Cyclic Dimeric GMP Signaling Regulates Intracellular Aggregation ...

3 downloads 4177 Views 1MB Size Report
May 1, 2011 - senger produced by a GGDEF domain-associated diguanylate cyclase (14, 17, 34, ... host cell-free systems, the involvement of c-di-GMP in the.
INFECTION AND IMMUNITY, Oct. 2011, p. 3905–3912 0019-9567/11/$12.00 doi:10.1128/IAI.05320-11 Copyright © 2011, American Society for Microbiology. All Rights Reserved.

Vol. 79, No. 10

Cyclic Dimeric GMP Signaling Regulates Intracellular Aggregation, Sessility, and Growth of Ehrlichia chaffeensis䌤§ Yumi Kumagai,1†‡ Junji Matsuo,1† Zhihui Cheng,1 Yoshihiro Hayakawa,2 and Yasuko Rikihisa1* Department of Veterinary Biosciences, The Ohio State University, 1925 Coffey Road, Columbus, Ohio 43210,1 and Department of Applied Chemistry, Faculty of Engineering, Aichi Institute of Technology, 1247 Yachigusa, Yakusa, Toyota 470-0392, Japan2 Received 1 May 2011/Returned for modification 25 May 2011/Accepted 8 July 2011

Cyclic dimeric GMP (c-di-GMP), a bacterial second messenger, is known to regulate bacterial biofilm and sessility. Replication of an obligatory intracellular pathogen, Ehrlichia chaffeensis, is characterized by formation of bacterial aggregates called morulae inside membrane-bound inclusions. When E. chaffeensis matures into an infectious form, morulae become loose to allow bacteria to exit from host cells to infect adjacent cells. E. chaffeensis expresses a sensor kinase, PleC, and a cognate response regulator, PleD, which can produce c-di-GMP. A hydrophobic c-di-GMP antagonist, 2ⴕ-O-di(tert-butyldimethysilyl)-c-diGMP (CDGA) inhibits E. chaffeensis internalization into host cells by facilitating degradation of some bacterial surface proteins via endogenous serine proteases. In the present study, we found that PleC and PleD were upregulated synchronously during exponential growth of bacteria, concomitant with increased morula size. While CDGA did not affect host cells, when infected cells were treated with CDGA, bacterial proliferation was inhibited, morulae became less compact, and the intracellular movement of bacteria was enhanced. Concurrently, CDGA treatment facilitated the extracellular release of bacteria with lower infectivity than those spontaneously released from sham-treated cells. Addition of CDGA to isolated inclusions induced dispersion of the morulae, degradation of an inclusion matrix protein TRP120, and bacterial intrainclusion movement, all of which were blocked by a serine protease inhibitor. These results suggest that c-di-GMP signaling regulates aggregation and sessility of E. chaffeensis within the inclusion through stabilization of matrix proteins by preventing the serine protease activity, which is associated with bacterial intracellular proliferation and maturation. permeable and (ii) most bacteria contain multiple GGDEF domain proteins. Ehrlichia chaffeensis causes a potentially fatal infectious disease, human monocytic ehrlichiosis, which has emerged primarily in the United States but is also occasionally seen in other parts of the world (7, 27). E. chaffeensis is an obligatory intracellular Gram-negative bacterium that belongs to the order Rickettsiales of Alphaproteobacteria. By electron microscopy, E. chaffeensis was shown to have a biphasic intracellular development cycle consisting of densely cored cells (DC) and reticulate cells (RC). DC invade monocytes/macrophages and differentiate into RC, which replicate in membrane-bound inclusions (42). Unlike Chlamydia and Chlamydophila spp. (26) and Coxiella burnetii (13), which also have a biphasic intracellular development cycle, Ehrlichia spp. replicate as intracellular aggregates of bacteria called morulae (“mulberry” in Latin), because they look like mulberries under the light microscope (31). Because by ordinary light microscopy, the inclusion membrane cannot be defined, “morulae” refer to bacterial aggregates with or without inclusion membranes. However, by electron microscopy, all intracellular bacteria are confined within the inclusion membrane. E. chaffeensis inclusions have characteristics of early endosomes, having early endosome antigen 1, Rab5, and transferring receptor, and exogenous transferrin reaches inside inclusion compartments (2, 25). Following multiplication, RC differentiate into DC, which have the ability to infect other cells, in parallel with morula dissolution (42). DC exit the host cell by exocytosis, by host cell

Cyclic dimeric GMP (c-di-GMP) is a bacterial second messenger produced by a GGDEF domain-associated diguanylate cyclase (14, 17, 34, 35, 39). GGDEF domain-containing proteins have been shown to downregulate the planktonic (motility) and upregulate communal or sessile (biofilm, exopolysaccharide, and stalk) traits in several bacteria, including Salmonella enterica serovar Typhimurium, Escherichia coli, Pseudomonas aeruginosa, Vibrio cholerae, and Caulobacter crescentus (14, 16–18, 34, 35, 39). Since most studies on c-di-GMP functions have been performed with free-living bacteria or in host cell-free systems, the involvement of c-di-GMP in the interaction of pathogenic bacteria with the host is unclear. c-di-GMP signaling is required for the virulence of S. Typhimurium in mouse (38); however, the contribution of c-di-GMP signaling to the establishment of S. Typhimurium intracellular infection remains to be elucidated. The study of the detailed molecular path by which c-di-GMP regulates bacterial cellular events is challenging, since (i) c-di-GMP is not membrane

* Corresponding author. Mailing address: Department of Veterinary Biosciences, College of Veterinary Medicine, The Ohio State University, 1925 Coffey Road, Columbus, OH 43210-1093. Phone: (614) 292-9677. Fax: (614) 292-6473. E-mail: [email protected]. † Y.K. and J.M. contributed equally to this study. ‡ Present address: Research Institute for Microbial Diseases, Osaka University, Yamada-oka 3-1, Suita, Osaka 5650871, Japan. § Supplemental material for this article may be found at http://iai .asm.org/. 䌤 Published ahead of print on 25 July 2011. 3905

3906

KUMAGAI ET AL.

