were successfully complemented with the pkdhAB plasmid carrying the kdhA and kdhB genes. LPS analysis showed that the complemented strain had only a single Kdo residue in its core LPS (see Table S1 in the supplemental material). In addition, complementation of kdhAB restored virulence to the wild-type (WT) level. The survival rates of BALB/cByJ mice (10 mice in each group) 28 days after an intranasal challenge with 5 103 F. tularensis LVS or a kdhAB mutant harboring either a vector control (pFNLTP6) or a plasmid carrying kdhAB genes (pkdhAB) were as follows: 0% for mice infected with LVS, 100% for mice infected with the kdhAB mutant carrying pFNLTP6, and 20% for mice infected with the kdhAB mutant carrying pkdhAB. These data indicate that Kdo hydrolase activity is important for the virulence of F. tularensis LVS. We also examined the ability of the kdhAB mutant to colonize the lung and to disseminate into other organs after intranasal challenge of mice. Six days after infection with 103 bacteria, both parental LVS and its kdhAB mutant had reached a similar level of colonization in the lungs, with counts of 3 106 CFU/g and 107 CFU/g, respectively (Fig. 2B). However, the numbers of kdhAB mutant bacteria in spleen, liver, and blood were 1.5 to 3 logarithmic units lower than the numbers of LVS bacteria (Fig. 2B) (Student’s t test, P 0.05). Overall, these results indicate that the kdhAB mutant is deﬁcient in its ability to disseminate systemically. To determine whether the observed defect in dissemination is related to bacterial growth kinetics, growth rates were compared under different conditions. There was no difference in growth between LVS and the kdhAB mutant when bacteria were cultured in broth medium or in J774A.1 macrophage-like and MH-S alveolar macrophage-like cell lines (see Fig. S1A and S1B in the supplemental material). Moreover, the kdhAB mutant was not sensitive to complement-mediated killing (Fig. S1C). Therefore, the defect of the kdhAB mutant in dissemination to systemic organs was not due to a growth defect or to serum sensitivity. The cytokine response to the F. tularensis LVS kdhAB mutant is enhanced. Within the ﬁrst day after infection, mutants of F. tularensis that cannot modify their lipid A elicit a stronger proinﬂammatory response than parental LVS did (8, 19). Since the KdhAB hydrolase is involved in the modiﬁcation of the LPS core oligosaccharide but not in the modiﬁcation of lipid A, we compared the proinﬂammatory responses to LVS and the kdhAB mutant. To this end, mice were challenged intranasally with 103 CFU of either LVS or the kdhAB mutant, and the levels of tumor necrosis factor alpha (TNF- ) and interleukin-1 (IL-1 ) in samples of bronchoalveolar lavage (BAL) ﬂuid were determined in an enzyme-linked immunosorbent assay (ELISA). Consistent with previous reports (6), the levels of TNF- and IL-1 in

2

®

mbio.asm.org

January/February 2013 Volume 4 Issue 1 e00638-12

Characterization of kdhAB in Francisella tularensis

FIG 2 The virulence of the F. tularensis LVS kdhAB mutant strain is attenuated in mice. (A) Survival of mice after F. tularensis intranasal challenge. BALB/cByJ mice (n 6) were challenged intranasally with LVS (103 CFU) or kdhAB bacteria (103, 104, or 105 CFU). (B) Bacterial burden in lung, spleen, liver, and blood of mice 6 days after infection. BALB/cByJ mice (n 3) were challenged intranasally with LVS bacteria (103 CFU) or kdhAB bacteria (103 CFU).

BAL ﬂuid 24 h after inoculation of LVS bacteria were not different from those in control mice given phosphate-buffered saline (PBS) (Fig. 3A, left panel). In contrast, the levels of TNF- and IL-1 were signiﬁcantly higher (~2-fold) in mice infected by the kdhAB mutant than in mice infected by LVS or control mice given PBS (Fig. 3A, left panel). The stronger proinﬂammatory response was not due to differences in bacterial burden, as the numbers of kdhAB and LVS bacteria in the lungs were similar (Fig. 3A, right panel). Therefore, in contrast to LVS, the kdhAB mutant elicited an early proinﬂammatory response that was detected 24 h after infection. To test whether the in vivo cytokine response also differs in ex vivo macrophages, bone marrow-derived macrophages (BMMs) from C57BL/6J mice were infected at different multiplicities of infection (MOIs) with either LVS or the kdhAB mutant. Supernatants were collected after 24 h, and secretion of TNF- and IL-1 was measured by ELISA. At all MOIs tested, TNF- secretion by BMMs stimulated by the kdhAB mutant was higher by a factor of at least 2 than that by BMMs stimulated by LVS (Fig. 3B, left panel). Results were similar for IL-1 secretion (Fig. 3B, right panel). Consistent with the cytokine secretion proﬁle, the expression of TNF- and IL-1 genes was higher in kdhAB mutantstimulated BMMs than in LVS-stimulated BMMs (Fig. 3C). The F. tularensis LVS kdhAB mutant activates the TLR2 signaling pathway. To investigate the basis for the enhanced innate immune response to the kdhAB mutant, we assessed the re-

