|Home | About | Journals | Submit | Contact Us | Français|
HD-GYP domain cyclic dimeric GMP (c-di-GMP) phosphodiesterases are implicated in motility and virulence in bacteria. Borrelia burgdorferi possesses a single set of c-di-GMP-metabolizing enzymes, including a putative HD-GYP domain protein, BB0374. Recently, we characterized the EAL domain phosphodiesterase PdeA. A mutation in pdeA resulted in cells that were defective in motility and virulence. Here we demonstrate that BB0374/PdeB specifically hydrolyzed c-di-GMP with a Km of 2.9 nM, confirming that it is a functional phosphodiesterase. Furthermore, by measuring phosphodiesterase enzyme activity in extracts from cells containing the pdeA pdeB double mutant, we demonstrate that no additional phosphodiesterases are present in B. burgdorferi. pdeB single mutant cells exhibit significantly increased flexing, indicating a role for c-di-GMP in motility. Constructing and analyzing a pilZ pdeB double mutant suggests that PilZ likely interacts with chemotaxis signaling. While virulence in needle-inoculated C3H/HeN mice did not appear to be altered significantly in pdeB mutant cells, these cells exhibited a reduced ability to survive in Ixodes scapularis ticks. Consequently, those ticks were unable to transmit the infection to naïve mice. All of these phenotypes were restored when the mutant was complemented. Identification of this role of pdeB increases our understanding of the c-di-GMP signaling network in motility regulation and the life cycle of B. burgdorferi.
The bacterial second messenger cyclic dimeric GMP (c-di-GMP) [bis(3′,5′)-cyclic diguanylic acid] has been associated with a wide range of adaptive processes, most notably, the regulation of virulence-related gene products and motility (17, 36, 76, 83, 92, 96). C-di-GMP exerts its regulatory function at the level of transcription, translation, protein activity, secretion, and/or protein stability via its interaction with various types of downstream effector molecules (36, 37, 80, 90). For example, in response to c-di-GMP, the Escherichia coli PilZ domain protein YcgR induces a counterclockwise (CCW) rotational bias by interacting with flagellar motor switch protein FliG, FliM, or MotA; however, in Caulobacter crescentus, a PilZ domain receptor, DgrA, was reported to destabilize the flagellar protein FliL, negatively affecting motility in response to c-di-GMP (7, 14, 22, 66, 80).
The levels of c-di-GMP are controlled by the opposing activities of diguanylate cyclase (DGC) and phosphodiesterase (PDE) enzymes. DGC enzymes synthesize c-di-GMP from two molecules of GTP, while PDEs hydrolyze c-di-GMP to pGpG or GMP. Open reading frames encoding DGC activity are characterized by the presence of a GGDEF protein domain. Proteins with PDE activity typically harbor either EAL or HD-GYP domains (36, 76, 83, 92). Although there are numerous reports on the characterization of GGDEF and EAL domain-containing proteins in diverse species of bacteria, only a few HD-GYP domain (HD-GYP; Pfam PF01966) proteins have so far been characterized, despite the fact that the HD-GYP domain proteins are also abundant in bacteria (28, 31). Comparative genome analyses indicated that a number of bacterial genomes encode proteins with the GGDEF domain but no identifiable EAL domain PDE (30, 31). A role for HD-GYP domain proteins in c-di-GMP hydrolysis was then proposed based on the distribution and numbers of GGDEF, EAL, and HD-GYP domains encoded by different bacteria and on the known activities of other members of the HD superfamily of enzymes as metal-dependent hydrolases (30). Experimental evidence of a role for the HD-GYP domain in c-di-GMP hydrolysis was provided through studies of Xanthomonas sp. RpfG (1, 77, 79). Two additional HD-GYP domain proteins, PA4108 and PA4718, from Pseudomonas aeruginosa also have been reported to be c-di-GMP PDEs (78). In each case, inactivation of these putative HD-GYP PDEs resulted in reduced motility and decreased virulence (1, 78, 79).
Synthesis of c-di-GMP in Borrelia burgdorferi, the Lyme disease spirochete, is controlled by the Hk1/Rrp1 two-component signal transduction system (31, 75), with Hk1 serving as the sensor histidine kinase component responsible for phosphorylating Rrp1, a diguanylate cyclase. Recent independent studies by Caimano et al. (10a), He et al. (35a), and Kostick et al. (45) indicate that these two gene products act cooperatively to promote spirochete survival in feeding ticks following acquisition. Interestingly, neither gene was required to establish infection in mice by needle inoculation. Moreover, spirochetes lacking Hk1 were able to migrate out of the bite site into feeding nymphs but were killed within 36 h following ingestion. These results indicate that the function of Hk1/Rrp1 and c-di-GMP is likely to be restricted in the arthropod phase of the enzootic cycle. The presence of two distinct PDEs, PdeA (BB0363) and PdeB (BB0374), implies that the ability to modulate the intracellular levels of c-di-GMP is an important component of the spirochete's pathogenic strategy. Indeed, inactivation of the EAL domain containing c-di-GMP-specific PDE (pdeA) resulted in cells that were avirulent in mice, presumably as a result of their altered swimming pattern (91).
The motility of B. burgdorferi results from coordinated rotation of the periplasmic flagella residing between the outer membrane and the cell cylinder (12, 13, 18, 34, 46, 50). The motile behavior of B. burgdorferi and other spirochetes is unique and complex (12, 13). In vitro, B. burgdorferi exhibits three different swimming modes, i.e., running, flexing, and reversing (5, 12, 50, 58), with running occurring when the periplasmic flagellar motors at either end of the spirochete cell body rotate in opposite directions. Thus, the periplasmic flagella of the anterior ribbon rotates CCW and those of the posterior end rotate clockwise (CW) (as a frame of reference, a periplasmic flagellum is viewed from its distal tip along the filament toward insertion into the motor) (12, 13, 18, 50). The flex is a nontranslational mode that is often associated with bending of the cell body (33, 34, 58). During a flex, the motors at both ends rotate in the same direction; i.e., both rotate either CW or CCW (12, 23, 34, 50, 58). The spirochete flex is thought to be equivalent to the tumbling behavior displayed by Escherichia coli and Salmonella enterica serovar Typhimurium (8, 12, 23, 34, 50, 58). A reversal occurs in translating cells when the motors at either end simultaneously reverse their direction of rotation (12, 50). Phosphorylated CheY (CheY-P) plays a major role in regulating the direction in which flagellar motors rotate. Thus, a B. burgdorferi cheY3 mutant (no CheY3-P) constantly runs whereas a CheY-P phosphatase cheX mutant (elevated CheY3-P) constantly flexes (58, 61, 67).
