sign in sign up
<<
Home , Bubo, Bubonic plague

Adaptive response of to extracellular effectors of innate immunity during bubonic

Florent Sebbane*†, Nadine Lemaıˆtre*‡§, Daniel E. Sturdevant¶, Roberto Rebeil*ʈ, Kimmo Virtaneva¶, Stephen F. Porcella¶, and B. Joseph Hinnebusch*,**

*Laboratory of Zoonotic Pathogens and ¶Genomics Core Facility, Rocky Mountain Laboratories, National Institute of Allergy and Infectious Diseases, National Institutes of Health, Hamilton, MT 59840; ‡Institut National de la Sante´et de la Recherche Me´dicale Unite´801 and Faculte´deMe´ decine Henri Warembourg, Universite´de Lille II, Lille F-59045, France; and §Institut Pasteur, Lille F-59021, France

Edited by John J. Mekalanos, Harvard Medical School, Boston, MA, and approved June 13, 2006 (received for review February 11, 2006) Yersinia pestis causes , characterized by an en- phagocytic oxidative burst that generates antimicrobial reactive larged, painful , termed a , that develops after oxygen species (ROS), down-regulate the normal proinflamma- bacterial dissemination from a fleabite site. In susceptible , tory response, and induce apoptosis of dendritic cells, NK cells, the rapidly escape containment in the lymph node, spread macrophages, and polymorphonuclear neutrophils (PMNs) (3, 4, systemically through the blood, and produce fatal sepsis. The 6–8). Although substantial, most of the experimental evidence fulminant progression of disease has been largely ascribed to the for these models comes from in vitro studies with the less virulent ability of Y. pestis to avoid phagocytosis and exposure to antimi- enteric pathogens Yersinia pseudotuberculosis and Yersinia en- crobial effectors of innate immunity. In vivo microarray analysis of terocolitica or from studies using attenuated strains of Y. pestis Y. pestis gene expression, however, revealed an adaptive response injected directly into the blood stream. to nitric oxide (NO)-derived reactive nitrogen species and to iron Bubonic plague does not invariably progress to septicemic limitation in the extracellular environment of the bubo. Polymor- plague, even in susceptible animals, indicating that the lymph phonuclear neutrophils recruited to the infected lymph node ex- node is the deciding arena of immune response success or failure. pressed abundant inducible NO synthase, and several Y. pestis To better understand –pathogen interactions in this key homologs of genes involved in the protective response to reactive arena, we investigated Y. pestis adaptation to the innate immune nitrogen species were up-regulated in the bubo. of one response and other host factors in the bubo by in vivo gene of these genes, which encodes the Hmp flavohemoglobin that expression profiling and virulence testing of select mutants using detoxifies NO, attenuated virulence. Thus, the ability of Y. pestis to a previously characterized model of bubonic plague (2). destroy immune cells and remain extracellular in the bubo appears to limit exposure to some but not all innate immune effectors. High Results NO levels induced during plague may also influence the developing Transcriptional Profile of Y. pestis in the Bubo. Within a few days adaptive immune response and contribute to septic shock. after injecting Y. pestis into the dermis of the upper pelvic region, develop classic signs of bubonic plague, including a grossly inducible nitric oxide synthase ͉ reactive nitrogen species enlarged inguinal lymph node and limping and reluctance to move the associated hind leg. The primary bubo in rats with 7 9 Y. pestis ersinia pestis, the agent of plague, primarily affects terminal plague contains 10 to 10 (2), and sufficient bacterial RNA was recovered from the contents of a single but is also one of the most invasive and virulent bacterial Y inguinal bubo for microarray analysis (Fig. 5, which is published pathogens of (1). Bubonic plague, the most common as supporting information on the PNAS web site). To identify form of the disease in rodents and humans, is usually acquired bacterial genes that are differentially expressed during disease, from the bite of an infected . Disease progression and we compared the transcriptional profiles of Y. pestis isolated histopathology of plague in the rat closely resemble those of from buboes and from laboratory cultures. The gene expression bubonic plague (2). From an intradermal inoculation site, patterns were reproducible and discrete for each of the different Y. pestis rapidly disseminates through afferent lymphatic vessels in vivo and in vitro conditions, with the bubonic profile most to the regional draining lymph nodes. After entering the sub- similar to that of stationary-phase bacteria cultured at 37°C capsular sinus of the node, the bacteria rapidly multiply and (Fig. 1). spread throughout the entire parenchyma of the node, resulting Among the Y. pestis genes most highly expressed in the bubo in lymphadenitis characterized by a painful, swollen lymph node were those for known virulence factors (1), such as the antiph- termed a bubo. Histologically, primary buboes of rats and agocytic F1 protein capsule, a plasminogen activator required for humans contain extensive masses of extracellular Y. pestis, systemic invasion from the lymph node, the iron acquisition necrotic lymphoid tissue, hemorrhage and fibrin, and are sur- system encoded on the Y. pestis pathogenicity island, and the rounded by an edematous perinodal capsule (2). Without early treatment, bubonic plague usually progresses rapidly to , a form of the disease characterized by Conflict of interest statement: No conflicts declared. bacteremia, systemic spread, and life-threatening Gram-negative This paper was submitted directly (Track II) to the PNAS office. sepsis. Hematogenous spread to the lungs can also result in Freely available online through the PNAS open access option. . Abbreviations: RNS, reactive nitrogen species; ROS, reactive oxygen species; iNOS, inducible Current models of plague pathogenesis emphasize the impor- nitric oxide synthase; TTSS, Type III secretion system; PMN, polymorphonuclear neutrophil; tance of the bacterial Type III secretion system (TTSS) in GSNO, nitrosoglutathione. evading innate immunity, the first line of defense against infec- Data deposition: The MIAME (Minimum Information About a Microarray Experiment) data tion (1, 3, 4). The Y. pestis TTSS consists of at least 42 genes have been submitted to the Gene Expression Omnibus repository (accession no. GSE3793). arranged in several operons on a 70.5-kb plasmid that is essential †Present address: Institut National de la Sante´ et de la Recherche Me´dicale Unite´ 801, for virulence (5). These genes encode virulence proteins, termed Universite de Lille II, and Institut Pasteur, Lille F-59021, France. Yops, and a delivery system to inject the Yops into eukaryotic ʈPresent address: Sandia National Laboratories, Albuquerque, NM 87185. cells (3). The translocated Yops disrupt a variety of eukaryotic **To whom correspondence should be addressed. E-mail: [email protected]. cell signaling pathways to inhibit phagocytosis, attenuate the © 2006 by The National Academy of Sciences of the USA

