(C) PLOS One
This story was originally published by PLOS One and is unaltered.
. . . . . . . . . .
A chemosensory-like histidine kinase is dispensable for chemotaxis in vitro but regulates the virulence of Borrelia burgdorferi through modulating the stability of RpoS [1]
['Ching Wooen Sze', 'Department Of Oral Craniofacial Molecular Biology', 'Virginia Commonwealth University', 'Richmond', 'Virginia', 'United States Of America', 'Kai Zhang', 'Michael J. Lynch', 'Department Of Chemistry', 'Chemical Biology']
Date: 2024-02
As an enzootic pathogen, the Lyme disease bacterium Borrelia burgdorferi possesses multiple copies of chemotaxis proteins, including two chemotaxis histidine kinases (CHK), CheA 1 and CheA 2 . Our previous study showed that CheA 2 is a genuine CHK that is required for chemotaxis; however, the role of CheA 1 remains mysterious. This report first compares the structural features that differentiate CheA 1 and CheA 2 and then provides evidence to show that CheA 1 is an atypical CHK that controls the virulence of B. burgdorferi through modulating the stability of RpoS, a key transcriptional regulator of the spirochete. First, microscopic analyses using green-fluorescence-protein (GFP) tags reveal that CheA 1 has a unique and dynamic cellular localization. Second, loss-of-function studies indicate that CheA 1 is not required for chemotaxis in vitro despite sharing a high sequence and structural similarity to its counterparts from other bacteria. Third, mouse infection studies using needle inoculations show that a deletion mutant of CheA 1 (cheA 1 mut ) is able to establish systemic infection in immune-deficient mice but fails to do so in immune-competent mice albeit the mutant can survive at the inoculation site for up to 28 days. Tick and mouse infection studies further demonstrate that CheA 1 is dispensable for tick colonization and acquisition but essential for tick transmission. Lastly, mechanistic studies combining immunoblotting, protein turnover, mutagenesis, and RNA-seq analyses reveal that depletion of CheA 1 affects RpoS stability, leading to reduced expression of several RpoS-regulated virulence factors (i.e., OspC, BBK32, and DbpA), likely due to dysregulated clpX and lon protease expression. Bulk RNA-seq analysis of infected mouse skin tissues further show that cheA 1 mut fails to elicit mouse tnf-α, il-10, il-1β, and ccl2 expression, four important cytokines for Lyme disease development and B. burgdorferi transmigration. Collectively, these results reveal a unique role and regulatory mechanism of CheA 1 in modulating virulence factor expression and add new insights into understanding the regulatory network of B. burgdorferi.
Lyme disease is an infectious disease caused by the bacterium Borrelia burgdorferi which is transmitted to humans through the bite of infected blacklegged ticks. It is the most commonly reported tick-borne illness in North America and Europe. B. burgdorferi is a highly invasive bacterium that can swim and penetrate deep into tissues to cause a wide range of disorders including arthritis, meningitis, and migratory musculoskeletal pain. This invasiveness is driven by the ability of the spirochete to sense and respond to its surrounding initiated by signal transduction coupled with a sophisticated array of sensory proteins located at both cell ends. The genome of B. burgdorferi encodes multiple copies of chemotaxis-related genes and the function of many of these genes remains unknown. In this report, we focus on deciphering the role of one of the chemotaxis histidine kinases, CheA 1 . Our results reveal a novel regulatory contribution of CheA 1 to the infectious life cycle and pathogenesis of B. burgdorferi.
The role of CheA 1 currently remains unknown. CheA 1 contains all the characteristic domains of a histidine kinase similar to CheA 2 but the P2 domain required for the binding of CheY response regulator is not as conserved as other known CheAs [ 11 ]. A recent in vivo study by Xu et al., as well as our group, indicated that cheY 2 and cheW 2 are dispensable for chemotaxis and motility of B. burgdorferi in vitro but essential for a productive infection within the vertebrate host [ 15 , 18 ]. In order to comprehend the function of the cheW 2 operon, and in particular CheA 1 , in the chemotaxis and pathogenicity of B. burgdorferi, a cheA 1 mutant was generated in the virulent B31 A3-68 ∆bbe02 background [ 19 ] and its phenotype was characterized in great detail using different in vitro and in vivo approaches. The ability of this mutant to complete the enzootic infection cycle was analyzed and its contribution to RpoS-mediated regulation was explored.
The genome of B. burgdorferi encodes for two cheA (cheA 1 and cheA 2 ), three cheY (cheY 1 , cheY 2 , and cheY 3 ), and three cheW (cheW 1 , cheW 2 , and cheW 3 ) [ 8 , 9 ]. The majority of these genes are located within two gene clusters: cheA 2 -cheW 3 -cheX-cheY 3 , and cheW 2 -bb0566-cheA 1 -cheB 2 -bb0569-cheY 2 [ 9 – 11 ]. The current evidence has revealed that all of the chemotaxis genes in the cheA 2 cluster are required for chemotaxis [ 11 – 13 ]. In contrast, the genes in the cheW 2 cluster that have been studied (with the exception of bb0569, which encodes for an atypical MCP [ 14 ]) are not involved in chemotaxis in vitro. For instance, cheW 2 and cheY 2 mutants have a similar in vitro phenotype as the wild type with no apparent defect in chemotactic function [ 11 , 13 , 15 , 16 ]. Our previous study showed that CheA 2 was essential for the chemotaxis of B. burgdorferi, e.g., the cheA 2 mutant failed to reverse, ran constantly in one direction, and was non-chemotactic to attractants [ 11 ]. In addition, in vivo studies showed that CheA 2 was essential for B. burgdorferi to establish infection in a mammalian host as well as to accomplish its enzootic life cycle [ 17 ].
Lyme disease (LD), also known as Lyme borreliosis, is the most commonly reported tick-borne illness in the United States and Europe, with more than 476,000 new cases annually in the U.S. [ 5 ]. LD is caused by the bacterium Borrelia burgdorferi (also known as Borreliella burgdorferi) which is transmitted to humans through the bite of infected Ixodes ticks. Upon dissemination from the bite site, highly motile spirochetes can invade other tissues and distal parts of the body. Early-stage LD can be treated effectively with antibiotics. Unrecognized LD, however, can have debilitating long-term effects, such as arthritis, meningitis, and migratory musculoskeletal pain, which can appear after months or even years of infection [ 6 , 7 ].
Chemotaxis allows motile bacteria to sense surrounding environments and accordingly navigate their swimming behaviors to promote their survival and ability to colonize and invade a host [ 1 ]. The bacterial chemotaxis signaling pathway relies on a two-component system (TCS) that contains a chemotaxis histidine kinase CheA and a response regulator CheY. In Escherichia coli, the histidine kinase CheA is a key component of the chemotaxis signaling pathway [ 1 – 4 ]. Chemotaxis is initiated when the membrane bound methyl-accepting chemotaxis proteins (MCPs) bind a ligand (i.e., attractants) leading to activation of signal transduction from the MCP to the histidine kinase CheA via a linker protein, CheW. Activated CheA through autophosphorylation then transfers its phosphate group to the chemotaxis response regulator CheY, which in turn binds to the flagellar motor switch proteins to modulate flagellar motor rotation allowing cells to swim toward or away from an environment.
