Introduction Arthropod-borne transmission of bacterial pathogens is somewhat rare but has evolved in a phylogenetically diverse group that includes the rickettsiae, Borrelia spirochetes, and the gram-negative bacteria Francisella tularensis and Yersinia pestis, the plague bacillus. Y. pestis circulates among many species of wild rodents, its primary reservoir hosts, via flea bite. As it alternates between fleas and mammals, it is postulated that Y. pestis regulates gene expression appropriately to adapt to the two disparate host environments, and that different sets of genes are required to produce a transmissible infection in the flea and disease in the mammal. Many important Y. pestis virulence factors that are required for plague in mammals have been identified, and most of them are induced by a temperature shift from 2-fold higher in fleas than in any other condition (Tables 1, S1, S3). PhoP and MgtC are established virulence factors known to be important for survival of Y. pestis and other gram-negative bacteria in macrophages and for resistance to cationic antimicrobial peptides (CAMPs) of the mammalian innate immune response [38],[39],[40]. The PhoPQ system is induced in low Mg2+ or low pH environments, or by exposure to CAMPs [41],[42],[43]. The Mg2+ concentration and pH of the flea digestive tract have not been defined, so the inducing stimulus is unknown, but CAMPs are induced and secreted into the gut by blood feeding insects when they take a blood meal containing bacteria [44],[45]. X. cheopis fleas encode homologs of the insect CAMPs cecropin and defensin, and mount an inducible antibacterial response to infection (unpublished data). Thus, the PhoPQ regulatory system may be induced by the flea's immune system in response to Y. pestis in the midgut. Despite the upregulation of phoP in the flea, with the notable exception of mgtC there was little correlation between predicted PhoP-regulated genes in vitro and genes upregulated in the flea [39],[46],[47]. Differential regulation of members of the PhoP regulon may occur depending on the inducing stimulus, however [48]. Induction of a phagocytosis-resistant phenotype in the flea Soon after transmission, Y. pestis would be expected to encounter rapidly-responding phagocytic cells in the dermis. To assess the overall effect of the flea-specific phenotype on this encounter, we compared the interaction of Y. pestis recovered from infected fleas and from in vitro cultures with murine bone marrow macrophages. Bacteria from fleas showed significantly lower levels of phagocytosis (Fig. 4A). We have previously reported analogous findings using human polymorphonuclear leukocytes (PMNs) [7]. 10.1371/journal.ppat.1000783.g004 Figure 4 Phagocytosis-resistant phenotype of Y. pestis isolated from fleas correlates with expression level of the yit-yip insecticidal-like toxin genes. (A) The percentage of extracellular Y. pestis KIM6+ 1 hour after addition to murine bone marrow macrophages are shown for bacteria from in vitro cultures (LB) or from infected fleas. The mean and SEM of five independent experiments done in duplicate are shown; P 10-fold in the flea, but its expression was not detected in the rat bubo (Table 1). The specific induction in the flea of yitR and genes in the adjacent Tc-like yit and yip loci suggests that they are involved in adaptation to and colonization of the flea. However, deletion of yitR or yitA-yipB (y0183–y0191) does not affect the ability of Y. pestis KIM6+ to infect or block fleas (data not shown). These observations, and the fact that the Yersinia Tc proteins have toxicity to certain eukaryotic cell lines in vitro [50],[51], prompted us to investigate a possible post-transmission antiphagocytic role for these proteins in the mammalian host. To determine if the insecticidal-like toxins were involved in resistance to phagocytosis, we repeated the macrophage experiments with a Y. pestis ΔyitR mutant, which as expected showed greatly reduced expression of the yit and yip genes in vitro and in the flea (Fig. 4B). Loss of yitR significantly reduced the increased resistance to phagocytosis of Y. pestis isolated from infected fleas (Fig. 4C). Since the yit and yip genes are not required for Y. pestis to produce a transmissible infection in fleas, it