11766–11771 ͉ PNAS ͉ August 1, 2006 ͉ vol. 103 ͉ no. 31 www.pnas.org͞cgi͞doi͞10.1073͞pnas.0601182103 Downloaded by guest on September 25, 2021 flavoglobin that is induced by NO and confers resistance to RNS in vitro and in macrophages (12–15), increased 10- to 20-fold in the bubo. In addition, metL, which is required for synthesis of the NO antagonist homocysteine and for Salmonella resistance to GSNO (16), and ytfE (dnrN), which is required for resistance to NO (14), were both up-regulated 3- to 12-fold. Up-regulation of hmp, ytfe (dnrN), and other genes that may be involved in protection against RNS, such as hcp, hcr, and tehB, has recently been attributed to repression of the NO-sensitive negative regulator encoded by yjeB (nsrR) (17, 18). Expression of the Y. pestis nsrR homolog (YPO0379) was down-regulated 1.7- to 4-fold in the bubo compared with stationary-phase, in vitro cultures, and hcr, tehB, hmp, and ytfE (dnrN) were up-regulated (Fig. 2B). Y. pestis homologs of other genes, which have been consistently found to be induced in E. coli by in vitro exposure to RNS (13–15), such as the nrdHIEF ribonucleotide reductase operon, were also up-regulated in the bubo (Fig. 2B). Fig. 1. Principal component analysis of microarray data from six biological replicate samples for each of the in vitro and in vivo conditions. The first two Localized and Systemic Induction of Inducible NO Synthase (iNOS) principal components show the separation of the bubo-specific expression During Bubonic Plague. NO-derived antimicrobial activity is gen- profile from the expression profiles in exponential and stationary-phase erated by macrophages, PMNs, and other host cells by the iNOS cultures grown at 21°C and 37°C. Ellipses encompass exponential, stationary, and rat samples. in response to and is an important component of the innate immune response to intracellular pathogens (19, 20). Both PMNs and macrophages are recruited to the rat bubo (2), and TTSS (Table 1, which is published as supporting information on we detected large numbers of iNOS-expressing cells, which the PNAS web site). The histopathology and transcriptome of Y. appeared to be primarily PMNs, in close association with pestis in the bubo are consistent with the predicted antiphago- extracellular masses of Y. pestis in infected lymph nodes (Fig. 3).