(A) Scatter plot of enriched KEGG pathway of differentially expressed genes between Sham and infected samples at 72hr p.i. Vertical coordinates represent pathway name, and horizontal coordinates represent Rich factor. The size and color of point represent the number of differential genes in the pathway and the range of different Q value, respectively. (B) Volcano plot of differentially expressed genes (DEG) between MT (cheA 1 mut ) vs WT samples at 72hr p.i.. Each dot represents a gene, and its position is determined by the log2 fold change (x-axis) and the -log10 (p-value) (y-axis). Genes with significant upregulation (red) or downregulation (green) are highlighted, while non-significant genes are shown in gray. (C) Detection of il-10, tnf, il-1β, and ccl-2 transcript levels in WT and cheA 1 mut infected mouse skin tissues using qRT-PCR. The β-actin transcript was used as an internal control. Data are expressed as mean fold change relative to WT ± SEM. *, significant difference (P < 0.05).
The attenuated phenotype of cheA 1 mut in establishing systemic infection even in immunodeficient mice implicates a defect in both innate and adaptive immune evasion. Interleukin 10 (IL-10) is a key mediator in innate immune responses during Lyme disease infection [ 60 ]. Timely induction of IL-10 expression is critical for the disease progression by dampening the host inflammatory response to favor the survival of B. burgdorferi during early infection. To dissect the cause behind cheA 1 mut failure to establish systemic infection in mice, RNA-seq analysis was performed using infected mouse skin tissues. Briefly, 100 μl of 1 x 10 7 WT or cheA 1 mut were needle inoculated into the dorsal side of BALB/c mice at three separate locations. Skin tissues were harvested from the inoculation sites at 72-hour post-injection (p.i.) for RNA extraction and RNA-seq analysis. Scatter plot of enriched KEGG pathway showed that distinct host genes were differentially induced between Sham vs WT and Sham vs MT (cheA 1 mut ) infected samples at 72 hrs p.i. ( Fig 11A ). In WT-infected samples, significant changes of numerous immune system related genes were observed, such as genes involved in leukocyte, granulocyte, and neutrophil migration. These genes were unchanged in the mutant infected samples. The transcript levels of several important cytokines and chemokines that have been reported to contribute to Lyme disease progression, such as IL-10, IL-1β, TNF-α, and CCL2 [ 61 – 68 ], were upregulated in the WT- but not MT-infected tissues ( Fig 11B ), which was validated by qRT-PCR ( Fig 11C ). Consistent with the RNA-seq analysis, qRT-PCR showed that the transcript levels of il-10, il-1β, and ccl-2 were significantly lower in the cheA 1 mut -infected mouse skin tissues as compared to those infected by WT ( Fig 11C ). The data obtained from RNA-seq analysis suggests that the inability of cheA 1 mut to establish systemic infection is partly due to its failure in eliciting timely expression of key host cytokines and chemokines to aid in dissemination and colonization.
(A) Detection of clpX, lon-1, and lon-2 transcript level in WT, cheA 1 mut , and cheA 1 com using qRT-PCR. The dnaK transcript was used as an internal control. Experiment was repeated with two independent replicates. Data are expressed as mean fold change relative to WT ± SEM. *, significant difference (P < 0.05). (B) Immunoblot analysis of WT, clpX mut , and lon-1 mut samples using specific antibodies against P66 [ 101 ], BBK32 [ 102 ], RpoS, OspC, and DbpA [ 86 ]. DnaK was used as an internal control, as previously described [ 97 ]. Experiments were repeated twice with independent replicates. Lon-1 mutant (lon-1 mut ) was kindly provided by the Ouyang’s lab as a positive control to show the up-regulation of RpoS as previously described [ 50 ].
The homeostasis and stability of proteins are governed by a set of proteases responsible for the regulated degradation of a given protein target [ 48 , 49 ]. The genome of B. burgdorferi encodes several proteases, including Lon-1 [ 50 ], Lon-2 [ 51 ], HtrA [ 52 – 55 ], CtpA [ 56 ], and Clp [ 9 ]. Lon-1 protease was shown to negatively affect RpoS expression at the transcriptional level [ 50 ] while Lon-2 protease positively regulates RpoS expression at the post-transcriptional level [ 51 ]. In other bacteria, such as E. coli, the protein stability of RpoS is governed by the ClpXP protease [ 57 – 59 ]. Since the stability of RpoS protein was affected in cheA 1 mut , the expression of these known proteases was examined in the mutant. qRT-PCR data from two independent biological replicates consistently showed that the transcripts of clpX, lon-1, and lon-2 were significantly reduced in cheA 1 mut as compared to the WT, which was successfully restored in the complemented strain ( Fig 10A ). Preliminary RNA-seq analysis also revealed a significant down-regulation of clpX, lon-1 and lon-2 expression in cheA 1 mut ( S8 Fig ). Our data suggest that depletion of CheA 1 leads to dysregulated expression of B. burgdorferi proteases, such as clpX and lon, which subsequently affects the expression and protein stability of RpoS in a negative manner. In congruence with this, deletion of clpX significantly impaired the expression of RpoS along with the various RpoS-regulated proteins ( Fig 10B ) while deletion of lon-1 led to a high level of RpoS expression as shown by Thompson et al. [ 50 ], further supporting the regulatory role of ClpX and Lon proteases on RpoS expression.
B. burgdorferi strains (WT, cheA 1 mut , and cheA 1 com ) were cultivated at 37°C/pH 6.8 to mimic the mammalian host condition and harvested at stationary phase (∼10 8 cells/ml) for (A,B) immunoblotting analysis, and (C) qRT-PCR analysis. (A) Similar amounts of whole-cell lysates were analyzed by stain-free SDS-PAGE. Gel image showed the total protein of WT, cheA 1 mut , and cheA 1 com analyzed on SDS-PAGE. (B) Immunoblot analysis of WT, cheA 1 mut , and cheA 1 com samples using specific antibodies against CheA 2 , CheA 1 [ 11 ], P66 [ 101 ], BBK32 [ 102 ], RpoS, OspC, and DbpA [ 86 ]. DnaK was used as an internal control, as previously described [ 97 ]. (C) Detection of rpoS, ospC, dbpA, and ospA transcript level in WT, cheA 1 mut , and cheA 1 com using qRT-PCR. The dnaK transcript was used as an internal control. Experiments were repeated at least twice with independent replicates. Data from two replicates are expressed as mean fold change relative to WT ± SEM. *, significant difference (P < 0.05). (D) RpoS protein turnover in WT and cheA 1 mut . The stability of RpoS protein was examined in stationary phase WT B31 A3-68 and B31 A3-68 cheA 1 mut upon protein synthesis arrest using spectinomycin. Samples were harvested at the indicated time points and probed using antibody against RpoS and DnaK (as a loading control). Experiments were repeated at least three times with independent biological replicates. A representative image is shown here.