sponses of murine macrophage-like cell lines in which various TLR signaling pathways were inactivated. Cells were infected with either LVS or the kdhAB mutant at an MOI of 100 and lysed 24 h later for measurement of IL-1 . Cell activation was calculated as the IL-1 ratio in infected cells versus uninfected cells. LVS bacteria elicited a proinﬂammatory response in WT, Unc93B / , and TLR4 / cells, but not in MyD88 / or TLR2 / cells (Fig. 4A). These data are in agreement with previous reports that the innate immune response to LVS is mediated through TLR2 signaling (3, 20). Interestingly, the kdhAB mutant showed a similar pattern of activation, a result indicating that TLR2— but not TLR3, -4, -7, or -9 —mediates the innate immune response to kdhAB mutant bacteria. Moreover, the proinﬂammatory response was signiﬁcantly greater in kdhAB mutant-infected cells than in LVSinfected cells (Fig. 4A). Control experiments conﬁrmed that WT cells and knockout cells for Unc93B (required for TLR3, TLR7, and TLR9 signaling), MyD88, TLR4, and TLR2 responded to various TLR agonists as expected (data not shown). To assess whether kdhAB mutant-mediated enhancement of the proinﬂammatory cytokine response is likely to be due to an increase in nuclear factor B (NF- B) activation, we used HEK293 cells expressing TLR2 (HEK-TLR2) or TLR4 and MD2 (HEK-TLR4/MD2). The stable cell lines were transiently transfected with the reporter plasmid pNF B-luc, which expresses a ﬁreﬂy luciferase gene under control of an NF- B promoter, and were incubated with LVS or kdhAB organisms. Reporter activity in HEK-TLR4/MD2 cells incubated with bacteria of either strain did not differ (Fig. 4B). On the other hand, reporter activity increased slightly but signiﬁcantly in HEK-TLR2 cells stimulated with the kdhAB mutant over that in HEK-TLR2 cells stimulated with LVS (P 0.05) (Fig. 4B). Control experiments using E. coli LPS and Pam3CSK, known TLR4 and TLR2 antagonists, respectively, showed high NF- B reporter activity as expected (Fig. 4B). Mechanistic basis for activation of TLR2 by the kdhAB mutant. One possible explanation for the increase in TLR2-mediated NF- B activity is that, unlike WT LVS LPS, the mutant LPS produced by the kdhAB mutant is a TLR2 agonist (9). To test this hypothesis, we assessed puriﬁed LPS preparations from each strain for their ability to activate NF- B in HEK-TLR2 cells. LPS from LVS and the kdhAB mutant showed similar levels of NF- B reporter activity, which was also similar to that elicited by whole LVS organisms (Fig. 4B). Additionally, BMMs did not respond to LPS from either LVS or the kdhAB mutant by secretion of TNFand IL-1 when the puriﬁed LPS was tested at concentrations of 50, 250, 1,250, and 10,000 ng/ml (see Fig. S2 in the supplemental material). Taken together, these results show that the kdhAB mutant elicits stronger NF- B activation in a TLR2-dependent manner, with a consequently greater proinﬂammatory response. However, the greater proinﬂammatory response induced by the mutant LPS probably is not due to activation of TLR2 by the mutant LPS per se, as the LPS of the kdhAB mutant alone does not trigger TLR2 signaling in vitro. To further investigate the mechanism for enhanced TLR2 signaling by the kdhAB mutant, we evaluated bacterial membrane integrity. A leaky cell membrane caused by kdhAB mutation would potentially release TLR2 ligands into the culture supernatant to be detected by host cells. Thus, the growth behaviors of LVS and the kdhAB mutant were tested in the presence of SDS and EDTA, agents that inhibit growth of bacterial cells with defective outer membranes. The similar growth characteristics of the two

January/February 2013 Volume 4 Issue 1 e00638-12

®

mbio.asm.org 3

Okan et al.

FIG 3 Enhanced proinﬂammatory cytokine response to the F. tularensis LVS kdhAB mutant strain. (A) BALB/cByJ mice (n 3) were challenged intranasally with LVS bacteria (103 CFU), kdhAB bacteria (103 CFU), or PBS (control). BAL ﬂuid was collected 24 h after infection, and the levels of TNF- and IL-1 were measured by ELISA (left panel). The bacterial burden in lung was measured 24 h after infection (right panel). (B) Levels of TNF- and IL-1 released by ex vivo BMMs after stimulation with LVS or kdhAB bacteria (MOI of 10, 20, or 50). Culture supernatants were analyzed by ELISA 24 h after infection. (C) TNF- and IL-1 gene expression in BMMs infected with LVS or kdhAB bacteria (MOI of 10). Gene expression was measured by real-time PCR, normalized with RP II gene expression, and reported as relative expression compared with that in uninfected macrophages. All data are presented as mean plus SD (error bars) values (n 3).

4

®

mbio.asm.org

January/February 2013 Volume 4 Issue 1 e00638-12

Characterization of kdhAB in Francisella tularensis

FIG 4 Bacteria of the F. tularensis LVS kdhAB mutant strain activate NF- B
through the TLR2 signaling pathway. (A) Macrophage-like cell lines were stimulated with LVS or kdhAB bacteria for 2 h (MOI 100). The IL-1 level in cell lysates was assessed by ELISA. Measurements were expressed as fold induction in infected versus uninfected cells (mean plus SD; n 3). (B) HEK293 cells stably expressing TLR2 or TLR4/MD2 were transiently transfected with an NF- B luciferase reporter plasmid and stimulated for 24 h with LVS or kdhAB bacteria (MOI 20) or with Pam3CSK (1 g/ml) or E. coli LPS (100 ng/ml) as positive controls. Puriﬁed LPSs from LVS or kdhAB bacteria (5 g/ml) were tested in a mixture with soluble CD14 recombinant protein (5 g/ml). The data are reported as relative luciferase units and represent fold induction of luciferase activity in the cell lysate compared with that in unstimulated cells (mean plus SD; n 3).

strains (Fig. 5A) strongly supported the notion that the kdhAB mutation does not compromise cell membrane integrity. In an alternative approach, we tested the ﬁltered supernatants of LVS or kdhAB cultures for their ability to activate BMMs in vitro. The culture supernatants were concentrated by Centricon ﬁlter devices and incubated with BMMs for 24 h. Macrophage activation was measured by the release of TNF- and IL-6. While the addition of LPS or PAM3CSK (as positive controls) to the BMMs did result in signiﬁcant activation of the cells, little activation was seen after the addition of the LVS- or kdhAB mutant-conditioned media (see Fig. S3 in the supplemental material). The similarly low levels of TNF- and IL-6 for both LVS and kdhAB organisms suggested that there is no leakage of potential TLR2 ligands due to the kdhAB mutation. Unable to detect a leaky-membrane phenotype, we next investigated changes in bacterial LPS and cell-surface protein availability. It has been proposed that fully formed LPS on the bacterial surface plays a vital role in preventing the binding of serum complement, antibodies, and receptors of the innate immune system (21–24). Toward this end, the analysis of total LPS level and a characteristic LPS banding pattern showed no difference between LVS and the kdhAB mutant (Fig. 6A). Additionally, matrixassisted laser desorption ionization time of ﬂight (MALDI-TOF) analysis of lipid A indicated that LVS and the kdhAB mutant