lysis, or by being transported into adjacent cells through filopodia, thereby allowing the bacteria to invade new sets of cells and establish a widespread infection (40). This suggests coordinated formation and dissolution of the morula during the E. chaffeensis infectious cycle; however, the biological significance and regulation of morula formation and dissolution are poorly understood. E. chaffeensis intracellular development is accompanied by differential expression of a bacterial surface-exposed protein, TRP120 (tandem repeat protein 120) (29, 42). Electron microscopy has shown that the E. chaffeensis and Ehrlichia canis inclusion matrix is filled with filamentous materials (29, 31), and E. chaffeensis TRP120 is also associated with filamentous materials in the matrix (29). E. chaffeensis and a related bacterium, Anaplasma phagocytophilum, have only a single GGDEF domain-containing protein, PleD, which is conserved among members of the order Rickettsiales (3, 33). PleD is a response regulator of the bacterial two-component regulatory system, and in E. chaffeensis and A. phagocytophilum, it is specifically phosphorylated by a cognate sensor kinase, PleC (19, 22). E. chaffeensis and A. phagocytophilum PleD activated by phosphorylation has a diguanylate cyclase activity that produces c-di-GMP (21, 22). Targeted gene knockout is not technically feasible for E. chaffeensis, and even if it were possible, genes essential for intracellular infection could not be knocked out because there is no way to recover the mutants other than by infection. We, therefore, have been investigating the functions of c-di-GMP by using a hydrophobic c-di-GMP antagonist, 2⬘-O-di(tert-butyldimethysilyl)-c-di-GMP (CDGA). The affinity of CDGA to c-di-GMP binding proteins is similar to that of c-di-GMP (21). We showed that CDGA inhibits well-defined c-di-GMP-regulated phenomena in S. Typhimurium, including cellulose synthesis, clumping, and upregulation of csgD and adrA mRNA (21). CDGA competitively inhibits c-di-GMP binding to recombinant PleD and to native proteins of E. chaffeensis (21). When host cell-free E. chaffeensis is preincubated with CDGA, bacterial internalization into host cells is impaired (21), indicating that c-di-GMP plays important roles in the host-bacterium interaction. We also showed that c-di-GMP signaling is involved in stabilization of E. chaffeensis surface-exposed proteins by preventing the degradation of certain surface proteins by a serine protease (21). In the present study, we analyzed c-di-GMP functions in the obligatory intracellular life cycle of E. chaffeensis by determining the effects of CDGA. We found the involvement of c-diGMP signaling in bacterial proliferation, maturation, and release from host cells, particularly in protease-dependent morula dispersion and bacterial intracellular movement. This is the first evidence for the involvement of c-di-GMP and protease in intracellular bacterial aggregation, which provides new insights into the E. chaffeensis infection cycle and c-diGMP signaling.

MATERIALS AND METHODS Bacterial strains and cell culture. E. chaffeensis Arkansas was propagated in THP-1 cells (ATCC, Manassas, VA) in RPMI 1640 medium supplemented with 10% fetal bovine serum (FBS) and 2 mM L-glutamine at 37°C in 5% CO2 and 95% air. Antibodies. Rabbit anti-PleC and anti-PleD antibodies were described previously (19). The other antibodies used were dog anti-E. chaffeensis (15), rabbit

INFECT. IMMUN. anti-TRP120 (29), rabbit anti-Hsp60 (43), horseradish peroxidase (HRP)-conjugated anti-rabbit IgG (KPL, Gaithersburg, MD), Alexa Fluor 555-conjugated goat anti-rabbit IgG (Invitrogen, Carlsbad, CA), Texas Red-conjugated goat anti-dog IgG (Rockland, Gilbersville, PA), and fluorescein isothiocyanate (FITC)-conjugated goat anti-dog IgG (Rockland). Synchronized culture of E. chaffeensis and expression of proteins at different growth stages. To synchronize the E. chaffeensis growth stages, host cell-free E. chaffeensis bacteria were isolated by sonication from highly infected THP-1 cells cultured at 28°C and used to infect THP-1 cells as described previously (20). The culture was maintained at 37°C. Bacterial growth was monitored every 6 to 12 h by Diff-Quik staining (Baxter Scientific Products, Obetz, OH), and infected cells were harvested as needed. Total DNA was isolated from each culture using the QIAamp DNA blood minikit (Qiagen, Valencia, CA), and the number of bacteria was calculated based on the results of quantitative real-time PCR using a Brilliant SYBR green QPCR core reagent kit (Agilent, La Jolla, CA) with primers for amplifying the 16S rRNA gene (single copy in the E. chaffeensis genome), as described previously (3). Total DNA samples were assayed in triplicate. The protein amount loaded onto an SDS-polyacrylamide gel was normalized by the bacterial number, and PleC and PleD proteins were analyzed by Western blotting using anti-PleC and PleD antibodies as previously described (19). EMSA. DNA fragments of 453 and 449 bp corresponding to the sequences upstream of the pleC and pleD start codons, respectively, were amplified by PCR. PCR products were biotinylated, and an electrophoretic mobility shift assay (EMSA) using recombinant EcxR (E. chaffeensis expression regulator) (rEcxR) was carried out as previously described (4). Effects of CDGA on E. chaffeensis proliferation in THP-1 cells, E. chaffeensis morulae, and uninfected THP-1 cells. THP-1 cells (4 ⫻ 105) infected with early-, mid-, or late-exponential-stage E. chaffeensis bacteria were incubated with 1.8% (wt/vol) NaCl, 0.6 mM c-di-GMP, 2% (vol/vol) dimethyl sulfoxide (DMSO), or 0.6 mM CDGA (synthesized at the Yoshihiro Hayakawa laboratory) (21) in 200 ␮l of RPMI 1640 medium without FBS for 2 h at 37°C. An equal amount of RPMI 1640 medium containing 10% FBS was added (final concentrations, 0.9% [wt/vol] NaCl, 0.3 mM c-di-GMP, 1% [vol/vol] DMSO, 0.3 mM CDGA, and 5% [vol/vol] FBS), and the cells were incubated at 37°C for an additional 16 to 46 h for the early- and mid- exponential-stage cultures and for an additional 3 h for the late-exponential-stage culture. E. chaffeensis morula morphology was monitored at each time point by Diff-Quik staining. The bacterial number per THP-1 cell was calculated based on the copy numbers of the E. chaffeensis 16S rRNA gene normalized by the copy numbers of the human glyceraldehyde-3-dehydrogenase (G3PDH) gene 24 (see Table S1 in the supplemental material for a list of the primers) determined by real-time PCR. The total number of released bacteria in the culture supernatant of infected cells was calculated based on the copy numbers of the E. chaffeensis 16S rRNA gene. To isolate E. chaffeensis morulae, 4 ⫻ 107 THP-1 cells containing mediumsized morulae (⬃3 ␮m) were harvested, suspended in 30 ml RPMI 1640 medium, and homogenized with 40 strokes in a 40-ml Dounce homogenizer. Unbroken cells and nuclei were pelleted twice by centrifugation at 500 ⫻ g for 5 min. Morulae were harvested by centrifugation of the resulting supernatant at 10,000 ⫻ g for 5 min. Cell lysis and morulae isolation were monitored by Diff-Quik staining. Isolated morulae were suspended in 500 ␮l SPK buffer (200 mM sucrose, 50 mM potassium phosphate buffer, pH 7.4) supplemented with 2 mM L-glutamine. Each aliquot (100 ␮l) was incubated with 1% (vol/vol) DMSO or 0.3 mM CDGA in the presence or absence of 10 mM di-isopropyl fluorophosphate (DFP) (Sigma, Saint Louis, MO) for 2 h at 37°C. Morulae were stained by Diff-Quik staining and by immunofluorescence labeling (see below). The amount of TRP120 was analyzed by Western blotting using E. chaffeensis Hsp60 as a loading control. Uninfected THP-1 cells were cultured in the presence of 0.3 mM CDGA or 1% (vol/vol) DMSO for 72 h at 37°C. The growth rate and cell morphology (visualized by Diff-Quik staining) were monitored, and cell viability was determined through a dye exclusion test using 0.4% trypan blue in phosphate-buffered saline (PBS; 137 mM NaCl, 2.7 mM KCl, 8.1 mM Na2HPO4, 1 mM KH2PO4, pH 7.4) every 12 h. To examine the induction of alpha interferon (IFN-␣) and IFN-␤ in host cells, uninfected and infected THP-1 cells were incubated with 0.3 mM CDGA or 1% (vol/vol) DMSO for 2 h at 37°C. Total RNA was purified with an RNeasy minikit (Qiagen), treated with DNase I, and reverse transcribed as described previously (3). The expression level of the IFN-␣ and -␤ genes was quantified by real-time PCR using the primers shown in Table S1 in the supplemental material as described previously (3), with mRNA of the human G3PDH gene as an internal control.