B. burgdorferi encodes several proteins that are essential for host tissue binding/interaction and immune evasion in mammalian hosts, e.g., OspC, BBK32, and DbpB/A [ 42 – 45 ]. The attenuated phenotype of cheA 1 mut observed in the mouse infection studies could be the result of altered expression of virulence-associated proteins. To test this possibility, B. burgdorferi cells were cultivated under elevated temperature and reduced pH (37°C/pH 6.8) to mimic the host environment. Total protein profiles of WT, cheA 1 mut and cheA 1 com were analyzed on SDS-PAGE ( Fig 9A ) and a number of B. burgdorferi virulence determinants were measured by immunoblotting analysis and compared to the level of DnaK, a loading control. The protein level of RpoS was significantly reduced in the cheA 1 mut mutant as compared to the WT and cheA 1 com strains ( Fig 9B ). Consistent with this finding, the levels of several RpoS-regulated proteins (i.e., BBK32, OspC, and DbpA) were also significantly reduced in the mutant and restored in the complemented strain. By contrast, the level of P66, a non-RpoS regulated protein [ 46 ], was not affected by the deletion of cheA 1 , suggesting that CheA 1 specifically affects RpoS and RpoS-regulon expression ( Fig 9B ). qRT-PCR analysis revealed that the transcript level of rpoS was slightly reduced in the mutant but not significant while the transcript levels of ospC and dbpA were significantly reduced by 65% and 75%, respectively ( Fig 9C ). In contrast, no significant change was observed in the mutant for ospA transcript, a RpoD-regulated gene [ 47 ]. In sum, deletion of cheA 1 leads to reduced expression of RpoS and RpoS-regulated virulence factors, which in turn could explain the attenuated infectivity observed with cheA 1 mut . Depletion of CheA 1 affects RpoS post-transcriptionally as no significant reduction in rpoS transcript was detected in cheA 1 mut ( Fig 9C ). To determine how CheA 1 affects RpoS protein expression, protein turnover assays were performed using WT and mutant cells cultivated at the condition mimicking the host environment to induce RpoS expression. Compared to the wild type, RpoS turnover was increased in the absence of CheA 1 . The level of RpoS was relatively stable in the stationary phase of the wild-type culture whereas over 50% turnover of RpoS was observed in cheA 1 mut as early as one hour after protein synthesis was arrested ( Fig 9D ). Complementation of CheA 1 successfully restored the stability of RpoS ( S7 Fig ). Thus, the reduced RpoS protein level observed in cheA 1 mut is due to reduced protein stability in the absence of CheA 1 .
The tick transmission experiment revealed that cheA 1 mut was able to survive in ticks via artificial infection but failed to transmit to naïve mammalian hosts ( Fig 8A and Table 4 ). To test the possible role of CheA 1 in the acquisition phase, mice were infected with WT or cheA 1 mut via needle inoculation. Two weeks post-injection, infection was confirmed via culturing of the skin biopsy from the injection site. Nymphs were placed and confined to the injection site and allowed to feed for 72 hours. Fed ticks were then collected for qPCR to determine the spirochete acquisition efficiency. As shown in Fig 8B , the flaB gene was successfully detected in nymphs that parasitized mutant-infected mice and its copy number was comparable to that detected in nymphs that fed on WT-infected mice ( Fig 8C ). It should be noted that in the majority of cheA 1 mut -infected ticks, acquisition was greatly attenuated relative to wild type-infected nymphs with 3 out of 12 nymphs showing no detectable flaB gene. This result suggests that CheA 1 is not absolutely required for spirochete acquisition by ticks from mice, but may play some role in the process.
The infectivity of cheA 1 mut in mice was first evaluated via needle inoculation. However, this method is different from the natural route of mammalian infection through tick bite. As cheA 1 mut is still able to establish infection in ticks ( Fig 8A ), we examined whether the mutant can be transmitted to mice when fed upon by infected ticks. Groups of naïve C3H mice were allowed to be parasitized by artificially infected ticks as detailed in the Material and Methods. Seven days after tick feeding, the mice were sacrificed and specimens from the skin, ear, heart, bladder and blood were collected and subjected to qRT-PCR analysis to determine the pathogen level. Consistent with the results obtained via needle inoculation ( Table 1 ), only the tissue specimens from mice fed on by WT- and cheA 1 com -infected ticks were positive for flaB transcripts via qRT-PCR analysis ( Table 4 ). flaB transcripts were only detected from the skin and heart of one mouse fed upon by cheA 1 com -infected ticks, consistent with the results observed by needle inoculation of mice in which only a partial complementation was achieved with cheA 1 com ( Table 1 ). No flaB transcript was detected in the specimens from mice fed upon by cheA 1 mut -infected ticks ( Table 4 ), indicating that the mutant failed to be transmitted to mice even though it could colonize the tick vector. Collectively, these results suggest that CheA 1 is essential for the transmission of B. burgdorferi to mice via tick bite.
(A) Detection of spirochete burdens in microinjected nymphal ticks after feeding. RNA samples were extracted from whole fed ticks (after repletion; 5 to 7 days) and subjected to qRT-PCR analysis. The bacterial burdens in ticks were measured by the number of copies of flaB transcript compared to the number of copies of tick β-actin transcript as previously described [ 17 ]. The data are presented as the means of relative levels of flaB transcript ± SEM for each strain (WT, cheA 1 mut , and cheA 1 com ). (B) Detection of spirochete burdens in mice infected via tick bite. At day 14 after tick feeding, mice were sacrificed and tissues from the skin, heart, joint, and bladder were harvested for qRT-PCR analysis as previously documented [ 17 , 107 ]. No trace of flaB transcript was detected in the mouse tissues fed upon by cheA 1 mut infected ticks while 1 in 3 mice fed upon by cheA 1 com showed positive results at the skin and heart tissues. *, significant difference (P < 0.05). (C) Detection of spirochete burdens in naïve nymphal ticks fed on infected mice. C3H mice were artificially infected with WT or cheA 1 mut strain via needle inoculation. Naïve nymphal ticks were confined to the injection site and allowed to feed to repletion. After 72 hours, fed nymphs were collected and individually tested for the presence of flaB via qPCR. Experiments were repeated twice and data are presented as mean flaB copies per fed nymph from both data sets ± SEM.
As cheA 1 mut failed to cause disseminated infection in wild-type mice, we used a microinjection-based infection procedure to artificially infect ticks [ 39 ]. For this study, Naïve I. scapularis nymphs were microinjected with equal amounts of WT, cheA 1 mut , or cheA 1 com strains. After the infection, ticks were allowed to feed to repletion on naïve C3H mice (5 ticks/ mouse and 3 mice for each strain). The spirochete burdens in engorged ticks were measured by qRT-PCR as described previously [ 17 , 40 , 41 ]. As shown in Fig 8 , the bacterial burden of cheA 1 mut in fed ticks was not significantly different from that of the WT and cheA 1 com ( Fig 8A ). This result indicates that CheA 1 is not required for B. burgdorferi colonization and survival in the tick vector.
B. burgdorferi is able to resist the bactericidal activity of serum [ 33 , 34 ] as it produces several virulence factors that protects the spirochetes from complement mediated killing, such as BBK32, OspC, and CRASP proteins [see recent reviews [ 35 – 38 ]]. The attenuated infectivity of cheA 1 mut in both the SCID mice and tail-vein infection model suggests that CheA 1 is important for the immune evasion of B. burgdorferi. To determine if deletion of CheA 1 renders the spirochete susceptible to serum bactericidal activity, WT, cheA 1 mut and cheA 1 com were treated with normal human serum (NHS) or heat-inactivated NHS (hiNHS) and the percentage of viable cells after treatment was normalized against the input cells. Our data showed that cheA 1 mut was more susceptible to serum killing with ~25% survival rate as compared to ~60% survival rate of the WT and cheA 1 com in NHS ( Fig 7B ). This observation is in congruence with the SCID mice study where only 50% of cheA 1 mut was able to survive in the blood of infected mice and the complemented strain fully restored the survival rate to WT level ( Table 2 ). The reduced survival of the mutant in the bloodstream can likely be attributed to its susceptibility to serum bactericidal activity.