Bacterial motility is important for the colonization/disease process for numerous species of pathogenic bacteria, including B. burgdorferi (9, 10, 42, 52, 55, 62). In vitro, B. burgdorferi is able to traverse viscous gel-like media in which most other flagellated bacteria slow down or stop (44). B. burgdorferi is able to disseminate out of the site of inoculation and colonize diverse host tissues, such as joints, the nervous system, and the heart (33, 44, 86), due in part to its unique motility (12, 18, 34, 50-52, 91). However, motility is likely not required for spirochetes to survive or replicate in ticks (our unpublished observation; 19).
To better understand the importance of c-di-GMP signaling in the life cycle of B. burgdorferi, we characterized the single HD-GYP domain protein PdeB. We demonstrate that PdeB is a functional c-di-GMP PDE that binds c-di-GMP with high affinity and hydrolyzes it. By inactivating both pdeA and pdeB and using cell extracts in PDE assays, we established that B. burgdorferi encodes no additional PDE. Dark-field analysis of spirochetes lacking PdeB demonstrated that these cells exhibit significantly increased flexing, indicating a role for this PDE in modulating motility. Furthermore, although loss of PdeB did not significantly affect virulence in needle-inoculated mice, the mutant displayed a reduced ability to survive in Ixodes scapularis ticks. Ticks infected with the pdeB mutant were unable to transmit the infection into naïve mice, resulting in a failure to complete the mouse-tick-mouse infection cycle. Normal motility and survival in ticks were restored by complementation. Together, these data suggest a role for PdeB in tick-mediated transmission and imply that decreased levels of c-di-GMP may be important for the spirochete's survival postfeeding/postrepletion. Inactivation of the c-di-GMP receptor PilZ protein PlzA in the pdeB mutant background suggests a role for c-di-GMP/PlzA in chemotaxis signaling. Possible mechanisms by which PdeB, PlzA, and c-di-GMP alter motility and infectivity are discussed.
Low-passage, virulent B. burgdorferi strain B31-A3-K2Δbbe002 (a kind gift from R. Rego and P. Rosa, Rocky Mountain Laboratories [RML], NIH) (72) was used as the wild type throughout this study. This strain is a derivative of A3-68 lacking circular plasmid 9 and linear plasmid 56 (lp56), with the bbe002 gene inactivated using a PflgB-kan cassette to increase the transformation frequency (43, 72). B31-A3-K2Δbbe002 is a derivative of strain B31 (21). The genome of the virulent B31 strain has been sequenced and was found to contain a total of 21 plasmids with 12 linear and 9 circular plasmids, in addition to the 960-kb linear chromosome (11, 25). B. burgdorferi was cultured in liquid Barbour-Stoenner-Kelly (BSK-II) medium; plating BSK-II was prepared using 0.6% agarose (57).
Targeted inactivation of pdeB (1,140 bp; gene locus bb0374) was achieved by homologous recombination using a streptomycin resistance gene (aadA) as a marker (24). The pdeB gene was inactivated by replacing the first 1,040 bp of pdeB with the 792-bp aadA coding sequence using overlapping PCRs, as we recently described in detail (59). PCR was used to amplify 3 regions of DNA in 3 steps (not shown; see reference 59). In the first step, each DNA region was amplified separately using PCR pairs P1-P2 (5′-flanking DNA, bb0373), P3-P4 (aadA coding sequence from plasmid pKFSS1), and P5-P6 (3′-flanking DNA, pfs). Primers P2, P3, P4, and P5 contain several overlapping base pairs (see below). During step 2, a PCR product was obtained using primers P1-P4 and the purified DNA products for bb0373 and aadA as templates. In step 3, the final PCR product was obtained using primers P1-P6 and purified DNA products bb0373-aadA and pfs as templates. The primer sequences (5′-3′) are as follows: P1, AGACTTTGCAAAAAAATAAT; P2, CGCTTCCCTCATTTATTCCCTTGAAATTAAGCTC; P3, TCAAGGGAATAAATGAGGGAAGCGGTGATCGCCG; P4, TGGTATAGATTGTTATTTGCCGACTACCTTGGTG; P5, GTCGGCAAATAACAATCTATACCAAATACAAATAC; P6, AATCAATATTTGTATAGCCTATATC1. The final PCR yielded a 2,891-bp product that was gel purified and cloned into the pGEM-T Easy vector (Promega Inc.). The deletion-insertion was in frame. The integrity of the pdeB inactivation plasmid was confirmed by PCR and restriction mapping. NotI-digested DNA was electroporated into wild-type competent cells and plated on plating BSK containing 200 μg ml−1 kanamycin plus 100 μg ml−1 streptomycin as described previously (57). Resistant clones were analyzed by PCR for the confirmation of homologous recombination. PCR-positive mutants were further examined for their plasmid contents using 21 sets of primers to detect 21 linear and circular plasmids (100).
To complement the pdeB mutant, the promoter of the pdeB operon (PpdeB; 345 bp upstream from the ATG start codon of the pdeB gene) (74) and the coding sequence of pdeB were amplified by PCR with primers Bb0373-BamHI-F (GGATCCGTGAATGCATCTTCCATCCCA) and Bb0374-Pst-R (CTGCAGTTATATAATATCTATTAAAGAA). BamHI and PstI sites are in bold font. The 1,499-bp PCR product (PpdeB-pdeB) was ligated into pGEM-T Easy. The PpdeB-pdeB DNA was then inserted into shuttle vector pBSV2G (41, 91) using PstI and BamHI restriction digestion, yielding pBSbb0374. Fifty micrograms of the resultant plasmid was electroporated into the pdeB::aadA competent cells. Transformants were selected on solid growth medium containing 200 μg ml−1 kanamycin, 100 μg ml−1 streptomycin, and 40 μg ml−1 gentamicin. Resistant transformants were analyzed by PCR for the presence of Strr, Gmr, and pdeB. To confirm that the plasmid was complementing in trans, the pBSVbb0374 shuttle vector was rescued from complemented pdeB+ cells, transformed, and then purified from E. coli and the integrity of PpdeB-pdeB was verified by restriction digestion. Positive transformants were further examined for their plasmid content by PCR.