VOL. 79, 2011

E. CHAFFEENSIS MORULA AND c-di-GMP

Immunofluorescence labeling. Isolated morulae incubated with DMSO or CDGA were fixed with 2 or 4% (wt/vol) paraformaldehyde and stained with dog anti-E. chaffeensis (1:100), followed by FITC–goat anti-dog IgG (1:100) antibodies as previously described (21). Infected cells incubated with DMSO or CDGA were stained with rabbit anti-TRP120 (1:100) and dog anti-E. chaffeensis (1:100) antibodies followed by Alexa Fluor 555–goat anti-rabbit (1:100) and FITC–goat anti-dog IgG (1:100) antibodies as described previously (21). Uninfected THP-1 cells were preincubated with 0.3 mM CDGA or 1% (vol/ vol) DMSO for 2 h, washed with PBS, mixed with host cell-free E. chaffeensis at a multiplicity of infection (MOI; the ratio of the number of bacteria to the number of host cells) of 10, and incubated for 3 h. After the cells were washed to remove unbound bacteria, noninternalized and internalized bacteria were differentially stained as previously described (21). Briefly, noninternalized bacteria were stained with dog anti-E. chaffeensis (1:100) antibody, followed by FITC–goat anti-dog IgG (1:100) antibody, the cells were permeabilized with saponin, and then both noninternalized and internalized bacteria were stained with dog anti-E. chaffeensis antibody followed by Texas Red-conjugated goat anti-dog IgG (1:100) antibody. Internalized bacteria (positive for Texas Red) per 100 cells were counted in merged images. As negative controls, fixed samples were incubated with normal dog or rabbit serum followed by the appropriate fluorochrome-conjugated secondary antibodies. The fluorescent images were analyzed using a Nikon Eclipse E400 fluorescence microscope with a xenonmercury light source (Nikon Instruments Inc., Melville, NY). Dark-field live imaging. To observe bacterial movement in THP-1 cells, cells infected with late-exponential-growth-stage E. chaffeensis were incubated with 0.6 mM CDGA or 2% DMSO solvent in RPMI 1640 medium without FBS for 2 h at 37°C. An equal volume of RPMI 1640 medium containing 10% FBS was added, and then the cells were incubated for an additional 1 h at 37°C. To observe bacterial movement in morulae, morulae were isolated as described above and treated with 1% DMSO or with 0.3 mM CDGA in the absence or presence of 10 ␮g/ml oligomycin (Sigma) or 10 mM DFP for 10 min. For dark-field imaging, the cell or morula suspension was dropped on a glass slide coated with poly L-lysine (Sigma) and covered with a coverslip. The movement of E. chaffeensis bacteria in THP-1 cells or in morulae was visualized with CytoViva dark-field illumination (CytoViva, Inc., Auburn, AL) at ⫻1,000 magnification. Real-time images were captured at 42.5 frames per second with a cooled chargecoupled-device (CCD) digital camera (model XLM; Dage-MTI, Michigan, IN). The movies were processed using Image J software (NIH, Bethesda, MD). The movement of individual bacteria in morulae in 42 frames (approximately 1 s) was tracked with the manual tracking plug-in. Statistical analysis. An unpaired Student t test and analysis of variance (ANOVA) with Bonferroni’s correction were used for statistical analysis. A P value of ⬍0.05 was considered significant.

RESULTS Expression of PleC and PleD is upregulated during the exponential growth stage. Similar to the findings of our previous study (4), when E. chaffeensis was synchronously cultured, bacteria became visually detectable after a long lag phase (⬃24 to 36 h postinfection [p.i.]). Small (⬍1-␮m) dense morulae were detected in ⬎80% of host cells at 57 h p.i. (early exponential growth stage) by Diff-Quik staining (Fig. 1A). After this, the average size of the morulae was increased: 1 to 3 ␮m at 72 h p.i. (mid-exponential growth stage), and ⬎3 ␮m at 80 h p.i. (late exponential growth stage) (Fig. 1A). PleC and PleD were expressed in parallel during the intracellular proliferation of E. chaffeensis. The amount of PleC and PleD per bacterium was highest at the mid-exponential growth stage, followed by a decline at the late exponential stage (Fig. 1B). The level of the two proteins was substantially reduced when cells began to lyse (later than 90 h p.i.) (data not shown). We previously reported that an E. chaffeensis transcription factor, EcxR, regulates the expression of genes encoding type IV secretion machinery during the exponential growth stage of E. chaffeensis (4). To examine whether pleC and pleD are coregulated by EcxR, EMSA was performed using rEcxR. Upon incubation of rEcxR with the biotinylated probes consisting of

3907

FIG. 1. PleC and PleD are upregulated during the exponential stage of synchronously cultured E. chaffeensis. (A) A representative cell at 57 h (early exponential), 72 h (mid-exponential), and 80 h (late exponential) stained by Diff-Quik staining. The bacteria or morulae are indicated by arrows. Scale bar, 10 ␮m. (B) PleC and PleD levels per bacterium at each time point were determined by Western blotting. Amounts loaded per lane were normalized to the bacterial genome equivalents as determined by real-time PCR. (C) EMSA for recombinant EcxR binding to the promoter regions of pleC and pleD. The length of the probe is shown above each panel. For each panel, the DNA probe (2 nM) was incubated alone (lane 1), with rEcxR (10 nM) (lane 2), or with rEcxR in the presence of a 50-fold excess of the corresponding unlabeled DNA probe (lane 3). Shifted bands are indicated by arrowheads. Data shown are representative of three independent experiments.