(A) To determine if high dose of inoculum can restore cheA 1 mut immune evasion and colonization barriers, 10 6 of WT, cheA 1 mut and non-motile flaB mut [ 32 ] strains were injected into wild type mice via tail-vein and sacrificed two weeks after infection. Tissues from the skin, ear, joint, heart and spleen were harvested to assess for borrelial burden using qRT-PCR as described [ 17 ]. The data are presented as mean flaB transcript copies over 10 5 of mouse β-actin ± SEM. (B) Serum bactericidal activity. 10 6 cells of WT, cheA 1 mut and cheA 1 com were incubated with 20% normal human serum (NHS) or heat-inactivated serum (hiNHS) for 2 hours. The percentage of viable cells in NHS treated cells were determined after normalization to their respective hiNHS-treated groups. The data shown here are means of three independent replicates ± SEM. *, significant difference (P < 0.05).
The above study showed that the mutant has reduced infectivity but is still able to disseminate into the bloodstream to cause systemic infection in SCID mice but not in immunocompetent mice. It is possible that in the presence of adaptive immunity, the low number of mutant cells that survived was unable to disseminate and cause systemic infection. To rule out this possibility, a 10-fold higher number (1 x 10 6 cells) of spirochetes were directly introduced into the bloodstream of BALB/c mice via tail-vein injection to determine if high doses of bacteria can overcome the dissemination and colonization barrier. As a control, a non-motile, non-infectious flaB mut strain was included [ 32 ]. 1 x 10 6 of WT, cheA 1 mut or flaB mut cells in 100 μl of PBS were injected into the tail-vein of BALB/c mice. Tissues were harvested two weeks post-injection to check for dissemination. All mice injected with the WT strain showed disseminated infection (3/3) whereas the flaB mut control was non-infectious as expected (0/3). For mice injected with the cheA 1 mut strain, live spirochetes were successfully recovered from the skin of only 1 out of 5 mice infected, indicating that the mutant is attenuated in tissue colonization even at a high inoculum introduced directly into the bloodstream ( Table 3 ). In addition, no mutant cells were recovered from the blood of any mice injected with cheA 1 mut . This result clearly indicates that CheA 1 is essential for the survival of B. burgdorferi in the blood. Consistent with live culture recovery, high levels of flaB transcripts were detected from mice infected with the WT strain whereas cheA 1 mut -infected mice had less than one copy of flaB transcript per 10 5 of mouse β-actin transcript overall (most likely due to residual B. burgdorferi that was initially introduced into the bloodstream ( Fig 7A )). These findings further substantiate the key role of CheA 1 in B. burgdorferi immune evasion and tissue colonization.
To determine if CheA 1 is required for B. burgdorferi immune evasion, needle infection study was repeated using severe combined immunodeficiency (SCIDs) mice. For this study, 10 5 of WT, cheA 1 mut and cheA 1 com strains were subcutaneously inoculated into SCID mice and sacrificed three weeks after infection. Skin tissues around and distal from the injection site were harvested along with tissues from the ear, joint, heart and spleen to assess for borrelial burden using qRT-PCR as described [ 17 ]. The data are presented as mean flaB transcript copies over 10 5 of mouse β-actin ± SEM. *, significant difference (P < 0.05).
We reasoned that the failure of the mutant to cause systemic infection could be due to the following three possibilities: 1) the mutant failed to disseminate from the injection site; 2) the mutant successfully disseminated from the injection site into the bloodstream but failed to spread to distal organs due to a defect in tissue tropism; or 3) the mutant successfully disseminated into the bloodstream but was cleared by the host immune system. To test these possibilities, the above animal study was repeated using severe combined immuno-deficiency (SCIDs) mice. Based on the tissue re-isolation results in Table 2 , WT and cheA 1 com were successfully recovered from all of the harvested tissues. In contrast, the cheA 1 mut mutant was able to disseminate and cause systemic infection in 50% of the injected SCID mice. Consistent with the result obtained using immunocompetent mice, cheA 1 mut was able to colonize and persist in the skin better than in other tissues ( Table 2 ). qRT-PCR analysis indicated that of those 50% SCID mice with systemic infection, no statistically significant difference in tissue spirochetal burdens was detected between the wild type, cheA 1 mut , and cheA 1 com with the exception of the ear and heart tissues ( Fig 6 ). This result strongly suggests that in the absence of an adaptive immune response, cheA 1 mut can disseminate from the injection site and cause systemic infection albeit to an attenuated degree. In the immunocompetent mice, cheA 1 mut most likely failed to evade the host adaptive immune response and was cleared from the animal before it was able to colonize and establish infection in distal organs. Furthermore, blood culture showed that 3 out of 6 mice were positive for cheA 1 mut when injected subcutaneously into SCID mice ( Table 2 ), further substantiating that the mutant is able to disseminate from the inoculation site and spread systemically through the bloodstream. Taken together, the mouse infection studies indicate that in the absence of CheA 1 , B. burgdorferi is attenuated in both innate and adaptive immune evasion and tissue colonization.
Intradermal infection of the mutant failed to cause systemic infection but the mutant could be re-isolated from the initial inoculation site, suggesting that the mutant is able to survive and persist at the site of injection and evade the host immune system. To determine how long the mutant can persist at the inoculation site, 1 x 10 5 cells of WT or cheA 1 mut were injected subcutaneously into BALB/c mice and the inoculation site was marked for bacterial re-isolation at different time points as indicated in the Material and Methods. qRT-PCR analysis showed that the mutant was not only able to persist but replicate at the inoculation site to the same extent as the WT as indicated by the increase in flaB transcript copies for up to 28 days ( Fig 5 ).
To determine if CheA 1 is essential for the virulence of B. burgdorferi BALB/c mice were infected intradermally with 1 x 10 5 cells of the WT, cheA 1 mut , and cheA 1 com strains (3–7 mice per strain). Tissues from the ear, skin, heart, and joint were harvested three weeks post-infection and transferred to BSK-II medium for re-isolation. The wild-type strain was successfully recovered from all of the infected animals (7/7) while cheA 1 com was successfully recovered from some tissues of five infected mice (5/5), indicating a partial complementation. Interestingly, cheA 1 mut was only recovered from the skin at the site of inoculation and not from distal organs ( Table 1 ). To rule out the potential of a spontaneous mutation in our mutant, a new deletion construct was produced to replace the entire cheA 1 gene in-frame with a kanamycin (kan) cassette ( S4D Fig ) and the mouse infection study was repeated. The newly constructed mutant (cheA 1 IFD ) was able to survive at the inoculation site but failed to establish systemic infection in immunocompetent mice ( Table 1 ), similar to cheA 1 mut . Taken together, these results suggest that CheA 1 is not essential for B. burgdorferi colonization of skin, but is required for dissemination from the initial inoculation site and/or to establish a systemic infection in mice.
(A) Bacteria tracking analysis. The swimming velocities of WT, cheA 1 mut , and cheA 1 com were measured using a computer-based motion tracking system as previously described [ 11 ]. At least 25 cells were tracked and the swimming velocities were measured. (B) Capillary tube chemotaxis assay. This experiment was carried out using 100 mM N-acetyl-D-glucosamine (NAG) as an attractant according to a previous report [ 31 ]. The results are expressed as fold increases in cell number entering the capillary tubes containing the attractant relative to the number entering control tubes without attractant (buffer alone). Results are expressed as means ± SEM for five tubes. A 2-fold increase is considered the threshold for a positive response. (C) and (D) Swimming plate assays. The assay was carried out on 0.35% agarose containing 1:10-diluted BSK-II medium as described previously [ 100 ]. The flgE mutant, a previously constructed nonmotile mutant [ 103 ], was included to determine the sizes of the inocula. The data are presented as mean diameters (in millimeters) of rings ± SEM for twelve plates. All assays were repeated at least twice with independent replicates and representative data are shown here. * significant difference (P < 0.05).