cytic activities of the TTSS. Y. pestis is extracellular in the bubo; iNOS induction and RNS production by rat PMNs have previ- MICROBIOLOGY thus, the bacteria successfully escape internalization and killing ously been shown to occur in vivo after exposure to LPS and ␥ by macrophages and PMNs and instead cause and IFN- (21), a cytokine present at high levels in the serum of rats apoptosis of immune cells in the lymph node (2). Although infected by Y. pestis (2). Serum NO levels also were elevated in ͞ highly expressed in the bubo and up-regulated compared with in rats with bubonic plague (mean, 260 mM ml NOx; range, ϭ vitro growth at 21°C, expression of the F1 capsule and the TTSS 64–605; n 5) compared with uninfected rats (mean, 74 ͞ ϭ genes was not up-regulated compared with in vitro growth at mM ml; range, 57–87; n 5). 37°C. F1 transcription appears to be regulated primarily by temperature, and the TTSS appears to be regulated by temper- Limited Oxidative Stress Response in the Bubo. In contrast with the ature and low calcium in vitro and by host cell contact in vivo (1, case for RNS, there was less evidence for exposure to ROS in the 3). The high TTSS expression levels we observed in vitro may be bubo. Bacteria respond to exogenous oxidative stress generated due to the fact that the culture medium was not supplemented by activated macrophages and PMNs by two major regulatory pathways. The adaptive response to H O or OClϪ follows with calcium. In addition, the TTSS of bacteria within the dense 2 2 oxidative activation of the OxyR transcriptional regulator, re- extracellular masses in the bubo may not be maximally activated sulting in the induction of bacterial catalase and peroxidase- because they are not in contact with host cells (2). detoxifying enzymes as well as the glutathione and thioredoxin systems (22). None of these Y. pestis genes were up-regulated in Evidence for Iron Starvation in the Bubo. Approximately 13% of Y. Ϫ the bubo (Fig. 2C). The response to superoxide (O )iscon- pestis genes were differentially expressed in the bubo: 119 were 2 trolled by the SoxRS two-component regulatory system in most up-regulated and 203 were down-regulated at least 2-fold Gram-negative bacteria, leading to the induction of superoxide compared with the four in vitro growth conditions (Tables 2 Ϫ dismutases (Sod), O2 -resistant isoforms of metabolic enzymes, and 3, which are published as supporting information on the and other protective enzymes and membrane proteins (22). Y. PNAS web site). The in vivo transcriptome indicated that Y. pestis does not appear to have an intact SoxRS response regu- pestis faces iron limitation in the bubo. Expression of 10 of the Ϫ lator (10, 11); but, of the O2 -responsive genes in other bacteria, 15 proven or potential iron or heme transport systems, includ- only sodA and other genes coinduced by RNS had higher ing three siderophore-based systems, four ATP-binding cas- expression levels in the bubo. Furthermore, genes specifically sette iron transporters, two iron permeases, and a heme implicated in the protective response to peroxynitrite (ONOOϪ), transport system was up-regulated Ͼ4-fold (Fig. 2A). Only the Ϫ an antimicrobial product of O2 and NO, were not up-regulated ferrous ion uptake system and the ATP-binding cassette-, Ϫ (Fig. 2B), suggesting that Y. pestis was not exposed to O2 in the hemophore-, and siderophore-dependent transport systems bubo. Fhu, Has, and Ynp, which are probably nonfunctional in Y. pestis (9), were not up-regulated in the bubo compared with in Role of the Y. pestis Hmp Flavoglobin in Virulence. We next com- vitro conditions. pared the relative effect on virulence of loss of protection against RNS and ROS. Y. pestis ⌬hmp and ⌬sodA mutants were 5- to Evidence for Nitric Oxide (NO)-Derived Nitrosative Stress in the Bubo. 15-fold more susceptible to RNS and ROS, respectively, and A marked number of the genes that have been implicated in the were equally susceptible to killing by macrophages (Fig. 4 A and bacterial protective response to the reactive nitrogen species B). Complementation of the ⌬hmp and ⌬sodA mutants restored (RNS) NO and nitrosoglutathione (GSNO) were up-regulated in wild-type resistance to RNS and ROS. Notably, deletion of hmp vivo (Fig. 2B). Y. pestis does not encode the NO or GSNO had a greater effect on virulence. The Y. pestis ⌬hmp mutant had reductases (nor and nrf genes) that many other bacteria use to a 10-fold higher LD50 (73, compared with 6 for wild type) and detoxify RNS (10, 11); but expression of hmp, which encodes a a longer incubation period in rats (Fig. 4C). In contrast, although