the pleC and pleD promoter sequences, shifted bands were detected for both probes (Fig. 1C), suggesting that these genes are coregulated by EcxR. E. chaffeensis proliferation and morula structure are changed when infected cells are treated with CDGA. Because E. chaffeensis PleD can generate c-di-GMP (21), and PleD expression was observed during the exponential growth stage (Fig. 1), we hypothesized that morula structure is dependent on c-di-GMP. To address this hypothesis, host cells infected with E. chaffeensis were incubated with 0.3 mM CDGA or solvent (DMSO) at early exponential (small morulae of ⬍1 ␮m were found in 88% of cells, 48 h p.i.), mid-exponential (1- to 3-␮m morulae were found in 86% of cells, 66 h p.i), and late exponential (3- to 5-␮m morulae were found in 70% of cells, 75 h p.i) growth stages. Bacterial growth rates for E. chaffeensis per THP-1 cell were somewhat variable batch to batch; however, the trends were the same. Bacterial growth tended to be slower in the presence of solvent DMSO than without it. We had previously determined the effective CDGA concentration for inhibition of c-di-GMP functions in E. chaffeensis and S. Typhimurium to be 0.3 mM (21). Infected cells were also incubated with solvent as a negative control. When cells infected with E. chaffeensis in the early exponential growth stage were incubated with CDGA for 22 h and 44 h,

3908

KUMAGAI ET AL.

FIG. 2. CDGA inhibits intracellular proliferation and maturation of E. chaffeensis. (A, B) E. chaffeensis was synchronously cultured in THP-1 cells. At the early, mid-, and late exponential stages, cells (4 ⫻ 105) were incubated with 1.8% (wt/vol) NaCl (NaCl), 0.6 mM c-diGMP (CDG), 2% (vol/vol) DMSO (DMSO), or 0.6 mM CDGA (CDGA) in 200 ␮l of RPMI 1640 medium without FBS for 2 h at 37°C. An equal volume of RPMI 1640 medium with 10% FBS was added, and cells were cultured for an additional 22 or 44 h for the earlyexponential, 16 or 46 h for the mid-exponential, and 3 h for the late-exponential-stage cultures. (A) Bacterial numbers per THP-1 cell were determined by quantitative real-time PCR. Values are presented relative to bacterial numbers in the presence of NaCl at 22 h and 16 h of incubation for the early and the mid-exponential stages, respectively. (B) A representative cell stained with Diff-Quik is shown after being incubated with solvent DMSO (CTL) or CDGA (CDGA) for 22 h and 44 h for the early exponential stage (early), for 16 h and 46 h for the mid-exponential stage (mid), and for 3 h at the late exponential stage (late). The bacteria or morulae are indicated by arrows. N, nucleus. Scale bar, 10 ␮m. (C) Cells infected with late-exponentialstage E. chaffeensis were cultured as described for panel A in the presence of NaCl solution (NaCl), c-di-GMP (CDG), solvent DMSO (DMSO), or CDGA (CDGA). The numbers of bacteria released in the

INFECT. IMMUN.

bacterial numbers per THP-1 cell were 38% and 16%, respectively, of those of solvent-treated cells (Fig. 2A). Consistent with this, the morulae in CDGA-treated cells were smaller than those in solvent-treated cells (Fig. 2B). When infected cells were incubated with CDGA in the mid-exponential growth stage for 16 h and 46 h, bacterial numbers per cell dropped to 20% and 31%, respectively, of those of cells treated with solvent alone (Fig. 2A). The morulae were less compact, and fewer morulae were found in the cells (Fig. 2B). When infected cells were treated with CDGA in the late exponential growth stage, E. chaffeensis morulae became less densely packed than in solvent-treated cells after an additional 3 h of treatment (Fig. 2B). More extracellular bacteria were detected in the culture medium of cells treated with CDGA than in the medium of solvent-treated cells (Fig. 2C). Furthermore, bacteria released from cells treated in the late exponential growth stage with CDGA were less infectious than bacteria released from solvent-treated cells when the released bacteria were incubated with uninfected THP-1 cells (Fig. 2D). Thus, CDGA inhibited the intracellular proliferation of E. chaffeensis, perturbed morula morphogenesis, and induced the extracellular release of noninfectious immature E. chaffeensis. When cells infected with E. chaffeensis at early exponential, mid-exponential, and late exponential growth stages were incubated with c-di-GMP or with its solvent (NaCl solution) as a negative control, no significant differences were observed in morula morphology (data not shown), bacterial intracellular proliferation (Fig. 2A), or the number of bacteria released from host cells during the late exponential growth stage (Fig. 2C). The infectivities of the bacteria released from c-di-GMPand solvent-treated cells were also comparable (data not shown). We also examined the influence of CDGA on host THP-1 cells. Uninfected cells that were cultured for 72 h in the presence of CDGA or solvent did not show any differences in growth rate or morphology (data not shown). Cell viabilities in the presence of CDGA or solvent, as determined by trypan blue exclusion, were similar. The levels of E. chaffeensis internalization into (Fig. 3A) and infection of (Fig. 3B) THP-1 cells that had been pretreated with CDGA for 2 h and washed were similar to those of cells pretreated with solvent and washed, indicating that CDGA does not act through host cells to inhibit E. chaffeensis internalization and infection. It has been reported that c-di-GMP induces IFN-␣ and -␤ in mouse bone marrow macrophages when c-di-GMP is delivered into the cells by Lipofectamine (23). We therefore examined the influence of CDGA on IFN-␣ and -␤ expression by host cells to determine whether the inhibitory effect of CDGA on E.

culture supernatant were determined by quantitative real-time PCR. (D) Infectivity of the bacteria released in the culture supernatant of cells treated with CDGA (CDGA) or solvent DMSO (CTL) shown in panel C. Bacteria were precipitated and mixed with THP-1 cells. Infectivity was determined by counting bacterial numbers in 100 THP-1 cells after Diff-Quik staining and normalized to the bacterial numbers calculated as described for panel C. (A, C, and D) The values are the means ⫾ standard deviations (n ⫽ 3). Asterisks indicate significant differences (P ⬍ 0.05). Data shown are representative of three independent experiments.

VOL. 79, 2011

E. CHAFFEENSIS MORULA AND c-di-GMP

3909

FIG. 4. CDGA induces downregulation of TRP120 in morulae. E. chaffeensis-infected THP-1 cells (4 ⫻ 105) were incubated with 2% (vol/vol) DMSO (CTL) or 0.6 mM CDGA (CDGA) in 200 ␮l of RPMI 1640 medium without FBS for 2 h at 37°C. An equal volume of RPMI 1640 medium with 10% FBS was added, and the cells were cultured for an additional 2 h. Cells were stained with rabbit anti-TRP120 and dog anti-E. chaffeensis antibodies, followed by Alexa Fluor 555–goat antirabbit and FITC–goat anti-dog IgG antibodies. Scale bar, 10 ␮m. Data shown are representative of three independent experiments.