B. burgdorferi has three different swimming modes: run, flex, and reverse [ 11 , 12 ]. Our previous work showed that deletion of cheA 1 did not have impacts on the swimming behavior of B. burgdorferi in vitro [ 11 ]. All of these studies, however, were conducted using a high-passage avirulent B31A strain. To test whether this is the case in the virulent strain background, the swimming behaviors of the wild-type, cheA 1 mut , and cheA 1 com cells were analyzed using a computer-assisted cell tracker as documented previously [ 11 ]. Similar to the WT, cheA 1 mut showed three swimming modes–runs interrupted by reverses/flexes, with a mean swimming velocity of 12.5 ± 0.5 μm/sec compared to 13.7 ± 0.3 and 12.9 ± 0.3 μm/sec of WT and cheA 1 com , respectively ( Fig 4A ). In addition, capillary tube assays using N-acetyl-D-glucosamine (NAG) as a chemoattractant [ 31 ] ( Fig 4B ) and swimming plate analysis ( Fig 4C–4D ) did not reveal significant differences among WT, cheA 1 mut , and cheA 1 com , further supporting that CheA 1 is dispensable for B. burgdorferi motility and chemotaxis in vitro.
(A) In vitro localization of CheA 1 . To localize CheA 1 , a CheA 1 -GFP reporter construct ( S3 Fig ) was used to complement cheA 1 mut . The obtained cells were cultivated under unfed tick (UF) condition at 23°C/pH 7.6 or routine laboratory culture condition at 34°C/pH 7.6. Cells were monitored for the presence of GFP signal and images were taken at early log phase at ×200 magnification using a Zeiss Axiostar Plus microscope. Scale bars represent 10 μm. A cheA 1 mut carrying a GFP reporter construct was used as a control to confirm that the polar localization seen in CheA 1 -GFP complemented strain is due to CheA 1 and not an artifact from GFP protein. (B) Time course images of CheA 1 -GFP strain cultivated under UF condition. 10 5 cells/ml of CheA 1 -GFP strain was inoculated into 10 ml fresh BSK-II and cultivated under UF condition. Images were taken every 2 days to monitor the localization of CheA 1 -GFP signals. As cells entered late log to early stationary phase, increasing numbers of discrete bright puncta were observed evenly distributed along the cell body, coincided with the zones of new peptidoglycan formation that mark the division sites for daughter cells [ 106 ].
As mentioned earlier, the genome of B. burgdorferi encodes for multiple copies of chemotaxis genes. It was speculated that B. burgdorferi may possess two chemotaxis pathways that function in different hosts during the infection cycle [ 8 , 11 , 13 , 18 , 29 ], i.e., CheA 2 , CheW 3 , and CheY 3 form a pathway that executes chemotaxis in mammalian hosts, whereas CheA 1 , CheW 2 , and CheY 2 form a distinct pathway that is activated in the tick vector. Consistent with this speculation, our previous work showed that inactivation of cheA 2 impaired the ability of the spirochete to establish infection in mice but not in ticks [ 17 ]. In order to decipher the function of CheA 1 in the enzootic cycle of B. burgdorferi, a GFP reporter construct ( S5A Fig ) was created and used to complement cheA 1 mut . The reporter strain was cultivated under laboratory conditions as well as conditions mimicking the tick milieu in vitro and observed microscopically. Surprisingly, we noticed that CheA 1 -GFP was localized to one cell pole when cultivated under the tick-phase condition but appeared diffused at elevated temperature ( Fig 3A ). The diffused pattern was not due to degradation of the GFP-fusion protein at elevated temperature as full-length CheA 1 -GFP fusion protein was detected with no excessive degradation ( S5C Fig ) . This pattern was not observed in a strain carrying only the empty GFP vector ( Fig 3A ) or in a strain expressing CheA 2 -GFP fusion protein ( S5C–S5D Fig ). This result indicates that the polar localization pattern is unique to CheA 1 ( Fig 3A ). In addition, the localization of CheA 1 appeared to be connected with cell division. During early log phase, the CheA 1 -GFP signal was localized to one cell pole. As the cells entered mid log phase, the signal of CheA 1 -GFP was seen distributed to the mid cell as well as both cell poles ( Fig 3B Day 7 onward). As the cells entered late log to stationary phase, CheA 1 -GFP signal was seen distributed along the cell body in specific puncta as indicated by the white arrows in Fig 3B Day 13–17 ( Fig 3B ). When the cells were cultivated under elevated temperature, the polar localization of CheA 1 -GFP disappeared ( S6 Fig ). At elevated temperature, during the early growth phase, CheA 1 -GFP signal appeared dispersed along the cell body and only when the cells entered stationary phase single pole localization reappeared as was the case with cells cultivated at 23°C ( S6B Fig ), suggesting that the localization of CheA 1 is regulated spatially and temporally. In contrast to CheA 1 -GFP, the localization of GFP and CheA 2 -GFP remained evenly distributed along the cells under all culture conditions and growth phases ( S6A and S6C Fig ). Collectively, these results strongly suggest that the localization of CheA 1 is dynamic throughout the growth of B. burgdorferi and that CheA 1 function may be more prominent during the tick phase of the enzootic cycle. In addition, pH (acidic when cell density increases and enters stationary phase [ 30 ]) or temperature may have an impact on the localization and function of CheA 1 .
(A) Deletion of cheA 1 has no impact on B. burgdorferi growth. Construction of CheA 1 /pBBE22 for the complementation of cheA 1 mut . bb0566-cheA 1 fragment was PCR amplified using primer pair P 1 /P 2 and cloned into pBBE22 [ 17 , 99 ], a shuttle vector of B. burgdorferi, yielding CheA 1 /pBBE22. (B) Immunoblot analysis of WT, cheA 1 mut and cheA 1 com strains. The same amounts of WT, cheA 1 mut and cheA 1 com whole-cell lysates were analyzed by SDS-PAGE and then probed with CheA 1 and DnaK antibodies, as previously documented [ 11 , 97 ]. Growth analysis of WT, cheA 1 mut and cheA 1 com under routine laboratory condition (C) (34°C/pH7.6), (D) tick-like condition (23°C/pH7.6), and (E) host condition (37°C/pH6.8). Growth curve analysis was carried out to determine if CheA 1 affects cell growth. 10 4 or 10 5 cells/ml of bacteria were inoculated into 10 ml BSK-II media and cultivated under 34°C/pH7.6, 23°C/pH7.6, or 37°C/pH6.8. Cell numbers were enumerated every 1–4 days until cells entered stationary phase. Cell counting was repeated in triplicate with at least two independent samples, and the results are expressed as means ± SEM.