have similar lipid A mass values (Fig. 6B). Although the kdhAB mutation does not affect the total LPS level or the characteristic banding pattern of LPS, we thought it is possible that the additional Kdo—a negatively charged sugar—in the kdhAB mutant alters the conformation of LPS molecules, thereby making other surface constituents more accessible to TLR2 receptors. To test this hypothesis, we evaluated the binding of FopA antibodies to the FopA surface protein in vitro. Live bacteria were incubated with mouse antibodies to FopA, and surface-bound antibodies were analyzed by Western blotting. Consistent with a previous study (24), fully formed LPS signiﬁcantly reduced the binding of antibodies to the LVS bacterial surface, whereas wbtA mutant LPS, which is devoid of O polysaccharide, did not (Fig. 5B). Strikingly, substantial binding of FopA antibodies to the kdhAB mutant bacterial surface was found, and this result suggested that a TLR2 ligand present on the bacterial cell surface would be potentially exposed for detection by host immune-cell TLR2 receptors. The accessibility of surface proteins might well explain the enhanced TLR2 signaling activity caused by the kdhAB mutation. Biological impact of TLR2 activation by the kdhAB mutant. We next addressed the biological importance of the TLR2mediated response to kdhAB bacteria in a mouse infection model. WT C57BL/6, TLR2 / , and TLR4 / mice were infected with various doses of the kdhAB mutant via the intranasal route and monitored for survival. Most strikingly, at a dose of 104 CFU, all WT and TLR4 / mice survived, while no TLR2 / mice survived (Table 1). The results suggest that TLR2-mediated responses play a signiﬁcant role in the survival of mice infected with the F. tularensis kdhAB mutant strain. Thus, the association of KdhAB Kdo hydrolase with enhanced TLR2-mediated inﬂammation points to a critical role for this enzyme complex in the virulence of LVS. Immunization with F. tularensis LVS kdhAB mutant protects against the highly virulent F. tularensis Schu S4 strain. Immunization with LVS is protective against the fully virulent F. tularensis type A strain Schu S4 in a mouse model of infection, and this protection involves TLR2-dependent innate immunity (25– 27). However, even in low numbers, LVS retains signiﬁcant virulence in mice. Our ﬁndings that the kdhAB mutant is further attenuated and elicits enhanced TLR2 signaling led us to investigate the protective capacity of this mutant strain as a live cell vaccine. To this end, BALB/cByJ mice were immunized with three intranasal doses of LVS, the kdhAB mutant, or (as a control) PBS. Four weeks after the last dose, the mice were challenged via the intranasal route with either LVS at a dose of 104 CFU (10 LD50) or the Schu S4 strain at a dose of 10 CFU (10 LD50). All mice that had been immunized with LVS survived LVS challenge, whereas all control mice that had received PBS succumbed within 9 days (Fig. 7A). The kdhAB mutant was also fully protective against LVS challenge (Fig. 7A). Furthermore, as expected, LVSimmunized mice survived the fully virulent Schu S4 challenge, while all PBS-treated control mice succumbed within 7 days (Fig. 7B). Strikingly, mice immunized with the kdhAB mutant were well protected; 90% of these mice survived an otherwise lethal challenge with Schu S4. Histologic examination of the livers from infected mice revealed that kdhAB mutant-immunized mice had much less tissue necrosis after challenge with Schu S4 (Fig. 7C) than did PBS-treated control mice. The protection conferred by the kdhAB mutant was also reﬂected by lower bacterial loads in the livers, spleens, and lungs from mutant-immunized

January/February 2013 Volume 4 Issue 1 e00638-12

®

mbio.asm.org 5

Okan et al.

Kdo from its LPS (Fig. 1), was impaired in its ability to kill mice after intranasal challenge (Fig. 2). This attenuation was not due to a growth defect or to sensitivity to complement-mediated killing. In H. pylori, Kdo hydrolase inactivation has downstream effects on subsequent steps in LPS biosynthesis, reducing the efﬁciency of lipid A modiﬁcation enzymes 4=phosphatase and 3=-O-deacylase and the amount of fully extended O antigen (17). Both 4=-phosphatase and O antigen are required for full virulence of F. tularensis in mice (10, 19, 29). However, Kdo hydrolase inactivation in F. tularensis LVS did not affect lipid A or O-antigen biosynthesis (Fig. 6). Therefore, virulence attenuation was not related to a defect in 4=phosphatase activity or O-antigen assembly. The avoidance of the TLR response by modiﬁcations in lipid A structure has been described in a numFIG 5 Mutation in kdhAB leads to enhanced accessibility of the bacterial surface. (A) ber of pathogens, including members of the genera Investigation of the leaky-membrane phenotype. Mid-logarithmic-phase cultures were Salmonella, Yersinia, Pseudomonas, and Francisella (8, spread onto CHAH plates, and a 10- l volume of SDS (20, 10, and 5%) or EDTA (500, 10, 30–32). These modiﬁcations include lipid A deacy250 and 125 mM) was spotted onto ﬁlter discs. Images were captured after 48 h of lation, dephosphorylation, and glycosylation. We thus bacterial growth. Black arrows indicate zone of inhibition around 6-mm ﬁlter discs. (B) F. tularensis LVS and kdhAB and wbtA mutant strains were tested for binding of hypothesized that F. tularensis removes side chain Kdo mouse anti-FopA antibodies to their surface FopA proteins in an antigen accessibility from LPS core oligosaccharide to avoid the triggering assay. Bacteria—along with surface-bound FopA antibodies—were lysed, and proteins of innate immunity. In agreement with this hypothewere resolved by 5 to 20% SDS-PAGE. Bound FopA antibodies were detected as IgG sis, the kdhAB mutant stimulated a stronger TLR2heavy chain (HC) and IgG light chain (LC). Total FopA and bacterioferritin (Bfr) protein levels are shown as protein loading controls. mediated proinﬂammatory response in macrophage monolayers than the parental LVS did (Fig. 3 and 4). In vivo, kdhAB bacteria induced a proinﬂammatory remice (Fig. 7D) than in these tissues from PBS-treated control sponse in the lungs 24 h after intranasal challenge, while LVS mice. The Schu S4 bacterial loads in organs were similar in LVS- bacteria did not (Fig. 3A, left panel). Attenuation of the virulence and kdhAB mutant-immunized mice. Finally, the levels of of kdhAB bacteria may be related to this early innate immune Francisella-speciﬁc IgG antibody were slightly higher in the sera of response. Indeed, several studies have shown that eliciting an early kdhAB mutant-immunized mice compared to LVS-immunized inﬂammatory response in mice by priming the animals with an mice (Fig. 7E). agonist of TLR3, TLR4, or TLR9 enhances survival to F. tularensis infection and decreases organ bacterial burdens (33–35). DISCUSSION Recent studies have indicated that Kdo residues contribute to Evasion of innate immunity is a key virulence factor for a number proinﬂammatory responses to LPS (36–38). Therefore, the presof pathogens, including F. tularensis (6, 7). To avoid detection by ence of a side chain Kdo in the kdhAB mutant, with direct recpattern recognition receptors such as TLRs, F. tularensis modiﬁes ognition of this modiﬁed LPS by the TLR2 signaling pathway, has its lipid A through the action of enzymes such as 4=-phosphatase, emerged as a potential mechanism for the observed phenotype. galactosamine transferase, and mannosyl transferase (8, 10). A However, our in vitro studies suggest that this is not the case, as novel LPS modiﬁcation was hypothesized in F. tularensis and puriﬁed LPSs from the kdhAB mutant and LVS elicit similarly H. pylori: the removal of side chain Kdo from Kdo2-lipid A pre- weak NF- B activation (Fig. 4B). Another possible explanation is cursors by a Kdo hydrolase (16, 17). Such a modiﬁcation was that the structural alteration of LPS exposes one or more TLR2 recently conﬁrmed, and the genes coding for a two-component ligands present on the bacterial surface. The electrostatic charge of Kdo hydrolase were identiﬁed in F. tularensis and H. pylori (14, 15, LPS is important for the folding of certain outer membrane pro17, 18). The biological signiﬁcance of the Kdo hydrolase, which is teins (39–42), and Kdo—as a negatively charged sugar— contribalso expressed by L. pneumophila (14), remained to be addressed. utes to the overall charge of LPS. Thus, the presence of an addiBy screening an F. tularensis subsp. novicida transposon mutant tional Kdo residue in kdhAB mutant LPS may alter the library, Weiss and colleagues showed that a kdhB transposon mu- conformation of the outer membrane, exposing TLR2 ligands tant was attenuated in mice, but no further characterization has such as lipoproteins (43). In support of this idea, we showed that been reported (28). Here we demonstrate that the two- FopA surface protein is more accessible to antibodies in the component Kdo hydrolase KdhAB is important in the pathogekdhAB mutant than in parental LVS (Fig. 5B). nicity of F. tularensis LVS bacteria and in the organism’s avoidance LVS is an attenuated F. tularensis subsp. holarctica strain (type of the TLR2-mediated immune response. We further show that a B) that was empirically derived in the former Soviet Union and Kdo hydrolase-deﬁcient mutant is protective against the fully vir- developed as a live attenuated vaccine against the highly virulent ulent F. tularensis type A strain Schu S4. Such protective capacity is F. tularensis. However, the LVS-based vaccine remains unlicensed obviously a key feature to be hoped for in an effective vaccine in the United States and several other nations for several reasons, strain. including the following: (i) an incomplete understanding of the The kdhAB deletion mutant, unable to remove side chain genetic basis for its attenuation; (ii) untoward reactions; and (iii)