Exponentially growing B. burgdorferi cells were observed under a dark-field microscope (Zeiss Axio Imager M1) connected to an AxioCam digital camera and video recorded. The flex rate was determined by determining the mean number of flexes per minute ± the standard deviation. At least 12 cells of each strain were analyzed. A paired Student t test was used to determine statistical significance (a P value for the difference between the wild type and the mutant). Swarm plate assays were performed as described previously (50, 52, 57, 58). Approximately 1 × 106 cells in a 5-μl volume were spotted onto a 0.35% agarose plate containing plating BSK-II medium diluted 1:10 in phosphate-buffered saline (PBS) without divalent cations. Because B. burgdorferi is a slow-growing organism with an 8- to 12-h generation time, swarm plates were incubated for 6 days at 35°C in a 2.5% CO2 humidified incubator (57, 91).
Exponentially growing cultures of B. burgdorferi were treated with RNAprotect, and then total RNA was isolated using the RNeasy minikit (Qiagen Inc.). Contaminating DNA in the RNA samples was removed by RNase-free Turbo DNase I (Ambion, Inc.) digestion for 3 h at 37°C, followed by RNeasy mini purification. For RT-PCR, cDNA was prepared from 1 μg of RNA using the Bio-Rad cDNA synthesis kit according to the manufacturer's protocol. The iCycler detection system (Bio-Rad Inc.) was used to measure the transcript level of the test genes in wild-type, pdeB mutant, and complemented cells according to the manufacturer's instructions. The B. burgdorferi gene for enolase was used as a reference (60). The gene-specific primers (5′-3′) were (i) BB0374 qRT F (TTATTCATTCGGTAAATACAGCTATC) and BB0374 qRT R (CTAGTTCTTCCTCAGTTAATGCTTCT) and (ii) RT-enolase-F (TGGAGCGTACAAAGCCAACATT) and RT-enolase-R (TGAAAAACCTCTGCTGCCATTC). The relative level of expression was calculated using the 2−ΔΔCT method (54, 84).
SDS-PAGE and Western blotting with an enhanced chemiluminescent detection method (GE Healthcare) were carried out according to the manufacturer's instructions. The concentration of proteins in the cell lysates was determined by a Bio-Rad protein assay kit. Unless noted otherwise, 10 μg of lysates was loaded in each lane for SDS-PAGE, transferred to polyvinylidene fluoride membranes, and subjected to Western blotting using specific antibodies. The following antibodies were kindly provided by other investigators: polyclonal anti-Pfs by B. Stevenson (University of Kentucky, Lexington, KY), monoclonal anti-FlaB (H9724) by A.G. Barbour (University of California, Irvine, CA), monoclonal anti-DnaK by J. Benach (SUNY, Stony Brook, NY), polyclonal anti-MotB by J. Carroll (NIH at RML, Hamilton, MT), polyclonal FliG1 and FliG2 by C. Li (SUNY, Buffalo, NY), and polyclonal FliM by D. Blair (University of Utah, Salt Lake City, UT). The specific reactivity of these antibodies to B. burgdorferi CheX, CheY3, Pfs, FlaB, MotB, FliG1, FliG2, and FliM has been reported previously (16, 50, 52, 56, 58, 60, 74, 82).
For overexpression and purification of recombinant PdeB (R-PdeB) protein, the coding sequence of pdeB was amplified without the first ATG start codon by PCR using primers R-374-Bam-F (GGATCCCAAAATTCTGAAAGCATTAT) and R-374-Pst R (CTGCAGTTATATAATATCTATTAAAGAATA), creating a BamHI and a PstI site (bold), respectively. For overexpression and purification of the HD-GYP domain (amino acids [aa] 90 to 305) of PdeB, the pdeB coding sequence was amplified without the first 89 and the last 75 aa residues by PCR using primers R-374-HDGYP-Bam-F (GGATCCTATGAAAATTGCAAAAGAAATAAA) and R-374-HDGYP-Pst-R (CTGCAGTAAAGTCTAAGGGACAAGAAGATAT), creating a BamHI and a PstI site (bold), respectively. The amplified DNA fragments were ligated into E. coli expression vector pMAL-c2X (New England BioLabs) using BamHI and PstI. The recombinant expression vectors constructed for PdeB and the HD-GYP domain were named pbb0374-MAL and pbbHDGYP-MAL, respectively. The recombinant expression vectors were analyzed by restriction digestion to confirm their integrity. E. coli cells containing pbb0374-MAL or pbbHDGYP-MAL were induced with 0.25 mM isopropyl-β-d-thiogalactopyranoside at 30°C for 4 h and purified on amylose resin (New England BioLabs). Purified proteins were concentrated and dialyzed in PDE buffer (75 mM Tris-HCl [pH 8], 10 mM MgCl2, 25 mM KCl, 250 mM NaCl). Purified R-PdeB or HD-GYP protein was resolved by SDS-PAGE to determine protein concentration and purity.