To investigate the role of CheA 1 in the infectious cycle of B. burgdorferi, we have attempted to inactivate cheA 1 in the virulent B31 A3-68 Δbbe02 strain [ 25 ]. pGA 1 kan, a previously constructed vector [ 11 ], was linearized and electro-transformed into wild-type competent cells. A previously described PCR analysis was first carried out to screen for clones with the desired targeted mutagenesis [ 11 ]. One clone (referred to as cheA 1 mut ) was selected for further analysis. For the complementation, the plasmid CheA 1 /pBBE22 ( Fig 2A ) was electro-transformed into the cheA 1 mut competent cells. The presence of CheA 1 /pBBE22 in antibiotic-resistant colonies was confirmed by immunoblots using anti-CheA 1 antibody and anti-DnaK antibody as a control. As shown in Fig 2B , a band of approximately 80 kDa was detected in the wild type and the cheA 1 com strain but was absent in cheA 1 mut , indicating that the cognate gene product was abrogated in the mutant and restored in the complemented strain ( Fig 2B ). B. burgdorferi B31 strain contains 21 linear and circular plasmids [ 26 ] of which some are essential for its infectivity but are easily lost during in vitro cultivation [ 27 ]. To check for the plasmid profile of our isolated strains, a previously developed PCR method was used to determine if the obtained positive clones contain the same plasmids as their parental strains [ 28 ]. Our results showed that cheA 1 mut ( S4B Fig ) and its isogenic complemented strain ( S4C Fig ) contained the same plasmid profile as the wild-type strain ( S4A Fig ). Growth analysis indicated that deletion of cheA 1 did not affect the fitness of the mutant at 23°C,34°C or 37°C ( Fig 2C–2E ) indicating that depletion of CheA 1 has no impact on B. burgdorferi growth in vitro.
To better contextualize the differences in P2 and P3 predicted for the BbCheAs, we performed multiple-sequence alignment analysis on discrete groups of CheA sequences within the spirochete phylum. We first examined all the CheA sequences in spirochete species that contained only one CheA gene, assuming that this single CheA would be associated with chemotaxis signaling. Within this category, we examined only those members from the order Treponematales. Multiple sequence alignments of these CheA sequences reveal they have a single P2 domain and an extended P3 domain ( S2 Fig ). These results indicate that while the extra P2 domain may or may not play a role in chemotaxis for other spirochete species, the extended P3 domain is a conserved structural motif for CheAs associated with chemotaxis in spirochetes. We next looked at all CheA 1 and CheA 2 sequences from species in the Borreliaceae family. Multiple alignment data from these sequences reveal that in species with two distinct CheA isoforms, all the CheA 1 proteins encode a single P2 domain and have a short P3 domain similar to BbCheA 1 and TmCheA ( S3 Fig ), and all CheA 2 proteins encode for a second P2 domain and an extended P3 domain ( S3 Fig ). These data suggest that while an extended P3 domain appears to be important for CheA proteins associated with chemotaxis in diverse spirochete species, the addition of a second P2 domain may be a Borrelia spp. and Borreliella spp. specific modification.
To gain more insight into the overall differences in the domain architecture of CheA 1 and CheA 2 , we superimposed domains P1-P5 from BbCheA 1 (pink) and BbCheA 2 (green) to the analogous domains of TmCheA (tan, Fig 1D ). Overall, the P1, P4, and P5 domains are quite similar in all three proteins (RMSDs < 1), whereas the analogous P2 and P3 domains vary substantially (RMSD ranging 0.71–10.2 Å) among the three CheAs. For the P2 domain superimposition, the domain cores are similar with most of the structural deviations arising from the loops connecting the β-strands, and the N- and C-termini connecting to the P1-P2 and P2-P3 linkers, respectively. Analysis of the putative CheY binding site for BbCheA 1 P2 and BbCheA 2 P2 (α and β) reveals that the structures of the P2 domains are generally well-conserved in this region, supporting sequence alignment data that indicates conservation in the CheY:P2 binding interface residues ( S1A–S1B Fig ). Superimposition of the P3 domains from TmCheA, BbCheA 1 , and BbCheA 2 AlphaFold models reveals the extent by which BbCheA 2 P3 domain differs from that of TmCheA and BbCheA 1 , whose P3 domains align with an RMSD of 0.7 Å. Compared to TmCheA P3, BbCheA 2 P3 aligns with an RMSD of 3.7 Å and is approximately 43 Å longer. The extended P3 domain may be important for maintaining the proper protein-protein interactions between CheA subunits within a homodimer and/or between CheA and the chemosensory receptors in the arrays [ 21 , 24 ].
( A) AlphaFold models of B. burgdorferi CheA 1 (BbCheA 1 ) and CheA 2 (BbCheA 2 ) subunits compared to Thermotoga maritima CheA (TmCheA). Protein structures are colored according to their pLDDT scores [ 20 ]. (B) Domain organization of BbCheA 1 , BbCheA 2 , and TmCheA. The first P2 domain of BbCheA2 (referred to here as P2 α ) extends from Glu153 to Lys265 and is connected to the second P2 domain (P2 β ) via an 18-residue long P2 α -P2 β linker. P2 β is comprised of residues Glu283 to Val387 and is connected to the dimerization domain (P3) via a 32-residue long P2 β -P3 linker. Overall, the P1-P2 α and P2 β -P3 linkers consist of approximately 50 residues, similar to the 43-residue long P2-P3 linker of TmCheA. (C) Superimposition of BbCheA P2 α (RMSD: 0.48 Å, cyan, C150-V267) and P2 β (gray, F279-V387). (D) Superimposition of phospho-transfer domain (P1), CheY-binding domain (P2), dimerization domain (P3), kinase domain (P4), and regulatory domain (P5) of TmCheA (tan), BbCheA 1 (pink), and BbCheA 2 (green). RMSD of each alignment is indicated in the lower region of each panel. All structural superimposition and figures were generated in PyMol [ 105 ].
To investigate why CheA 1 is dispensable for chemotaxis in vitro while CheA 2 is required [ 11 ], we generated AlphaFold [ 20 ] models of the two CheA proteins (BbCheA 1 and BbCheA 2 ) and compared them to the solved structure of CheA from Thermotoga maritima (Tm) [ 21 , 22 ] ( Fig 1A ). The models show high confidence (pLDDT scores >80) for all domains in each monomer, with pLDDT scores < 50 only for the residues that comprise the P1-P2 and P2-P3 linkers. The domain composition of TmCheA is similar to BbCheA 1 and BbCheA 2 with some notable differences ( Fig 1B ). BbCheA 1 has an analogous domain arrangement compared to TmCheA with the only major difference being an extended P2-P3 linker (73 residues-long compared to 42 residues of TmCheA). In contrast, BbCheA 2 differs from BbCheA 1 and TmCheA in several distinct ways. First, BbCheA 2 has two P2 domains (CheY binding domains, referred to here as P2 α and P2 β ) compared to a single P2 domain in TmCheA and BbCheA 1 . Superimposition of the P2 α and P2 β domains reveals a high degree of structural similarity despite a mere 26% identity ( Fig 1C ). Sequence alignment suggests that both P2 domains bind to CheY, which could potentially act to increase the local phospho-CheY concentration within the chemosensory arrays of B. burgdorferi ( S1A Fig ), as has been shown for the P2 domain of the canonical E. coli CheA [ 23 ]. The second major difference between BbCheA 1 and BbCheA 2 is that BbCheA 2 has an extended P3 domain ( Fig 1B ). CheA homodimerizes via P3-P3’ interactions, wherein two adjacent P3 domains come together to form a four-helix bundle [ 21 ]. In BbCheA 2 , the 122 residues in P3 domain is significantly longer than the ~53 residues domain found in TmCheA and BbCheA 1 due to inserts between residues Ile446-Leu515 ( Fig 1A and 1B ).