FIG. 3. CDGA treatment of THP-1 cells neither blocks E. chaffeensis internalization and infection nor induces IFN-␣ and -␤. (A, B) THP-1 cells were preincubated with 0.3 mM CDGA or solvent DMSO (1% [vol/vol]) (CTL) for 2 h, washed, and mixed with host cell-free E. chaffeensis. (A) After incubation with bacteria for 3 h, cells were washed, and noninternalized and internalized bacteria were differentially stained. The numbers of internalized bacteria in 100 cells were counted. (B) Bacterial infectivity in 100 cells was determined after Diff-Quik staining at 2 and 3 days postinfection (p.i.). (C) Uninfected cells (THP-1) and E. chaffeensis-infected cells (THP-1/Ec) were incubated with 0.3 mM CDGA or control DMSO (1% [vol/vol]) (CTL) for 2 h. The mRNA levels of the IFN-␣ and -␤ genes were quantified by real-time reverse transcription (RT)-PCR. The mean expression level in CDGA-treated cells relative to that in solvent-treated cells is shown. All data were normalized to the mRNA level of the human G3PDH gene. The values are the means ⫾ standard deviations (n ⫽ 3). Data shown are representative of three independent experiments.

chaffeensis is through the induction of these cytokines. The transcription level of IFN-␣ and -␤ in uninfected THP-1 cells and in E. chaffeensis-infected THP-1 cells in the presence of CDGA was comparable to those in the presence of solvent (Fig. 3C). Thus, it is unlikely that the inhibitory effects of CDGA on E. chaffeensis are mediated via IFN-␣ and -␤ induction. CDGA reduces the restriction of E. chaffeensis movement. Because c-di-GMP signaling directs bacteria such as C. crescentus, S. Typhimurium, and P. aeruginosa to assume a sessile status in the extracellular environment (14, 16, 17, 34, 39), we hypothesized that c-di-GMP regulates E. chaffeensis sessility within host cells. To address this, we observed the movement of E. chaffeensis inside THP-1 cells in the presence of either CDGA or control solvent using a dark-field imaging system with live bacteria. First, we examined the mobility of exponential-growth-stage E. chaffeensis in THP-1 cells that had been

treated with solvent for 3 h. A rapid vibratory motion of bacteria (brightly reflective numerous dots of approximately 1 ␮m) was seen in the peripheral cytoplasm (see Movie S1 in the supplemental material). Uninfected THP-1 cells lacked these structures (data not shown). When infected cells were treated with CDGA, the bacterial movement was enhanced compared to that observed in solvent-treated cells, including the exhibition of a slightly longer range of movement (see Movie S2 in the supplemental material). After the short treatment of cells in this study, some dense bacterial aggregates remained to be seen in both solvent- and CDGA-treated cells, and bacterial movement was restricted in these aggregates. Since it was difficult to focus and track individual bacteria within aggregates, they were excluded from the movement analysis. Nevertheless, these results suggest that c-di-GMP is involved in restricting the movement of E. chaffeensis in host cells, which was counteracted by CDGA treatment. CDGA induces serine protease-dependent dissolution of host cell-free morulae. Although the mechanisms of biofilm formation are not fully understood, the components of biofilm have been identified in some bacteria: S. Typhimurium produces cellulose and curli fimbriae (32), and P. aeruginosa produces alginate, Psl, and Pel polysaccharides (36) to form biofilm. In contrast, there is no information about the substance which glues E. chaffeensis into tight morulae: the genome does not encode a set of enzymes for extracellular polysaccharide biosynthesis or modification. However, E. chaffeensis inclusions have been shown to be filled with filamentous materials (29, 31) by electron microscopy, and a bacterial surface-exposed protein, TRP120, is associated with filamentous materials in the matrix (29). Therefore, TRP120 might be a component of the matrix, and thus we examined the amount of TRP120 protein in morulae when infected cells were incubated with CDGA. By double immunofluorescence labeling, TRP120 was clearly detected within E. chaffeensis inclusions in cells incubated with solvent, whereas in CDGA-treated cells, morulae were loosely expanded and TRP120 was either reduced or became diffuse and fell largely below the detection limit of immunofluorescence (Fig. 4). To further investigate the effect of CDGA on morula morphol-

3910

INFECT. IMMUN.

KUMAGAI ET AL.

FIG. 5. CDGA induces serine protease-dependent morula dissolution and downregulation of TRP120. E. chaffeensis morulae were isolated and incubated with 0.3 mM CDGA (CDGA) or with solvent (1% [vol/vol] DMSO) (CTL) for 2 h in the presence or absence of 10 mM DFP. (A) Morulae were stained with dog anti-E. chaffeensis antibodies followed by FITC–goat anti-dog IgG antibodies. Scale bar, 5 ␮m. (B) The amounts of TRP120 and Hsp60 (as a loading control) were analyzed by Western blotting. Data shown are representative of three independent experiments.

ogy in the absence of host cell influence, morulae were isolated from infected cells and treated with CDGA. When isolated morulae were incubated with control solvent for 2 h, most morulae (⬎80%) remained as dense aggregates and stained dark purple by Diff-Quik staining (data not shown), and individual bacteria were clearly stained with dog anti-E. chaffeensis antibody. In contrast, CDGA-incubated morulae appeared swollen and bacterial aggregates were not detectable (Fig. 5A). Since in the case of host cell-free bacteria, CDGA is known to induce the degradation of TRP120 by an endogenous bacterial serine protease (21), we examined the involvement of the serine protease in the degradation of TRP120 and dissolution of morulae. When morulae were incubated with CDGA in the presence of a cell-permeable serine protease inhibitor, DFP, the bacteria remained aggregated (Fig. 5A). Furthermore, Western blot analysis revealed that CDGA induced the degradation of TRP120 in isolated morulae, which was also blocked by DFP (Fig. 5B). To observe bacterial movement in the absence of host cell influences, we examined bacterial motility in isolated morulae after 10 min of treatment. Inclusion membranes could be seen more clearly in isolated morulae than in infected cells, perhaps due to the absence of cytoplasmic density and the thinness of the structure. Bacterial movement was also observed in the isolated morulae, and greater motility was seen with CDGA treatment than without it (see Movies S3 and S4 in the supplemental material) (Fig. 6A). Statistical analysis showed that the distance of bacterial movement was significantly increased with CDGA treatment (P ⬍ 0.05) (Fig. 6B). When isolated morulae were treated with an ATP synthesis inhibitor, oligomycin, in the presence of CDGA, bacterial motility was reduced, indicating that the CDGA-enhanced motility is ATP dependent (see Movie S5 in the supplemental material) (Fig. 6). The bacterial motility was also significantly reduced by DFP treatment, indicating that the CDGA-enhanced motility is serine protease dependent (see Movie S6 in the supplemental material) (Fig. 6). After the brief