PDE assays of wild-type, pdeB mutant, complemented pdeB+, and pdeA pdeB double mutant cell extracts were performed as described previously (15, 91), with minor modifications. Briefly, cell pellets were washed once with PBS and resuspended in PDE assay buffer (250 mM NaCl, 25 mM Tris-HCl [pH 8], 10 mM MgCl2, 25 mM KCl, 10% glycerol). Cells were lysed by a combination of French press and sonication. Lysed cells were centrifuged at 100,000 × g for 1 h, and the protein concentration in the supernatants was determined using a Bio-Rad protein assay kit. Synthesis of radiolabeled c-di-GMP substrate for PDE assays was performed as described previously (49, 91) and approved by the East Carolina University Radiation Safety Subcommittee. An E. coli clone carrying P. aeruginosa His6-WspR, which has high diguanylate cyclase activity, was affinity purified as described previously (47, 49, 91). 33P-labeled c-di-GMP was prepared by incubating 20 μg of His6-WspR, 100 mCi of [α-33P]GTP (3,000 Ci mmol−1; 10 mCi ml−1; Perkin-Elmer) in 50 mM Tris-HCl (pH 7.5)-250 mM NaCl-10 mM MgCl2-1 mM dithiothreitol for 4 h at 25°C. The reaction mixture was then incubated with 20 units of calf intestine alkaline phosphatase (New England BioLabs) for 1 h at 25°C to hydrolyze unreacted GTP. The reaction was stopped with 0.5 M EDTA, and the product was passed through an Ultrafree MC-5000 column (Millipore Inc.). PDE assays were performed as described previously (91, 94). R-PdeB protein (2 μg) or B. burgdorferi crude extract (4 μg) was incubated with 33P-labeled c-di-GMP with or without 50 mM MnCl2 for different periods of time. A no-enzyme or unrelated-protein (maltose binding protein [MBP]-FlhF) negative control and a Vibrio cholerae PDE His6-CdpA (1.5 μg) positive control were used (91, 93). Reaction products (1 μl) were spotted and separated by thin-layer chromatography (TLC) using polyethyleneimine-cellulose plates in 1.5 M KH2PO4 (pH 3.65) buffer. The plates were air dried, and then phosphorimaging and volume analysis using Image Quant TL v2003 were done. To determine the Km, R-PdeB was incubated with 2 to 70 nM 33P-labeled c-di-GMP for up to 30 min, reactions were stopped every 5 min, the products were subjected to TLC, and the results were analyzed using GraphPad Prism 5 (Michaelis-Menten equation) as described previously (15, 91, 94). To determine the substrate specificity of R-PdeB, 15 to 30 nM 3H-labeled cGMP (Perkin-Elmer) was incubated with 6 mg of R-PdeB for 30 min to overnight and the reaction product was separated by TLC as described previously (91, 94).
Five-week-old female C3H/HeN mice were purchased from Charles River Laboratories, Durham, NC, and housed in the East Carolina University animal facility at the Brody School of Medicine according to the institutional guidelines for the care and use of laboratory animals. For infection via needle, 5 × 102 to 5 × 104 in vitro-grown spirochetes were injected subcutaneously as described previously (21, 40, 73). The number of spirochetes was determined using a Petroff-Hausser chamber, and each clone was verified for retention of lp25, lp28-1, and lp36 plasmids. Mice were bled 2 weeks postinfection, and sera against B. burgdorferi antigens were tested to determine infectivity as described previously (21, 85). Reisolation of spirochetes from mouse ears, joints, and bladders was performed at 4 weeks postinfection as previously described (91) to assess the ability of spirochetes to infect mice. Mouse tissues were incubated in BSK-II growth medium for up to 35 days, and the presence of spirochetes was determined by dark-field microscopy. Reisolated clones were genotyped using PCR as described above.
For tick infection studies, naïve I. scapularis larvae were purchased from Oklahoma State University. Two independent experiments were performed with ticks. Ticks were kept at 23 to 24°C under a 14/10 light/dark photoperiod in a humidified chamber with 85 to 90% relative humidity. Approximately 150 larval ticks were fed to repletion on spirochete-infected mice for 5 to 7 days, allowed to fall off, and collected. One subset of larvae was dissected 7 days after repletion, and the isolated midguts were analyzed by indirect immunofluorescence assays (IFA) for the presence of spirochetes (87, 91). A second subset of fed larvae were surface sterilized using 3% H2O2, followed by 70% ethanol; crushed in BSK-II medium; and plated to determine the number of CFU per tick. The remaining fed larvae were allowed to molt to nymphs. Those nymphs were then fed to repletion on naïve mice (five-week-old female C3H/HeN mice), allowed to fall off, and collected. Subsets of nymphs were dissected 7 to 9 days after repletion, and the isolated midguts were analyzed by IFA for the presence of spirochetes. Fed nymphs were surface sterilized as described above, crushed in BSK-II medium, and plated to determine the number of CFU per tick. Mice were bled 2 weeks after nymph infection for immunoblot analysis of mouse sera, and reisolation of B. burgdorferi from mouse ear, bladder, and joint tissues was performed 4 weeks postinfection.
The dose required to infect 50% of the mice inoculated was experimentally determined for the wild-type and pdeB mutant strains as described previously (40, 68, 71). The data from the ID50 infection experiment and the single-dose infection experiment for each strain were combined for estimation of the ID50. Comparison of strain ID50 values was done using a generalized linear model with a probit link function. This method is also known as probit regression (40), and in it we assume identical slopes in the response/log-dose relationship but different intercepts for each strain. Graphically, those assumptions manifest themselves as dose-response curves with lateral shifts corresponding to the changes in intercept. Additionally, an overdispersion parameter was fitted in order to accommodate greater homogeneity of infection rates than would otherwise be permitted by the model. All calculations were carried out using JMP V9 software (SAS Institute Inc., Cary, NC).
Approximately 150 larval ticks were artificially infected by immersion (in duplicate) in equal-density exponential-phase B. burgdorferi cultures as described previously (6, 69, 91), except that larval ticks were equilibrated to a lower relative humidity overnight before immersion to enhance spirochete uptake. Ticks were fed to repletion on separate naïve mice for 5 to 7 days, allowed to fall off, and collected. One subset of larvae were crushed 7 days after repletion and analyzed by IFA for the presence of spirochetes (87, 91). A second subset of fed larvae were surface sterilized using 3% H2O2, followed by 70% ethanol; crushed in BSK-II medium; and plated to determine the number of CFU per tick. Serum was collected at 2 weeks after repletion, and mouse tissues were harvested at 4 weeks after repletion as described above. In determining the total number of spirochetes per fed tick, 5 ticks were analyzed for each strain and the results are expressed as the mean ± the standard deviation.
Ticks were dissected in10 μl PBS-5 mM MgCl2 on Teflon-coated microscope slides, mixed by pipetting, and then air dried (91). To avoid quenching by hemin in the blood, dissected tick contents were 10-fold serially diluted (P. Policastro, RML, NIH, personal communication). Slides were blocked with 0.75% bovine serum albumin in PBS-5 mM MgCl2 and washed with PBS-5 mM MgCl2. Spirochetes were detected using a 1:100 dilution of goat anti-B. burgdorferi antiserum labeled with fluorescein isothiocyanate (Kirkegaard & Perry Laboratories). Images were captured using a Zeiss Axio Imager M1 microscope coupled with a digital camera (91).