Sebbane et al. PNAS ͉ August 1, 2006 ͉ vol. 103 ͉ no. 31 ͉ 11767 Downloaded by guest on September 25, 2021 Fig. 3. iNOS expression by PMNs in the bubo. Shown are sections of a bubo (A–D) or uninfected lymph node (E and F) stained to detect iNOS (A, C, and E) or PMNs (B, D, and F; yellow arrowheads). PMNs producing iNOS (dark brown) are indicated by red arrowheads. Masses of extracellular bacteria adjacent to PMNs are indicated by blue arrowheads. (Magnification: A and B, ϫ100; D, ϫ400; C, E, and F, ϫ600.)

SodA is required for full virulence of Y. enterocolitica (23), loss of sodA in Y. pestis did not increase the LD50 or the survival rate of infected rats and resulted only in a slightly longer time to disease (Fig. 4D).

Discussion Generation of RNS is considered to be more important for defense against intracellular pathogens that reside in macro- phages, whereas ROS are primarily associated with killing of bacteria that are ingested by PMNs (20). Y. pestis is a facultative intracellular pathogen, but uptake and replication in appears to be limited to initial stages of infection at the dermal infection site (24). Resistance to macrophage-generated ROS (25) and RNS (24, 26) is likely to be important during this early intracellular phase (Fig. 4B). By the time the bacteria dissemi- nate to the lymph node, however, they are predominantly extracellular, presumably because of the antiphagocytic proper- ties of the TTSS and F1 capsule, which are in fact highly expressed in the bubo (Table 1).

or to peroxynitrite are highlighted in yellow and green, respectively. Genes highlighted in gray, along with hmp and ytfE, are predicted to be up- regulated in response to RNS by repression of the NO-sensitive negative regulator encoded by nsrR (17, 18); the other genes were shown to be Fig. 2. Differential expression in the bubo of genes implicated in the up-regulated at least 2-fold in response to NO or GSNO stress in three different response to iron limitation (A) (9), nitrosative stress (B) (13–20), and oxidative in vitro microarray studies (13–15). The YPO and Y numbers refer to the Y. stress (C) (22, 39). The mean fold increase or decrease in individual gene pestis CO92 and KIM genome annotations, respectively (10, 11). *, not anno- expression in the bubo compared with exponential (E) and stationary (S) phase tated in Y. pestis CO92; a, pseudogene; b, the Y. pestis response to superoxide cultures at 21°C and 37°C is indicated by the color scale bar. For the nitrosative may not occur via SoxRS regulation because no soxRS homologs were identi- stress gene list (B), genes specifically required for resistance to NO and GSNO fied in the genome (10, 11).

11768 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.0601182103 Sebbane et al. Downloaded by guest on September 25, 2021 common to both infection models. However, only 2 of the 10 iron acquisition systems that were up-regulated in the bubo were up-regulated in the lung. Among the nitrosative stress response genes, hmp, the met genes, and tehB were up- regulated in the lung, but ytfE (dnrN), hcr, and the nrd genes were not. Furthermore, hmp expression was increased only 2.5-fold in the lung (31) but 10 to 20-fold in the bubo (Table 2) compared with stationary-phase growth at 37°C in vitro, suggesting that iron deprivation and NO-induced stress are more severe in the rat bubo than in the mouse lung. Much of the antibacterial activity of NO is due to nitrosy- lation of cysteine thiols and transition metal centers of proteins required for respiratory metabolism, DNA synthesis, and transcriptional regulation (20). These effects could reinforce the shift to anaerobic metabolism (Fig. 6, which is published as supporting information on the PNAS web site) and the robust iron-acquisition response we observed in the bubo. For example, NO inhibits the ferric ion uptake regulator, Fur, leading to induction of iron-acquisition pathways (13, 32). Two of the Y. pestis iron-acquisition systems have been shown to be Fig. 4. Decreased resistance of Y. pestis to RNS attenuates virulence more required for bubonic plague pathogenesis, however, indicating than decreased resistance to ROS. (A and B) Susceptibility of Y. pestis wild type (black bars), ⌬hmp (gray bars), and ⌬sodA (white bars) to RNS and ROS in vitro that iron limitation does occur in vivo (9). NO has also been (A) and to intracellular killing by macrophages (B). The mean and standard shown to activate OxyR in E. coli (33), but we did not detect error of three experiments are shown. (C and D) Incidence of plague in rats evidence for this (Fig. 2C), in agreement with recent studies injected intradermally with Ϸ150 Y. pestis wild type (black circles), ⌬hmp (gray showing that the transcriptional responses to RNS and ROS circles), or ⌬sodA (white circles). are largely distinct (13).