6

®

mbio.asm.org

January/February 2013 Volume 4 Issue 1 e00638-12

Characterization of kdhAB in Francisella tularensis

FIG 6 The F. tularensis kdhAB mutant strain is not altered in the biosynthesis of O antigen or lipid A. (A) SDS-PAGE analysis of F. tularensis LVS and kdhAB
mutant outer membranes. LPS was visualized by zinc staining. (B) Negative-ion MALDI-TOF analysis of lipid A from LVS and kdhAB bacteria.

concerns about possible reversion back to the virulent form. The mechanisms of LVS-mediated protection against tularemia involve TLR2-dependent innate immunity (25–27). Therefore, an LVS mutant that exhibits attenuated virulence in mice and more strongly elicits activation of TLR2 signaling might prove to be a safer and more effective vaccine candidate than the original LVS. Here we show that the kdhAB LVS mutant may be such a vaccine candidate: its virulence is attenuated, its ability to activate TLR2 is enhanced, and its protective efﬁcacy against F. tularensis type A strain Schu S4 is similar to that of LVS. Mutants unable to modify lipid A by carbohydrate incorporation were attenuated in mice, activated proinﬂammatory responses, and were protective against F. tularensis subsp. novicida (8). However, protection against the highly virulent F. tularensis type A strain Schu S4 has not been investigated. Few previously assessed mutants have been shown to induce protective immunity against type A strains (for a review, see reference 44). These mutants have had deletions in genes including FTT1103, FTT0918, sodB, wzy, and ggt (45, 46). We report a new mutation that is relevant for tularemia vaccine design.

Herein we demonstrate that the two-component Kdo hydrolase KdhAB is involved in the virulence of F. tularensis LVS. A kdhAB mutant is attenuated in mice, elicits a stronger proinﬂammatory cytokine response in a TLR2-dependent manner, and induces protective immunity against the highly virulent Schu S4 strain of F. tularensis. Our data suggest that a novel LPS modiﬁcation—the removal of side chain Kdo—is important in the evasion of innate immunity by F. tularensis. Kdo hydrolase is also produced by two other pathogens, H. pylori and L. pneumophila (14, 15, 17, 18). It will be particularly interesting to study the role of this enzyme in the virulence of these organisms.
MATERIALS AND METHODS
Ethics statement. All work with animals was conducted according to the recommendations in the Guide for the Care and Use of Laboratory Animals (47). The protocol was approved by the Standing Committee on Animals of the Harvard Medical Area (protocol 04723). Bacterial strains and plasmids. Francisella tularensis LVS (kindly provided by Karen Elkins, U.S. Food and Drug Administration, Bethesda, MD) was grown in cysteine heart agar supplemented with 1% hemoglobin (CHAH) for 72 h at 37°C and 5% CO2, in modiﬁed Muller-Hinton broth (Becton, Dickinson, Franklin Lakes, NJ) supplemented with ferric pyrophosphate and IsoVitaleX (Becton, Dickinson, Franklin Lakes, NJ), or in tryptic soy broth (Difco, Detroit, MI). Escherichia coli strain DH5 was used for general cloning and was grown either in Luria-Bertani (LB) broth or on LB agar plates. When appropriate, ampicillin (100 g/ml) and kanamycin (50 g/ml) were added to broth medium, and kanamycin (10 g/ ml) was added to CHAH. Construction of the F. tularensis LVS kdhAB deletion mutant. The kdhAB mutant was constructed by allelic exchange (48) by a procedure described previously (14). The whole kdhA gene was deleted. As the last

TABLE 1 Survival of WT C57BL/6, TLR2 / , and TLR4 / mice 21 days after an intranasal challenge with F. tularensis kdhAB mutant
Survival rate (%) of mice: Challenge dose (CFU) 102 103 104 C57BL/6 (n 10) 100 100 100 TLR2 / (n 10) 100 40 0 TLR4 / (n 5) 100 100 100

January/February 2013 Volume 4 Issue 1 e00638-12

®

mbio.asm.org 7

Okan et al.

FIG 7 Bacteria of the F. tularensis LVS kdhAB mutant strain elicit protective immunity against F. tularensis LVS and type A strain Schu S4. (A and B) Groups
of BALB/cByJ mice (10 mice in each group) were immunized intranasally with 3 doses of LVS or kdhAB bacteria; 28 days after the last dose, the mice were challenged intranasally with either 104 CFU of LVS (10 LD50) (A) or 10 CFU of strain Schu S4 (10 LD50) (B). (C) Hematoxylin and eosin (H&E) staining of liver tissue on day 5 after Schu S4 challenge. The black arrowhead indicates extensive tissue necrosis (10 ). (D) Bacterial burden in lung, spleen, and liver of mice 5 days after Schu S4 challenge (mean plus SD; n 3). (E) Titers of serum antibody to F. tularensis LPS were determined by ELISA. Sera from LVSimmunized, kdhAB mutant-immunized, and unimmunized control mice were collected on the day before challenge. ELISA was performed as described in Materials and Methods.