B. burgdorferi genome organization, as well as RT-PCR analyses, indicated that pdeB is cotranscribed with other genes in an operon, pdeB-pfs-metK-luxS (25, 38, 74). A σ70 promoter has also been located directly upstream of pdeB (74). In this operon, pdeB is predicted to be a PDE (see below), pfs encodes a nucleosidase which is essential for depleting toxic metabolic by-products, metK is essential for normal cell growth and septation, and the luxS-encoded enzyme synthesizes 4,5-dihydroxy-2,3-pentanedione/autoinducer 2 (4, 38, 65, 74). A luxS mutant was constructed and reported to be not essential for infecting I. scapularis ticks or mice or transmission between these hosts (38). The expression of genes in the pdeB operon has also been reported to be modulated as a function of the growth phase (74). However, whether the HD-GYP domain-containing protein PdeB is a functional PDE has not been reported. To determine if pdeB functions in the B. burgdorferi enzootic life cycle, we inactivated this gene using allelic-exchange mutagenesis (Fig. 1) (59) by replacing a 1,040-bp fragment of the gene with the streptomycin resistance aadA gene (24) without its promoter. We found that this strategy prevents polar effects on the expression of downstream genes (59). PCR analysis confirmed that the Strr marker (aadA coding sequence) was correctly inserted within the pdeB gene, as expected (Fig. 2, left panel). To assess whether insertion of aadA exerted an unexpected polar effect on the expression of pfs (immediately downstream from pdeB), Western blot analysis was conducted using antiserum raised against Pfs (74). Western blotting results indicated that the synthesis of Pfs was not altered in the pdeB mutant (Fig. 2, right panel). These results indicated that the in-frame insertion of the aadA/Strr marker did not induce a polar effect.
To demonstrate that the pdeB mutant phenotype (see below) was due solely to the mutation and not due to a secondary alteration elsewhere, we complemented the mutant in trans using a shuttle vector, pBSV2G (Fig. 1, bottom) (20, 88, 91). Real-time PCR analysis indicated that pdeB transcripts were detected in wild-type cells but not in mutant cells, confirming inactivation of pdeB; pdeB transcripts in complemented pdeB+ cells were elevated approximately 5-fold above the level in wild-type cells, indicating overexpression from the multicopy vector (data not shown; see below) (95). To confirm retention of B. burgdorferi endogenous plasmids (63, 70) in the pdeB mutant and complemented strains, PCR-based plasmid profiling using each of the plasmid-specific primers was employed (70, 100). The plasmid profiles of the wild-type, pdeB mutant, and pdeB+ strains were the same, confirming the retention of all of the plasmids required for infectivity (not shown; see below).
To determine if a mutation in pdeB altered B. burgdorferi motility, the mutant cells were analyzed using dark-field microscopy and swarm plate motility assays (Table 1; see Fig. S1 in the supplemental material). As shown in Table 1, the mutant cells ran, flexed, and reversed as did the wild-type cells; however, their rate of flexing was significantly higher than that of the wild-type cells (5 versus 10 flexes/min; P value, 0.002). When a B. burgdorferi cell flexes, it becomes distorted or bent at the center (see videos S1 to S3 in the supplemental material for wild-type, pdeB mutant, and complemented pdeB+ cells) (33, 34, 58). In the swarm plate assays, the mutant swarm diameters were not significantly (P = 0.074) different from those of wild-type cells (see Fig. S1). As expected, the motility of the complemented pdeB+ cells was restored to the wild-type phenotype (Table 1; see Fig. S1 and videos S1 to S3). Together, these results indicate a role for pdeB in motility and that the mutant's phenotype was due solely to this mutation and not to a secondary alteration elsewhere.
In other species of bacteria, PilZ proteins were reported to control motility in response to changes in the level of c-di-GMP (3, 7, 14, 22, 66). B. burgdorferi possesses a c-di-GMP binding PilZ protein, PlzA (26, 68). To determine if PlzA can function in response to c-di-GMP to regulate B. burgdorferi motility, we constructed a plzA mutant in the pdeB mutant background by introducing one inactivated plasmid (plzA::Pl-kan) into the other mutant cells (pdeB::aadA). The plzA single mutant exhibits a normal swimming pattern but reduced swarming motility (68). The B. burgdorferi plzA pdeB double mutant cells exhibited a constant-flexing motility phenotype similar to the phenotype of cheX mutant cells (see the constant-flexing cheX mutant and plzA pdeB double mutant cell videos) (58). To determine if the increased flexing motility exhibited by the pdeB single mutant or the plzA pdeB double mutant was due to diminished levels of a motility/chemotaxis protein, we performed Western blotting using specific antiserum. As shown in Fig. 3, the expression of prominent motility or chemotaxis protein FlaB, FliG2, FliG1, FliM, MotB, CheY3, or CheX was not altered in the pdeB or plzA pdeB mutant cells. These results suggest that the altered motility exhibited by those mutants is not due to the reduced expression of major motility/chemotaxis proteins. However, it is possible that the constant-flexing phenotype exhibited by the double mutant cells is due to inhibition of CheX activity.