Conclusions MICROBIOLOGY The extracellular niche and antiinflammatory activities of Y. Our results implicate NO-derived RNS as an important innate pestis have been predicted to limit its exposure to - immune effector mechanism that Y. pestis encounters and re- derived ROS and RNS in the lymph node, thus enabling rapid sponds to in the bubo, where it is primarily extracellular. multiplication and systemic spread (3, 4, 6). Our results Inhibition of phagocytosis by the TTSS and other known viru- indicate that, although ROS are not a significant factor in the lence factors appears to limit exposure to ROS, but not RNS, bubo, exposure to RNS is sustained throughout bubonic and during bubonic plague. Thus, although best known as an intra- septicemic plague. Of the bacterial genes with proven roles in cellular effector, release of NO into infected tissue may also be protection against RNS, only adhC, the glutathione-dependent an important innate immune mechanism against extracellular formaldehyde dehydrogenase that reduces GSNO, was not pathogens. In addition to its antibacterial role in innate immu- up-regulated in the bubo compared with in vitro growth nity, NO has known concentration-dependent immunosuppres- conditions. However, adhC does not appear to be induced by sive and other regulatory effects on the developing adaptive NO or nitrosative stress (12–15). Expression of the Y. pestis immune response (19). Systemic iNOS activation during plague ripA gene, which is required for reduction of NO levels and also likely contributes to septic shock, the ultimate cause of replication within activated macrophages in vitro (26), was not plague mortality, as it does in other Gram-negative septicemias up-regulated in the bubo. Although best known as an intra- (34). Further characterization of the effects of NO on the cellular effector, NO as well as bacteriostatic iron-chelating bacteria and the host during plague will lead to a better agents, such as lactoferrin and lipocalin 2 (NGAL), may be understanding of Y. pestis pathogenesis and a potential new released into the extracellular environment from degranulat- target for anti-plague therapy. ing or degrading PMNs (21, 27–29). Extracellular NO-related antimicrobial stress could also derive from macrophages (30) Materials and Methods and from the high serum NO levels present at the terminal Y. pestis Strains and in Vitro Growth Conditions. The fully virulent stage of disease. Endogenous NO produced as a minor by- Y. pestis strain 195͞P was used in this study (2). RNA was product of bacterial anaerobic respiration using nitrate as a extracted from exponential and stationary-phase Y. pestis terminal electron acceptor may also contribute to the RNS grown at 21°C or 37°C in LB͞0.1 M Mops, pH 7.4. From frozen response (18). However, the Y. pestis respiratory nitrate re- stocks, bacteria were first serially cultured in brain–heart ductases were not up-regulated in the bubo. infusion (BHI) and LB broth for 18 h at 28°C with aeration. The transcriptional profile of Y. pestis recovered from rat Four milliliters of the LB culture was added to 50 ml of buboes shows similarities and differences with that of Y. pestis LB͞Mops and incubated with aeration at 21°C (flea temper- recovered from bronchoalveolar washings of mice with pri- ature) or 37°C (rat temperature). When the cultures reached mary pneumonic plague (31). In keeping with the differences exponential (A600 ϭ 0.5) phase and3hafterentering station- in lymph node and lung environments, there was limited ary phase (A600 ϭ 1.3), one volume of culture was added to two overlap between the up- and down-regulated genes. Only 34 of volumes of RNAprotect Bacteria Reagent according to the the 119 Y. pestis genes that were up-regulated, and 19 of the 203 manufacturer’s protocol (Qiagen). Bacteria were harvested genes that were down-regulated in the rat bubo had similar from six independent replicate cultures for each of the four in differential regulation in the mouse lung. The Y. pestis TTSS vitro conditions. and F1 capsule genes were highly expressed in both in vivo Individual hmpϪ and sodAϪ mutant strains, in which a 1,118- environments, whereas the plasminogen activator was down- or 573-bp internal segment, respectively, of the single-gene regulated in the lung but not in the lymph node. Notably, operons was replaced with the aph kanamycin-resistance cassette up-regulation of genes required for iron acquisition and for from pUC4K, were produced by allelic exchange (35). The aph resistance to nitrosative stress, but not oxidative stress, was cassette was introduced in the same orientation as the deleted