Discussion
B. burgdorferi lives a complex enzootic life cycle alternating between arthropod vector and mammalian reservoir. As such, highly coordinated expressions of tick- and host-specific transcriptomes are indispensable to ensure the survival and persistence of the spirochete in nature. Two two-component systems (TCS), the Hk2-Rrp2-RpoN-RpoS and HK1-Rrp1 signaling pathways, are largely responsible for these extensive transcriptional adjustments (see reviews [69–71]). Among the enzootic cycle stage-specific differentially expressed genes, in addition to multiple crucial surface lipoproteins, many chemotaxis-related genes are also found to have distinct expression profiles when B. burgdorferi alternates between different hosts, which strongly indicates that the composition of the B. burgdorferi chemotaxis machinery differs between ticks and mammals [72]. The genome of B. burgdorferi encodes for orthologs of chemotaxis-related genes, many of which are present in several copies, such as two chemotaxis histidine kinases (cheA 1 , cheA 2 ) and three chemotaxis response regulators (cheY 1 , cheY 2 , cheY 3 ) [9]. Some of these genes have no known function in chemotaxis as mutagenesis studies revealed no apparent contribution to chemotaxis in vitro. For example, a mutant deficient in CheA 1 exhibited a normal chemotaxis response whereas loss of CheA 2 resulted in a non-chemotactic phenotype in vitro [11,17]. Hence, understanding how these chemotaxis-related proteins are coordinated during the natural enzootic cycle of B. burgdorferi is crucial to fully comprehend the significance of each protein in the modulation of chemotaxis responses in vivo.
Interestingly, even though CheA 1 is not required for chemotaxis in vitro, it has a domain arrangement and a local domain architecture more similar to TmCheA than the chemotactic CheA 2 . AlphaFold models reveal that CheA 1 has a single P2 domain, while CheA 2 has two P2 domains (α and β, Fig 1A–1B). Sequence alignment, structural superimposition, and previous biochemical data suggest that these domains all bind to CheY-like proteins [13] (S1A–S1B Fig), although whether different P2 domains are specific for one CheY isoform over another (i.e. CheY 1-3 ) is not currently known. In addition to the P2 domain, the other major difference between CheA 1 and CheA 2 is the length of the P3 domains (Fig 1D). CheA 1 has a P3 domain very similar in length to TmCheA, whereas in CheA 2 the P3 is over 68 residues and 43 Å longer than CheA 1 and TmCheA. Previously, it has been recognized that B. burgdorferi and Treponema denticola have extended P3 domain structures [21,24]. It was hypothesized that these extended P3 domains may stabilize the CheA 2 homodimer by increasing the binding interface surface area of the two CheA subunits. Another possibility could be that the longer P3-P3’ four-helix bundle is an adaptation to allow better engagement with the chemosensory arrays of spirochetes, which have distinct architectures owing to the high membrane curvature [21,24]. Although the exact molecular mechanism behind why Borreliaceae CheA 2 encodes for a second P2 domain and an extended P3 domain has yet to be determined, it is nonetheless interesting that these two features are conserved across the family of Borreliaceae (S3 Fig).
Motility and chemotaxis contribute significantly to the pathogenesis of B. burgdorferi as loss of motility or chemotaxis resulted in non-infectious spirochetes that failed to colonize tick vectors and/or mammalian hosts [17,73,74]. A previous study from our group established that chemotaxis-signaling mediated by the CheA 2 histidine kinase is required for the transmission of spirochetes from ticks to mammals, as well as for establishing an infection in a vertebrate host, but is not essential for its survival in ticks [17]. Based on the organization of the chemotaxis gene clusters [75] as well as mutagenesis studies of genes in the cheA 1 [11,14–16] and cheA 2 [12,13,17,73] operons, we hypothesized that B. burgdorferi may utilize different chemosensory machineries in different environments during its infectious cycle, i.e., CheA 2 modulates chemotaxis during transmission/persistence in mammalian hosts while CheA 1 constitutes a second chemotaxis pathway to guide the chemotactic response in the tick vector and accordingly mediate B. burgdorferi transmission from infected ticks to mammals. In line with our hypothesis, a recent study by Iyer et al. showed that CheA 1 and CheY 2 are highly induced in transmitting ticks which indicates that these two proteins may form a functional chemotaxis pathway that is activated when an infected tick takes a blood meal so as to prepare dormant spirochetes for migration and transmission to a new host [72]. Lastly, a recent study by Caimano et al. further substantiated the role of CheA 1 during transmission whereby the level of cheA 1 is highly induced by RpoS while in fed ticks and mammals [76].
One interesting observation from the current study revealed that CheA 1 has a dynamic localization during growth. Importantly, the intense single pole localization of CheA 1 when cultivated under condition mimicking the unfed tick milieu versus elevated temperature indicates that CheA 1 may play a more prominent role in the arthropod vector as compared to the vertebrate host (Figs 3 and S6). CheA 1 may be strategically placed near the cell pole in proximity to the chemoreceptor arrays [16], ready to detect stimuli from the blood meal in order to initiate transmigration out from the tick midgut to the salivary glands for transmission [77,78]. To answer this question, in vivo visualization of cheA 1 mut cells after the tick blood meal is necessary to reveal whether the mutant’s inability to be transmitted via tick bite (Table 4) is due to failure in migration from the hemolymph to the tick salivary glands. Additionally, the increase in puncta localization of CheA 1 -GFP as the cells transition from early to late log phase has a striking resemblance to ParB localization. ParB is a chromosome segregation protein that has been shown to recruit the structural maintenance of the chromosome (SMC) complex to the origin of replication in B. burgdorferi [79]. It would be interesting to test if the localization of CheA 1 is associated with the DNA partitioning Par proteins perhaps to ensure equal segregation of CheA 1 protein among the daughter cells during cell division.
Given that the cheW 2 operon is positively regulated by RpoS and BosR [18,76], the failure of cheA 1 mut during transmission could be partly due to defect in transmigration from the tick gut to the salivary glands during the blood meal, similar to the phenotype of a rpoS mutant in the feeding tick [80]. Most significantly, the possible association of CheA 1 with the Par proteins implies that the established pool of CheA 1 needs to be distributed equally among the daughter cells. This unique partitioning was also observed in cells cultured at elevated temperature but only during the late stationary phase (S6B Fig). Once these late stationary phase cells were re-inoculated into fresh BSK-II media, CheA 1 localization reverted to the diffused pattern as seen during day 1. This protein localization is specific to CheA 1 as we did not observe this with CheA 2 (S5D and S6C Fig). The need for CheA 1 to have spatial and temporal localization is currently unknown. Given that the tick-mouse infection study points toward a role for CheA 1 during tick transmission and hematogenous dissemination, it would be interesting to monitor the localization of CheA 1 in feeding ticks and in mammals to determine if CheA 1 alters its localization in different cellular milieus and whether such localization contributes to its function.