Recently, we reported that the EAL domain-containing protein PdeA is a functional c-di-GMP-specific PDE (91). HD-GYP domain proteins have been reported to possess c-di-GMP PDE activity in Xanthomonas campestris and P. aeruginosa (77, 78). To determine if PdeB is an active PDE, we performed PDE assays using purified R-PdeB, as well as B. burgdorferi mutant cell extracts, as we and others previously described (15, 91, 94). Purified (greater than 90% pure; see Fig. 4B, right panel) recombinant full-length PdeB (379 aa) or only the HD-GYP domain (aa 90 to 305) and crude cell extracts prepared from wild-type, pdeB mutant, and pdeB+ complemented cells were tested for the ability to hydrolyze 33P-labeled c-di-GMP. As shown in Fig. 4A, the wild-type crude extract hydrolyzed c-di-GMP to GMP, indicating PDE activity (29 ± 3 fmol/mg/min; wild-type lanes). Addition of 50 mM MnCl2 to the reaction mixture enhanced the PDE activity, which is indicative of metal-dependent PDE activity (all lanes). pdeB mutant extracts also hydrolyzed c-di-GMP to GMP but at a lower rate (19 ± 1.9 fmol/mg/min) than the wild-type extracts (Fig. 4A, lanes pdeB), indicating the presence of another PDE in these mutants. pdeB+ complemented cell extracts rapidly hydrolyzed 33P-labeled c-di-GMP (67 ± 1.7 fmol/mg/min; lanes pdeB+), indicating that the reduction of 33P-labeled c-di-GMP hydrolysis in pdeB mutant cells was restored upon complementation. pdeB+ cell extracts exhibited greater-than-wild-type PDE activity, likely because pdeB was overexpressed in the complemented cells. These results indicated that PdeB is a functional PDE. To further confirm that PdeB is a c-di-GMP PDE, purified full-length PdeB or only the HD-GYP protein was assayed for the ability to hydrolyze 33P-labeled c-di-GMP (Fig. 4B and C). The full-length protein or only the functional domain hydrolyzed 33P-labeled c-di-GMP to GMP (Fig. 4B and C). The addition of MnCl2 enhanced the PDE activity, establishing that PdeB is a metal-dependent c-di-GMP PDE (all lanes).
Full-length R-PdeB was assayed with different concentrations of 33P-labeled c-di-GMP, and the Km value was calculated to be 2.9 nM, indicating a high affinity for c-di-GMP. The positive control, CdpA, rapidly hydrolyzed c-di-GMP, and negative controls without added protein or with an unrelated FlhF protein failed to hydrolyze c-di-GMP (Fig. 4A and B) (91, 93). Furthermore, 3H-labeled cGMP was used to determine if PdeB is a c-di-GMP-specific PDE. 3H-labeled cGMP was not hydrolyzed by R-PdeB, even when the reaction mixture was incubated for a prolonged period of time (overnight; data not shown), suggesting that PdeB is a c-di-GMP-specific PDE.
To determine if B. burgdorferi possesses only two PDEs (PdeA and PdeB), we constructed a pdeA::Pl-kan-pdeB::aadA double mutant by introducing one inactivated plasmid into the other mutant cells. The pdeA pdeB double mutant cell motility phenotype was indistinguishable from that of the pdeB single mutant. Extracts from pdeA pdeB double mutant cells failed to hydrolyze 33P-labeled c-di-GMP, even when the reaction mixture was incubated for extended times (up to 4 h) (Fig. 4A, pdeA pdeB lanes; data not shown). Therefore, it is likely that PdeB and PdeA are the only two active c-di-GMP PDEs in B. burgdorferi.
In order to evaluate the infection potential of the pdeB mutant, groups of C3H/HeN mice were challenged by subcutaneous injection with 10-fold increasing doses of wild-type, pdeB mutant, or isogenic complemented pdeB+ cells to determine the ID50. At 2 weeks postinoculation, the mice were bled and their sera were assessed for reactivity with B. burgdorferi antigens. B. burgdorferi membrane protein A, also known as P39, was used as a marker of infection in animals (40, 85). To confirm the serological results, mice were sacrificed at 4 weeks postinoculation and tissue (ear, joint, and bladder) samples were aseptically isolated and assessed for the presence/absence of B. burgdorferi using dark-field microscopy. Serology results correlated well with reisolation of spirochetes from the tissues examined (Table 2and not shown). pdeB mutant cells did not show a significant increase in the ID50 compared to parental wild-type cells (P = 0.3976; Table 2). These results indicate that the pdeB mutant is not significantly attenuated in virulence relative to the parental wild-type strain or the complemented pdeB+ strain.
The B. burgdorferi infectious life cycle includes persistent infection of and survival within tick and mammalian hosts. Because persistent infection of the mammalian host represents only one facet of the spirochete's enzootic life cycle, a more comprehensive evaluation of the behavior of pdeB mutant strains in the tick vector was warranted. To perform such an evaluation, naïve larval ticks were allowed to feed on mice that were infected with the wild-type, mutant, or complemented strain. Upon larval repletion, the majority of the ticks were allowed to advance naturally through the month-long molting process. At 7 days postfeeding, a subset of fed larvae was assessed for the survivability of spirochetes using IFA as well as squashing and plating of triturated ticks to count viable spirochetes (Table 3and data not shown; see Fig. S2 in the supplemental material). While the numbers of spirochetes in ticks that fed on the wild-type and complemented pdeB+ strains were similar, the number of viable spirochetes was 5-fold lower in ticks that fed on the mutant-infected mice (Table 3). The percentage of ticks infected with the wild-type, mutant, or complemented strain was approximately 50%; these results indicated that a mutation in pdeB resulted in cells that were deficient in survival in ticks (Table 3). The subset of fed larvae that molted to nymphs was examined for the spirochetes' ability to migrate from the ticks to naïve mice to complete the mouse-tick-mouse infection cycle (below).
In the normal B. burgdorferi infectious cycle, mice are infected when spirochete-laden ticks bite them. We determined whether infected nymphs are able to transmit the spirochetes to naïve mice. Thirty nymphs of each group of wild-type, pdeB mutant, and complemented pdeB+ strains were allowed to feed on a naïve mouse (n = 4) to determine spirochete transmission from ticks to the mammalian hosts (Table 3). Although the percentage of ticks infected with all 3 strains was approximately 50%, the mutant-infected ticks failed to transmit the infection to any mice, whereas wild-type-infected ticks were able to transmit the infection to all mice, as confirmed by reisolation of spirochetes from sacrificed mouse tissue (ear, joint, and bladder) specimens (Table 3). The pdeB+ spirochete-infected ticks were also able to transmit the infection to 2/4 naïve mice (Table 3). The pdeB+ cells' partially restored virulence likely resulted from the overexpression of pdeB, or a small population of the complemented pdeB+ cells may have lost an endogenous plasmid(s). These results indicated that pdeB mutant cells failed to complete the mouse-tick-mouse infection cycle.