Sebbane et al. PNAS ͉ August 1, 2006 ͉ vol. 103 ͉ no. 31 ͉ 11769 Downloaded by guest on September 25, 2021 gene, and the rho-independent transcription terminator of each gene expression. A 2-fold difference (P Յ 0.05 by two-tailed t gene was maintained in the deleted alleles to avoid polar effects. test with Benjamini and Hochberg false discovery rate cor- were verified by PCR. Complemented versions of the rection) was considered significant. Y. pestis hmp and sodA mutants were produced by transforma- tion with recombinant pACYC177 plasmids containing a wild- Real-Time RT-PCR. In vivo microarray results for hmp and 11 type copy of the respective gene. randomly selected Y. pestis genes that covered the range of expression level in the bubo were verified by TaqMan RT-PCR Rat Model of Bubonic Plague and Isolation of Y. pestis from the Bubo. analysis using an ABI 7700 TaqMan instrument and RNA For in vivo RNA samples, 12 female Brown Norway rats, each 9 isolated from the buboes of six additional rats (Fig. 7, which is weeks old (Charles River Laboratories), were injected intrader- published as supporting information on the PNAS web site) (37). mally in the lower back with Ϸ500 Y. pestis as previously The quantity of each mRNA was determined relative to that of described (2). Signs of terminal plague (roughcast fur, watery the reference gene crr (YPO2995, encoding a phosphotransfer- eyes, hunched posture, reluctance to move, and limping of the ase system component), whose expression level was not affected hind leg adjacent to the bubo) occurred between 48–72 h after by in vivo or in vitro growth conditions. infection. At this point, each rat was anesthetized, and the inguinal bubo was quickly dissected and disrupted on a cell Virulence Determination and Histopathology. Groups of 10 female strainer (70 ␮m mesh), simultaneously releasing the bubo con- Brown Norway rats, each 8 to 10 weeks old, were infected by tents into 9 ml of RNAprotect. The filter was washed three times intradermal injection of 50 ␮l of PBS containing different with 1 ml of RNAprotect. Cell suspensions were centrifuged for numbers of Y. pestis. Rats were killed upon signs of terminal 10 min. at 1,500 ϫ g at 20°C and resuspended in 600 ␮lofTE plague, and the quantity of Y. pestis in heart blood and spleen buffer (10 mM Tris͞1 mM EDTA, pH 8.0) containing 2 mg͞ml were determined by counting colony-forming units (2). LD50 lysozyme. After 10 min at room temperature, RNA was extracted values were calculated by using the Reed–Muench equation, and by using an RNeasy kit (Qiagen). RNA samples were DNase- survival rate differences were determined by logrank test. Tissue treated twice with the DNA-free kit (Ambion), then purified and sections of the inguinal lymph node proximal to the inoculation concentrated to 14 ␮l with the RNeasy cleanup kit (Qiagen). site were stained by Naphthol AS-D chloroacetate esterase RNA quality and quantity was assayed on a Bioanalyzer 2100 (Sigma) to detect PMNs or by immunohistochemistry using using RNA LabChips (Agilent Technologies) and by spectro- anti-iNOS antibody 15323 (Abcam) and a peroxidase- photometry. Absence of DNA was verified by real-time TaqMan conjugated secondary antibody. All experiments were per- PCR using primers and probe specific for the Y. pestis proS gene formed at 3 and were approved by the Rocky (Table 4, which is published as supporting information on the Mountain Laboratories Biosafety and Care and Use PNAS web site). Committees in accordance with National Institutes of Health guidelines. Microarray Analysis. A custom antisense oligonucleotide array Susceptibility to ROS and RNS. Minimum inhibitory concentrations representing 149,856 perfect match and mismatch probes, ͞ targeting 4,683 ORFs of the Y. pestis CO92 and KIM strains of GSNO, diethylenetriamine NO adduct, H2O2, and paraquat (10, 11), and manufactured by Affymetrix was used as previ- (Sigma) were determined in LB at 37°C. Susceptibility of Y. pestis ously described (36). Biotin–ddUTP-labeled cDNA was pre- grown at 37°C to killing by J774A.1 macrophages was deter- pared from purified RNA samples from six biological repli- mined as previously described (38). The ratio of viable intracel- cates of each in vitro culture condition and from the buboes of lular bacteria at6hcompared with2hafterinfection was six rats. The cDNA was then fragmented, hybridized to determined by colony-forming unit counts from three replicate Affymetrix chips (one chip per sample; six biological replicates wells in three independent experiments. for each sample type), and scanned per manufacturer proto- NO Assay. Cumulative NO production was evaluated by measur- cols. Data were analyzed with GeneChip Operating Software ing the total nitrite͞nitrate (NO ) in the sera of uninfected and v1.1 (Affymetrix), GeneSpring v6.0 (Agilent Technologies), x diseased rats by the Griess reaction (Total Nitric Oxide Assay and Partek Pro (Partek, Inc.). Data were normalized with the kit; Endogen). mean signal from the Y. pestis probe sets multiplied by a scale factor to obtain the target signal. Default detection thresholds Ͻ Ն We thank F. DeLeo, F. Gherardini, and D. Erickson for the critical of P 0.06 and P 0.06 were used to call individual genes as reading of the manuscript. This work was supported by the Division of present or absent, respectively. The mean signal intensities of Intramural Research, National Institute of Allergy and Infectious Dis- all genes called present in three or more of the six replicate eases (National Institutes of Health) and by a Ellison Medical Founda- microarray experiments were used to calculate differential tion New Scholars Award in Global Infectious Diseases (to B.J.H.).