FIG. 6. CDGA enhances bacterial movement in isolated morulae. Isolated morulae were treated with solvent (1% [vol/vol] DMSO) (CTL), 0.3 mM CDGA (CDGA), or the combination of 0.3 mM CDGA and 10 ␮g/ml oligomycin (OLM) or 10 mM DFP (DFP) for 10 min. (A) Movement of four bacteria tracked in 42 frames (each frame lasts 23.5 ms, approximately 1 s in total) using the manual tracking plug-in of Image J software is shown. Scale bar, 5 ␮m. (B) The movement of a total of 10 bacteria in three morulae for each group was tracked in 42 frames (each frame lasts 23.5 ms, approximately 1 s in total), and the mean distance of bacterial movement was calculated. The values are the means ⫾ standard deviations (n ⫽ 10). Asterisks indicate significant differences with Bonferroni’s correction (P ⬍ 0.05). Data shown are representative of three independent experiments.

treatment used in this study, some bacterial aggregates remained in both solvent- and CDGA-treated morulae, and bacterial movement appeared restricted in these aggregates. Bacteria in these aggregates could not be tracked and therefore were not included in the analysis. Taken together, these results suggest that c-diGMP regulates bacterial aggregation and sessility as well as the level of TRP120 in morulae in a serine protease-dependent manner. DISCUSSION In bacteria having multiple GGDEF domain proteins and c-diGMP phosphodiesterases that degrade c-di-GMP, such as S. Typhimurium and C. crescentus, it has been suggested that c-diGMP signaling temporally and spatially regulates various bacterial biological events. Distinct from these bacteria, since in E. chaffeensis, PleD is the only enzyme that produces c-di-GMP and no c-di-GMP phosphodiesterase homolog is encoded (33), E. chaffeensis c-di-GMP signaling is expected be simpler. In E. chaffeensis, multiple events in the interaction of bacteria and host cell are expected to be controlled through fluctuation of the bacterial c-di-GMP level that is determined solely by PleD activity. Despite the pleC and pleD genes not being in the same operon (3, 19), PleC and PleD protein levels are up- and downregulated in parallel during E. chaffeensis intracellular development. The EcxR transcription factor may be involved in the coregulation, as it binds the promoter regions of pleC and pleD. Hence, PleD is perhaps activated by PleC upon PleD synthesis, and PleD activity is likely determined by the amount of PleC, although we cannot deny the possibility that PleC sensor kinase activation by un-

VOL. 79, 2011

known ligands adds another layer of regulation. Further studies are required to clarify how E. chaffeensis coordinately controls upand downregulation of PleC and PleD and diguanylate cyclase activity during bacterial intracellular development. The present study revealed that in E. chaffeensis, bacterial sessility, intracellular aggregation, and proliferation are linked to c-di-GMP signaling. Although E. chaffeensis morulae are formed inside eukaryotic host cells, in a sense, this formation is analogous to the biofilm formed by extracellular bacteria, since the insides of membrane-bound inclusions are topologically outside host cells, having access to exogenous iron-loaded transferrin. E. chaffeensis replicates in sessile morulae, similar to C. crescentus, a free-living alphaproteobacterium; only the nonmotile sessile “stalked cell” adherent to the substratum is able to replicate chromosomal DNA and produce a daughter cell (6). C. crescentus PleD generates c-di-GMP (28). A C. crescentus pleD null mutant is motile, and this mutation is recessive to the pleD⫹ allele (12), implying that PleD is a negative regulator of motility. Although PleD is involved in flagellum ejection and stalk formation in C. crescentus (1), the molecular mechanisms for PleD-associated sessility have not been clarified, despite some c-di-GMP binding proteins having been reported (5, 8). E. chaffeensis PleD, a sole c-di-GMPproducing protein (3, 33), may also actively suppress bacterial movement during the exponential growth stage when the PleD level is high. The CDGA-induced degradation of E. chaffeensis surface TRP120 is dependent on a surface serine protease HtrA (21). Similarly, in host cell-free morulae, it is likely that c-di-GMP signaling contributes to the maintenance of the matrix protein by repressing a serine protease(s) such as HtrA, since CDGA-induced TRP120 degradation was also blocked by DFP. Further investigation is required to fully understand the function of TRP120 in morula morphogenesis. A connection between c-diGMP and protease activity has been reported previously. In C. crescentus, c-di-GMP regulates CtrA degradation during the cell cycle by controlling the dynamic sequestration of PopA, a cyclic di-GMP effector protein, to the cell pole (8). In contrast to the case with E. chaffeensis morulae, protease positively regulates biofilm formation in P. aeruginosa: inner membrane proteases are involved in the cleavage of anti-sigma factor, which results in the induction of the alginate biosynthetic operon. However, whether c-di-GMP is involved in this process is unknown (30). In a previous work, we have shown that CDGA inhibits c-di-GMP-activated processes in S. Typhimurium and a c-diGMP-repressed process in E. chaffeensis, the degradation of E. chaffeensis surface TRP120 (21). Mechanistically, to inhibit c-diGMP-activated processes, CDGA may bind to the I-site (allosteric inhibitory binding site for c-di-GMP) (4a) of functional diguanylate cyclases, leading to the inhibition of c-di-GMP production, or alternatively, may bind to c-di-GMP receptors in a nonproductive way (without any effect on downstream processes). On the other hand, to inhibit c-di-GMP-repressed processes, CDGA needs to bind to c-di-GMP receptors in a productive way. To understand in more detail the effect of CDGA binding, it would be worth investigating in the future the effect of CDGA on the well-known c-di-GMP-repressed process of the inhibition of bacterial motility that is dependent on PilZ domain protein YcgR in S. Typhimurium (37) and E. coli (9), taking into consideration the presence of multiple GGDEF domain proteins and c-di-GMP phosphodiesterases in these bacteria.