Tick immersion studies were performed to confirm the extent to which wild-type and mutant B. burgdorferi strains are able to infect, survive, and transmit the infection from arthropod tick vectors to a mammalian host (87, 89, 91). Tick immersion studies allow direct artificial tick infection and serve two purposes, i.e., (i) to optimally infect naïve ticks with the wild-type and mutant spirochetes and determine their colonization and the ability to survive within the tick vector and (ii) to examine for the potential of spirochetes to migrate from the arthropod host vector to the mammalian host. Three-month-old tick larvae were artificially inoculated with exponentially growing (5 × 107 cells/ml) cultures of wild-type, pdeB mutant, or complemented pdeB+ B. burgdorferi cells. As ticks ingest B. burgdorferi cultures, their intestines are inoculated with spirochetes. After tick immersion, approximately 150 spirochete-laden ticks of each strain were allowed to feed on a naïve mouse (n = 2 or 4) (Table 4). Fed ticks were analyzed for the presence of spirochetes by IFA (see Fig. S2 in the supplemental material), and crushed ticks were plated and viable spirochetes were counted. While the spirochete burdens of the ticks inoculated with wild-type or complemented pdeB+ cells were approximately the same, the mutant strain spirochete burden was 5-fold lower when ticks were analyzed 7 days after blood acquisition (Table 4). The percentage of fed ticks infected with the wild-type, mutant, or complemented strain was the same (~80%; Table 4). These data and the mouse-tick-mouse results demonstrate that PdeB is likely required for B. burgdorferi to colonize and/or survive in the tick vector. Mice on which ticks fed were sacrificed to determine the extent to which spirochetes could be isolated from tissue specimens. Consistent with mouse-tick-mouse infection cycle studies (Table 3), spirochetes were isolated only from tissue specimens collected from mice on which ticks infected with the wild-type or the complemented pdeB+ strain fed. No spirochetes were isolated from any of the tissue specimens from mice on which ticks infected with the pdeB mutant cells fed (Table 4). These results confirm that the pdeB mutant has a reduced ability to colonize ticks and is unable to transmit an infection from ticks to naïve mice.
HD-GYP domains belong to the HD superfamily of metal-dependent hydrolases (2, 29-31) and are often encountered fused to other signal transduction domains such as REC, GGDEF, and PAS domains (27, 31). Based on these observations, Galperin et al (31) initially proposed that HD-GYP domain-containing proteins may function as PDEs directed toward either the dephosphorylation of phosphoproteins or the hydrolysis of cyclic nucleotides, especially c-di-GMP. HD-GYP domain-containing proteins subsequently have been shown to indeed possess c-di-GMP-hydrolyzing activity (77, 78). Although more than 590 HD-GYP domains in over 140 genomes have been reported (www.ncbi.nlm.nih.gov/complete_genomes/signalcensus.html), only a handful have been characterized (1, 35, 77, 78). The genome of B. burgdorferi encodes a single diguanylate cyclase and two PDEs, one of the EAL type and one of the HD-GYP type (25, 31). Purified recombinant forms of the diguanylate cyclase Rrp1 and the EAL domain-containing PdeA have been shown to synthesize and hydrolyze c-di-GMP in vitro, respectively (81, 91). In PDE assays, extracts from pdeA mutant cells exhibited a reduced ability to hydrolyze 33P-labeled c-di-GMP, relative to that of wild-type cells. However, pdeA mutant cell extracts retained considerable c-di-GMP-hydrolyzing activity, leading us to propose that B. burgdorferi expresses more than one functional PDE (91). Several lines of evidence led us to conclude that the HD-GYP domain-containing protein PdeB functions as a c-di-GMP-specific PDE (Fig. 4A to C). First, we constructed a pdeB mutant of B. burgdorferi and compared the abilities of extracts from wild-type, mutant, and complemented cells to hydrolyze radiolabeled c-di-GMP in vitro. The wild-type and complemented cell extracts readily hydrolyzed 33P-labeled c-di-GMP, and this activity was reduced in extracts from pdeB mutants (Fig. 4A). Second, purified recombinant full-length protein and a truncated form of PdeB containing the HD-GYP domain exhibit c-di-GMP-specific PDE activity (Fig. 4B and C); as with other HD-GYP PDEs, the activity of PdeB was enhanced by the addition of Mn2+ (77, 78, 94). It is notable that when the HD-GYP domain of RpfG of X. campestris was mutated to HA-GYP, the c-di-GMP-hydrolyzing activity was abolished (77). B. burgdorferi PdeB contains an imperfect HK-GYP domain instead of HD-GYP; it is not known if a mutation in the HD-GYP domain of RpfG to HK-GYP would alter enzyme activity. However, our PDE assays clearly demonstrate that the HK-GYP domain-containing PdeB hydrolyzes c-di-GMP and exhibits a high affinity for c-di-GMP (Fig. 4A to C).
To date, only three HD-GYP-type PDEs have been shown to hydrolyze c-di-GMP in bacteria; however, Km values for those PDEs have not been reported (77, 78). Here, we determined that the Km for PdeB is 2.9 nM, a value which is substantially lower than those previously reported for EAL-type PDEs (60 nM to a few micromolar) (36, 91, 94). Bioinformatic analysis suggests that PdeA and PdeB are the only PDEs in B. burgdorferi (31). Consistent with this notion, cell extracts prepared from the wild type, the single mutants (pdeA and pdeB), or their complements hydrolyzed 33P-labeled c-di-GMP within 2 to 20 min of incubation, while extracts prepared from spirochetes lacking both PdeA and PdeB failed to hydrolyze the nucleotide even when the reaction mixtures were incubated for up to 4 h (Fig. 4Aand reference 91).