1. Perry, R. D. & Fetherston, J. D. (1997) Clin. Microbiol. Rev. 10, 35–66. 11. Deng, W., Burland, V., Plunkett, G., Boutin, A., Mayhew, G. F., Liss, P., Perna, 2. Sebbane, F., Gardner, D., Long, D., Gowen, B. B. & Hinnebusch, B. J. (2005) N. T., Rose, D. J., Mau, B., Zhou, S., et al. (2002) J. Bacteriol. 184, 4601–4611. Am. J. Pathol. 166, 1427–1439. 12. Poole, R. K. (2005) Biochem. Soc. Trans. 33, 176–180. 3. Cornelis, G. R., Boland, A., Boyd, A. P., Geuijen, C., Iriarte, M., Neyt, C., Sory, 13. Mukhopadhyay, P., Zheng, M., Bedzyk, L. A., LaRossa, R. A. & Storz, G. M. P. & Stainier, I. (1998) Microbiol. Mol. Biol. Rev. 62, 1315–1352. (2004) Proc. Natl. Acad. Sci. USA 101, 745–750. 4. Navarro, L., Alto, N. M. & Dixon, J. E. (2005) Curr. Opin. Microbiol. 8, 21–27. 14. Justino, M. C., Vicente, J. B., Teixeira, M. & Saraiva, L. M. (2005) J. Biol. Chem. 5. Une, T. & Brubaker, R. R. (1984) Infect. Immun. 43, 895–900. 280, 2636–2643. 6. Brubaker, R. R. (2003) Infect. Immun. 71, 3673–3681. 15. Flatley, J., Barrett, J., Pullan, S. T., Hughes, M. N., Green, J. & Poole, R. K. 7. Kerschen, E. J., Cohen, D. A., Kaplan, A. M. & Straley, S. C. (2004) Infect. (2005) J. Biol. Chem. 280, 10065–10072. Immun. 72, 4589–4602. 16. De Groote, M. A., Testerman, T., Xu, Y., Stauffer, G. & Fang, F. C. (1996) 8. Marketon, M. M., DePaolo, R. W., DeBord, K. L., Jabri, B. & Schneewind, O. Science 272, 414–417. (2005) Science 309, 1739–1741. 17. Bodenmiller, D. M. & Spiro, S. (2006) J. Bacteriol. 188, 874–881. 9. Perry, R. D. (2004) in Iron Transport in Bacteria, eds. Crosa, J. H., Mey, A. R. & 18. Rodionov, Q. A., Dubchak, I. L., Arkin, A. P., Alm, E. J. & Gelfand, M. S. Payne, S. M. (Am. Soc. Microbiol., Washington, DC), pp. 219–240. (2005) PLoS Comp. Biol. 1, 415–431. 10. Parkhill, J., Wren, B. W., Thomson, N. R., Titball, R. W., Holden, M. T. G., 19. Bogdan, C. (2001) Nat. Immunol. 2, 907–916. Prentice, M. B., Sebhaihia, M., James, K. D., Churcher, C., Mungall, K. L., et al. 20. De Groote, M. A. & Fang, F. C. (1999) in Nitric Oxide and Infection, ed. Fang, (2001) Nature 413, 523–527. F. C. (Kluwer Academic͞Plenum, New York), pp. 231–261.