eight bases of kdhA overlap with kdhB, the ﬁrst eight bases of kdhB were also deleted. The following primers were used: 5=-ACCCCCGGGCACTT TCGTGCTTAGATATCT-3= (XmaI site underlined) and 5=-TGTGGTAC CAGTTTTAAAAATTAAGTAAGT-3= (KpnI site underlined) to amplify the 1-kb upstream region; 5=-ACCGGTACCTAATTATAACCTTTTGAC AAC-3= (KpnI site underlined) and 5=-TGTGAATTCGAAAGTAGGTG ATATGATTGC-3= (EcoRI site underlined) to amplify the 1-kb downstream region. For complementation studies, plasmid pkdhAB encoding kdhAB genes under a groEL promoter was constructed. In brief, a DNA sequence

containing open reading frames of both kdhA and kdhB genes was ampliﬁed by PCR with primers 5=-TGGAATTCATGAAACATAAACTAAAGC TAG-3= (EcoRl site underlined) and 5=-TCCTCGAGTTAGTAATATAT AATCTTATCTT-3= (Xhol site underlined). After restriction endonuclease digestion, the kdhAB DNA fragment was ligated into the EcoRI and XhoI sites of plasmid pTH17 (49). Bacterial DNA/RNA preparation. Plasmids were introduced into F. tularensis LVS by electroporation, and miniplasmids for E. coli DH5 and F. tularensis LVS were prepared with a QIAprep spin miniprep kit according to the manufacturer’s instructions (Qiagen, Venlo, Nether-

8

®

mbio.asm.org

January/February 2013 Volume 4 Issue 1 e00638-12

Characterization of kdhAB in Francisella tularensis

lands). Restriction enzymes and DNA-modifying enzymes were purchased from New England Biolabs. DNA ligations and transformations of chemically competent E. coli were performed by standard protocols (50). LPS puriﬁcation. LPS was puriﬁed from F. tularensis WT and kdhAB strains as described elsewhere (49). A modiﬁcation of the hot phenolwater method was used (51). Mouse infection and immunization studies. Male BALB/cByJ, C57BL/6, TLR2 / , and TLR4 / speciﬁc-pathogen-free mice (6 to 8 weeks old; Jackson Laboratory, Bar Harbor, ME) were housed in an animal facility at Harvard Medical School. All procedures conformed to protocols approved by the Institutional Animal Care and Use Committee. Bacterial virulence was assessed by intranasal challenge of BALB/cByJ mice (n 6) with 103 CFU of F. tularensis LVS bacteria or with 103, 104, or 105 CFU of kdhAB bacteria (inoculum volume, 50 l per mouse). Mice were lightly anesthetized with isoﬂurane (Baxter, Deerﬁeld, IL). The importance of the TLR2 signaling pathway in mouse survival after intranasal challenge with kdhAB mutant bacteria (102, 103, or 104 CFU) was assessed in C57BL/6 WT, TLR2 / , and TLR4 / mice. In all these instances, survival was monitored over a 26-day period unless otherwise indicated. For immunization studies, groups of BALB/cByJ mice (n 10) were immunized by the intranasal route with three doses of bacteria at 2-week intervals; the doses were 10, 10, and 102 CFU for F. tularensis LVS and 103, 104, and 105 CFU for the kdhAB mutant. A mock immunization group that received PBS served as a control. Four weeks after the last dose, mice were challenged intranasally with either 104 CFU of LVS (10 LD50) or 10 CFU of strain Schu S4 (10 LD50). Mice were challenged with the F. tularensis Schu S4 strain at the animal biosafety level 3 (ABSL3) core facility at Harvard Medical School. Titers of serum antibody to LPS were measured by ELISA as described previously (26). Organ bacterial burden and histopathology. BALB/cByJ mice were challenged intranasally with 103 CFU of F. tularensis LVS or kdhAB bacteria. Six days after infection, mice were killed, blood was collected, and tissues were weighed, hand mashed, and homogenized with a Stomacher 80-paddle action blender (Seward, Port Saint Lucie, FL). Serial 10-fold dilutions were prepared with sterile 1 PBS supplemented with 2% fetal bovine serum. A 10- l volume of each dilution was plated onto CHAH plates and incubated at 37°C in 5% CO2 until colonies became visible (~72 h). Data were expressed as CFU per gram of tissue or per milliliter of blood. Histological analysis was performed on the livers of F. tularensis Schu S4-challenged mice that had been immunized with F. tularensis LVS or kdhAB mutant or had been given PBS as a control. Livers from each group of mice were ﬁxed in 10% formalin solution, embedded in parafﬁn, and stained with hematoxylin and eosin after sectioning. Bronchoalveolar lavage. BALB/cByJ mice were challenged intranasally with 103 CFU of F. tularensis LVS or kdhAB bacteria. After 24 h, the mice were killed, the tracheae were exposed, and the lungs were lavaged with 0.5 ml of PBS supplemented with 0.5 mM EDTA. TNF- and IL-1 were assayed with ELISA Duoset kits (R&D Systems, Minneapolis, MN). BMM infection and measurements of cytokine production and gene expression. Bone marrow was ﬂushed with PBS from the femurs of C57BL/6J mice (6 to 8 weeks old; Jackson Laboratory), spun down, resuspended in BMM growth medium (Dulbecco modiﬁed Eagle medium [DMEM] containing D-glucose [4.5 mg/liter], L-glutamine [4 mM], and sodium pyruvate [110 mg/liter] [Invitrogen, Carlsbad, CA] supplemented with 10% heat-inactivated fetal bovine serum [FBS] [Invitrogen] and 30% cell line L929-conditioned DMEM with nonessential amino acids). The cells from each femur were plated in two deep petri dishes, and the plates were incubated at 37°C in 5% CO2 in the presence of gentamicin (50 g/ml). After 3 days, BMM growth medium was added to petri dishes. The cells were incubated for an additional 2 to 3 days. For measurements of cytokine release, BMMs were seeded in 24-well plates at a density of 5 105 cells/well, and the plates were incubated at