E. CHAFFEENSIS MORULA AND c-di-GMP

3911

When E. chaffeensis at the late exponential stage was incubated with CDGA, more bacteria were released from the host cells, but the released bacteria were less infectious than those from solventtreated infected cells. These results suggest that by maintaining bacterial sessility, c-di-GMP plays a role(s) in preventing the premature release of E. chaffeensis. This delay would allow E. chaffeensis sufficient time to mature from the replicative RC to the infectious DC that are able to withstand the harsh extracellular environment and initiate a new round of infection. Because PleC and PleD protein levels were reduced at the late exponential stage, the level of c-di-GMP is also expected to be reduced, thus allowing the intracellular bacteria to be freed from the host cells. However, c-di-GMP is also required for bacterial internalization, as CDGA treatment of isolated bacteria impaired their internalization (21). Even though at the time of host cell rupture the levels of PleC and PleD are reduced and the amount of c-di-GMP inside the bacteria is expected to be low, the residual c-di-GMP may be required for the next cycle of infection. Therefore, finetuning of the c-di-GMP level would be expected to be essential for coordinating E. chaffeensis proliferation, maintenance of sessility, and maturation within the host cell environment. Members of the order Rickettsiales lack flagella; however, it has been shown that the spotted fever group of Rickettsia spp. assemble actin tails and move inside host cells (10, 41) with the aid of a bacterial surface formin-like factor, Sca2 (11). Rickettsia spp. encode PleD and c-di-GMP phosphodiesterase. We speculate that c-di-GMP signaling is also involved in rickettsial intracellular proliferation and movement by controlling the level of surface proteins such as Sca2. E. chaffeensis does not encode a Sca2 homolog. This may be related to the different intracellular lifestyles between Rickettsia spp. and E. chaffeensis. Rickettsia spp. escape from the phagosome and are free to interact with actin in the cytoplasm, where they replicate, whereas E. chaffeensis is confined within membrane-bound inclusions. Oligomycin blocked CDGA-enhanced bacterial movement, suggesting that enhanced E. chaffeensis movement in morulae is ATP dependent; however, the mechanisms for ATP-dependent movement remain to be elucidated. c-di-GMP induces IFN-␣ and -␤ in mouse bone marrow macrophages when delivered into cells using Lipofectamine (23). CDGA is hydrophobic and expected to penetrate through membranes; indeed, CDGA impairs E. chaffeensis internalization and S. Typhimurium cellulose synthesis and mRNA expression when incubated with the bacteria without Lipofectamine (21). We therefore incubated THP-1 cells and E. chaffeensis-infected THP-1 cells with CDGA without Lipofectamine. CDGA did not influence THP-1 cell viability or its ability to support E. chaffeensis infection. Furthermore, CDGA did not induce IFN-␣ or -␤ in THP-1 cells, presumably because CDGA is not sensed by an unknown eukaryotic intracellular receptor for c-di-GMP or no receptor for CDGA is present. Compounds related to c-di-GMP, including pGpG, the product of bacterial enzymatic hydrolysis of c-di-GMP, are unable to induce IFN (23). Although we cannot exclude the possibility that CDGA affects host cells in another way, inhibition of E. chaffeensis intracellular proliferation and maturation by CDGA is unlikely to be due to an effect(s) on host cells, because the levels of E. chaffeensis internalization into and infection of host cells pretreated with CDGA or solvent control were observed to be comparable. Moreover, perturbation of morula morphology caused by CDGA is also not due to an effect

3912

KUMAGAI ET AL.

on host cells, because CDGA induced similar morphological changes in host cell-free isolated morulae. Taken together, these results are the first example of a role(s) for c-di-GMP in coordinating the formation and dissolution of bacterial aggregates which are associated with bacterial intracellular proliferation and maturation. c-di-GMP is not essential for the extracellular growth of S. Typhimurium, since S. Typhimurium mutants lacking all 12 GGDEF domain proteins grow similarly to wild-type S. Typhimurium (38). In contrast to the case with S. Typhimurium, c-diGMP signaling is likely essential for the intracellular survival of E. chaffeensis, hinting at the therapeutic potential of c-di-GMP antagonists. ACKNOWLEDGMENTS We thank Xue-jie Yu at the University of Texas Medical Branch at Galveston (Galveston, TX) for anti-TRP120. We appreciate Mingqun Lin’s instruction in CytoViva microscopy. This work was supported by National Institutes of Health grant R01 AI054476. REFERENCES 1. Aldridge, P., and U. Jenal. 1999. Cell cycle-dependent degradation of a flagellar motor component requires a novel-type response regulator. Mol. Microbiol. 32:379–391. 2. Barnewall, R. E., Y. Rikihisa, and E. H. Lee. 1997. Ehrlichia chaffeensis inclusions are early endosomes which selectively accumulate transferrin receptor. Infect. Immun. 65:1455–1461. 3. Cheng, Z., Y. Kumagai, M. Lin, C. Zhang, and Y. Rikihisa. 2006. Intraleukocyte expression of two-component systems in Ehrlichia chaffeensis and Anaplasma phagocytophilum and effects of the histidine kinase inhibitor closantel. Cell Microbiol. 8:1241–1252. 4. Cheng, Z., X. Wang, and Y. Rikihisa. 2008. Regulation of type IV secretion apparatus genes during Ehrlichia chaffeensis intracellular development by a previously unidentified protein. J. Bacteriol. 190:2096–2105. 4a.Christen, B., et al. 2006. Allosteric control of cyclic di-GMP signaling. J. Biol. Chem. 281:32015–32024. 5. Christen, M., et al. 2007. DgrA is a member of a new family of cyclic diguanosine monophosphate receptors and controls flagellar motor function in Caulobacter crescentus. Proc. Natl. Acad. Sci. U. S. A. 104:4112–4117. 6. Curtis, P. D., and Y. V. Brun. 2010. Getting in the loop: regulation of development in Caulobacter crescentus. Microbiol. Mol. Biol. Rev. 74:13–41. 7. Demma, L. J., R. C. Holman, J. H. McQuiston, J. W. Krebs, and D. L. Swerdlow. 2006. Human monocytic ehrlichiosis and human granulocytic anaplasmosis in the United States, 2001-2002. Ann. N. Y. Acad. Sci. 1078:118– 119. 8. Duerig, A., et al. 2009. Second messenger-mediated spatiotemporal control of protein degradation regulates bacterial cell cycle progression. Genes Dev. 23:93–104. 9. Fang, X., and M. Gomelsky. 2010. A post-translational, c-di-GMP-dependent mechanism regulating flagellar motility. Mol. Microbiol. 76:1295–1305. 10. Gouin, E., et al. 1999. A comparative study of the actin-based motilities of the pathogenic bacteria Listeria monocytogenes, Shigella flexneri and Rickettsia conorii. J. Cell Sci. 112:1697–1708. 11. Haglund, C. M., J. E. Choe, C. T. Skau, D. R. Kovar, and M. D. Welch. 2010. Rickettsia Sca2 is a bacterial formin-like mediator of actin-based motility. Nat. Cell Biol. 12:1057–1063. 12. Hecht, G. B., and A. Newton. 1995. Identification of a novel response regulator required for the swarmer-to-stalked-cell transition in Caulobacter crescentus. J. Bacteriol. 177:6223–6229. 13. Heinzen, R. A., T. Hackstadt, and J. E. Samuel. 1999. Developmental biology of Coxiella burnettii. Trends Microbiol. 7:149–154. 14. Hengge, R. 2009. Principles of c-di-GMP signalling in bacteria. Nat. Rev. Microbiol. 7:263–273. 15. Huang, H., et al. 2008. Proteomic analysis of and immune responses to Ehrlichia chaffeensis lipoproteins. Infect. Immun. 76:3405–3414. 16. Jenal, U. 2004. Cyclic di-guanosine-monophosphate comes of age: a novel secondary messenger involved in modulating cell surface structures in bacteria? Curr. Opin. Microbiol. 7:185–191.