11770 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.0601182103 Sebbane et al. Downloaded by guest on September 25, 2021 21. Fierro, I. M., Nascimento-DaSilva, V., Arruda, M. A., Freitas, M. S., Plotkowski, 31. Lathem, W. W., Crosby, S. D., Miller, V. L. & Goldman, W. E. (2005) Proc. M. C., Cunha, F. Q. & Barja-Fidalgo, C. (1999) J. Leukocyte Biol. 65, 508–514. Natl. Acad. Sci. USA 102, 17786–17791. 22. Storz, G. & Zheng, M. (2000) in Bacterial Stress Responses, eds. Storz, G. 32. D’Autreaux, B., Touati, D., Bersch, B., Latour, J. M. & Michaud-Soret, I. & Hengge-Aronis, R. (Am. Soc. Microbiol., Washington, DC), pp. 47–59. (2002) Proc. Natl. Acad. Sci. USA 99, 16619–16624. 23. Roggenkamp, A., Bittner, T., Leitritz, L., Sing, A. & Heesemann, J. (1997) 33. Hausladen, A., Privalle, C. T., Keng, T., DeAngelo, J. & Stamler, J. S. (1996) Infect. Immun. 65, 4705–4710. Cell 86, 719–729. 24. Pujol, C. & Bliska, J. B. (2005) Clin. Immunol. 114, 216–226. 34. Titheradge, M. A. (1999) Biochim. Biophys. Acta 1411, 437–455. 25. Oyston, P. C. F., Dorrell, N., Williams, K., Li, S.-R., Green, M., Titball, R. W. 35. Sebbane, F., Jarrett, C. O., Linkenhoker, J. R. & Hinnebusch, B. J. (2004) & Wren, B. W. (2000) Infect. Immun. 68, 3419–3425. Infect. Immun. 72, 7334–7337. 26. Pujol, C., Grabenstein, J. P., Perry, R. D. & Bliska, J. B. (2005) Proc. Natl. Acad. 36. Graham, M. R., Virtaneva, K., Porcella, S. F., Barry, W. T., Gowen, B. B., Sci. USA 102, 12909–12914. Johnson, C. R., Wright, F. A. & Musser, J. M. (2005) Am. J. Pathol. 166, 27. Brinkmann, V., Reichard, U., Goosmann, C., Fauler, B., Uhlemann, Y., Weiss, D. S., Weinrauch, Y. & Zychlinsky, A. (2004) Science 303, 1532–1535. 455–465. 28. Faurschou, M. & Borregaard, N. (2003) Microbes Infect. 5, 1317–1327. 37. Chaussee, M. S., Watson, R. O., Smoot, J. C. & Musser, J. M. (2001) Infect. 29. Goetz, D. H., Holmes, M. A., Borregaard, N., Bluhm, M. E., Raymond, K. N. Immun. 69, 822–831. & Strong, R. K. (2002) Mol. Cell 10, 1033–1043. 38. Pujol, C. & Bliska, J. B. (2003) Infect. Immun. 71, 5892–5899. 30. Webb, J. L., Harvey, M. W., Holden, D. W. & Evans, T. J. (2001) Infect. Immun. 39. Zheng, M., Wang, X., Templeton, L. J., Smulski, D. R., LaRossa, R. A. & Storz, 69, 6391–6400. G. (2001) J. Bacteriol. 183, 4562–4570. MICROBIOLOGY

Sebbane et al. PNAS ͉ August 1, 2006 ͉ vol. 103 ͉ no. 31 ͉ 11771 Downloaded by guest on September 25, 2021