37°C in 5% CO2 overnight. The macrophages were then infected with mid-logarithmic-phase F. tularensis LVS or kdhAB bacteria at an MOI of 10, 20, or 50 (ratio of bacteria to macrophages) for 24 h. TNF- and IL-1 in cell supernatants were measured with ELISA Duoset kits according to the manufacturer’s recommendations (R&D Systems). For measurements of gene expression, BMMs were seeded in 6-well plates at a density of 2 106 cells/well, and the plates were incubated at 37°C in 5% CO2 overnight. The macrophages were then infected with mid-log F. tularensis LVS or kdhAB bacteria at an MOI of 10. At various time points, cells were lysed in 0.4 ml TRIzol (Invitrogen), and RNA was isolated according to the manufacturer’s instructions. Reverse transcription was performed with a qScript Flex cDNA synthesis kit (Quanta Biosciences, Gaithersburg, MD) according to the manufacturer’s instructions. Real-time PCR was conducted with Perfecta SYBR green supermix (Quanta Biosciences) in a Bio-Rad iCycler iQ system (45 cycles, with 1 cycle consisting of 15 s at 95°C, 45 s at 60°C, and 30 s at 72°C). Gene expression was normalized to the expression of the RNA polymerase II (RP II) gene, which is considered a stably expressed housekeeping gene (52). The following primer sets were used: for TNF- , sense (5=-TATGG CTCAGGGTCCAACTC-3=) and antisense (5=-CTCCCTTTGCAGAACT CAGG-3=); for IL-1 , sense (5=-GGGCCTCAAAGGAAAGAATC-3=) and antisense (5=-TACCAGTTGGGGAACTCTGC-3=); and for RP II, sense (5=-GCACCACGTCCAATGATAT-3=) and antisense (5=-CTGGCGGTT GACCCCATGAC-3=). Relative gene expression was calculated by the CT method (53) and reported as fold induction over the uninfected control level (mean standard deviation [SD] values for samples from three independent experiments). Murine macrophage-like cell line infection. WT, Unc93B / , MyD88 / , TLR4 / , and TLR2 / murine macrophage-like cell lines, kindly provided by Douglas T. Golenbock (University of Massachusetts Medical School, Worcester, MA), were seeded in 24-well plates at a density of 5 105 cells/well, and the plates were incubated overnight at 37°C in 5% CO2. The macrophages were then infected with mid-log F. tularensis LVS or kdhAB bacteria at an MOI of 100. After 2 h, cells were washed three times with Dulbecco’s PBS and lysed in 60 l of lysis buffer (kit BF14100; R&D Systems). IL-1 in lysates was measured with an ELISA Duoset kit (R&D Systems). NF- B reporter luciferase assay. Stable HEK293 cell lines expressing TLR2 receptor were kindly provided by Douglas T. Golenbock. HEKTLR2 and HEK-TLR4/MD2 cells were seeded into 12-well tissue culture plates at a density of 3 105 cells/well. The following day, cells were transiently transfected with luciferase reporter genes; Genejuice (EMD Chemicals, Darmstadt, Germany) was used per the manufacturer’s recommendations. To assess NF- B activation, an NF- B luciferase reporter gene (consisting of an artiﬁcial promoter NF- B site driving the ﬁreﬂy luciferase gene) was cotransfected with a constitutively active Renilla luciferase reporter gene, phRL-TK (Promega, Fitchburg, WI); phRL-TK was used to normalize transfection efﬁciency. The following day, the cells were stimulated as indicated in Results. During stimulation with LPSs from F. tularensis LVS or kdhAB bacteria, the mixture contained soluble CD14 recombinant protein (5 g/ml) (ProSci Inc., Poway, CA). After 24 h of stimulation, the cells were lysed in passive lysis buffer (Promega), and reporter gene activity was measured with a plate reader luminometer (Victor2; PerkinElmer Life Sciences) and the Dual-Luciferase reporter assay system (Promega). Surface antigen-binding assay. Evaluation of FopA antibody binding to the bacterial surface was performed as described previously (24), with some modiﬁcations. In brief, cultures of WT F. tularensis LVS, kdhAB, and wbtA bacteria were grown overnight in brain heart infusion (BHI) broth. Cultures were diluted in fresh BHI medium to an optical density at 620 nm (OD620) of 0.07, and bacteria were further grown to an OD620 of 0.5. After the bacteria were harvested, 6 108 bacteria from each strain (in a 50- l volume) were mixed with monoclonal FopA antibodies (ﬁnal concentration, 0.1 mg/ml) (54) in PBS containing 0.1% glucose and 0.5% bovine serum albumin (BSA). After incubation for 1 h at 37°C, bacteria

January/February 2013 Volume 4 Issue 1 e00638-12

®

mbio.asm.org 9

Okan et al.

were washed three times with PBS containing 0.1% glucose. After the ﬁnal wash, pellets were resuspended in Laemmli sample buffer (Bio-Rad, Hercules, CA). Protein samples were resolved on 5 to 20% Protean TGX gels (Bio-Rad) and transferred to nitrocellulose membranes. Alkaline phosphatase (AP)-conjugated rabbit anti-mouse antibody (Abcam, Cambridge, MA) was used to detect heavy- and light-chain subunits of bound FopA antibodies. LPS analysis by SDS-PAGE and MALDI-TOF mass spectrometry. Outer membranes from F. tularensis LVS and kdhAB bacteria were isolated, resolved by Tricine SDS-PAGE, and stained with zinc as previously described (14). Lipid A from LVS and kdhAB bacteria was isolated and analyzed by negative-ion matrix-assisted laser desorption ionizationtime of ﬂight (MALDI-TOF) mass spectrometry (14). Data analysis. Data are expressed as mean standard deviation (SD) values. Unpaired Student’s t test was used for statistical analysis. A P value of 0.05 was considered signiﬁcant.

9. 10. 11.

12.

SUPPLEMENTAL MATERIAL
Supplemental material for this article may be found at http://mbio.asm.org /lookup/suppl/doi:10.1128/mBio.00638-12/-/DCSupplemental. Figure S1, EPS ﬁle, 0.8 MB. Figure S2, JPG ﬁle, 0.5 MB. Figure S3, JPG ﬁle, 0.5 MB. Table S1, DOCX ﬁle, 0.1 MB.

13. 14.

15. 16.

ACKNOWLEDGMENTS
We are grateful to Colette Cywes-Bentley (Brigham and Women’s Hospital, Boston, MA) and Marie Charrel-Dennis (University of Massachusetts Medical School, Worcester, MA) for helpful discussions and to Julie McCoy for help in preparation of the manuscript. We thank Douglas T. Golenbock (University of Massachusetts Medical School, Worcester, MA) for providing macrophage-like cell lines (WT, TLR2 / , TLR4 / , Unc93B / , and MyD88 / ), HEK-TLR2 and HEK-TLR4/MD2 cell lines, and plasmid pNF B-luc. We thank Anne G. Savitt (Stony Brook University) for providing FopA and Bfr antibodies. We also thank Jessica Pinkham, Sean Fitzgerald, John Warner, Alec Derian, and Bella Printseva (Harvard Medical School, Boston, MA) for technical assistance. This study was supported by the New England Center of Excellence for Biodefense and Emerging Infectious Diseases under grant AI057159 from the National Institute of Allergy and Infectious Diseases. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
17.

18. 19.

20. 21.