Similar to their EAL-type counterparts (39, 94, 96), HD-GYP-type PDEs have been implicated in motility regulation in P. aeruginosa and X. campestris (77, 78). Specifically, mutations in the P. aeruginosa PDEs PA4108 and PA4781 and X. campestris pv. campestris PDE rpfG resulted in reduction of swarming motility (77, 78). Recently, we reported that inactivation of the B. burgdorferi EAL-type PdeA rendered spirochetes unable to reverse their direction of swimming (91). Interestingly, the aberrant motility displayed by spirochetes lacking PdeB differed from that of the pdeA mutant; PdeB-deficient cells exhibit significantly increased flexing (Table 1). This phenotype was attributed solely to the pdeB mutation, since complementation in trans restored the wild-type phenotype (Table 1; see Fig. S1 and videos in the supplemental material). Two conditions might be expected to lead to increased flexing, (i) overexpression of CheY3 and (ii) inhibition of CheX activity or decreased expression of CheX (58, 61, 67). Based on Western blotting, however, both CheY3 and CheX were expressed at wild-type levels in lysates prepared from pdeB single and plzA pdeB double mutants (Fig. 3), ruling out a direct role for c-di-GMP in regulating the expression of these chemotaxis proteins. Alternatively, c-di-GMP may regulate bacterial motility via an interaction with PilZ proteins (7, 14, 66, 91, 96). B. burgdorferi possesses a c-di-GMP-binding PilZ domain protein, PlzA (26, 68). Inactivation of plzA results in reduced swarming motility, although the swimming pattern of this mutant was indistinguishable from that of its parent (68). Recently, we demonstrated that the alteration of motility displayed by the pdeA mutant (91) is modulated by a PlzA-independent mechanism (68). However, spirochetes lacking PlzA in the pdeB mutant background constantly flex, a phenotype that is indistinguishable from that of a cheX mutant (see the cheX single mutant and plzA pdeB double mutant cell videos in the supplemental material) (58). These results suggest that c-di-GMP/PlzA may interact with the B. burgdorferi chemotaxis signaling pathway via CheX or CheY3, most likely by modulating the activity of CheX. We hypothesize that when it is not bound to c-di-GMP, PlzA may stimulate CheX activity. Thus, in the pdeB mutant, c-di-GMP is elevated, more PlzA-c-di-GMP complex is formed, and CheX activity is diminished, resulting in elevated CheY3-P and increasing the rate of flexing. In the latter scenario, the plzA pdeB double mutant would exhibit very low CheX activity, greatly increasing CheY3-P and causing the constant-flexing phenotype. It is possible to test these possibilities using in vitro biochemical assays.
It is interesting that B. burgdorferi apparently has only two functional c-di-GMP PDEs with disparate motility and virulence phenotypes (91; this study). While both PDEs appear to impact motility, it is unclear how they are able to alter motility in such different ways. Because of their distinct motility phenotypes, it is tempting to speculate that PdeA or PdeB may localize at different poles of a cell, thereby causing the periplasmic flagellar motor rotation to be CW or CCW by interacting with a flagellar protein or modulating the activity of a chemotaxis protein(s) in a manner that is not yet understood. A mutation in B. burgdorferi pdeA or E. coli/S. enterica PDE (yhjH) resulted in a biased direction of rotation of flagellar motors and/or decreased flagellar motor rotation. One possibility is that PdeA functions analogously to the EAL-type PDE YhjH in E. coli and S. enterica. Studies in different laboratories indicated that YhjH alters motility via the c-di-GMP effector protein YcgR, which, when complexed with c-di-GMP, acts as a “brake” on the flagellar motor switch complex (7, 22, 66, 91). However, to our knowledge, this is the first report to demonstrate that pdeB mutant cells exhibit a flexing motility phenotype without altering the speed of motor rotations.
In addition to their role in regulating bacterial motility, HD-GYP-type PDEs have also been linked to virulence (76, 79, 96). Recently, we demonstrated that spirochetes lacking PdeA were unable to infect mice following needle or tick inoculation, presumably as a result of their failure to reverse swimming direction (91). Loss of PdeB, on the other hand, had no significant effect on virulence (P = 0.3976), further supporting our hypothesis that PdeA and PdeB exert their regulatory effects by independent mechanisms. Recently, independent studies by Caimano et al. (10a), He et al. (35a), and Kostick et al. (45) demonstrated that hk1 and rrp1 are required for survival within fed tick midguts. Survival of the rrp1 mutant could be partially restored in feeding ticks by constitutive expression of genes involved in glycerol uptake and utilization (glp genes) (X. F. Yang, personal communication) (35a). Interestingly, spirochetes lacking pdeB also displayed a survival defect within fed ticks, although the mechanism(s) underlying the phenotype observed with our pdeB mutant differs from that of hk1 and rrp1 mutants; unlike hk1/rrp1 mutants, which are killed early (within 36 h) during the blood meal, spirochetes lacking PdeB appear to be killed postrepletion. Not surprisingly, loss of pdeB had no effect on glp gene (bb0241 and bb0243) expression in vitro (our unpublished observation). Together, these data may indicate that a precise modulation of c-di-GMP levels is central to the tick phase of the spirochete's enzootic cycle, with increased and decreased levels of c-di-GMP being required for adaptation to the fed and flat tick midgut environments, respectively. Somewhat surprisingly, nymphal ticks infected with the pdeB mutant failed to transmit the infection to naïve mice during engorgement, despite containing substantial numbers of viable spirochetes (Table 3). The latter data suggest a potential role for decreased c-di-GMP levels in promoting the migration of spirochetes out of the midgut during feeding. The decreased ability of PdeB-deficient spirochetes to survive within replete ticks may suggest that the degradation of c-di-GMP controls the expression and/or activity of B. burgdorferi gene products required to withstand growth within the flat midgut environment. Recently, constructed bb0323, bb0365, guaAB, dps, ospA/B, and bptA mutant strains exhibited defects in the ability to survive in ticks (41, 53, 64, 68, 73, 98, 99). Furthermore, the bba52, bba64, or bba07 mutant was reported to be unable to transmit spirochetes from ticks to naïve mice, although each of the mutants was able to infect mice by needle inoculation (32, 48, 97). Whether c-di-GMP affected the expression of one or more of these or other B. burgdorferi genes remains to be investigated. Clearly, additional studies are required to understand the unusual and complicated signaling produced by c-di-GMP-metabolizing proteins.
We thank Melissa Caimano, Justin Radolf, and X. Frank Yang for sharing their unpublished results and Nyles Charon, Justin Radolf, and Melissa Caimano for comments on the manuscript. We thank C. Li, R. Rego, P. Rosa, B. Stevenson, V. T. Lee, A. Camilli, D. Blair, A. Barbour, J. Benach, and J. Carroll for providing reagents. We also thank D. Akins for B. burgdorferi plasmid primer sequence information.
This research is sponsored by an East Carolina University Research and Development “start-up” fund to M.A.M.
†Supplemental material for this article may be found at http://iai.asm.org/.
Published ahead of print on 13 June 2011.