Following initial deposition, either via tick bite or needle inoculation, B. burgdorferi needs to first establish a local infection before disseminating to peripheral organs. Therefore, hematogenous dissemination is a critical pathogenic event for early infection [81]. Once in the bloodstream, the spirochetes must transmigrate via vascular extravasation to spread and colonize distal organs [82,83]. Several findings elucidate the dissemination event in which CheA 1 is most likely to be involved. First, the cheA 1 mutant is able to survive at the injection site of immune-competent and–deficient mice after needle inoculation, indicating that CheA 1 is not required for skin tissue survival. Needle inoculation into SCID mice allowed partial dissemination of the cheA 1 mutant to the bloodstream and subsequent colonization of skin regions (i.e., ears) distal from the injection site as well as other organs, but failed to do so in immunocompetent mice. This result indicates that CheA 1 is required for the spirochete survival in the bloodstream most likely by aiding the bacteria in immune evasion. In line with this, immunoblot analysis revealed that deletion of CheA 1 negatively impacted the expression of several essential surface virulence factors, such as OspC and BBK32 (Fig 9B). OspC has been shown to interact with host immune components, such as the complement component 4b (C4b), and contributes to the survival of B. burgdorferi in the bloodstream [42]. This can explain the finding in SCID mice where only 50% of the mice had systemic infection as the complement system remains intact in the immune-deficient mice and cheA 1 mut is not as protected in the bloodstream as compared to the wild type (Table 2 and Fig 6). This was corroborated by serum bactericidal assays, as deletion of CheA 1 renders the spirochetes more susceptible to serum killing (Fig 7B). B. burgdorferi infection by tail-vein further supported this proposition, as the mutant was cleared from the bloodstream of wild-type mice within 14 days and was severely attenuated in hematogenous dissemination. Tail-vein injection with high doses of spirochetes was not able to rescue the survival and systemic dissemination defect of cheA 1 mutant indicating that in addition to attenuated innate and adaptive immune evasion, the mutant may have a defect in tissue colonization or an attenuated ability to extravasate due to dysregulated surface protein expression (Fig 9), such as BBK32 [84], which led to the clearance of spirochetes in the bloodstream by the host immune system. The ability of cheA 1 mut to transmigrate from the vasculature will need to be examined via intravital imaging in order to determine if the reduced ability of cheA 1 mut to colonize and disseminate from the bloodstream of mice is partially due to a defect in extravasation in addition to impaired immune evasion. Artificial infection of Ixodes ticks showed no significant variation in the survival between wild type and mutant (Fig 8A). Furthermore, the cheA 1 mutant can be acquired from the skin of infected mice suggesting that CheA 1 is dispensable for acquisition. In addition, cheA 1 mut failed to be transmitted via infected tick bites, suggesting that CheA 1 is required for tick transmission, consistent with the observed increase in cheA 1 transcripts during transmission [72]. Immunoblot and qRT-PCR analyses (Fig 9) suggest that the attenuated infectivity in mice upon depletion of CheA 1 is likely due to dysregulated RpoS and RpoS-regulon expression, which are essential for B. burgdorferi survival, immune evasion, and persistence in mammalian hosts [42,43,45,69,78,85,86]. Protein turnover assays revealed that the stability of RpoS was significantly reduced in the absence of CheA 1 (Fig 9D) and that this is likely due to dysregulated protease expression (Fig 10).
In other bacteria such as E. coli and Bacillus subtilis, in addition to the σ70 house keeping sigma factor, , multiple alternative sigma factors are employed to effectively regulate gene expression based on changes in the growth condition, such as stress response, starvation, and transition between motile and sessile stages [87,88]. The activity of each σ is controlled by the binding of anti-σ proteins such that the proper σ can be released for interacting with the RNA polymerase (RNAP) when required [89–91]. The genome of B. burgdorferi encodes two alternative sigma factors, σ54 (RpoN) and σ38 (RpoS) [9]. How sigma factor partner switching occurs in B. burgdorferi remains unclear as no anti-sigma factors have yet been discovered. Preliminary RNA-seq analysis performed using in vitro cultured samples revealed significant up-regulation of many RpoS-repressed genes [76], such as bb_d18 and bb_a68 [76,92] in the cheA 1 mut mutant (S7 Fig). As shown in Fig 10A, the level of bb_d18 was significantly higher in cheA 1 mut as compared to WT when validated using qRT-PCR. Notably, the expression of bb_d18 remained elevated in the complemented strain cheA 1 com, which may explain why only a partial complementation was achieved in cheA 1 com (Table 1). The untimely expression of RpoS-repressed genes may have diluted the pool of RNAP-RpoD complex available for transcription of core genes, such as clpX and lon proteases, leading to global dysregulation of the CheA 1 mutant transcriptome (S8 Fig). Additionally, BBD18 is known as a negative regulator of RpoS [93–95] and upregulation of this gene can further repress the expression of RpoS-regulon in the mutant (Fig 9). In sum, the data strongly suggest that CheA 1 affects the pathogenicity and transcriptome of B. burgdorferi by influencing the stability of RpoS potentially through ClpX and Lon proteases.
As discussed above, CheA 1 may contribute to chemotaxis within the tick. It is also possible that it has a role to play in host dissemination. We did not detect any chemotaxis defect in the cheA 1 mut in vitro (Fig 4); however, the significance of CheA 1 in chemotaxis may only be observable in vivo as lab cultivation rarely accurately represents any aspect of the natural infectious cycle of B. burgdorferi [96]. We had previously shown that CheW 2 contributes to the full pathogenicity of B. burgdorferi as depletion of CheW 2 resulted in attenuated dissemination and colonization of the spirochetes [18]. Given that CheW 2 preferentially interacts with CheA 1 [16], CheA 1 , CheW 2 , and CheY 2 may constitute an active chemotaxis pathway during the vector transmission and host dissemination phases to guide the chemotactic responses and migration of B. burgdorferi. Notably, the phenotype of cheA 1 mut differs from what was observed in an infectious cheY 2 and cheW 2 mutant; Xu et al. showed that cheY 2 mutant was cleared from the inoculation site after two weeks and failed to cause systemic infection [15]. In contrast, cheA 1 mut persisted at the initial injection site and was able to replicate as shown by our time course study (Fig 5). As for CheW 2 , even though deletion of CheW 2 led to attenuated infectivity in immune-competent mice, its infectivity was not affected in SCID mice, which differs significantly from the in vivo phenotype of cheA 1 mut [18]. However, all three mutants exhibited wild-type chemotaxis behavior in vitro which raises the question as to whether CheA 1 -CheW 2 -CheY 2 is required for chemotaxis under specific circumstances in vivo. cheA 1 and cheY 2 mutant survival in the tick vector was not significantly different from the wild type (Fig 8A), emphasizing that CheA 1 and CheY 2 [15] are not required for the persistence and survival of B. burgdorferi in the arthropod vector. How is cheA 1 mut able to persist and spread in the skin tissue but not to distal organs? Given that B. burgdorferi can spread through direct tissue spreading [81], it is possible that once cheA 1 mut establishes localized infection at the injection site, it is able to spread within the skin tissue to distal skin regions (e.g., ears) and avoid clearance by the host immune system via non-hematogenous spreading.
In order to colonize and survive in a mammalian host, B. burgdorferi has evolved to alter its surface antigen expression for distinct tissue tropism as well as in response to host immune reaction [45]. The disparity in host immune repones elicited by WT vs cheA 1 mut (Fig 11) is in congruence with the mouse infection results. The lack of key inflammatory mediators necessary for spirochete dissemination may dampen the progression of cheA 1 mut infection and dissemination outcome, leading to attenuated infectivity as observed (Tables 1, 2 and 3). The anti-inflammatory cytokine IL-10 plays a key role in the progression of B. burgdorferi infection [60–62]. The failure of cheA 1 mut to elicit timely expression of IL-10 in mice may have contributed to its failure in dissemination and colonization in both immune-competent and–deficient mice (Tables 1 and 2). Despite this, the mutant was able to persist at the inoculation site (Fig 5) without the protection of an IL-10 cytokine effect which implies that additional bacterial and host factors may be required. Lack of an activated host immune reaction following cheA 1 mut infection (Fig 11B) could potentially allow the spirochetes to evade immune surveillance and persist at the inoculation site.
[END]
---
[1] Url:
https://journals.plos.org/plospathogens/article?id=10.1371/journal.ppat.1011752
Published and (C) by PLOS One
Content appears here under this condition or license: Creative Commons - Attribution BY 4.0.
via Magical.Fish Gopher News Feeds:
gopher://magical.fish/1/feeds/news/plosone/