was possible to compare the virulence of wild-type and ΔyitR Y. pestis following transmission by fleabite. The incidence rate and time to disease onset were identical for both Y. pestis strains, demonstrating that expression of yit and yip is not essential for flea-borne transmission or disease (data not shown). On average, the mice challenged with Y. pestis ΔyitR-infected fleas, both those that developed disease and those that did not, received a higher cumulative number of bites from blocked fleas than the mice challenged with Y. pestis-infected fleas, but this difference was not statistically significant (Fig. 5). However, it was not possible to detect any relatively minor difference in LD50 because the number of bacteria transmitted by a blocked flea varies widely [1],[52]. Even a small decrease in LD50 provided by the Yit-Yip proteins would be significant at the ecological level in the maintenance of plague transmission cycles, because the transmission efficiency of blocked fleas is very low– often only a few or no bacterial cells are transmitted in an individual fleabite [52]. Because phoP is required by Y. pestis to produce a transmissible infection in fleas (unpublished data), it was not possible to similarly assess the effect on disease transmission of phoP induction in the flea. 10.1371/journal.ppat.1000783.g005 Figure 5 Mean and range of the cumulative number of blocked flea bites received by mice. Circles and squares indicate individual mice challenged by fleas infected with wt or ΔyitR Y. pestis 195/P, respectively. Filled symbols indicate mice that developed terminal plague; open symbols indicate mice that did not develop disease. Does transit through the flea vector preadapt Y. pestis to resist mammalian innate immunity? When Y. pestis is transmitted into the dermis by an infected flea, it is immediately exposed to the mammalian innate immune system. The most important antiphagocytic virulence factors, the cytotoxic Yersinia outer proteins (Yops), part of the T3SS encoded by the Y. pestis virulence plasmid and the F1 capsule encoded by the pMT1 plasmid, are not present at this initial stage of infection. Their expression is strictly temperature-regulated and are not produced in vivo until 3–5 hours after the temperature shift to 37°C that accompanies transmission [1],[3],[53],[54]. Consequently, Y. pestis grown at <28°C in vitro are initially susceptible to in vivo uptake and killing by phagocytes until the Yop and F1 virulence factors are produced, effectively preventing further phagocytosis [53],[54]. Our results indicate that Y. pestis entering the mammal from an infective flea is relatively resistant to macrophages, as well as PMNs [7]; a vector-specific phenotype that is not related to the T3SS or capsule. Coming from the flea, Y. pestis is also associated with the biofilm ECM, identical or closely related to the poly-β-1,6-N-acetyl glucosamine ECM of staphylococcal biofilms, which has been shown to provide protection from innate immune components [55],[56]. In addition, although the antiphagocytic F1 capsule and Psa fimbriae do not appear to be produced in the flea, upregulation in the flea of most F1 genes in the cafRcaf1M1A1 locus and the Psa usher protein gene psaC (Tables 1, S1) suggests that components of the F1 and Psa translocation system are made, which may prime Y. pestis for rapid secretion of these extracellular virulence factors after transmission. The upregulation of the innate immunity resistance genes phoP and mgtC suggest that those Y. pestis that are phagocytized may be prepared for resistance to CAMPs and intracellular survival while still in the flea vector. Finally, the major essential virulence factors yadBC and pla, essential for Y. pestis dissemination from the dermis, were maximally or very highly expressed in the flea (Tables 1, S3). Besides degrading plasminogen, the Pla protease may also inactivate CAMPs, particularly when the F1 capsule is not present [57], which matches the phenotype of Y. pestis in the flea. In summary, Y. pestis appears to be prepared for pathogenesis in the mammal while still in the flea vector. The biofilm phenotype of Y. pestis and the virulence factors upregulated or highly expressed in the flea may enhance the earliest stages of plague