REFERENCES
1. Bosio CM, Bielefeldt-Ohmann H, Belisle JT. 2007. Active suppression of the pulmonary immune response by Francisella tularensis Schu4. J. Immunol. 178:4538 – 4547. 2. Hall JD, Woolard MD, Gunn BM, Craven RR, Taft-Benz S, Frelinger JA, Kawula TH. 2008. Infected-host-cell repertoire and cellular response in the lung following inhalation of Francisella tularensis Schu S4, LVS, or U112. Infect. Immun. 76:5843–5852. 3. Malik M, Bakshi CS, Sahay B, Shah A, Lotz SA, Sellati TJ. 2006. Toll-like receptor 2 is required for control of pulmonary infection with Francisella tularensis. Infect. Immun. 74:3657–3662. 4. Mares CA, Ojeda SS, Morris EG, Li Q, Teale JM. 2008. Initial delay in the immune response to Francisella tularensis is followed by hypercytokinemia characteristic of severe sepsis and correlating with upregulation and release of damage-associated molecular patterns. Infect. Immun. 76: 3001–3010. 5. Moreland JG, Hook JS, Bailey G, Ulland T, Nauseef WM. 2009. Francisella tularensis directly interacts with the endothelium and recruits neutrophils with a blunted inﬂammatory phenotype. Am. J. Physiol. Lung Cell. Mol. Physiol. 296:L1076 –L1084. 6. Cowley SC. 2009. Proinﬂammatory cytokines in pneumonic tularemia: too much too late? J. Leukoc. Biol. 86:469 – 470. 7. Woolard MD, Frelinger JA. 2008. Outsmarting the host: bacteria modulating the immune response. Immunol. Res. 41:188 –202. 8. Kanistanon D, Hajjar AM, Pelletier MR, Gallagher LA, Kalhorn T, 22. 23.

24.

25.

26. 27. 28.

Shaffer SA, Goodlett DR, Rohmer L, Brittnacher MJ, Skerrett SJ, Ernst RK. 2008. A Francisella mutant in lipid A carbohydrate modiﬁcation elicits protective immunity. PLoS Pathog. 4:e24. http://dx.doi.org/10.1371 /journal.ppat.0040024. Sandström G, Sjöstedt A, Johansson T, Kuoppa K, Williams JC. 1992. Immunogenicity and toxicity of lipopolysaccharide from Francisella tularensis LVS. FEMS Microbiol. Immunol. 5:201–210. Wang X, Ribeiro AA, Guan Z, Abraham SN, Raetz CR. 2007. Attenuated virulence of a Francisella mutant lacking the lipid A 4=-phosphatase. Proc. Natl. Acad. Sci. U. S. A. 104:4136 – 4141. Wang X, Ribeiro AA, Guan Z, McGrath SC, Cotter RJ, Raetz CR. 2006. Structure and biosynthesis of free lipid A molecules that replace lipopolysaccharide in Francisella tularensis subsp. novicida. Biochemistry 45: 14427–14440. Galanos C, Lüderitz O, Rietschel ET, Westphal O, Brade H, Brade L, Freudenberg M, Schade U, Imoto M, Yoshimura H, Kusumpyo S, Shiba T. 1985. Synthetic and natural Escherichia coli free lipid A express identical endotoxic activities. Eur. J. Biochem. 148:1–5. Raetz CR, Whitﬁeld C. 2002. Lipopolysaccharide endotoxins. Annu. Rev. Biochem. 71:635–700. Chalabaev S, Kim TH, Ross R, Derian A, Kasper DL. 2010. 3-DeoxyD-manno-octulosonic acid (Kdo) hydrolase identiﬁed in Francisella tularensis, Helicobacter pylori, and Legionella pneumophila. J. Biol. Chem. 285: 34330 –34336. Zhao J, Raetz CR. 2010. A two-component Kdo hydrolase in the inner membrane of Francisella novicida. Mol. Microbiol. 78:820 – 836. Wang X, Karbarz MJ, McGrath SC, Cotter RJ, Raetz CR. 2004. MsbA transporter-dependent lipid A 1-dephosphorylation on the periplasmic surface of the inner membrane: topography of Francisella novicida LpxE expressed in Escherichia coli. J. Biol. Chem. 279:49470 – 49478. Stead C, Tran A, Ferguson D, Jr, McGrath S, Cotter R, Trent S. 2005. A novel 3-deoxy-D-manno-octulosonic acid (Kdo) hydrolase that removes the outer Kdo sugar of Helicobacter pylori lipopolysaccharide. J. Bacteriol. 187:3374 –3383. Stead CM, Zhao J, Raetz CR, Trent MS. 2010. Removal of the outer Kdo from Helicobacter pylori lipopolysaccharide and its impact on the bacterial surface. Mol. Microbiol. 78:837– 852. Raynaud C, Meibom KL, Lety MA, Dubail I, Candela T, Frapy E, Charbit A. 2007. Role of the wbt locus of Francisella tularensis in lipopolysaccharide O-antigen biogenesis and pathogenicity. Infect. Immun. 75: 536 –541. Katz J, Zhang P, Martin M, Vogel SN, Michalek SM. 2006. Toll-like receptor 2 is required for inﬂammatory responses to Francisella tularensis LVS. Infect. Immun. 74:2809 –2816. Russo TA, Beanan JM, Olson R, MacDonald U, Cope JJ. 2009. Capsular polysaccharide and the O-speciﬁc antigen impede antibody binding: a potential obstacle for the successful development of an extraintestinal pathogenic Escherichia coli vaccine. Vaccine 27:388 –395. Schneider MC, Exley RM, Ram S, Sim RB, Tang CM. 2007. Interactions between Neisseria meningitidis and the complement system. Trends Microbiol. 15:233–240. Wu MF, Yang CY, Lin TL, Wang JT, Yang FL, Wu SH, Hu BS, Chou TY, Tsai MD, Lin CH, Hsieh SL. 2009. Humoral immunity against capsule polysaccharide protects the host from magA Klebsiella pneumoniae-induced lethal disease by evading Toll-like receptor 4 signaling. Infect. Immun. 77:615– 621. Zarrella TM, Singh A, Bitsaktsis C, Rahman T, Sahay B, Feustel PJ, Gosselin EJ, Sellati TJ, Hazlett KR. 2011. Host-adaptation of Francisella tularensis alters the bacterium’s surface-carbohydrates to hinder effectors of innate and adaptive immunity. PLoS One 6:e22335. http://dx.doi.org /10.1371/journal.pone.0022335. Cole LE, Mann BJ, Shirey KA, Richard K, Yang Y, Gearhart PJ, Chesko KL, Viscardi RM, Vogel SN. 2011. Role of TLR signaling in Francisella tularensis-LPS-induced, antibody-mediated protection against Francisella tularensis challenge. J. Leukoc. Biol. 90:787–797. Kim TH, Pinkham JT, Heninger SJ, Chalabaev S, Kasper DL. 2012. Genetic modiﬁcation of the O-polysaccharide of Francisella tularensis results in an avirulent live attenuated vaccine. J. Infect. Dis. 205:1056 –1065. Kirimanjeswara GS, Olmos S, Bakshi CS, Metzger DW. 2008. Humoral and cell-mediated immunity to the intracellular pathogen Francisella tularensis. Immunol. Rev. 225:244 –255. Weiss DS, Brotcke A, Henry T, Margolis JJ, Chan K, Monack DM. 2007.

10

®

mbio.asm.org

January/February 2013 Volume 4 Issue 1 e00638-12

Characterization of kdhAB in Francisella tularensis

29.