Editor: A. J. Ba¨umler

INFECT. IMMUN. 17. Jenal, U., and J. Malone. 2006. Mechanisms of cyclic-di-GMP signaling in bacteria. Annu. Rev. Genet. 40:385–407. 18. Kulasakara, H., et al. 2006. Analysis of Pseudomonas aeruginosa diguanylate cyclases and phosphodiesterases reveals a role for bis-(3⬘-5⬘)-cyclic-GMP in virulence. Proc. Natl. Acad. Sci. U. S. A. 103:2839–2844. 19. Kumagai, Y., Z. Cheng, M. Lin, and Y. Rikihisa. 2006. Biochemical activities of three pairs of Ehrlichia chaffeensis two-component regulatory system proteins involved in inhibition of lysosomal fusion. Infect. Immun. 74:5014– 5022. 20. Kumagai, Y., H. Huang, and Y. Rikihisa. 2008. Expression and porin activity of P28 and OMP-1F during intracellular Ehrlichia chaffeensis development. J. Bacteriol. 190:3597–3605. 21. Kumagai, Y., J. Matsuo, Y. Hayakawa, and Y. Rikihisa. 2010. Cyclic di-GMP signaling regulates invasion by Ehrlichia chaffeensis of human monocytes. J. Bacteriol. 192:4122–4133. 22. Lai, T. H., Y. Kumagai, M. Hyodo, Y. Hayakawa, and Y. Rikihisa. 2009. The Anaplasma phagocytophilum PleC histidine kinase and PleD diguanylate cyclase two-component system and role of cyclic Di-GMP in host cell infection. J. Bacteriol. 191:693–700. 23. McWhirter, S. M., et al. 2009. A host type I interferon response is induced by cytosolic sensing of the bacterial second messenger cyclic-di-GMP. J. Exp. Med. 206:1899–1911. 24. Miura, K., and Y. Rikihisa. 2009. Liver transcriptome profiles associated with strain-specific Ehrlichia chaffeensis-induced hepatitis in SCID mice. Infect. Immun. 77:245–254. 25. Mott, J., R. E. Barnewall, and Y. Rikihisa. 1999. Human granulocytic ehrlichiosis agent and Ehrlichia chaffeensis reside in different cytoplasmic compartments in HL-60 cells. Infect. Immun. 67:1368–1378. 26. Moulder, J. W. 1985. Comparative biology of intracellular parasitism. Microbiol. Rev. 49:298–337. 27. Paddock, C. D., and J. E. Childs. 2003. Ehrlichia chaffeensis: a prototypical emerging pathogen. Clin. Microbiol. Rev. 16:37–64. 28. Paul, R., et al. 2004. Cell cycle-dependent dynamic localization of a bacterial response regulator with a novel di-guanylate cyclase output domain. Genes Dev. 18:715–727. 29. Popov, V. L., X. Yu, and D. H. Walker. 2000. The 120 kDa outer membrane protein of Ehrlichia chaffeensis: preferential expression on dense-core cells and gene expression in Escherichia coli associated with attachment and entry. Microb. Pathog. 28:71–80. 30. Qiu, D., V. M. Eisinger, D. W. Rowen, and H. D. Yu. 2007. Regulated proteolysis controls mucoid conversion in Pseudomonas aeruginosa. Proc. Natl. Acad. Sci. U. S. A. 104:8107–8112. 31. Rikihisa, Y. 1991. The tribe Ehrlichieae and ehrlichial diseases. Clin. Microbiol. Rev. 4:286–308. 32. Ro ¨mling, U. 2005. Characterization of the rdar morphotype, a multicellular behaviour in Enterobacteriaceae. Cell Mol. Life Sci. 62:1234–1246. 33. Ro ¨mling, U. 2009. Cyclic di-GMP (c-Di-GMP) goes into host cells—cDi-GMP signaling in the obligate intracellular pathogen Anaplasma phagocytophilum. J. Bacteriol. 191:683–686. 34. Ro ¨mling, U., and D. Amikam. 2006. Cyclic di-GMP as a second messenger. Curr. Opin. Microbiol. 9:218–228. 35. Ro ¨mling, U., M. Gomelsky, and M. Y. Galperin. 2005. c-di-GMP: the dawning of a novel bacterial signalling system. Mol. Microbiol. 57:629–639. 36. Ryder, C., M. Byrd, and D. J. Wozniak. 2007. Role of polysaccharides in Pseudomonas aeruginosa biofilm development. Curr. Opin. Microbiol. 10: 644–648. 37. Ryjenkov, D. A., R. Simm, U. Romling, and M. Gomelsky. 2006. The PilZ domain is a receptor for the second messenger c-di-GMP: the PilZ domain protein YcgR controls motility in enterobacteria. J. Biol. Chem. 281:30310– 30314. 38. Solano, C., et al. 2009. Genetic reductionist approach for dissecting individual roles of GGDEF proteins within the c-di-GMP signaling network in Salmonella. Proc. Natl. Acad. Sci. U. S. A. 106:7997–8002. 39. Tamayo, R., J. T. Pratt, and A. Camilli. 2007. Roles of cyclic diguanylate in the regulation of bacterial pathogenesis. Annu. Rev. Microbiol. 61:7997– 8002. 40. Thomas, S., V. L. Popov, and D. H. Walker. 2010. Exit mechanisms of the intracellular bacterium Ehrlichia. PLoS One 5:e15775. 41. Van Kirk, L. S., S. F. Hayes, and R. A. Heinzen. 2000. Ultrastructure of Rickettsia rickettsii actin tails and localization of cytoskeletal proteins. Infect. Immun. 68:4706–4713. 42. Zhang, J. Z., V. L. Popov, S. Gao, D. H. Walker, and X. J. Yu. 2007. The developmental cycle of Ehrlichia chaffeensis in vertebrate cells. Cell Microbiol. 9:610–618. 43. Zhang, Y., N. Ohashi, E. H. Lee, A. Tamura, and Y. Rikihisa. 1997. Ehrlichia sennetsu groE operon and antigenic properties of the GroEL homolog. FEMS Immunol. Med. Microbiol. 18:39–46.