pathogenesis while the full complement of temperature-shift-regulated virulence factors is still being induced. Increased resistance to innate immunity that is preinduced in the flea vector may be critical to productive transmission because blocked fleas transmit relatively few bacteria, often below the LD50 of Y. pestis grown in vitro at <28°C [1],[52]. Materials and Methods Ethics statement All animals were handled in strict accordance with good animal practice as defined by NIH animal care and use policies and the Animal Welfare Act, USPHS; and all animal work was approved by the Rocky Mountain Laboratories Animal Care and Use Committee. Bacterial strains and growth conditions for in vitro transcriptome analyses Y. pestis KIM6+, which lacks the 70-kb virulence plasmid that is not required for flea infection or blockage, was used for gene expression analyses. A KIM6+ strain with an in-frame deletion that eliminated amino acids 28–281 of the predicted 291 amino acid residue yitR (y0181) gene product was produced by allelic exchange, using the pCVD442 suicide vector system [11]. This mutant was complemented by electroporation with a recombinant pWKS130 plasmid containing the wild-type yitR promoter and orf. The ΔyitR mutant was also transformed with pWKS130 alone to generate an empty vector control strain. For in vitro planktonic samples, bacteria were grown from frozen stocks in brain heart infusion (BHI) medium at 28°C, followed by two successive transfers in Luria Bertani broth supplemented with 100 mM MOPS, pH 7.4 (LB/MOPS) at 21°C. An inoculum of 104 cells/ml was added to 50 ml of LB/MOPS and incubated at 21°C with shaking at 250 rpm until exponential (OD600 = 2.5) or stationary phase (OD600 = 4.5). Approximately 0.5 ml of the exponential phase culture and 0.25 ml of the stationary phase culture was resuspended in 1 ml and 0.5 ml, respectively, of RNAprotect bacterial reagent (Qiagen; Valencia, CA), incubated for 5 min at room temperature, and centrifuged at 21°C for 5 min prior to RNA extraction. For in vitro biofilms, 400 µl of a 107/ml bacterial suspension was injected into a flowcell (Stovall; Greensboro, NC) that was connected to a reservoir of LB/MOPS at 21°C. Following a 30 min incubation period to allow the bacteria to adhere to the glass surface of the flow cell, LB/MOPS was pumped through the flow cell at a rate of 0.3 ml/min. After 48 hours, the flowcell was disconnected and the thick Y. pestis biofilm was harvested and treated with 0.5ml of RNAprotect similarly to the planktonic cultures. Flea infections and collection of samples for in vivo transcriptome analyses X. cheopis fleas were infected with Y. pestis KIM6+ by using a previously described artificial feeding system [3]. The infectious blood meal was prepared by growing Y. pestis KIM6+ overnight at 37°C in BHI medium, without aeration. A cell pellet containing 109 bacterial cells was resuspended in 1 ml PBS and added to 5 ml heparinized mouse blood. Fleas that took a blood meal were maintained at 21°C and 75% relative humidity, fed twice weekly on uninfected mice, and monitored for proventricular blockage as previously described [3]. On the day blockage was diagnosed, the digestive tract was dissected out and macerated in RNAprotect, a process that required about 1 min. Thirty midguts from blocked fleas were pooled for each of the two biological replicates. Midguts from 60 uninfected fleas were also collected as controls to assess background hybridization of flea RNA to the microarray. A flea-borne transmission model [58] was used to determine Y. pestis infectivity after challenge by flea bite. Fleas were infected with Y. pestis 195/P, a fully virulent wild-type strain, or with a Y. pestis 195/P ΔyitR mutant constructed as described above. Between 2–3 weeks after infection, the time required for Y. pestis to block fleas with a proventricular biofilm, groups of 20–40 fleas were applied to a restrained mouse and allowed to feed for 60 min. The fleas were then recovered and examined under a dissecting microscope to determine how many had taken a normal blood meal (unblocked or non-infective fleas) and how many were blocked (infective fleas). After challenge, mice were monitored and euthanized upon the appearance of signs of terminal illness. Mice