30.

31.

32.

33.

34.

35.

36.

37.

38.

39.

40.

In vivo negative selection screen identiﬁes genes required for Francisella virulence. Proc. Natl. Acad. Sci. U. S. A. 104:6037– 6042. Sebastian S, Dillon ST, Lynch JG, Blalock LT, Balon E, Lee KT, Comstock LE, Conlan JW, Rubin EJ, Tzianabos AO, Kasper DL. 2007. A deﬁned O-antigen polysaccharide mutant of Francisella tularensis live vaccine strain has attenuated virulence while retaining its protective capacity. Infect. Immun. 75:2591–2602. Ernst RK, Adams KN, Moskowitz SM, Kraig GM, Kawasaki K, Stead CM, Trent MS, Miller SI. 2006. The Pseudomonas aeruginosa lipid A deacylase: selection for expression and loss within the cystic ﬁbrosis airway. J. Bacteriol. 188:191–201. Kawahara K, Tsukano H, Watanabe H, Lindner B, Matsuura M. 2002. Modiﬁcation of the structure and activity of lipid A in Yersinia pestis lipopolysaccharide by growth temperature. Infect. Immun. 70:4092– 4098. Kawasaki K, Ernst RK, Miller SI. 2004. 3-O-deacylation of lipid A by PagL, a PhoP/PhoQ-regulated deacylase of Salmonella typhimurium, modulates signaling through Toll-like receptor 4. J. Biol. Chem. 279: 20044 –20048. Pyles RB, Jezek GE, Eaves-Pyles TD. 2010. Toll-like receptor 3 agonist protection against experimental Francisella tularensis respiratory tract infection. Infect. Immun. 78:1700 –1710. Rozak DA, Gelhaus HC, Smith M, Zadeh M, Huzella L, Waag D, Adamovicz JJ. 2010. CpG oligodeoxyribonucleotides protect mice from Burkholderia pseudomallei but not Francisella tularensis Schu S4 aerosols. J. Immune Based Ther. Vaccines 8:2. West TE, Pelletier MR, Majure MC, Lembo A, Hajjar AM, Skerrett SJ. 2008. Inhalation of Francisella novicida delta mglA causes replicative infection that elicits innate and adaptive responses but is not protective against invasive pneumonic tularemia. Microbes Infect. 10:773–780. Gaekwad J, Zhang Y, Zhang W, Reeves J, Wolfert MA, Boons GJ. 2010. Differential induction of innate immune responses by synthetic lipid A derivatives. J. Biol. Chem. 285:29375–29386. Muroi M, Tanamoto K. 2002. The polysaccharide portion plays an indispensable role in Salmonella lipopolysaccharide-induced activation of NFkappaB through human Toll-like receptor 4. Infect. Immun. 70: 6043– 6047. Zughaier SM, Tzeng YL, Zimmer SM, Datta A, Carlson RW, Stephens DS. 2004. Neisseria meningitidis lipooligosaccharide structuredependent activation of the macrophage CD14/Toll-like receptor 4 pathway. Infect. Immun. 72:371–380. Brandenburg K, Andrä J, Müller M, Koch MH, Garidel P. 2003. Physicochemical properties of bacterial glycopolymers in relation to bioactivity. Carbohydr. Res. 338:2477–2489. de Cock H, Brandenburg K, Wiese A, Holst O, Seydel U. 1999. Nonlamellar structure and negative charges of lipopolysaccharides required

41.

42. 43. 44. 45.

46. 47. 48. 49. 50. 51. 52. 53. 54. 55.

for efﬁcient folding of outer membrane protein PhoE of Escherichia coli. J. Biol. Chem. 274:5114 –5119. Kramer RA, Brandenburg K, Vandeputte-Rutten L, Werkhoven M, Gros P, Dekker N, Egmond MR. 2002. Lipopolysaccharide regions involved in the activation of Escherichia coli outer membrane protease OmpT. Eur. J. Biochem. 269:1746 –1752. Snyder S, Kim D, McIntosh TJ. 1999. Lipopolysaccharide bilayer structure: effect of chemotype, core mutations, divalent cations, and temperature. Biochemistry 38:10758 –10767. Thakran S, Li H, Lavine CL, Miller MA, Bina JE, Bina XR, Re F. 2008. Identiﬁcation of Francisella tularensis lipoproteins that stimulate the Tolllike receptor (TLR) 2/TLR1 heterodimer. J. Biol. Chem. 283:3751–3760. Conlan JW. 2011. Tularemia vaccines: recent developments and remaining hurdles. Future Microbiol. 6:391– 405. Ireland PM, Lebutt H, Thomas RM, Oyston PC. 2011. A Francisella tularensis SCHU S4 mutant deﬁcient in gamma-glutamyl transferase activity induces protective immunity: characterisation of an attenuated vaccine candidate. Microbiology 157:3172–3179. Qin A, Scott DW, Thompson JA, Mann BJ. 2009. Identiﬁcation of an essential Francisella tularensis subsp. tularensis virulence factor. Infect. Immun. 77:152–161. National Research Council. 2011. Guide for the care and use of laboratory animals, 8th ed. National Academy Press, Washington, DC. Lauriano CM, Barker JR, Nano FE, Arulanandam BP, Klose KE. 2003. Allelic exchange in Francisella tularensis using PCR products. FEMS Microbiol. Lett. 229:195–202. Kim TH, Sebastian S, Pinkham JT, Ross RA, Blalock LT, Kasper DL. 2010. Characterization of the O-antigen polymerase (Wzy) of Francisella tularensis. J. Biol. Chem. 285:27839 –27849. Sambrook J, Fritsch EF, Maniatis T. 1989. Molecular cloning: a laboratory manual, 2nd ed. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. Westphal O, Jann K. 1965. Bacterial lipopolysaccharides. Extraction with phenol-water and further applications of the procedure. Methods Carbohydr. Chem. 5:83–91. Radonic A, Thulke S, Mackay IM, Landt O, Siegert W, Nitsche A. 2004. ´ Guideline to reference gene selection for quantitative real-time PCR. B 7iochem. Biophys. Res. Commun. 313:856 – 862. Livak KJ, Schmittgen TD. 2001. Analysis of relative gene expression data using real-time quantitative PCR and the 2(-Delta delta C(T)) method. Methods 25:402– 408. Savitt AG, Mena-Taboada P, Monsalve G, Benach JL. 2009. Francisella tularensis infection-derived monoclonal antibodies provide detection, protection, and therapy. Clin. Vaccine Immunol. 16:414 – 422. Vinogradov E, Perry MB, Conlan JW. 2002. Structural analysis of Francisella tularensis lipopolysaccharide. Eur. J. Biochem. 269:6112– 6118.

January/February 2013 Volume 4 Issue 1 e00638-12

®

mbio.asm.org 11