that did not develop any symptoms after one week following a challenge were re-challenged. A total of 9–10 BALB/cAnN and 10 RML Swiss-Webster mice were challenged with each strain. RNA isolation, amplification, and microarray RNA was isolated from six independent samples from in vitro and flow cell cultures and two independent samples from pooled blocked fleas (Fig. S1) using the RNeasy Mini Kit (Qiagen). Flea-derived RNA samples were secondarily split into three technical replicates each. RNA integrity was verified on a Bioanalyzer 2100 (Agilent Technologies; Santa Clara, CA). Total RNA (100 ng) was amplified and labeled with modified biotin-11-CTP (Perkin Elmer; Waltham, MA) and biotin-16-UTP (Roche Molecular Biochemicals, Pleasanton, CA) by using the Message-Amp II-Bacteria amplified antisense RNA (aRNA) kit (Ambion; Austin, TX). Amplified RNA was then fragmented using Ambion's Fragmentation reagent (Applied Biosystems), hybridized to the RML custom Affymetrix GeneChip that contains sequences for all Y. pestis predicted ORFs, and scanned. The amplification step did not affect the relative transcript signals obtained by microarray (data not shown). Microarray data analysis Affymetrix GeneChip Operating Software (GCOS v1.4, GEO platform GPL2129, http://www.affymetrix.com) was used for initial analysis of the microarray data at the probe-set level. All *.cel files, representing individual biological replicates, were scaled to a trimmed mean of 500 using a scale mask consisting of only the Yersinia pestis KIM6+ probe-sets to produce the *.chp files. A pivot table with all samples was created including calls, call p-value and signal intensities for each gene. The pivot table was then imported into GeneSpring GX 7.3 (http://www.chem.agilent.com), where hierarchical clustering (condition tree) using a Pearson correlation similarity measure with average linkage was used to produce the dendrogram indicating that biological replicates grouped together. The pivot table was also imported into Partek Genomics Suite software (Partek Inc.; St. Louis, MO) to produce a principal components analysis (PCA) plot as a second statistical test for the grouping of biological replicates. ANOVA was run from this data set to produce a false discovery rate report producing false positive reduced p-values for each comparison of interest. The correlated replicates of all test conditions and controls were combined, and quality filters based upon combined calls and signal intensities were used to further evaluate individual gene comparisons. Present and marginal calls were treated as the same whereas absent calls were negatively weighted and eliminated from calculations. Ratios of test/control values and associated t-test and ANOVA p-values values of all individual genes passing the above filters were determined using GeneSpring, SAM, and Partek software. The microarray data have been deposited in the NCBI GEO public database (accession number GSE16493). To compare differential in vivo gene expression patterns in the flea and the rat, the average hybridization signal for each individual Y. pestis gene was divided by the average signal of all 4,683 genes on the microarray for both the flea microarray (this study) and the rat bubo microarray [11] data sets. Gene by gene comparisons of these normalized expression data sets were used for Fig. 3 and Tables 1, S5, and S6). Macrophage phagocytosis assay Murine bone marrow-derived macrophages were prepared as described [59],[60] and cultured in Dulbecco's Modified Eagles medium (DMEM) supplemented with 5 mM L-glutamine, 25 mM HEPES, 10% heat-inactivated fetal bovine serum, 5 mM non-essential amino acids, and 10 ng/ml CSF-1 (PeproTech; Rocky Hills, NJ). 1-ml suspensions of Y. pestis KIM6+ containing pAcGFP1 (Clontech; Mountain View, CA) from 21°C stationary phase LB/MOPS cultures, or from triturated midguts dissected from fleas 2 to 3 weeks after infection were treated for 15 sec in a FastPrep FP120 using lysing matrix H (Qbiogene; Carlsbad, CA) to disrupt bacterial aggregates, quantified by Petroff-Hausser direct count, and diluted in DMEM to ∼1×106 bacteria/ml. 0.1 ml of bacterial suspension was added to tissue culture plate wells containing ∼1×105 differentiated primary macrophages cultured on 12 mm glass coverslips in 1 ml DMEM. The plates were not centrifuged after addition of the bacteria, and midgut triturate from an equivalent number of uninfected fleas was added to the in vitro-derived bacterial suspensions used for these experiments. After 1 h incubation at 37°C and 5% CO2, the medium was removed and the cells washed, fixed in 2.5% paraformaldehyde for 10 min at 37°C, and then rewashed. Extracellular bacteria were labelled by indirect immunofluorescence as described [60] using a 1∶50,000 dilution of hyperimmune rabbit anti-Y. pestis polyclonal antibody [7] and a 1∶400 dilution of AlexaFluor 568-conjugated goat anti-rabbit antibody (Invitrogen; Carlsbad, CA). The percentage of extracellular bacteria was determined by dividing the number of red-fluorescent bacteria by the total number (red- and green only-fluorescent) bacteria associated with individual macrophages. To calculate differential resistance to phagocytosis for a given strain, the average percent extracellular LB-grown bacteria was subtracted from the average percent extracellular flea-derived bacteria. Results from 2–3 independent experiments performed in triplicate were analyzed by unpaired two-tailed t-test. Quantitative RT-PCR Independent RNA samples were prepared from blocked fleas and in vitro biofilm and planktonic cultures as described for the microarray experiments, except that the RNA was not amplified. Samples were treated with rDnase I (Ambion) and confirmed by PCR to be free of genomic DNA contamination. cDNA was synthesized from the RNA and used for quantitative PCR on an ABI Prism 7900 sequence detection system (Taqman, Applied Biosystems). The reactions contained oligonucleotide primers and probes designed using Primer Express version 2.0 software (Applied Biosystems) and the Taqman Universal PCR Master Mix (Applied Biosystems). For each primer-probe set assay, a standard curve was prepared using known concentrations of Y. pestis KIM6+ genomic DNA and used to transform CT values into relative DNA quantity. The quantity of cDNA for each experimental gene was normalized relative to the quantity of the reference gene crr (y1485), and the ratio of the normalized quantity of each gene in the flea samples to the normalized quantity in the in vitro samples was calculated (Fig. S2). Primer and probe sets used are listed in Table S7. Supporting Information Figure S1 Representative electrophoretograms of total RNA extracted from dissected flea digestive tracts. Electrophoretograms derived from uninfected (A) and blocked (B) flea digestive tracts are shown, with prokaryotic and eukaryotic rRNA peaks indicated. (0.97 MB TIF) Click here for additional data file. Figure S2 Quantitative reverse transcription (QRT) PCR confirmation of microarray results. The quantity of each mRNA was determined relative to that of the reference gene crr (y1485). Fold-differences in transcript levels of the 12 Y. pestis genes in the flea compared to (A) in vitro biofilm, (B) exponential phase planktonic cultures, and (C) stationary phase planktonic cultures are shown as determined by microarray (grey bars) and QRT-PCR (black bars). *gabT transcript was detected by microarray in the flea samples only. (1.92 MB TIF) Click here for additional data file. Table S1 Y. pestis genes upregulated ≥2-fold in the flea relative to all in vitro conditions. (0.32 MB DOC) Click here for additional data file. Table S2 Y. pestis genes downregulated ≥2-fold in the flea relative to all in vitro conditions. (0.14 MB DOC) Click here for additional data file. Table S3 The 100 most highly expressed Y. pestis genes in the flea. (0.22 MB DOC) Click here for additional data file. Table S4 Y. pestis genes upregulated ≥2-fold in the flea and flowcell biofilms relative to planktonic culture conditions. (0.12 MB DOC) Click here for additional data file. Table S5 Y. pestis genes with significantly higher relative expression levels in the flea gut than in the rat bubo. (0.28 MB DOC) Click here for additional data file. Table S6 Y. pestis genes with significantly higher relative expression levels in the rat bubo than in the flea. (0.51 MB DOC) Click here for additional data file. Table S7 Primers and probes used for quantitative RT-PCR. (0.06 MB DOC) Click here for additional data file.
